Move from only supporting the default of pre-multiplied
alpha to supporting user specified blend mode using the
standardised property.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
We are using mode->crt_clock here which is filled by drm_mode_set_crtcinfo()
which is called right after .mode_valid.
Use mode->clock which is valid here.
Fixes: 624d93a4f0 ("drm/vc4: hdmi: Move clock calculation into its own function")
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Still under investigation, but the conditions under which the HVS
will accept values written to the gamma PWL are not straightforward.
Disable gamma on HVS5 again until it can be resolved to avoid
gamma being enabled with an incorrect table.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Replace the cfi directives with the UNWIND equivalents. This prevents
the .eh_frame section from being created, eliminating the warnings.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
With the automatic VL805 support being removed from the standard
CM4 dtb (since most CM4 carriers don't have a VL805), retain support
on those that do by creating a "vl805" overlay that restores the
deleted "usb@0,0" node.
The "vl805" overlay will be loaded automatically (after an upcoming
firmware update) on CM4 boards where the EEPROM config includes the
setting VL805=1.
See: https://forums.raspberrypi.com/viewtopic.php?t=326088
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
To allow for the cases where a simple panel does have a GPIO
controlled backlight. Defaults to having no backlight defined.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Use GitHubs issue form for bug reports.
- modern look
- user don't need to mess with given markdown parts while filling the issue template
Setup config.yml for general questions and problems with the Raspbian distribution packages.
The driver did allow the exposure to go up to VTS - 4 lines,
but this would produce a visible line on 1280x800, and a stall of
the sensor at 640x480.
Whilst it appears to work with a difference of 5, the datasheet states
there should be at least 25 lines difference between VTS and exposure,
so use that value.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The drm edid parser doesn't signal RGB support on DVI monitors
with old edid versions, leading to 8-bit RGB mode being rejected
and no video on DVI monitors.
As 8-bit RGB is mandatory on HDMI and DVI monitors anyways we can
simply drop the RGB format check, aligning vc4 with other drivers.
Signed-off-by: Matthias Reichl <hias@horus.com>
When vc4_hdmi_connector_detect() was called in
connector_status_connected state it incorrectly cleared the
hdmi_monitor flag, leading to no audio on RPi3.
Fix this by clearing hdmi_monitor only when the hpd check
indicated no connection or if reading the edid failed.
Signed-off-by: Matthias Reichl <hias@horus.com>
This fixes a regression introduced in 131f132203
(see also https://github.com/raspberrypi/linux/issues/4791).
The tc358743 driver refused to bind to the device. The irs1125
driver is likely behaving similarly.
The new unified cam1_clk node that represents the fixed on-board
oscillator is marked as disabled by default. These overlays didn't
expect this and so the clock nodes were stuck in disabled state.
This commit just adds the required status = "okay" line. Other sensor
drivers do this too.
Commit 311a839a1a upstream.
Add documentation for the V4L2_CID_NOTIFY_GAINS control.
This control is required by sensors that need to know what colour
gains will be applied to pixels by downstream processing (such as by
an ISP), though the sensor does not apply these gains itself.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Commit a9c80593ff upstream.
We add a new control V4L2_CID_NOTIFY_GAINS which allows the sensor to
be notified what gains will be applied to the different colour
channels by subsequent processing (such as by an ISP), even though the
sensor will not apply any of these gains itself.
For Bayer sensors this will be an array control taking 4 values which
are the 4 gains arranged in the fixed order B, Gb, Gr and R,
irrespective of the exact Bayer order of the sensor itself. The use of
an array makes it straightforward to extend this control to non-Bayer
sensors (for example, sensors with an RGBW pattern) in future.
The units are in all cases linear with the default value indicating a
gain of exactly 1.0. For example, if the default value were reported as
128 then the value 192 would represent a gain of exactly 1.5.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
This path actually occurs when audio is started during a hdmi mode set.
As the data will be written by vc4_hdmi_set_infoframes when packet RAM
is enabled again, don't treat as an error
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
In addition to the RGB444 output, the BCM2711 HDMI controller supports
the YUV444 and YUV422 output formats.
Let's add support for them in the driver, but still use RGB as the
preferred format.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Currently we take the max_bpc property as the bpc value and do not try
anything else.
However, what the other drivers seem to be doing is that they would try
with the highest bpc allowed by the max_bpc property and the hardware
capabilities, test if it results in an acceptable configuration, and if
not decrease the bpc and try again.
Let's use the same logic.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The current code only base its decision for whether the scrambler must be
enabled or not on the pixel clock of the mode, but doesn't take the bits
per color into account.
Let's leverage the new function to compute the clock rate in the
scrambler setup code.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In the function that validates that the clock isn't too high, we've only
taken our controller limitations into account so far.
However, the sink can have a limit on the maximum TMDS clock it can deal
with too which is exposed through the EDID and the drm_display_info.
Make sure we check it.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The code to compute our clock rate for a given setup will be called in
multiple places in the next patches, so let's create a separate function
for it.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Our code is doing the same clock rate validation in multiple instances.
Let's create a helper to share the rate validation.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In order to support the YUV output, we'll need the atomic state to know
what is the state of the associated property in the CSC setup callback.
Let's change the prototype of that callback to allow us to access it.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The current CSC setup code for the BCM2711 uses a sequence of register
writes to configure the CSC depending on whether we output using a full
or limited range.
However, with the upcoming introduction of the YUV output, we're going
to add new matrices to perform the conversions, so we should switch to
something a bit more flexible that takes the matrix as an argument and
programs the CSC accordingly.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
On BCM2711, the HDMI_CSC_CTL register value has been hardcoded to an
opaque value. Let's replace it with properly defined values.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
On the BCM2711, the HDMI_VEC_INTERFACE_XBAR register configuration
depends on whether we're using an RGB or YUV output. Let's move that
configuration to the CSC setup.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The CSC callbacks takes a boolean as an argument to tell whether we're
using the full range or limited range RGB.
However, with the upcoming YUV support, the logic will be a bit more
complex. In order to address this, let's make the callbacks take the
entire mode, and call our new helper to tell whether the full or limited
range RGB should be used.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
We're going to need to tell whether we want to run with a full or
limited range RGB output in multiple places in the code, so let's create
a helper that will return whether we need with full range or not.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The drm_hdmi_avi_infoframe_colorspace() function actually sets the
colorimetry and extended_colorimetry fields in the hdmi_avi_infoframe
structure with DRM_MODE_COLORIMETRY_* values.
To make things worse, the hdmi_avi_infoframe structure also has a
colorspace field used to signal whether an RGB or YUV output is being
used.
Let's remove the inconsistency and allow for the colorspace usage by
renaming the function.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The current code, when parsing the EDID Deep Color depths, that the
YUV422 cannot be used, referring to the HDMI 1.3 Specification.
This specification, in its section 6.2.4, indeed states:
For each supported Deep Color mode, RGB 4:4:4 shall be supported and
optionally YCBCR 4:4:4 may be supported.
YCBCR 4:2:2 is not permitted for any Deep Color mode.
This indeed can be interpreted like the code does, but the HDMI 1.4
specification further clarifies that statement in its section 6.2.4:
For each supported Deep Color mode, RGB 4:4:4 shall be supported and
optionally YCBCR 4:4:4 may be supported.
YCBCR 4:2:2 is also 36-bit mode but does not require the further use
of the Deep Color modes described in section 6.5.2 and 6.5.3.
This means that, even though YUV422 can be used with 12 bit per color,
it shouldn't be treated as a deep color mode.
This deviates from the interpretation of the code and comment, so let's
fix those.
Fixes: d0c94692e0 ("drm/edid: Parse and handle HDMI deep color modes.")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
bcm2835_isp_remove is called on an initialisation failure, but at that
point the drvdata hasn't been set. This causes a crash when e.g. using
the cutdown firmware (gpu_mem=16).
Move platform_set_drvdata before the instance probing loop to avoid the
problem.
See: https://github.com/raspberrypi/linux/issues/4774
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This reverts commit f814bfc5f4.
The commit caused media stalls with kodi and stateful
v4l2 video decode.
John is now using a different way of limiting latency
through stateful v4l2 so this is not required.
Change the maximum sample rate for the amplifier to
192KHz as given in the Infineon specification.
Signed-off-by: Joerg Schambacher <joerg@hifiberry.com>
Parameterise the overlays so that they can have an optional
cam0 parameter to switch to i2c_vc and csi0.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
CM4S follows CM1/3, so based on the documentation cameras/displays
connect to 0/1 and 28/29, not 0/1 and 44/45.
Likewise the camera regulator controls are independent as on CM1/3.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Not all implementations wire up the enable GPIO and may just tie
it to a supply rail.
Make it optional.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Fixing up shutdown GPIOs via overrides is ugly, and doesn't work
on eg CM4 where both cameras share the same shutdown GPIO.
The driver is now updated to use the regulator framework, so switch
to using that instead.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The driver supported using GPIOs to control the shutdown line,
but no regulator control.
Add regulator hooks.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Now that we have regulators and clocks defined in the base DT for
image sensors, switch the overlays to use them.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Unloading regulators through dynamic device tree doesn't work
as the regulators will unregister whilst clients are still
registered. Whilst the regulator framework does WARN when that
happens, the client putting the regulator then typically results
in a NULL dereference and badness.
Instead of creating regulators and clocks from the overlays,
create regulators and clocks for the sensors in the base DT.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The VL805 fetches up to 4 transfer TRBs at a time. TRB reads don't cross
a 64B boundary, and if a TRB is fetched and is not on a 64B boundary,
the read is sized up to the next 64B boundary.
However the VL805 implements a readahead prefetch for TRBs on a transfer
ring. This fetches the next 64B after any TRB read has happened. Near
the end of a ring segment, the prefetcher can read the first 64B of the
next page in physical memory and this is where the behaviour causes a
bug.
The controller does not tag reads with which endpoint they are for, so
if the start of the next page is a ring segment used by a victim
endpoint, and the victim endpoint is about to fetch TRBs from the start
of the segment, the victim endpoint will read from the prefetched data
and not perform a read to main memory. If the data is stale, the ring
cycle state bit may not be correct and the endpoint will silently halt.
Adjust trbs_per_seg for transfer rings allocated for this controller.
See https://github.com/raspberrypi/linux/issues/4685
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
In anticipation of adjusting the number of utilised TRBs in a ring
segment, add trbs_per_seg to struct xhci_ring and use this instead
of a compile-time define.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
Certain transfer ring access patterns can cause the controller to hang
fetching TRBs for a USB2.0 endpoint.
- If two USB2.0 endpoints are active at once and
- Both endpoints are traversing a Link TRB where the following segment
has a lower page address and
- One of the endpoints is a Bulk IN and
- The other endpoint is an Interrupt IN
Then the Interrupt IN endpoint can end up not getting polled.
It is unclear what the precise failure mode is, as the controller seems to
haphazardly and repeatedly fetch TRBs for both endpoints but does not
advance the Interrupt endpoint transfer.
As a workaround, add a quirk that initially constrains all USB2.0 transfer
rings to a single segment in size. If for any reason a device driver queues
up enough outstanding transfers to fill the ring segment, then the ring
will be expanded. This has not been seen to occur with UMS or UVC drivers,
which aggressively queue buffers.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
Don't calculate space based on the number of TRBs in the current segment,
as it's OK to wrap to the start (and flip the cycle state bit).
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
In a couple of error/incomplete configuration cases, the
DPI_FORMAT bits wouldn't get set.
Adopt a default of RGB666 in all these cases.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
DRM provides flags for inverting pixel clock and output enable
signals, but these were not mapped to the relevant registers.
Add those mappings.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The "panel-dpi" compatible string configures panel from device tree,
but it doesn't provide any way of configuring the bus format (colour
representation), nor does it populate it.
Add a DT parameter "bus-format" that allows the MEDIA_BUS_FMT_xxx value
to be specified from device tree.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
When using the "panel-dpi" compatible string, desc as passed
from DT is a dumy entry, and panel_dpi_probe allocates a new
one and attaches it to the panel.
However panel_simple_probe has already taken a local copy of
the variable, which means all the validation is done against
the empty dummy structure, not the configured data.
Update the local variable after panel_dpi_probe.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
panel-dpi doesn't know the bit depth, so in the same way that
DPI is guessed for the connector type, guess that it'll be 8bpc.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The clk-bcm2835 handling of the pixel clock does not function
correctly when the HDMI power domain is disabled.
The firmware supports it correctly, and the firmware clock
driver now supports it, so switch the vc4-hdmi driver to use
the firmware clock driver.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The clk-bcm2835 handling of the pixel clock does not function
correctly when the HDMI power domain is disabled.
The firmware supports it correctly, so add it to the
firmware clock driver.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Pi0-3 have power domains attached to the pm_runtime hooks
for the HDMI block. Initialisation done in the reset called
from bind is therefore lost if all users of the domain are
suspended.
The VEC shares the same lowest level clock/power gating as
the HDMI block, so whilst that is enabled the block is never
actually powered down, but if it isn't enabled then we lose
the state.
Reset and initialise the HDMI block from pm_resume.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The commit that changed from dev_err to dev_err_probe left the %d in the
format string, but removed the parameter, leading to a compile
warning.
Fixes: "6505412df625 drm/vc4: Use dev_err_probe when logging error registering HDMI audio"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Register 0x02 in the FT5x06 is TD_STATUS containing the number
of valid touch points being reported.
Iterate over that number of points rather than all that are
supported on the device.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
This reverts commit 0c41710df5.
The regulator driver is now hopefully fixed, therefore revert the
workaround that dropped the I2C frequency.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The I2C to the Atmel is very fussy, and locks up easily on
Pi0-3 particularly on reads.
If running at 100kHz on Pi3, reading the ID register generally
locks up the Atmel, but splitting the register select write and
read into two transactions is reliable.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The I2C to the Atmel is very fussy, and locks up easily on
Pi0-3 particularly on reads.
The LCD power status is controlled solely by this driver, so
rather than reading it back from the Atmel, use the cached
status last set.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If a dummy buffer is still active on a frame start, it indicates that this frame
will be dropped. The explicit logging helps users identify performance issues.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Deinterlace and decode aren't affected in the same way as encode
and ISP by the alignment requirement on 3 plane YUV420.
Decode would be affected, but it always aligns the height up to
a macroblock, and uses the selection API to reflect that.
Add in the facility to set the bytesperline alignment per role.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Remove min_buffers_needed=1 from src queue init. Src buffers are bound
to media requests therefore this setting is not needed and generates
a WARN in kernel 5.16.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Matching https://github.com/raspberrypi/linux/pull/4419, the ISP
block (which is also used on the input of the encoder, and output
of the decoder) needs the base address of all planes to be aligned
to multiples of 32. This includes the chroma planes of YUV420 and
YVU420.
If the height is only a multiple of 2 (not 4), then you get an odd
number of lines in the second plane, which means the 3rd plane
starts at a multiple of bytesperline/2.
Set the minimum bytesperline alignment to 64 for those formats
so that the plane alignment is always right.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Avoid logging a spurious error message with error -517
(-EPROBE_DEFER) whilst trying to load HDMI audio.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Whilst investigations continue as to why the 7" DSI screen doesn't
work on Pi3 with KMS, drop the I2C baudrate to 50kHz as that seems
to be a reliable workaround.
This will be reverted once the full issue is understood.
https://github.com/raspberrypi/linux/issues/4686
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Whilst all the datasheets describe FT5x06 as supporting "up to
10 points of absolution X and Y coordinates", the driver
implementation for the compatible string "edt,edt-ft5x06" only
allows for 5.
Switch to the "edt,edt-ft5506" compatible string which allows for
10 points with no other differences.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The ft5x06 is unreliable in sending touch up events, so some
touch IDs can become stuck in the detected state.
Ensure that IDs that are unreported by the controller are
released.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The Pi 4B dts file has had numerous updates since the Pi 400 and CM4
dts files were written. Apply those updates to the other files to
minimise the differences. The change is largely cosmetic, except for
the PCI "device-type" to "device_type" rename, and the correction of
the labels on the Pi 400 GPIO expander pins.
Also set the wifi-2.4ghz-coexistence property on RPi 4 B, which appears
to have been missed.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The existing logic was flawed in that it could try reading the
2711 specific registers for HPD on a CM1/3 where the HPD GPIO
hadn't been defined in DT.
Ensure we don't do the 2711 register read on invalid hardware,
and then
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The hardware can combine H&V syncs onto the output enable line
as composite syncs, so add the relevant configuration to do that.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add these missing V4L2 controls. Tested binned and full resolution
modes in all four orientations using Raspberry Pi running libcamera.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
commit 98481f3d72 upstream.
The PCIe host bridge has two interrupt lines, one that goes towards it
PCIE_INTR2 second level interrupt controller and one for its MSI second
level interrupt controller. The first interrupt line is not currently
managed by the driver, which is why it was not a functional problem.
The interrupt-map property was also only listing the PCI_INTA interrupts
when there are also the INTB, C and D.
Reported-by: Jim Quinlan <jim2101024@gmail.com>
Fixes: d5c8dc0d4c ("ARM: dts: bcm2711: Enable PCIe controller")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Add a second (identical) set of device nodes to allow concurrent use of the ISP
hardware by another user. This change effectively creates a second state
structure (struct bcm2835_isp_dev) to maintain independent state for the second
user. Node and media entity names are appened with the instance index
appropriately.
Further users can be added by changing the BCM2835_ISP_NUM_INSTANCES define.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
This reverts commit 063a326cba.
Reverting temporarily because not enabling the composite output has
been seen to break HDMI on older Pis.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
It referenced fragment@13 which used to be part of edt-ft5406.dtsi,
but has now been removed.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The code that set the scdc_enabled flag to ensure it was
disabled at boot time also ran on Pi0-3 where there is no
SCDC support. This lead to a warning in vc4_hdmi_encoder_post_crtc_disable
due to vc4_hdmi_disable_scrambling being called and trying to
read (and write) register HDMI_SCRAMBLER_CTL which doesn't
exist on those platforms.
Only set the flag should the interface be configured to support
more than HDMI 1.4.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add a check to vc4_hvs_gamma_check to ensure a new non-empty
gamma LUT is of the correct length before accepting it.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
edt-ft5406.dtsi is included from vc4-kms-dsi-7inch which was
also setting i2c0mux and i2c0if status fields. This meant that
dtoverlay wouldn't apply the overlay due to multiple fragments
changing the same parameter.
Move the enable from edt-ft5406.dtsi to edt-ft5406-overlay.dts
for when it should be needed as an independent overlay.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Two calls were made to drm_crtc_enable_color_mgmt to add gamma
and CTM, however they were both set to add the gamma properties,
so they ended up added twice.
Fixes: 766cc6b1f7 "drm/vc4: Add CTM support"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
With HVS5 the gamma block is now only reprogrammed with
a disable/enable. Loading the table from vc4_hvs_init_channel
(called from vc4_hvs_atomic_enable) appears to be at an
invalid point in time and so isn't applied.
Switch to enabling and disabling the gamma table instead. This
isn't safe if the pipeline is running, but it isn't now.
For HVS4 it is safe to enable and disable dynamically, so
adopt that approach there too.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The code falls through to "fail" under all conditions, so there is no
need for the drm_property_blob_put if the gamma lut hasn't been changed.
Fixes: 9cca26674a "drm: Check whether the gamma lut has changed before updating"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
On a Pi 4, enabling composite video disables the HDMI output. As a
consequence, the composite output is disabled by default. Change the
vc4-kms-v3d overlay used on older Pis to also disable composite by
default, replacing the "nocomposite" parameter with a "composite"
parameter.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
We currently rely on two functions, vc4_hdmi_supports_scrambling() and
vc4_hdmi_mode_needs_scrambling() to determine if we should enable and
disable the scrambler for any given mode.
Since we might need to disable the controller at boot, we also always
run vc4_hdmi_disable_scrambling() and thus call those functions without
a mode yet, which in turns need to make some special casing in order for
it to work.
Instead of duplicating the check for whether or not we need to take care
of the scrambler in both vc4_hdmi_enable_scrambling() and
vc4_hdmi_disable_scrambling(), we can do that check only when we enable
it and store whether or not it's been enabled in our private structure.
We also need to initialize that flag at true to make sure we disable the
scrambler at boot since we can't really know its state yet.
This allows to simplify a bit that part of the driver, and removes one
user of our copy of the CRTC adjusted mode outside of KMS (since
vc4_hdmi_disable_scrambling() might be called from the hotplug interrupt
handler).
It also removes our last user of the legacy encoder->crtc pointer.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
We currently poke at encoder->crtc in the ALSA code path to determine
whether the HDMI output is enabled or not, and thus whether we should
allow the audio output.
However, that pointer is deprecated and shouldn't really be used by
atomic drivers anymore. Since we have the infrastructure in place now,
let's just create a flag that we toggle to report whether the controller
is currently enabled and use that instead of encoder->crtc in ALSA.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Even though we already check that the encoder->crtc pointer is there
during in startup(), which is part of the open() path in ASoC, nothing
guarantees that our encoder state won't change between the time when we
open the device and the time we prepare it.
Move the sanity checks we do in startup() to a helper and call it from
prepare().
Fixes: 91e99e1139 ("drm/vc4: hdmi: Register HDMI codec")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Accessing the crtc->state pointer from outside the modesetting context
is not allowed. We thus need to copy whatever we need from the KMS state
to our structure in order to access it.
However, in the vc4 HDMI driver we do use that pointer in the ALSA code
path, and potentially in the hotplug interrupt handler path.
These paths both need access to the CRTC adjusted mode in order for the
proper dividers to be set for ALSA, and the scrambler state to be
reinstated properly for hotplug.
Let's copy this mode into our private encoder structure and reference it
from there when needed. Since that part is shared between KMS and other
paths, we need to protect it using our mutex.
Link: https://lore.kernel.org/all/YWgteNaNeaS9uWDe@phenom.ffwll.local/
Fixes: bb7d785688 ("drm/vc4: Add HDMI audio support")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The vc4 HDMI controller registers into the KMS, CEC and ALSA
frameworks.
However, no particular care is done to prevent the concurrent execution
of different framework hooks from happening at the same time.
In order to protect against that scenario, let's introduce a mutex that
relevant ALSA and KMS hooks will need to take to prevent concurrent
execution.
CEC is left out at the moment though, since the .get_modes and .detect
KMS hooks, when running cec_s_phys_addr_from_edid, can end up calling
CEC's .adap_enable hook. This introduces some reentrancy that isn't easy
to deal with properly.
The CEC hooks also don't share much state with the rest of the driver:
the registers are entirely separate, we don't share any variable, the
only thing that can conflict is the CEC clock divider setup that can be
affected by a mode set.
However, after discussing it, it looks like CEC should be able to
recover from this if it was to happen.
Fixes: bb7d785688 ("drm/vc4: Add HDMI audio support")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The vc4 HDMI driver has multiple path shared between the CEC, ALSA and
KMS frameworks, plus two interrupt handlers (CEC and hotplug) that will
read and modify a number of registers.
Even though not bug has been reported so far, it's definitely unsafe, so
let's just add a spinlock to protect the register access of the HDMI
controller.
Fixes: c8b75bca92 ("drm/vc4: Add KMS support for Raspberry Pi.")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Accessing the crtc->state pointer from outside the modesetting context
is not allowed. We thus need to copy whatever we need from the KMS state
to our structure in order to access it.
In VC4, a number of users of that pointers have crept in over the years,
and the previous commits removed them all but the HVS channel a CRTC has
been assigned.
Let's move this channel in struct vc4_crtc at atomic_begin() time, drop
it from our private state structure, and remove our use of crtc->state
from our vblank handler entirely.
Link: https://lore.kernel.org/all/YWgteNaNeaS9uWDe@phenom.ffwll.local/
Fixes: 87ebcd42fb ("drm/vc4: crtc: Assign output to channel automatically")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In some situation, we can end up being stuck on a non-blocking that went
through properly.
The situation that seems to trigger it reliably is to first start a
non-blocking commit, and then right after, and before we had any vblank
interrupt), start a blocking commit.
This will lead to the first commit workqueue to be scheduled, setup the
display, while the second commit is waiting for the first one to be
completed.
The vblank interrupt will then be raised, vc4_crtc_handle_vblank() will
run and will compare the active dlist in the HVS channel to the one
associated with the crtc->state.
However, at that point, the second commit is waiting using
drm_atomic_helper_wait_for_dependencies that occurs after
drm_atomic_helper_swap_state has been called, so crtc->state points to
the second commit state. vc4_crtc_handle_vblank() will compare the two
dlist addresses and since they don't match will ignore the interrupt.
The vblank event will never be reported, and the first and second commit
will wait for the first commit completion until they timeout.
The underlying reason is that it was never safe to do so. Indeed,
accessing the ->state pointer access synchronization is based on
ownership guarantees that can only occur within the functions and hooks
defined as part of the KMS framework, and obviously the irq handler
isn't one of them. The rework to move to generic helpers only uncovered
the underlying issue.
However, since the code path between
drm_atomic_helper_wait_for_dependencies() and
drm_atomic_helper_wait_for_vblanks() is serialised and we can't get two
commits in that path at the same time, we can work around this issue by
setting a variable associated to struct drm_crtc to the dlist we expect,
and then using it from the vc4_crtc_handle_vblank() function.
Since that state is shared with the modesetting path, we also need to
introduce a spinlock to protect the code shared between the interrupt
handler and the modesetting path, protecting only our new variable for
now.
Link: https://lore.kernel.org/all/YWgteNaNeaS9uWDe@phenom.ffwll.local/
Fixes: 56d1fe0979 ("drm/vc4: Make pageflip completion handling more robust.")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Accessing the crtc->state pointer from outside the modesetting context
is not allowed. We thus need to copy whatever we need from the KMS state
to our structure in order to access it.
In VC4, a number of users of that pointers have crept in over the years,
the first one being whether or not the downstream controller of the
pixelvalve is our writeback controller.
Fortunately for us, Since commit 39fcb28083 ("drm/vc4: txp: Turn the
TXP into a CRTC of its own") this is no longer something that can change
from one commit to the other and is hardcoded.
Let's set this flag in struct vc4_crtc if we happen to be the TXP, and
drop the flag from our private state structure.
Link: https://lore.kernel.org/all/YWgteNaNeaS9uWDe@phenom.ffwll.local/
Fixes: 008095e065 ("drm/vc4: Add support for the transposer block")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
It is important to reinitialise the firmware array pointers to protect
against the case that the brcmfmac driver is reprobed without first
being unloaded.
The potential hazard was noticed while investigating
https://github.com/raspberrypi/firmware/issues/1644 .
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
drm_crtc_legacy_gamma_set updates the gamma_lut blob unconditionally,
which leads to unnecessary reprogramming of hardware.
Check whether the blob contents has actually changed before
signalling that it has been updated.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add the ability to load the names of alternative firmwares from the
Device Tree node. This permits separate firmwares for 43436s and 43438
and allows downstream firmwares to coexist with upstream.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This reverts commit c52581ffa4.
Replace the hardcoded alternate firmware names with mappings provided
from Device Tree.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Add an option to enable configuration via the Media Controller API
(rather than the video-node-centric /dev/videoN) as about to
be used by libcamera as it enables more complex pipelines to be
handled.
Any source that has a libcamera tuning merged has MC enabled by
default.
Sources with no libcamera tuning merged have it disabled by
default.
In either case it can be overridden with the overlay parameter
"media-controller".
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
BCM43430/2 may be BCM43430B0 or BCM43436P, and BCM43430/1 can be either
BCM43430A1 or BCM43436S, the former being upstream names and the
latter downstream names for differently-sourced sister parts.
Make the choice of firmwares board-specific (without making the actual
firmwares board-specific) by placing the alternative firmware names for
each part in the DT node.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Add the ability to load the names of alternative firmwares from the
Device Tree node. This permits separate firmwares for 43436s and 43438
and allows downstream firmwares to coexist with upstream.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This reverts commit c52581ffa4.
Replace the hardcoded alternate firmware names with mappings provided
from Device Tree.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
An unwanted side effect of enabling the BRCMDBG config setting is
redefining brcmf_info to be brcmf_err. This can be alarming to users
and makes it harder to spot real errors, so don't do it.
See: https://github.com/raspberrypi/linux/issues/4663
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Retain the old names for backwards compatibility for a while, while the
necessary firmware change rolls out.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Add predefined modelines for the 240p (NTSC) and 288p (PAL) progressive
modes, and report them through vc4_vec_connector_get_modes().
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Make vc4_vec_encoder_atomic_check() accept arbitrary modelines, as long
as they result in somewhat sane output from the VEC. The bounds have
been determined empirically. Additionally, add support for the
progressive 262-line and 312-line modes.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
trigger_mode == 0 (default) => no effect / no registers written
trigger_mode == 1 => source
trigger_mode == 2 => sink
This can be set e.g. in /boot/cmdline.txt as imx477.trigger_mode=N
Signed-off-by: Jonas Jacob <jonas.jacob@neocortexvision.com>
commit aaac373317 upstream.
Replace the open coded calculations of the actual physical address
of the KVM stub vector table with a single adr_l invocation.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 3bcf906b19 upstream.
Replace the open coded arithmetic with a simple adr_l/sub pair. This
removes some open coded arithmetic involving virtual addresses, avoids
literal pools on v7+, and slightly reduces the footprint of the code.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit d74d2b2250 upstream.
Replace the open coded PC relative offset calculations with adr_l and
ldr_l invocations. This removes some open coded PC relative arithmetic,
avoids literal pools on v7+, and slightly reduces the footprint of the
code. Note that ALT_SMP() expects a single instruction so move the macro
invocation after it.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 59d2f2827d upstream.
Now that calling __do_fixup_smp_on_up() can be done without passing
the physical-to-virtual offset in r3, we can replace the open coded
PC relative offset calculations with a pair of adr_l invocations. This
removes some open coded arithmetic involving virtual addresses, avoids
literal pools on v7+, and slightly reduces the footprint of the code.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 450abd38fe upstream.
Currently, the .alt.smp.init section contains the virtual addresses
of the patch sites. Since patching may occur both before and after
switching into virtual mode, this requires some manual handling of
the address when applying the UP alternative.
Let's simplify this by using relative offsets in the table entries:
this allows us to simply add each entry's address to its contents,
regardless of whether we are running in virtual mode or not.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 91580f0dbf upstream.
Replace the open coded PC relative offset calculations with adr_l
and ldr_l invocations. This removes some open coded arithmetic
involving virtual addresses, avoids literal pools on v7+, and slightly
reduces the footprint of the code.
Note that it also removes a stale comment about the contents of r6.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 172c34c9ff upstream.
Replace the open coded PC relative offset calculations involving
__turn_mmu_on and __turn_mmu_on_end with a pair of adr_l invocations.
This removes some open coded arithmetic involving virtual addresses,
avoids literal pools on v7+, and slightly reduces the footprint of the
code.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 62c4a2e202 upstream.
Replace the open coded PC relative offset calculations with a pair of
adr_l invocations. This removes some open coded arithmetic involving
virtual addresses, avoids literal pools on v7+, and slightly reduces
the footprint of the code.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 9443076e43 upstream.
The ARM kernel's linear map starts at PAGE_OFFSET, which maps to a
physical address (PHYS_OFFSET) that is platform specific, and is
discovered at boot. Since we don't want to slow down translations
between physical and virtual addresses by keeping the offset in a
variable in memory, we implement this by patching the code performing
the translation, and putting the offset between PAGE_OFFSET and the
start of physical RAM directly into the instruction opcodes.
As we only patch up to 8 bits of offset, yielding 4 GiB >> 8 == 16 MiB
of granularity, we have to round up PHYS_OFFSET to the next multiple if
the start of physical RAM is not a multiple of 16 MiB. This wastes some
physical RAM, since the memory that was skipped will now live below
PAGE_OFFSET, making it inaccessible to the kernel.
We can improve this by changing the patchable sequences and the patching
logic to carry more bits of offset: 11 bits gives us 4 GiB >> 11 == 2 MiB
of granularity, and so we will never waste more than that amount by
rounding up the physical start of DRAM to the next multiple of 2 MiB.
(Note that 2 MiB granularity guarantees that the linear mapping can be
created efficiently, whereas less than 2 MiB may result in the linear
mapping needing another level of page tables)
This helps Zhen Lei's scenario, where the start of DRAM is known to be
occupied. It also helps EFI boot, which relies on the firmware's page
allocator to allocate space for the decompressed kernel as low as
possible. And if the KASLR patches ever land for 32-bit, it will give
us 3 more bits of randomization of the placement of the kernel inside
the linear region.
For the ARM code path, it simply comes down to using two add/sub
instructions instead of one for the carryless version, and patching
each of them with the correct immediate depending on the rotation
field. For the LPAE calculation, which has to deal with a carry, it
patches the MOVW instruction with up to 12 bits of offset (but we only
need 11 bits anyway)
For the Thumb2 code path, patching more than 11 bits of displacement
would be somewhat cumbersome, but the 11 bits we need fit nicely into
the second word of the u16[2] opcode, so we simply update the immediate
assignment and the left shift to create an addend of the right magnitude.
Suggested-by: Zhen Lei <thunder.leizhen@huawei.com>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit e8e00f5afb upstream.
In preparation for reducing the phys-to-virt minimum relative alignment
from 16 MiB to 2 MiB, switch to patchable sequences involving MOVW
instructions that can more easily be manipulated to carry a 12-bit
immediate. Note that the non-LPAE ARM sequence is not updated: MOVW
may not be supported on non-LPAE platforms, and the sequence itself
can be updated more easily to apply the 12 bits of displacement.
For Thumb2, which has many more versions of opcodes, switch to a sequence
that can be patched by the same patching code for both versions. Note
that the Thumb2 opcodes for MOVW and MVN are unambiguous, and have no
rotation bits in their immediate fields, so there is no need to use
placeholder constants in the asm blocks.
While at it, drop the 'volatile' qualifiers from the asm blocks: the
code does not have any side effects that are invisible to the compiler,
so it is free to omit these sequences if the outputs are not used.
Suggested-by: Russell King <linux@armlinux.org.uk>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 0e3db6c9d7 upstream.
Declutter the code in __fixup_pv_table() by using the new adr_l/str_l
macros to take PC relative references to external symbols, and by
using the value of PHYS_OFFSET passed in r8 to calculate the p2v
offset.
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 2730e8eaa4 upstream.
Free up a register in the p2v patching code by switching to relative
references, which don't require keeping the phys-to-virt displacement
live in a register.
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 7a94849e81 upstream.
The big and little endian versions of the ARM p2v patching routine only
differ in the values of the constants, so factor those out into macros
so that we only have one version of the logic sequence to maintain.
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 4b16421c3e upstream.
The ARM and Thumb2 versions of the p2v patching loop have some overlap
at the end of the loop, so factor that out. As numeric labels are not
required to be unique, and may therefore be ambiguous, use named local
labels for the start and end of the loop instead.
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit eae78e1a97 upstream.
Move the phys2virt patching code into a separate .S file before doing
some work on it.
Suggested-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
commit 22f2d23098 upstream.
When using the new adr_l/ldr_l/str_l macros to refer to external symbols
from modules, the linker may emit place relative ELF relocations that
need to be fixed up by the module loader. So add support for these.
Reviewed-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The HVS Gamma block can only be updated when idle, so we need to disable
the HVS channel when the gamma property is set in an atomic commit.
Since the pixelvalve cannot have its assigned channel halted without
stalling unless it's disabled as well, in our case that means forcing a
full disable / enable cycle on the pipeline.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
BCM2711 changes from a 256 entry lookup table to a 16 point
piecewise linear function as the pipeline bitdepth has increased
to make a LUT unwieldy.
Implement a simple conversion from a 256 entry LUT that userspace
is likely to expect to 16 evenly spread points in the PWL. This
could be improved with curve fitting at a later date.
Co-developed-by: Juerg Haefliger <juergh@canonical.com>
Signed-off-by: Juerg Haefliger <juergh@canonical.com>
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
clk-2835 is deprecated and gets an innacurate clock for VEC (107MHz).
Switch to clk-raspberrypi which uses the right PLL to get an accurate 108MHz.
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Both the firmware audio driver and the vc4-kms-v3d driver are capable
of providing HDMI audio, but only one should be active at any time.
The vc4-kms-v3d overlays disable the firmware audio driver, but they
also have a noaudio parameter that as well as disabling the ARM-side
HDMI audio also re-enables the firmware HDMI audio. This is not
guaranteed to work and has been seen to break the display completely.
Modify the noaudio parameters so that the firmware HDMI audio support
remains disabled.
See: https://github.com/raspberrypi/linux/issues/4651
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Limit the number of allowed video callbacks. This helps with limiting
the size of the coded input FIFO which in turn helps to control latency.
Choose -5 as the magic number as it translates to DPB+5 buffers which
has been proven to be a good number in the past.
Ideally coded buffers would not be returned to the user until they
had been decoded into the DPB or been discarded as bad, but that grade
of control is unavailable to us.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Commit db8e94e7cf upstream
The existing link format validation failure debug message in media-entity.c
helped to pinpoint the point of failure but provided no additional
information what's wrong. Tell the user exactly why the validation failed.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Adds Media Controller API support for more complex pipelines.
libcamera is about to switch to using this mechanism for configuring
sensors.
This can be enabled by either a module parameter, or device tree.
Various functions have been moved to group video-centric and
mc-centric functions together.
Based on a similar conversion done to ti-vpe.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The driver was making big assumptions about the source device
using pad 0 and 1, which doesn't follow for more complex
devices where Unicam's source device may be a sink device for
something else.
Read the pad numbers through media controller, and reference
them appropriately.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
RAW color spaces are more usually reported as having full range
quantization.
Tested using libcamera.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
The get() method does not understand the on-the-wire encoding of the
remote GPIO states, thinking they are simple on/off bits when they are
really pairs of 16-bit counts. Rewrite the get() handler to return the
value last written, which will eventually match the actual GPIO state
if there are no other changes.
See: https://github.com/raspberrypi/linux/issues/4638
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
[ Upstream commit ebc69e897e ]
This reverts commit 2d52c58b9c.
We have had several folks complain that this causes hangs for them, which
is especially problematic as the commit has also hit stable already.
As no resolution seems to be forthcoming right now, revert the patch.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=214503
Fixes: 2d52c58b9c ("block, bfq: honor already-setup queue merges")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
In order to resolve a potential startup order bug, the vcio driver has
been rewritten as a platform driver that depends on a DT node for
its instantiation and to locate the firmware driver.
Add that DT node.
See: https://github.com/raspberrypi/linux/issues/4620
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The old vcio driver is a simple character device that manually locates
the firmware driver. Initialising it before the firmware driver causes
a failure, and no retries are attempted.
Rewrite vcio as a platform driver that depends on a DT node for its
instantiation and the location of the firmware driver, making use of
the miscdevice framework to reduce the code size.
N.B. Using miscdevice changes the udev SUBSYSTEM string, so a change
to the companion udev rule is required in order to continue to set
the correct device permissions, e.g.:
KERNEL="vcio", GROUP="video", MODE="0660"
See: https://github.com/raspberrypi/linux/issues/4620
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Power-on reset after the insertion of a battery does not always complete
successfully, leading to corrupted register content. The EXT_TEST bit
will stop the clock from running, but currently the driver will never
recover.
Safely handle the erroneous state by clearing EXT_TEST as part of the
usual set_time method.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Using the rpi_firmware API as intended allows proper reference counting
of the firmware device and means we can remove a downstream patch to
the firmware driver.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Using the rpi_firmware API as intended allows proper reference counting
of the firmware device and means we can remove a downstream patch to
the firmware driver.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
For cases where spare PWM outputs are available, but are desired
to be addressed a standard outputs instead.
Wraps a PWM channel as a new GPIO chip with the one output.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
commit e8f71f8923 upstream.
nvkm test builds fail with the following error.
drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c: In function 'nvkm_control_mthd_pstate_info':
drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c:60:35: error: overflow in conversion from 'int' to '__s8' {aka 'signed char'} changes value from '-251' to '5'
The code builds on most architectures, but fails on parisc where ENOSYS
is defined as 251.
Replace the error code with -ENODEV (-19). The actual error code does
not really matter and is not passed to userspace - it just has to be
negative.
Fixes: 7238eca4cf ("drm/nouveau: expose pstate selection per-power source in sysfs")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 9848417926 ]
The intel powerclamp driver will setup a per-CPU worker with RT
priority. The worker will then invoke play_idle() in which it remains in
the idle poll loop until it is stopped by the timer it started earlier.
That timer needs to expire in hard interrupt context on PREEMPT_RT.
Otherwise the timer will expire in ksoftirqd as a SOFT timer but that task
won't be scheduled on the CPU because its priority is lower than the
priority of the worker which is in the idle loop.
Always expire the idle timer in hard interrupt context.
Reported-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210906113034.jgfxrjdvxnjqgtmc@linutronix.de
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 884f0e84f1 ]
The pending timer has been set up in blk_throtl_init(). However, the
timer is not deleted in blk_throtl_exit(). This means that the timer
handler may still be running after freeing the timer, which would
result in a use-after-free.
Fix by calling del_timer_sync() to delete the timer in blk_throtl_exit().
Signed-off-by: Li Jinlin <lijinlin3@huawei.com>
Link: https://lore.kernel.org/r/20210907121242.2885564-1-lijinlin3@huawei.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d44084c934 ]
A consumer is expected to disable a PWM before calling pwm_put(). And if
they didn't there is hopefully a good reason (or the consumer needs
fixing). Also if disabling an enabled PWM was the right thing to do,
this should better be done in the framework instead of in each low level
driver.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 9d768cd7fd ]
A consumer is expected to disable a PWM before calling pwm_put(). And if
they didn't there is hopefully a good reason (or the consumer needs
fixing). Also if disabling an enabled PWM was the right thing to do,
this should better be done in the framework instead of in each low level
driver.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c68eb29c8e ]
A consumer is expected to disable a PWM before calling pwm_put(). And if
they didn't there is hopefully a good reason (or the consumer needs
fixing). Also if disabling an enabled PWM was the right thing to do,
this should better be done in the framework instead of in each low level
driver.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a6c849012b ]
Currently there is no validity check for event ID received from F/W,
Thus exposing driver to memory overrun.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5f5dec07ac ]
Patch series "nilfs2: fix incorrect usage of kobject".
This patchset from Nanyong Sun fixes memory leak issues and a NULL
pointer dereference issue caused by incorrect usage of kboject in nilfs2
sysfs implementation.
This patch (of 6):
Reported by syzkaller:
BUG: memory leak
unreferenced object 0xffff888100ca8988 (size 8):
comm "syz-executor.1", pid 1930, jiffies 4294745569 (age 18.052s)
hex dump (first 8 bytes):
6c 6f 6f 70 31 00 ff ff loop1...
backtrace:
kstrdup+0x36/0x70 mm/util.c:60
kstrdup_const+0x35/0x60 mm/util.c:83
kvasprintf_const+0xf1/0x180 lib/kasprintf.c:48
kobject_set_name_vargs+0x56/0x150 lib/kobject.c:289
kobject_add_varg lib/kobject.c:384 [inline]
kobject_init_and_add+0xc9/0x150 lib/kobject.c:473
nilfs_sysfs_create_device_group+0x150/0x7d0 fs/nilfs2/sysfs.c:986
init_nilfs+0xa21/0xea0 fs/nilfs2/the_nilfs.c:637
nilfs_fill_super fs/nilfs2/super.c:1046 [inline]
nilfs_mount+0x7b4/0xe80 fs/nilfs2/super.c:1316
legacy_get_tree+0x105/0x210 fs/fs_context.c:592
vfs_get_tree+0x8e/0x2d0 fs/super.c:1498
do_new_mount fs/namespace.c:2905 [inline]
path_mount+0xf9b/0x1990 fs/namespace.c:3235
do_mount+0xea/0x100 fs/namespace.c:3248
__do_sys_mount fs/namespace.c:3456 [inline]
__se_sys_mount fs/namespace.c:3433 [inline]
__x64_sys_mount+0x14b/0x1f0 fs/namespace.c:3433
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x3b/0x90 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
If kobject_init_and_add return with error, then the cleanup of kobject
is needed because memory may be allocated in kobject_init_and_add
without freeing.
And the place of cleanup_dev_kobject should use kobject_put to free the
memory associated with the kobject. As the section "Kobject removal" of
"Documentation/core-api/kobject.rst" says, kobject_del() just makes the
kobject "invisible", but it is not cleaned up. And no more cleanup will
do after cleanup_dev_kobject, so kobject_put is needed here.
Link: https://lkml.kernel.org/r/1625651306-10829-1-git-send-email-konishi.ryusuke@gmail.com
Link: https://lkml.kernel.org/r/1625651306-10829-2-git-send-email-konishi.ryusuke@gmail.com
Reported-by: Hulk Robot <hulkci@huawei.com>
Link: https://lkml.kernel.org/r/20210629022556.3985106-2-sunnanyong@huawei.com
Signed-off-by: Nanyong Sun <sunnanyong@huawei.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8f96a5bfa1 ]
We update the ctime/mtime of a block device when we remove it so that
blkid knows the device changed. However we do this by re-opening the
block device and calling filp_update_time. This is more correct because
it'll call the inode->i_op->update_time if it exists, but the block dev
inodes do not do this. Instead call generic_update_time() on the
bd_inode in order to avoid the blkdev_open path and get rid of the
following lockdep splat:
======================================================
WARNING: possible circular locking dependency detected
5.14.0-rc2+ #406 Not tainted
------------------------------------------------------
losetup/11596 is trying to acquire lock:
ffff939640d2f538 ((wq_completion)loop0){+.+.}-{0:0}, at: flush_workqueue+0x67/0x5e0
but task is already holding lock:
ffff939655510c68 (&lo->lo_mutex){+.+.}-{3:3}, at: __loop_clr_fd+0x41/0x660 [loop]
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #4 (&lo->lo_mutex){+.+.}-{3:3}:
__mutex_lock+0x7d/0x750
lo_open+0x28/0x60 [loop]
blkdev_get_whole+0x25/0xf0
blkdev_get_by_dev.part.0+0x168/0x3c0
blkdev_open+0xd2/0xe0
do_dentry_open+0x161/0x390
path_openat+0x3cc/0xa20
do_filp_open+0x96/0x120
do_sys_openat2+0x7b/0x130
__x64_sys_openat+0x46/0x70
do_syscall_64+0x38/0x90
entry_SYSCALL_64_after_hwframe+0x44/0xae
-> #3 (&disk->open_mutex){+.+.}-{3:3}:
__mutex_lock+0x7d/0x750
blkdev_get_by_dev.part.0+0x56/0x3c0
blkdev_open+0xd2/0xe0
do_dentry_open+0x161/0x390
path_openat+0x3cc/0xa20
do_filp_open+0x96/0x120
file_open_name+0xc7/0x170
filp_open+0x2c/0x50
btrfs_scratch_superblocks.part.0+0x10f/0x170
btrfs_rm_device.cold+0xe8/0xed
btrfs_ioctl+0x2a31/0x2e70
__x64_sys_ioctl+0x80/0xb0
do_syscall_64+0x38/0x90
entry_SYSCALL_64_after_hwframe+0x44/0xae
-> #2 (sb_writers#12){.+.+}-{0:0}:
lo_write_bvec+0xc2/0x240 [loop]
loop_process_work+0x238/0xd00 [loop]
process_one_work+0x26b/0x560
worker_thread+0x55/0x3c0
kthread+0x140/0x160
ret_from_fork+0x1f/0x30
-> #1 ((work_completion)(&lo->rootcg_work)){+.+.}-{0:0}:
process_one_work+0x245/0x560
worker_thread+0x55/0x3c0
kthread+0x140/0x160
ret_from_fork+0x1f/0x30
-> #0 ((wq_completion)loop0){+.+.}-{0:0}:
__lock_acquire+0x10ea/0x1d90
lock_acquire+0xb5/0x2b0
flush_workqueue+0x91/0x5e0
drain_workqueue+0xa0/0x110
destroy_workqueue+0x36/0x250
__loop_clr_fd+0x9a/0x660 [loop]
block_ioctl+0x3f/0x50
__x64_sys_ioctl+0x80/0xb0
do_syscall_64+0x38/0x90
entry_SYSCALL_64_after_hwframe+0x44/0xae
other info that might help us debug this:
Chain exists of:
(wq_completion)loop0 --> &disk->open_mutex --> &lo->lo_mutex
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&lo->lo_mutex);
lock(&disk->open_mutex);
lock(&lo->lo_mutex);
lock((wq_completion)loop0);
*** DEADLOCK ***
1 lock held by losetup/11596:
#0: ffff939655510c68 (&lo->lo_mutex){+.+.}-{3:3}, at: __loop_clr_fd+0x41/0x660 [loop]
stack backtrace:
CPU: 1 PID: 11596 Comm: losetup Not tainted 5.14.0-rc2+ #406
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-2.fc32 04/01/2014
Call Trace:
dump_stack_lvl+0x57/0x72
check_noncircular+0xcf/0xf0
? stack_trace_save+0x3b/0x50
__lock_acquire+0x10ea/0x1d90
lock_acquire+0xb5/0x2b0
? flush_workqueue+0x67/0x5e0
? lockdep_init_map_type+0x47/0x220
flush_workqueue+0x91/0x5e0
? flush_workqueue+0x67/0x5e0
? verify_cpu+0xf0/0x100
drain_workqueue+0xa0/0x110
destroy_workqueue+0x36/0x250
__loop_clr_fd+0x9a/0x660 [loop]
? blkdev_ioctl+0x8d/0x2a0
block_ioctl+0x3f/0x50
__x64_sys_ioctl+0x80/0xb0
do_syscall_64+0x38/0x90
entry_SYSCALL_64_after_hwframe+0x44/0xae
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a6d37ccdd2 ]
capsnaps will take inode references via ihold when queueing to flush.
When force unmounting, the client will just close the sessions and
may never get a flush reply, causing a leak and inode ref leak.
Fix this by removing the capsnaps for an inode when removing the caps.
URL: https://tracker.ceph.com/issues/52295
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b11ed50346 ]
The current code will update the mtime and then try to get caps to
handle the write. If we end up having to request caps from the MDS, then
the mtime in the cap grant will clobber the updated mtime and it'll be
lost.
This is most noticable when two clients are alternately writing to the
same file. Fw caps are continually being granted and revoked, and the
mtime ends up stuck because the updated mtimes are always being
overwritten with the old one.
Fix this by changing the order of operations in ceph_write_iter to get
the caps before updating the times. Also, make sure we check the pool
full conditions before even getting any caps or uninlining.
URL: https://tracker.ceph.com/issues/46574
Reported-by: Jozef Kováč <kovac@firma.zoznam.sk>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Luis Henriques <lhenriques@suse.de>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit aac6c0f907 ]
The xilinx dma driver uses the consistent allocations, so for correct
operation also set the DMA mask for coherent APIs. It fixes the below
kernel crash with dmatest client when DMA IP is configured with 64-bit
address width and linux is booted from high (>4GB) memory.
Call trace:
[ 489.531257] dma_alloc_from_pool+0x8c/0x1c0
[ 489.535431] dma_direct_alloc+0x284/0x330
[ 489.539432] dma_alloc_attrs+0x80/0xf0
[ 489.543174] dma_pool_alloc+0x160/0x2c0
[ 489.547003] xilinx_cdma_prep_memcpy+0xa4/0x180
[ 489.551524] dmatest_func+0x3cc/0x114c
[ 489.555266] kthread+0x124/0x130
[ 489.558486] ret_from_fork+0x10/0x3c
[ 489.562051] ---[ end trace 248625b2d596a90a ]---
Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Reviewed-by: Harini Katakam <harini.katakam@xilinx.com>
Link: https://lore.kernel.org/r/1629363528-30347-1-git-send-email-radhey.shyam.pandey@xilinx.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c3811a50ad ]
Currently, iommu_init_ga() checks and disables IOMMU VAPIC support
(i.e. AMD AVIC support in IOMMU) when GAMSup feature bit is not set.
However it forgets to clear IRQ_POSTING_CAP from the previously set
amd_iommu_irq_ops.capability.
This triggers an invalid page fault bug during guest VM warm reboot
if AVIC is enabled since the irq_remapping_cap(IRQ_POSTING_CAP) is
incorrectly set, and crash the system with the following kernel trace.
BUG: unable to handle page fault for address: 0000000000400dd8
RIP: 0010:amd_iommu_deactivate_guest_mode+0x19/0xbc
Call Trace:
svm_set_pi_irte_mode+0x8a/0xc0 [kvm_amd]
? kvm_make_all_cpus_request_except+0x50/0x70 [kvm]
kvm_request_apicv_update+0x10c/0x150 [kvm]
svm_toggle_avic_for_irq_window+0x52/0x90 [kvm_amd]
svm_enable_irq_window+0x26/0xa0 [kvm_amd]
vcpu_enter_guest+0xbbe/0x1560 [kvm]
? avic_vcpu_load+0xd5/0x120 [kvm_amd]
? kvm_arch_vcpu_load+0x76/0x240 [kvm]
? svm_get_segment_base+0xa/0x10 [kvm_amd]
kvm_arch_vcpu_ioctl_run+0x103/0x590 [kvm]
kvm_vcpu_ioctl+0x22a/0x5d0 [kvm]
__x64_sys_ioctl+0x84/0xc0
do_syscall_64+0x33/0x40
entry_SYSCALL_64_after_hwframe+0x44/0xae
Fixes by moving the initializing of AMD IOMMU interrupt remapping mode
(amd_iommu_guest_ir) earlier before setting up the
amd_iommu_irq_ops.capability with appropriate IRQ_POSTING_CAP flag.
[joro: Squashed the two patches and limited
check_features_on_all_iommus() to CONFIG_IRQ_REMAP
to fix a compile warning.]
Signed-off-by: Wei Huang <wei.huang2@amd.com>
Co-developed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Link: https://lore.kernel.org/r/20210820202957.187572-2-suravee.suthikulpanit@amd.com
Link: https://lore.kernel.org/r/20210820202957.187572-3-suravee.suthikulpanit@amd.com
Fixes: 8bda0cfbdc ("iommu/amd: Detect and initialize guest vAPIC log")
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 907872baa9 ]
parisc build test images fail to compile with the following error.
drivers/parisc/dino.c:160:12: error:
'pci_dev_is_behind_card_dino' defined but not used
Move the function just ahead of its only caller to avoid the error.
Fixes: 5fa1659105 ("parisc: Disable HP HSC-PCI Cards to prevent kernel crash")
Cc: Helge Deller <deller@gmx.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4b92d4add5 ]
DEFINE_SMP_CALL_CACHE_FUNCTION() was usefel before the CPU hotplug rework
to ensure that the cache related functions are called on the upcoming CPU
because the notifier itself could run on any online CPU.
The hotplug state machine guarantees that the callbacks are invoked on the
upcoming CPU. So there is no need to have this SMP function call
obfuscation. That indirection was missed when the hotplug notifiers were
converted.
This also solves the problem of ARM64 init_cache_level() invoking ACPI
functions which take a semaphore in that context. That's invalid as SMP
function calls run with interrupts disabled. Running it just from the
callback in context of the CPU hotplug thread solves this.
Fixes: 8571890e15 ("arm64: Add support for ACPI based firmware tables")
Reported-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Will Deacon <will@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/871r69ersb.ffs@tglx
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b3dc549986 ]
Due to high latency in PCIE clock switching on RKL platforms,
switching the PCIE clock dynamically at runtime can lead to HDMI/DP
audio problems. On newer asics this is handled in the SMU firmware.
For SMU7-based asics, disable PCIE clock switching to avoid the issue.
AMD provide a parameter to disable PICE_DPM.
modprobe amdgpu ppfeaturemask=0xfff7bffb
It's better to contorl PCIE_DPM in amd gpu driver,
switch PCI_DPM by determining intel RKL platform for SMU7-based asics.
Fixes: 1a31474cdb ("drm/amd/pm: workaround for audio noise issue")
Ref: https://lists.freedesktop.org/archives/amd-gfx/2021-August/067413.html
Signed-off-by: Koba Ko <koba.ko@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit fb83610762 ]
There are two pairs of declarations for thermal_cooling_device_register()
and thermal_of_cooling_device_register(), and only one set was changed
in a recent patch, so the other one now causes a compile-time warning:
drivers/net/wireless/mediatek/mt76/mt7915/init.c: In function 'mt7915_thermal_init':
drivers/net/wireless/mediatek/mt76/mt7915/init.c:134:48: error: passing argument 1 of 'thermal_cooling_device_register' discards 'const' qualifier from pointer target type [-Werror=discarded-qualifiers]
134 | cdev = thermal_cooling_device_register(wiphy_name(wiphy), phy,
| ^~~~~~~~~~~~~~~~~
In file included from drivers/net/wireless/mediatek/mt76/mt7915/init.c:7:
include/linux/thermal.h:407:39: note: expected 'char *' but argument is of type 'const char *'
407 | thermal_cooling_device_register(char *type, void *devdata,
| ~~~~~~^~~~
Change the dummy helper functions to have the same arguments as the
normal version.
Fixes: f991de53a8 ("thermal: make device_register's type argument const")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Jean-Francois Dagenais <jeff.dagenais@gmail.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210722090717.1116748-1-arnd@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b4002173b7 ]
The first thing metric_delayed_work does is check mdsc->stopping,
and then return immediately if it's set. That's good since we would
have already torn down the metric structures at this point, otherwise,
but there is no locking around mdsc->stopping.
It's possible that the ceph_metric_destroy call could race with the
delayed_work, in which case we could end up with the delayed_work
accessing destroyed percpu variables.
At this point in the mdsc teardown, the "stopping" flag has already been
set, so there's no benefit to flushing the work. Move the work
cancellation in ceph_metric_destroy ahead of the percpu variable
destruction, and eliminate the flush_delayed_work call in
ceph_mdsc_destroy.
Fixes: 18f473b384 ("ceph: periodically send perf metrics to MDSes")
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d453ceb654 ]
Add trace event to report samples and their timestamp coming from the
EC. It allows to check if the timestamps are correct and the filter is
working correctly without introducing too much latency.
To enable these events:
cd /sys/kernel/debug/tracing/
echo 1 > events/cros_ec/enable
echo 0 > events/cros_ec/cros_ec_request_start/enable
echo 0 > events/cros_ec/cros_ec_request_done/enable
echo 1 > tracing_on
cat trace_pipe
Observe event flowing:
irq/105-chromeo-95 [000] .... 613.659758: cros_ec_sensorhub_timestamp: ...
irq/105-chromeo-95 [000] .... 613.665219: cros_ec_sensorhub_filter: dx: ...
Signed-off-by: Gwendal Grignou <gwendal@chromium.org>
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 020162d6f4 upstream.
This fixes a race condition: After pwmchip_add() is called there might
already be a consumer and then modifying the hardware behind the
consumer's back is bad. So reset before calling pwmchip_add().
Note that reseting the hardware isn't the right thing to do if the PWM
is already running as it might e.g. disable (or even enable) a backlight
that is supposed to be on (or off).
Fixes: 4dce82c1e8 ("pwm: add pwm-mxs support")
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3d2813fb17 upstream.
This fixes a race condition: After pwmchip_add() is called there might
already be a consumer and then modifying the hardware behind the
consumer's back is bad. So set the default before.
(Side-note: I don't know what this register setting actually does, if
this modifies the polarity there is an inconsistency because the
inversed polarity isn't considered if the PWM is already running during
.probe().)
Fixes: acfd92fdfb ("pwm: lpc32xx: Set PWM_PIN_LEVEL bit to default value")
Cc: Sylvain Lemieux <slemieux@tycoint.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 4a9344cd0a upstream.
There are variables(power.may_skip_resume and dev->power.must_resume)
and DPM_FLAG_MAY_SKIP_RESUME flags to control the resume of devices after
a system wide suspend transition.
Setting the DPM_FLAG_MAY_SKIP_RESUME flag means that the driver allows
its "noirq" and "early" resume callbacks to be skipped if the device
can be left in suspend after a system-wide transition into the working
state. PM core determines that the driver's "noirq" and "early" resume
callbacks should be skipped or not with dev_pm_skip_resume() function by
checking power.may_skip_resume variable.
power.must_resume variable is getting set to false in __device_suspend()
function without checking device's DPM_FLAG_MAY_SKIP_RESUME settings.
In problematic scenario, where all the devices in the suspend_late
stage are successful and some device can fail to suspend in
suspend_noirq phase. So some devices successfully suspended in suspend_late
stage are not getting chance to execute __device_suspend_noirq()
to set dev->power.must_resume variable to true and not getting
resumed in early_resume phase.
Add a check for device's DPM_FLAG_MAY_SKIP_RESUME flag before
setting power.must_resume variable in __device_suspend function.
Fixes: 6e176bf8d4 ("PM: sleep: core: Do not skip callbacks in the resume phase")
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 98e2e409e7 upstream.
When the refcount is decreased to 0, the resource reclamation branch is
entered. Before CPU0 reaches the race point (1), CPU1 may obtain the
spinlock and traverse the rbtree to find 'root', see
nilfs_lookup_root().
Although CPU1 will call refcount_inc() to increase the refcount, it is
obviously too late. CPU0 will release 'root' directly, CPU1 then
accesses 'root' and triggers UAF.
Use refcount_dec_and_lock() to ensure that both the operations of
decrease refcount to 0 and link deletion are lock protected eliminates
this risk.
CPU0 CPU1
nilfs_put_root():
<-------- (1)
spin_lock(&nilfs->ns_cptree_lock);
rb_erase(&root->rb_node, &nilfs->ns_cptree);
spin_unlock(&nilfs->ns_cptree_lock);
kfree(root);
<-------- use-after-free
refcount_t: underflow; use-after-free.
WARNING: CPU: 2 PID: 9476 at lib/refcount.c:28 \
refcount_warn_saturate+0x1cf/0x210 lib/refcount.c:28
Modules linked in:
CPU: 2 PID: 9476 Comm: syz-executor.0 Not tainted 5.10.45-rc1+ #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), ...
RIP: 0010:refcount_warn_saturate+0x1cf/0x210 lib/refcount.c:28
... ...
Call Trace:
__refcount_sub_and_test include/linux/refcount.h:283 [inline]
__refcount_dec_and_test include/linux/refcount.h:315 [inline]
refcount_dec_and_test include/linux/refcount.h:333 [inline]
nilfs_put_root+0xc1/0xd0 fs/nilfs2/the_nilfs.c:795
nilfs_segctor_destroy fs/nilfs2/segment.c:2749 [inline]
nilfs_detach_log_writer+0x3fa/0x570 fs/nilfs2/segment.c:2812
nilfs_put_super+0x2f/0xf0 fs/nilfs2/super.c:467
generic_shutdown_super+0xcd/0x1f0 fs/super.c:464
kill_block_super+0x4a/0x90 fs/super.c:1446
deactivate_locked_super+0x6a/0xb0 fs/super.c:335
deactivate_super+0x85/0x90 fs/super.c:366
cleanup_mnt+0x277/0x2e0 fs/namespace.c:1118
__cleanup_mnt+0x15/0x20 fs/namespace.c:1125
task_work_run+0x8e/0x110 kernel/task_work.c:151
tracehook_notify_resume include/linux/tracehook.h:188 [inline]
exit_to_user_mode_loop kernel/entry/common.c:164 [inline]
exit_to_user_mode_prepare+0x13c/0x170 kernel/entry/common.c:191
syscall_exit_to_user_mode+0x16/0x30 kernel/entry/common.c:266
do_syscall_64+0x45/0x80 arch/x86/entry/common.c:56
entry_SYSCALL_64_after_hwframe+0x44/0xa9
There is no reproduction program, and the above is only theoretical
analysis.
Link: https://lkml.kernel.org/r/1629859428-5906-1-git-send-email-konishi.ryusuke@gmail.com
Fixes: ba65ae4729 ("nilfs2: add checkpoint tree to nilfs object")
Link: https://lkml.kernel.org/r/20210723012317.4146-1-thunder.leizhen@huawei.com
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e1fbbd0731 upstream.
Keno Fischer reported that when a binray loaded via ld-linux-x the
prctl(PR_SET_MM_MAP) doesn't allow to setup brk value because it lays
before mm:end_data.
For example a test program shows
| # ~/t
|
| start_code 401000
| end_code 401a15
| start_stack 7ffce4577dd0
| start_data 403e10
| end_data 40408c
| start_brk b5b000
| sbrk(0) b5b000
and when executed via ld-linux
| # /lib64/ld-linux-x86-64.so.2 ~/t
|
| start_code 7fc25b0a4000
| end_code 7fc25b0c4524
| start_stack 7fffcc6b2400
| start_data 7fc25b0ce4c0
| end_data 7fc25b0cff98
| start_brk 55555710c000
| sbrk(0) 55555710c000
This of course prevent criu from restoring such programs. Looking into
how kernel operates with brk/start_brk inside brk() syscall I don't see
any problem if we allow to setup brk/start_brk without checking for
end_data. Even if someone pass some weird address here on a purpose then
the worst possible result will be an unexpected unmapping of existing vma
(own vma, since prctl works with the callers memory) but test for
RLIMIT_DATA is still valid and a user won't be able to gain more memory in
case of expanding VMAs via new values shipped with prctl call.
Link: https://lkml.kernel.org/r/20210121221207.GB2174@grain
Fixes: bbdc6076d2 ("binfmt_elf: move brk out of mmap when doing direct loader exec")
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Reported-by: Keno Fischer <keno@juliacomputing.com>
Acked-by: Andrey Vagin <avagin@gmail.com>
Tested-by: Andrey Vagin <avagin@gmail.com>
Cc: Dmitry Safonov <0x7f454c46@gmail.com>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Pavel Tikhomirov <ptikhomirov@virtuozzo.com>
Cc: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 4a86d41404 upstream.
Currently perf saves a build-id with size but old versions assumes the
size of 20. In case the build-id is less than 20 (like for MD5), it'd
fill the rest with 0s.
I saw a problem when old version of perf record saved a binary in the
build-id cache and new version of perf reads the data. The symbols
should be read from the build-id cache (as the path no longer has the
same binary) but it failed due to mismatch in the build-id.
symsrc__init: build id mismatch for /home/namhyung/.debug/.build-id/53/e4c2f42a4c61a2d632d92a72afa08f00000000/elf.
The build-id event in the data has 20 byte build-ids, but it saw a
different size (16) when it reads the build-id of the elf file in the
build-id cache.
$ readelf -n ~/.debug/.build-id/53/e4c2f42a4c61a2d632d92a72afa08f00000000/elf
Displaying notes found in: .note.gnu.build-id
Owner Data size Description
GNU 0x00000010 NT_GNU_BUILD_ID (unique build ID bitstring)
Build ID: 53e4c2f42a4c61a2d632d92a72afa08f
Let's fix this by allowing trailing zeros if the size is different.
Fixes: 39be8d0115 ("perf tools: Pass build_id object to dso__build_id_equal()")
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lore.kernel.org/lkml/20210910224630.1084877-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 099ec97ac9 upstream.
clang warns:
drivers/staging/rtl8192u/r8192U_core.c:4268:20: warning: bitwise and of
boolean expressions; did you mean logical and? [-Wbool-operation-and]
bpacket_toself = bpacket_match_bssid &
^~~~~~~~~~~~~~~~~~~~~
&&
1 warning generated.
Replace the bitwise AND with a logical one to clear up the warning, as
that is clearly what was intended.
Fixes: 8fc8598e61 ("Staging: Added Realtek rtl8192u driver to staging")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20210814235625.1780033-1-nathan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ef6c8d6ccf upstream.
When SCTP handles an INIT chunk, it calls for example:
sctp_sf_do_5_1B_init
sctp_verify_init
sctp_verify_param
sctp_process_init
sctp_process_param
handling of SCTP_PARAM_SET_PRIMARY
sctp_verify_init() wasn't doing proper size validation and neither the
later handling, allowing it to work over the chunk itself, possibly being
uninitialized memory.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b6ffe7671b upstream.
In one of the fallbacks that SCTP has for identifying an association for an
incoming packet, it looks for AddIp chunk (from ASCONF) and take a peek.
Thing is, at this stage nothing was validating that the chunk actually had
enough content for that, allowing the peek to happen over uninitialized
memory.
Similar check already exists in actual asconf handling in
sctp_verify_asconf().
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fcf044891c upstream.
We do not need a SWIOTLB unless we have DRAM that is addressable beyond
the arm_dma_limit. Compare max_pfn with arm_dma_pfn_limit to determine
whether we do need a SWIOTLB to be initialized.
Fixes: ad3c7b18c5 ("arm: use swiotlb for bounce buffering on LPAE configs")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a8b92b8c1e upstream.
We should not walk/touch page tables outside of VMA boundaries when
holding only the mmap sem in read mode. Evil user space can modify the
VMA layout just before this function runs and e.g., trigger races with
page table removal code since commit dd2283f260 ("mm: mmap: zap pages
with read mmap_sem in munmap").
find_vma() does not check if the address is >= the VMA start address;
use vma_lookup() instead.
Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Fixes: dd2283f260 ("mm: mmap: zap pages with read mmap_sem in munmap")
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3a2b2eb556 upstream.
The Linux console's VT102 implementation already consumes OSC
("Operating System Command") sequences, probably because that's how
palette changes are transmitted.
In addition to OSC, there are three other major clases of ANSI control
strings: APC ("Application Program Command"), PM ("Privacy Message"),
and DCS ("Device Control String"). They are handled similarly to OSC in
terms of termination.
Source: vt100.net
Add three new enumerated states, one for each of these types. All three
are handled the same way right now--they simply consume input until
terminated. I hope to expand upon this firmament in the future. Add
new predicate ansi_control_string(), returning true for any of these
states. Replace explicit checks against ESosc with calls to this
function. Transition to these states appropriately from the escape
initiation (ESesc) state.
This was motivated by the following Notcurses bugs:
https://github.com/dankamongmen/notcurses/issues/2050https://github.com/dankamongmen/notcurses/issues/1828https://github.com/dankamongmen/notcurses/issues/2069
where standard VT sequences are not consumed by the Linux console. It's
not necessary that the Linux console *support* these sequences, but it
ought *consume* these well-specified classes of sequences.
Tested by sending a variety of escape sequences to the console, and
verifying that they still worked, or were now properly consumed.
Verified that the escapes were properly terminated at a generic level.
Verified that the Notcurses tools continued to show expected output on
the Linux console, except now without escape bleedthrough.
Link: https://lore.kernel.org/lkml/YSydL0q8iaUfkphg@schwarzgerat.orthanc/
Signed-off-by: nick black <dankamongmen@gmail.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jirislaby@kernel.org>
Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 02319bf15a upstream.
After d12e1c4649 ("net: dsa: b53: Set correct number of ports in the
DSA struct") we stopped setting dsa_switch::num_ports to DSA_MAX_PORTS,
which created an off by one error between the statically allocated
bcm_sf2_priv::port_sts array (of size DSA_MAX_PORTS). When
dsa_is_cpu_port() is used, we end-up accessing an out of bounds member
and causing a NPD.
Fix this by iterating with the appropriate port count using
ds->num_ports.
Fixes: d12e1c4649 ("net: dsa: b53: Set correct number of ports in the DSA struct")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit eca4cf12ac upstream.
The recent patch has introduced a regression by not reading the reset
count in the ERROR_RECOVERY async event handler. We may have just
gone through a reset and the reset count has just incremented. If
we don't update the reset count in the ERROR_RECOVERY event handler,
the health check timer will see that the reset count has changed and
will initiate an unintended reset.
Restore the unconditional update of the reset count in
bnxt_async_event_process() if error recovery watchdog is enabled.
Also, update the reset count at the end of the reset sequence to
make it even more robust.
Fixes: 1b2b918319 ("bnxt_en: Fix possible unintended driver initiated error recovery")
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 81065b35e2 upstream.
There are two cases for machine check recovery:
1) The machine check was triggered by ring3 (application) code.
This is the simpler case. The machine check handler simply queues
work to be executed on return to user. That code unmaps the page
from all users and arranges to send a SIGBUS to the task that
triggered the poison.
2) The machine check was triggered in kernel code that is covered by
an exception table entry. In this case the machine check handler
still queues a work entry to unmap the page, etc. but this will
not be called right away because the #MC handler returns to the
fix up code address in the exception table entry.
Problems occur if the kernel triggers another machine check before the
return to user processes the first queued work item.
Specifically, the work is queued using the ->mce_kill_me callback
structure in the task struct for the current thread. Attempting to queue
a second work item using this same callback results in a loop in the
linked list of work functions to call. So when the kernel does return to
user, it enters an infinite loop processing the same entry for ever.
There are some legitimate scenarios where the kernel may take a second
machine check before returning to the user.
1) Some code (e.g. futex) first tries a get_user() with page faults
disabled. If this fails, the code retries with page faults enabled
expecting that this will resolve the page fault.
2) Copy from user code retries a copy in byte-at-time mode to check
whether any additional bytes can be copied.
On the other side of the fence are some bad drivers that do not check
the return value from individual get_user() calls and may access
multiple user addresses without noticing that some/all calls have
failed.
Fix by adding a counter (current->mce_count) to keep track of repeated
machine checks before task_work() is called. First machine check saves
the address information and calls task_work_add(). Subsequent machine
checks before that task_work call back is executed check that the address
is in the same page as the first machine check (since the callback will
offline exactly one page).
Expected worst case is four machine checks before moving on (e.g. one
user access with page faults disabled, then a repeat to the same address
with page faults enabled ... repeat in copy tail bytes). Just in case
there is some code that loops forever enforce a limit of 10.
[ bp: Massage commit message, drop noinstr, fix typo, extend panic
messages. ]
Fixes: 5567d11c21 ("x86/mce: Send #MC singal from task work")
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: <stable@vger.kernel.org>
Link: https://lkml.kernel.org/r/YT/IJ9ziLqmtqEPu@agluck-desk2.amr.corp.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 0341d5e3d1 ]
The cur_tx counter must be incremented after TACT bit of
txdesc->status was set. However, a CPU is possible to reorder
instructions and/or memory accesses between cur_tx and
txdesc->status. And then, if TX interrupt happened at such a
timing, the sh_eth_tx_free() may free the descriptor wrongly.
So, add wmb() before cur_tx++.
Otherwise NETDEV WATCHDOG timeout is possible to happen.
Fixes: 86a74ff21a ("net: sh_eth: add support for Renesas SuperH Ethernet")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit cdff1eda69 ]
One MIPS platform (mach-rc32434) defines GPIOBASE. This macro
conflicts with one of the same name in lpc_sch.c. Rename the latter one
to prevent the build error.
../drivers/mfd/lpc_sch.c:25: error: "GPIOBASE" redefined [-Werror]
25 | #define GPIOBASE 0x44
../arch/mips/include/asm/mach-rc32434/rb.h:32: note: this is the location of the previous definition
32 | #define GPIOBASE 0x050000
Cc: Denis Turischev <denis@compulab.co.il>
Fixes: e82c60ae7d ("mfd: Introduce lpc_sch for Intel SCH LPC bridge")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 922e8ce883 ]
The IRQ support for SCH GPIO is not specific to the Intel Quark SoC.
Moreover the IRQ routing is quite interesting there, so while it's
needs a special support, the driver haven't it anyway yet.
Due to above remove basically redundant code of IRQ support.
This reverts commit ec689a8a81.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1b2b918319 ]
If error recovery is already enabled, bnxt_timer() will periodically
check the heartbeat register and the reset counter. If we get an
error recovery async. notification from the firmware (e.g. change in
primary/secondary role), we will immediately read and update the
heartbeat register and the reset counter. If the timer for the next
health check expires soon after this, we may read the heartbeat register
again in quick succession and find that it hasn't changed. This will
trigger error recovery unintentionally.
The likelihood is small because we also reset fw_health->tmr_counter
which will reset the interval for the next health check. But the
update is not protected and bnxt_timer() can miss the update and
perform the health check without waiting for the full interval.
Fix it by only reading the heartbeat register and reset counter in
bnxt_async_event_process() if error recovery is trasitioning to the
enabled state. Also add proper memory barriers so that when enabling
for the first time, bnxt_timer() will see the tmr_counter interval and
perform the health check after the full interval has elapsed.
Fixes: 7e914027f7 ("bnxt_en: Enable health monitoring.")
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit f4d95c3c19 ]
We currently only log the error recovery settings if it is enabled.
In some cases, firmware disables error recovery after it was
initially enabled. Without logging anything, the user will not be
aware of this change in setting.
Log it when error recovery is disabled. Also, change the reset count
value from hexadecimal to decimal.
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a44daa8fcb ]
Firmware is capable of generating asynchronous debug notifications.
The event data is opaque to the driver and is simply logged. Debug
notifications can be enabled by turning on hardware status messages
using the ethtool msglvl interface.
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 6fdab8a3ad ]
The current asic.rev is incomplete and does not include the metal
revision. Add the metal revision and decode the complete asic
revision into the more common and readable form (A0, B0, etc).
Fixes: 7154917a12 ("bnxt_en: Refactor bnxt_dl_info_get().")
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 63f8428b40 ]
Broadcom's b53 switches have one IMP (Inband Management Port) that needs
to be programmed using its own designed register. IMP port may be
different than CPU port - especially on devices with multiple CPU ports.
For that reason it's required to explicitly note IMP port index and
check for it when choosing a register to use.
This commit fixes BCM5301x support. Those switches use CPU port 5 while
their IMP port is 8. Before this patch b53 was trying to program port 5
with B53_PORT_OVERRIDE_CTRL instead of B53_GMII_PORT_OVERRIDE_CTRL(5).
It may be possible to also replace "cpu_port" usages with
dsa_is_cpu_port() but that is out of the scope of thix BCM5301x fix.
Fixes: 967dd82ffc ("net: dsa: b53: Add support for Broadcom RoboSwitch")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8a0ed250f9 ]
The GRE tunnel device can pull existing outer headers in ipge_xmit.
This is a rare path, apparently unique to this device. The below
commit ensured that pulling does not move skb->data beyond csum_start.
But it has a false positive if ip_summed is not CHECKSUM_PARTIAL and
thus csum_start is irrelevant.
Refine to exclude this. At the same time simplify and strengthen the
test.
Simplify, by moving the check next to the offending pull, making it
more self documenting and removing an unnecessary branch from other
code paths.
Strengthen, by also ensuring that the transport header is correct and
therefore the inner headers will be after skb_reset_inner_headers.
The transport header is set to csum_start in skb_partial_csum_set.
Link: https://lore.kernel.org/netdev/YS+h%2FtqCJJiQei+W@shredder/
Fixes: 1d011c4803 ("ip_gre: add validation for csum_start")
Reported-by: Ido Schimmel <idosch@idosch.org>
Suggested-by: Alexander Duyck <alexander.duyck@gmail.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 9ddbc2a00d ]
Previous commit 68233c583a removes the qlcnic_rom_lock()
in qlcnic_pinit_from_rom(), but remains its corresponding
unlock function, which is odd. I'm not very sure whether the
lock is missing, or the unlock is redundant. This bug is
suggested by a static analysis tool, please advise.
Fixes: 68233c583a ("qlcnic: updated reset sequence")
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c7c5e6ff53 ]
syzbot found that forcing a big quantum attribute would crash hosts fast,
essentially using this:
tc qd replace dev eth0 root fq_codel quantum 4294967295
This is because fq_codel_dequeue() would have to loop
~2^31 times in :
if (flow->deficit <= 0) {
flow->deficit += q->quantum;
list_move_tail(&flow->flowchain, &q->old_flows);
goto begin;
}
SFQ max quantum is 2^19 (half a megabyte)
Lets adopt a max quantum of one megabyte for FQ_CODEL.
Fixes: 4b549a2ef4 ("fq_codel: Fair Queue Codel AQM")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 730affed24 ]
Bug reported by KASAN:
BUG: KASAN: use-after-scope in inet6_ehashfn (net/ipv6/inet6_hashtables.c:40)
Call Trace:
(...)
inet6_ehashfn (net/ipv6/inet6_hashtables.c:40)
(...)
nf_sk_lookup_slow_v6 (net/ipv6/netfilter/nf_socket_ipv6.c:91
net/ipv6/netfilter/nf_socket_ipv6.c:146)
It seems that this bug has already been fixed by Eric Dumazet in the
past in:
commit 78296c97ca ("netfilter: xt_socket: fix a stack corruption bug")
But a variant of the same issue has been introduced in
commit d64d80a2cd ("netfilter: x_tables: don't extract flow keys on early demuxed sks in socket match")
`daddr` and `saddr` potentially hold a reference to ipv6_var that is no
longer in scope when the call to `nf_socket_get_sock_v6` is made.
Fixes: d64d80a2cd ("netfilter: x_tables: don't extract flow keys on early demuxed sks in socket match")
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Benjamin Hesmans <benjamin.hesmans@tessares.net>
Reviewed-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d12e1c4649 ]
Setting DSA_MAX_PORTS caused DSA to call b53 callbacks (e.g.
b53_disable_port() during dsa_register_switch()) for invalid
(non-existent) ports. That made b53 modify unrelated registers and is
one of reasons for a broken BCM5301x support.
This problem exists for years but DSA_MAX_PORTS usage has changed few
times. It seems the most accurate to reference commit dropping
dsa_switch_alloc() in the Fixes tag.
Fixes: 7e99e34701 ("net: dsa: remove dsa_switch_alloc helper")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit cdb067d31c ]
It isn't true that CPU port is always the last one. Switches BCM5301x
have 9 ports (port 6 being inactive) and they use port 5 as CPU by
default (depending on design some other may be CPU ports too).
A more reliable way of determining number of ports is to check for the
last set bit in the "enabled_ports" bitfield.
This fixes b53 internal state, it will allow providing accurate info to
the DSA and is required to fix BCM5301x support.
Fixes: 967dd82ffc ("net: dsa: b53: Add support for Broadcom RoboSwitch")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit ecdc28defc ]
If the network devices connected to the system beyond
HSO_MAX_NET_DEVICES. add_net_device() in hso_create_net_device()
will be failed for the network_table is full. It will lead to
business failure which rely on network_table, for example,
hso_suspend() and hso_resume(). It will also lead to memory leak
because resource release process can not search the hso_device
object from network_table in hso_free_interface().
Add failure handler for add_net_device() in hso_create_net_device()
to solve the above problems.
Fixes: 72dc1c096c ("HSO: add option hso driver")
Signed-off-by: Ziyang Xuan <william.xuanziyang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit bfd862a7e9 ]
'$cin' and '$sin' variables are local to a function: they are then not
available from the cleanup trap.
Instead, we need to use '$large' and '$small' that are not local and
defined just before setting the trap.
Without this patch, running this script in a loop might cause a:
write: No space left on device
issue.
Fixes: 1a418cb8e8 ("mptcp: simult flow self-tests")
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0e90dfa7a8 ]
I noticed that only port 0 worked on the RTL8366RB since we
started to use custom tags.
It turns out that the format of egress custom tags is actually
different from ingress custom tags. While the lower bits just
contain the port number in ingress tags, egress tags need to
indicate destination port by setting the bit for the
corresponding port.
It was working on port 0 because port 0 added 0x00 as port
number in the lower bits, and if you do this the packet appears
at all ports, including the intended port. Ooops.
Fix this and all ports work again. Use the define for shifting
the "type A" into place while we're at it.
Tested on the D-Link DIR-685 by sending traffic to each of
the ports in turn. It works.
Fixes: 86dd9868b8 ("net: dsa: tag_rtl4_a: Support also egress tags")
Cc: DENG Qingfang <dqfext@gmail.com>
Cc: Mauri Sandberg <sandberg@mailfence.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 889a1b3f35 ]
If an error occurs after a 'gpiochip_add_data()' call it must be undone by
a corresponding 'gpiochip_remove()' as already done in the remove function.
To simplify the code a fix a leak in the error handling path of the probe,
use the managed version instead (i.e. 'devm_gpiochip_add_data()')
Fixes: 698b8eeaed ("gpio/mpc8xxx: change irq handler from chained to normal")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 555bda42b0 ]
Commit 698b8eeaed ("gpio/mpc8xxx: change irq handler from chained to normal")
has introduced a new 'goto err;' at the very end of the function, but has
not updated the error handling path accordingly.
Add the now missing 'irq_domain_remove()' call which balances a previous
'irq_domain_create_linear() call.
Fixes: 698b8eeaed ("gpio/mpc8xxx: change irq handler from chained to normal")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit edf7b4a2d8 ]
The build on fedora:35 and fedora:rawhide with clang is failing with:
49 41.00 fedora:35 : FAIL clang version 13.0.0 (Fedora 13.0.0~rc1-1.fc35)
bench/inject-buildid.c:351:6: error: variable 'len' set but not used [-Werror,-Wunused-but-set-variable]
u64 len = 0;
^
1 error generated.
make[3]: *** [/git/perf-5.14.0-rc7/tools/build/Makefile.build:139: bench] Error 2
50 41.11 fedora:rawhide : FAIL clang version 13.0.0 (Fedora 13.0.0~rc1-1.fc35)
bench/inject-buildid.c:351:6: error: variable 'len' set but not used [-Werror,-Wunused-but-set-variable]
u64 len = 0;
^
1 error generated.
make[3]: *** [/git/perf-5.14.0-rc7/tools/build/Makefile.build:139: bench] Error 2
That 'len' variable is not used at all, so just make sure all the
synthesize_RECORD() routines return ssize_t to propagate the writen()
return, as it may fail, ditch the 'ret' var and bail out if those
routines fail.
Fixes: 0bf02a0d80 ("perf bench: Add build-id injection benchmark")
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: http://lore.kernel.org/lkml/CAM9d7cgEZNSor+B+7Y2C+QYGme_v5aH0Zn0RLfxoQ+Fy83EHrg@mail.gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 6b5ff0405e ]
0day bot reports a build error:
ERROR: modpost: "clear_user_page" [drivers/media/v4l2-core/videobuf-dma-sg.ko] undefined!
so export it in arch/arc/ to fix the build error.
In most ARCHes, clear_user_page() is a macro. OTOH, in a few
ARCHes it is a function and needs to be exported.
PowerPC exported it in 2004. It looks like nds32 and nios2
still need to have it exported.
Fixes: 4102b53392 ("ARC: [mm] Aliasing VIPT dcache support 2/4")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: linux-snps-arc@lists.infradead.org
Signed-off-by: Vineet Gupta <vgupta@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 6826c6849b ]
The CPU_ON PSCI call takes a payload that KVM uses to configure a
destination vCPU to run. This payload is non-architectural state and not
exposed through any existing UAPI. Effectively, we have a race between
CPU_ON and userspace saving/restoring a guest: if the target vCPU isn't
ran again before the VMM saves its state, the requested PC and context
ID are lost. When restored, the target vCPU will be runnable and start
executing at its old PC.
We can avoid this race by making sure the reset payload is serviced
before userspace can access a vCPU's state.
Fixes: 358b28f09f ("arm/arm64: KVM: Allow a VCPU to fully reset itself")
Signed-off-by: Oliver Upton <oupton@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210818202133.1106786-3-oupton@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 6654f9dfcb ]
KVM correctly serializes writes to a vCPU's reset state, however since
we do not take the KVM lock on the read side it is entirely possible to
read state from two different reset requests.
Cure the race for now by taking the KVM lock when reading the
reset_state structure.
Fixes: 358b28f09f ("arm/arm64: KVM: Allow a VCPU to fully reset itself")
Signed-off-by: Oliver Upton <oupton@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210818202133.1106786-2-oupton@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a89d69a44e ]
Since 2431c4f5b4 ("mtd: Implement mtd_{read,write}() as wrappers
around mtd_{read,write}_oob()") don't allow _write|_read and
_write_oob|_read_oob existing at the same time, we should check the
existence of callbacks "_read and _write" from subdev's master device
(We can trust master device since it has been registered) before
assigning, otherwise following warning occurs while making
concatenated device:
WARNING: CPU: 2 PID: 6728 at drivers/mtd/mtdcore.c:595
add_mtd_device+0x7f/0x7b0
Fixes: 2431c4f5b4 ("mtd: Implement mtd_{read,write}() around ...")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20210817114857.2784825-3-chengzhihao1@huawei.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a946506c48 ]
The driver was registering IRQ 0 when no IRQ was set. This leads to
warnings with newer kernels.
Clear the resource flags, so no resource is registered at all in this
case.
Fixes: 2f17dd34ff ("mfd: tqmx86: IO controller with I2C, Wachdog and GPIO")
Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e3245a7b7b ]
Syzbot hit use-after-free in nf_tables_dump_sets. The problem was in
missing lock protection for nft_ct_pcpu_template_refcnt.
Before commit f102d66b33 ("netfilter: nf_tables: use dedicated
mutex to guard transactions") all transactions were serialized by global
mutex, but then global mutex was changed to local per netnamespace
commit_mutex.
This change causes use-after-free bug, when 2 netnamespaces concurently
changing nft_ct_pcpu_template_refcnt without proper locking. Fix it by
adding nft_ct_pcpu_mutex and protect all nft_ct_pcpu_template_refcnt
changes with it.
Fixes: f102d66b33 ("netfilter: nf_tables: use dedicated mutex to guard transactions")
Reported-and-tested-by: syzbot+649e339fa6658ee623d3@syzkaller.appspotmail.com
Signed-off-by: Pavel Skripkin <paskripkin@gmail.com>
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit daa3736151 ]
Remove interrupt disablement during backlight setting. It is
way to dangerous and makes platforms instable by having it
miss vblank IRQs leading to the graphics derailing.
The code is using ndelay() which is not available on
platforms such as ARM and will result in 32 * udelay(1)
which is substantial.
Add some code to detect if an interrupt occurs during the
tight loop and in that case just redo it from the top.
Fixes: 5317f37e48 ("backlight: Add Kinetic KTD253 backlight driver")
Cc: Stephan Gerhold <stephan@gerhold.net>
Reported-by: newbyte@disroot.org
Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit f949a9ebce ]
On Cherry Trail devices with an AXP288 PMIC the external SD-card slot
used the AXP's DLDO2 as card-voltage and either DLDO3 or GPIO1LDO
(GPIO1 pin in low noise LDO mode) as signal-voltage.
These regulators are turned on/off and in case of the signal-voltage
also have their output-voltage changed by the _PS0 and _PS3 power-
management ACPI methods on the MMC-controllers ACPI fwnode as well as
by the _DSM ACPI method for changing the signal voltage.
The AML code implementing these methods is directly accessing the
PMIC through ACPI I2C OpRegion accesses, instead of using the special
PMIC OpRegion handled by drivers/acpi/pmic/intel_pmic_xpower.c .
This means that the contents of the involved PMIC registers can change
without the change being made through the regmap interface, so regmap
should not cache the contents of these registers.
Mark the regulator power on/off, the regulator voltage control and the
GPIO1 control registers as volatile, to avoid regmap caching them.
Specifically this fixes an issue on some models where the i915 driver
toggles another LDO using the same on/off register on/off through
MIPI sequences (through intel_soc_pmic_exec_mipi_pmic_seq_element())
which then writes back a cached on/off register-value where the
card-voltage is off causing the external sdcard slot to stop working
when the screen goes blank, or comes back on again.
The regulator register-range now marked volatile also includes the
buck regulator control registers. This is done on purpose these are
normally not touched by the AML code, but they are updated directly
by the SoC's PUNIT which means that they may also change without going
through regmap.
Note the AXP288 PMIC is only used on Bay- and Cherry-Trail platforms,
so even though this is an ACPI specific problem there is no need to
make the new volatile ranges conditional since these platforms always
use ACPI.
Fixes: dc91c3b6fe ("mfd: axp20x: Mark AXP20X_VBUS_IPSOUT_MGMT as volatile")
Fixes: cd53216625 ("mfd: axp20x: Fix axp288 volatile ranges")
Reported-and-tested-by: Clamshell <clamfly@163.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 1511df6f5e upstream.
EMIT6_PCREL() macro assumes that the previous pass generated 6 bytes
of code, which is not the case if branch shortening took place. Fix by
using jit->prg, like all the other EMIT6_PCREL_*() macros.
Reported-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Fixes: 4e9b4a6883 ("s390/bpf: Use relative long branches")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6e61dc9da0 upstream.
The JIT uses agfi for subtracting constants, but -(-0x80000000) cannot
be represented as a 32-bit signed binary integer. Fix by using algfi in
this particular case.
Reported-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Fixes: 0546231057 ("s390/bpf: Add s390x eBPF JIT compiler backend")
Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit db7bee6538 upstream.
Currently the JIT completely removes things like `reg32 += 0`,
however, the BPF_ALU semantics requires the target register to be
zero-extended in such cases.
Fix by optimizing out only the arithmetic operation, but not the
subsequent zero-extension.
Reported-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Fixes: 0546231057 ("s390/bpf: Add s390x eBPF JIT compiler backend")
Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 0097ae5f7a ]
When the function IS_ALIGNED() returns false, the value of ret is 0.
So, we set ret to -EINVAL to indicate this error.
Clean up smatch warning:
drivers/ntb/test/ntb_perf.c:602 perf_setup_inbuf() warn: missing error
code 'ret'.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Reviewed-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 319f83ac98 ]
When the value of nm->isr_ctx is false, the value of ret is 0.
So, we set ret to -ENOMEM to indicate this error.
Clean up smatch warning:
drivers/ntb/test/ntb_msi_test.c:373 ntb_msit_probe() warn: missing
error code 'ret'.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Reviewed-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7db8263a12 ]
When adapter->registered_device_map is NULL, the value of err is
uncertain, we set err to -EINVAL to avoid ambiguity.
Clean up smatch warning:
drivers/net/ethernet/chelsio/cxgb/cxgb2.c:1114 init_one() warn: missing
error code 'err'
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 2d52c58b9c ]
The function bfq_setup_merge prepares the merging between two
bfq_queues, say bfqq and new_bfqq. To this goal, it assigns
bfqq->new_bfqq = new_bfqq. Then, each time some I/O for bfqq arrives,
the process that generated that I/O is disassociated from bfqq and
associated with new_bfqq (merging is actually a redirection). In this
respect, bfq_setup_merge increases new_bfqq->ref in advance, adding
the number of processes that are expected to be associated with
new_bfqq.
Unfortunately, the stable-merging mechanism interferes with this
setup. After bfqq->new_bfqq has been set by bfq_setup_merge, and
before all the expected processes have been associated with
bfqq->new_bfqq, bfqq may happen to be stably merged with a different
queue than the current bfqq->new_bfqq. In this case, bfqq->new_bfqq
gets changed. So, some of the processes that have been already
accounted for in the ref counter of the previous new_bfqq will not be
associated with that queue. This creates an unbalance, because those
references will never be decremented.
This commit fixes this issue by reestablishing the previous, natural
behaviour: once bfqq->new_bfqq has been set, it will not be changed
until all expected redirections have occurred.
Signed-off-by: Davide Zini <davidezini2@gmail.com>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Link: https://lore.kernel.org/r/20210802141352.74353-2-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit aabbdc67f3 ]
Add quirk CDC_MBIM_FLAG_AVOID_ALTSETTING_TOGGLE for Telit LN920
0x1061 composition in order to avoid bind error.
Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 21274aa178 ]
Check one more time before exiting the API with an error.
Fix API to poll at least twice, in case there are other high priority
tasks and this API doesn't get CPU cycles for multiple jiffies update.
In addition, increase timeout from usecs_to_jiffies(10000) to
usecs_to_jiffies(20000), to prevent the case that for CONFIG_100HZ
timeout will be a single jiffies.
A single jiffies results actual timeout that can be any time between
1usec and 10msec. To solve this, a value of usecs_to_jiffies(20000)
ensures that timeout is 2 jiffies.
Signed-off-by: Smadar Fuks <smadarf@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit dbe80cf471 ]
We must not pet a running watchdog when handle_boot_enabled is off
because this will kick off automatic triggering before userland is
running, defeating the purpose of the handle_boot_enabled control.
Furthermore, don't ping in case watchdog_set_last_hw_keepalive was
called incorrectly when the hardware watchdog is actually not running.
Fixed: cef9572e9a ("watchdog: add support for adjusting last known HW keepalive time")
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/93d56386-6e37-060b-55ce-84de8cde535f@web.de
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 32837d8a8f ]
Some Cavium endpoints are implemented as multi-function devices without ACS
capability, but they actually don't support peer-to-peer transactions.
Add ACS quirks to declare DMA isolation for the following devices:
- BGX device found on Octeon-TX (8xxx)
- CGX device found on Octeon-TX2 (9xxx)
- RPM device found on Octeon-TX3 (10xxx)
Link: https://lore.kernel.org/r/20210810122425.1115156-1-george.cherian@marvell.com
Signed-off-by: George Cherian <george.cherian@marvell.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit f1de58802f ]
J7200 has the same PCIe IP as in J721E with minor changes in the
wrapper. J7200 allows byte access of bridge configuration space
registers and the register field for LINK_DOWN interrupt is different.
J7200 also requires "quirk_detect_quiet_flag" to be set. Configure these
changes as part of driver data applicable only to J7200.
Link: https://lore.kernel.org/r/20210811123336.31357-4-kishon@ti.com
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 09c24094b2 ]
PCIe fails to link up if SERDES lanes not used by PCIe are assigned to
another protocol. For example, link training fails if lanes 2 and 3 are
assigned to another protocol while lanes 0 and 1 are used for PCIe to
form a two lane link. This failure is due to an incorrect tie-off on an
internal status signal indicating electrical idle.
Status signals going from SERDES to PCIe Controller are tied-off when a
lane is not assigned to PCIe. Signal indicating electrical idle is
incorrectly tied-off to a state that indicates non-idle. As a result,
PCIe sees unused lanes to be out of electrical idle and this causes
LTSSM to exit Detect.Quiet state without waiting for 12ms timeout to
occur. If a receiver is not detected on the first receiver detection
attempt in Detect.Active state, LTSSM goes back to Detect.Quiet and
again moves forward to Detect.Active state without waiting for 12ms as
required by PCIe base specification. Since wait time in Detect.Quiet is
skipped, multiple receiver detect operations are performed back-to-back
without allowing time for capacitance on the transmit lines to
discharge. This causes subsequent receiver detection to always fail even
if a receiver gets connected eventually.
Add a quirk flag "quirk_detect_quiet_flag" to program the minimum
time the LTSSM should wait on entering Detect.Quiet state here.
This has to be set for J7200 as it has an incorrect tie-off on unused
lanes.
Link: https://lore.kernel.org/r/20210811123336.31357-3-kishon@ti.com
Signed-off-by: Nadeem Athani <nadeem@cadence.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8e242060c6 ]
Since kprobe_events and uprobe_events only check whether the
other same-type probe event has the same name or not, if the
user gives the same name of the existing tracepoint event (or
the other type of probe events), it silently fails to create
the tracefs entry (but registered.) as below.
/sys/kernel/tracing # ls events/task/task_rename
enable filter format hist id trigger
/sys/kernel/tracing # echo p:task/task_rename vfs_read >> kprobe_events
[ 113.048508] Could not create tracefs 'task_rename' directory
/sys/kernel/tracing # cat kprobe_events
p:task/task_rename vfs_read
To fix this issue, check whether the existing events have the
same name or not in trace_probe_register_event_call(). If exists,
it rejects to register the new event.
Link: https://lkml.kernel.org/r/162936876189.187130.17558311387542061930.stgit@devnote2
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 9ff80e2de3 ]
Although irq_create_mapping() is able to deal with duplicate
mappings, it really isn't supposed to be a substitute for
irq_find_mapping(), and can result in allocations that take place
in atomic context if the mapping didn't exist.
Fix the handful of MFD drivers that use irq_create_mapping() in
interrupt context by using irq_find_mapping() instead.
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Lee Jones <lee.jones@linaro.org>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e1e71c1688 ]
There is a potential race between fuse_read_interrupt() and
fuse_request_end().
TASK1
in fuse_read_interrupt(): delete req->intr_entry (while holding
fiq->lock)
TASK2
in fuse_request_end(): req->intr_entry is empty -> skip fiq->lock
wake up TASK3
TASK3
request is freed
TASK1
in fuse_read_interrupt(): dereference req->in.h.unique ***BAM***
Fix by always grabbing fiq->lock if the request was ever interrupted
(FR_INTERRUPTED set) thereby serializing with concurrent
fuse_read_interrupt() calls.
FR_INTERRUPTED is set before the request is queued on fiq->interrupts.
Dequeing the request is done with list_del_init() but FR_INTERRUPTED is not
cleared in this case.
Reported-by: lijiazi <lijiazi@xiaomi.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit ec343111c0 ]
These are the actual frequencies reported by the PLL, so let's
report these. The roundoffs are inappropriate, we should round
to the frequency that the clock will later report.
Drop some whitespace at the same time.
Cc: phone-devel@vger.kernel.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 427900d27d upstream.
Currently, the VF does not clear the interrupt source immediately after
receiving the interrupt. As a result, if the second interrupt task is
triggered when processing the first interrupt task, clearing the
interrupt source before exiting will clear the interrupt sources of the
two tasks at the same time. As a result, no interrupt is triggered for
the second task. The VF detects the missed message only when the next
interrupt is generated.
Clearing it immediately after executing check_evt_cause ensures that:
1. Even if two interrupt tasks are triggered at the same time, they can
be processed.
2. If the second task is triggered during the processing of the first
task and the interrupt source is not cleared, the interrupt is reported
after vector0 is enabled.
Fixes: b90fcc5bd9 ("net: hns3: add reset handling for VF when doing Core/Global/IMP reset")
Signed-off-by: Jiaran Zhang <zhangjiaran@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b81d894874 upstream.
The firmware will not disable mac in flr process. Therefore, the driver
needs to proactively disable mac during flr, which is the same as the
function reset.
Fixes: 35d93a3004 ("net: hns3: adjust the process of PF reset")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1dc839ec09 upstream.
Currently, affinity_mask is set to a single cpu. As a result,
irqbalance becomes invalid in SUBSET or EXACT mode. To solve
this problem, change affinity_mask to numa node range. In this
way, irqbalance can be performed on the cpu of the numa node.
Fixes: 0812545487 ("net: hns3: add interrupt affinity support for misc interrupt")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d18e81183b upstream.
The hardware cannot handle short tunnel frames below 65 bytes,
and will cause vlan tag missing problem. So pads packet size to
65 bytes for tunnel frames to fix this bug.
Fixes: 3db084d28dc0("net: hns3: Fix for vxlan tx checksum bug")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1affc01fdc upstream.
The call to bnxt_free_mem(..., false) in the bnxt_half_open_nic() error
path will deallocate ring descriptor memory via bnxt_free_?x_rings(),
but because irq_re_init is false, the ring info itself is not freed.
To simplify error paths, deallocation functions have generally been
written to be safe when called on unallocated memory. It should always
be safe to call dev_close(), which calls bnxt_free_skbs() a second time,
even in this semi- allocated ring state.
Calling bnxt_free_skbs() a second time with the rings already freed will
cause NULL pointer dereference. Fix it by checking the rings are valid
before proceeding in bnxt_free_tx_skbs() and
bnxt_free_one_rx_ring_skbs().
Fixes: 975bc99a4a ("bnxt_en: Refactor bnxt_free_rx_skbs().")
Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 267cdfa213 upstream.
POWER9 DD2.2 and 2.3 hardware implements a "fake-suspend" mode where
certain TM instructions executed in HV=0 mode cause softpatch interrupts
so the hypervisor can emulate them and prevent problematic processor
conditions. In this fake-suspend mode, the treclaim. instruction does
not modify registers.
Unfortunately the rfscv instruction executed by the guest do not
generate softpatch interrupts, which can cause the hypervisor to lose
track of the fake-suspend mode, and it can execute this treclaim. while
not in fake-suspend mode. This modifies GPRs and crashes the hypervisor.
It's not trivial to disable scv in the guest with HFSCR now, because
they assume a POWER9 has scv available. So this fix saves and restores
checkpointed registers across the treclaim.
Fixes: 7854f7545b ("KVM: PPC: Book3S: Rework TM save/restore code and make it C-callable")
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210908101718.118522-2-npiggin@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 273c29e944 upstream.
If a failover occurs before a login response is received, the login
response buffer maybe undefined. Check that there was no failover
before accessing the login response buffer.
Fixes: 032c5e8284 ("Driver for IBM System i/p VNIC protocol")
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e50e711351 upstream.
Turn udp_tunnel_nic work-queue to an ordered work-queue. This queue
holds the UDP-tunnel configuration commands of the different netdevs.
When the netdevs are functions of the same NIC the order of
execution may be crucial.
Problem example:
NIC with 2 PFs, both PFs declare offload quota of up to 3 UDP-ports.
$ifconfig eth2 1.1.1.1/16 up
$ip link add eth2_19503 type vxlan id 5049 remote 1.1.1.2 dev eth2 dstport 19053
$ip link set dev eth2_19503 up
$ip link add eth2_19504 type vxlan id 5049 remote 1.1.1.3 dev eth2 dstport 19054
$ip link set dev eth2_19504 up
$ip link add eth2_19505 type vxlan id 5049 remote 1.1.1.4 dev eth2 dstport 19055
$ip link set dev eth2_19505 up
$ip link add eth2_19506 type vxlan id 5049 remote 1.1.1.5 dev eth2 dstport 19056
$ip link set dev eth2_19506 up
NIC RX port offload infrastructure offloads the first 3 UDP-ports (on
all devices which sets NETIF_F_RX_UDP_TUNNEL_PORT feature) and not
UDP-port 19056. So both PFs gets this offload configuration.
$ip link set dev eth2_19504 down
This triggers udp-tunnel-core to remove the UDP-port 19504 from
offload-ports-list and offload UDP-port 19056 instead.
In this scenario it is important that the UDP-port of 19504 will be
removed from both PFs before trying to add UDP-port 19056. The NIC can
stop offloading a UDP-port only when all references are removed.
Otherwise the NIC may report exceeding of the offload quota.
Fixes: cc4e3835ef ("udp_tunnel: add central NIC RX port offload infrastructure")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 20e100f527 upstream.
Handle MFW (management FW) error response in order to avoid a crash
during recovery flows.
Changes from v1:
- Add "Fixes tag".
Fixes: tag 5e7ba042fd ("qed: Fix reading stale configuration information")
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: Shai Malin <smalin@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1b704b27be upstream.
If altname deletion of the short alternative name fails, the error
message printed is: "Failed to add short alternative name".
This is obviously a typo, as we are testing altname deletion.
Fix this using a proper error message.
Fixes: f95e6c9c46 ("selftest: net: add alternative names test")
Signed-off-by: Andrea Claudi <aclaudi@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 4f884f3962 upstream.
Commit 10d3be5692 ("tcp-tso: do not split TSO packets at retransmit
time") may directly retrans a multiple segments TSO/GSO packet without
split, Since this commit, we can no longer assume that a retransmitted
packet is a single segment.
This patch fixes the tp->undo_retrans accounting in tcp_sacktag_one()
that use the actual segments(pcount) of the retransmitted packet.
Before that commit (10d3be5692), the assumption underlying the
tp->undo_retrans-- seems correct.
Fixes: 10d3be5692 ("tcp-tso: do not split TSO packets at retransmit time")
Signed-off-by: zhenggy <zhenggy@chinatelecom.cn>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a69ae291e1 upstream.
Commit 865c50e1d2 ("x86/uaccess: utilize CONFIG_CC_HAS_ASM_GOTO_OUTPUT")
added an optimised version of __get_user_asm() for x86 using 'asm goto'.
Like the non-optimised code, the 32-bit implementation of 64-bit
get_user() expands to a pair of 32-bit accesses. Unlike the
non-optimised code, the _original_ pointer is incremented to copy the
high word instead of loading through a new pointer explicitly
constructed to point at a 32-bit type. Consequently, if the pointer
points at a 64-bit type then we end up loading the wrong data for the
upper 32-bits.
This was observed as a mount() failure in Android targeting i686 after
b0cfcdd9b9 ("d_path: make 'prepend()' fill up the buffer exactly on
overflow") because the call to copy_from_kernel_nofault() from
prepend_copy() ends up in __get_kernel_nofault() and casts the source
pointer to a 'u64 __user *'. An attempt to mount at "/debug_ramdisk"
therefore ends up failing trying to mount "/debumdismdisk".
Use the existing '__gu_ptr' source pointer to unsigned int for 32-bit
__get_user_asm_u64() instead of the original pointer.
Cc: Bill Wendling <morbo@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Reported-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixes: 865c50e1d2 ("x86/uaccess: utilize CONFIG_CC_HAS_ASM_GOTO_OUTPUT")
Signed-off-by: Will Deacon <will@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6a52e73368 upstream.
DSA supports connecting to a phy-handle, and has a fallback to a non-OF
based method of connecting to an internal PHY on the switch's own MDIO
bus, if no phy-handle and no fixed-link nodes were present.
The -ENODEV error code from the first attempt (phylink_of_phy_connect)
is what triggers the second attempt (phylink_connect_phy).
However, when the first attempt returns a different error code than
-ENODEV, this results in an unbalance of calls to phylink_create and
phylink_destroy by the time we exit the function. The phylink instance
has leaked.
There are many other error codes that can be returned by
phylink_of_phy_connect. For example, phylink_validate returns -EINVAL.
So this is a practical issue too.
Fixes: aab9c4067d ("net: dsa: Plug in PHYLINK support")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/20210914134331.2303380-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 04f08eb44b upstream.
syzbot reported another data-race in af_unix [1]
Lets change __skb_insert() to use WRITE_ONCE() when changing
skb head qlen.
Also, change unix_dgram_poll() to use lockless version
of unix_recvq_full()
It is verry possible we can switch all/most unix_recvq_full()
to the lockless version, this will be done in a future kernel version.
[1] HEAD commit: 8596e589b7
BUG: KCSAN: data-race in skb_queue_tail / unix_dgram_poll
write to 0xffff88814eeb24e0 of 4 bytes by task 25815 on cpu 0:
__skb_insert include/linux/skbuff.h:1938 [inline]
__skb_queue_before include/linux/skbuff.h:2043 [inline]
__skb_queue_tail include/linux/skbuff.h:2076 [inline]
skb_queue_tail+0x80/0xa0 net/core/skbuff.c:3264
unix_dgram_sendmsg+0xff2/0x1600 net/unix/af_unix.c:1850
sock_sendmsg_nosec net/socket.c:703 [inline]
sock_sendmsg net/socket.c:723 [inline]
____sys_sendmsg+0x360/0x4d0 net/socket.c:2392
___sys_sendmsg net/socket.c:2446 [inline]
__sys_sendmmsg+0x315/0x4b0 net/socket.c:2532
__do_sys_sendmmsg net/socket.c:2561 [inline]
__se_sys_sendmmsg net/socket.c:2558 [inline]
__x64_sys_sendmmsg+0x53/0x60 net/socket.c:2558
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
read to 0xffff88814eeb24e0 of 4 bytes by task 25834 on cpu 1:
skb_queue_len include/linux/skbuff.h:1869 [inline]
unix_recvq_full net/unix/af_unix.c:194 [inline]
unix_dgram_poll+0x2bc/0x3e0 net/unix/af_unix.c:2777
sock_poll+0x23e/0x260 net/socket.c:1288
vfs_poll include/linux/poll.h:90 [inline]
ep_item_poll fs/eventpoll.c:846 [inline]
ep_send_events fs/eventpoll.c:1683 [inline]
ep_poll fs/eventpoll.c:1798 [inline]
do_epoll_wait+0x6ad/0xf00 fs/eventpoll.c:2226
__do_sys_epoll_wait fs/eventpoll.c:2238 [inline]
__se_sys_epoll_wait fs/eventpoll.c:2233 [inline]
__x64_sys_epoll_wait+0xf6/0x120 fs/eventpoll.c:2233
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
value changed: 0x0000001b -> 0x00000001
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 25834 Comm: syz-executor.1 Tainted: G W 5.14.0-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Fixes: 86b18aaa2b ("skbuff: fix a data race in skb_queue_len()")
Cc: Qian Cai <cai@lca.pw>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3c4cea8fa7 upstream.
If the sendmsg() call in vhost_tx_batch() fails, both the 'batched_xdp'
and 'done_idx' indexes are left unchanged. If such failure happens
when batched_xdp == VHOST_NET_BATCH, the next call to
vhost_net_build_xdp() will access and write memory outside the xdp
buffers area.
Since sendmsg() can only error with EBADFD, this change addresses the
issue explicitly freeing the XDP buffers batch on error.
Fixes: 0a0be13b8f ("vhost_net: batch submitting XDP buffers to underlayer sockets")
Suggested-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ec783c7cb2 upstream.
We need to import the 'sys' package since the script has called
sys.exit() method.
Fixes: 6ad7cbc015 ("Makefile: Add clang-tidy and static analyzer support to makefile")
Signed-off-by: Kortan <kortanzh@gmail.com>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b5c102238c upstream.
There is an off-by-one problem in ipa_table_init_add(), when
initializing filter tables.
In that function, the number of filter table entries is determined
based on the number of set bits in the filter map. However that
count does *not* include the extra "slot" in the filter table that
holds the filter map itself. Meanwhile, ipa_table_addr() *does*
include the filter map in the memory it returns, but because the
count it's provided doesn't include it, it includes one too few
table entries.
Fix this by including the extra slot for the filter map in the count
computed in ipa_table_init_add().
Note: ipa_filter_reset_table() does not have this problem; it resets
filter table entries one by one, but does not overwrite the filter
bitmap.
Fixes: 2b9feef2b6 ("soc: qcom: ipa: filter and routing tables")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b89a05b21f upstream.
In perf_event_addr_filters_apply, the task associated with
the event (event->ctx->task) is read using READ_ONCE at the beginning
of the function, checked, and then re-read from event->ctx->task,
voiding all guarantees of the checks. Reuse the value that was read by
READ_ONCE to ensure the consistency of the task struct throughout the
function.
Fixes: 375637bc52 ("perf/core: Introduce address range filtering")
Signed-off-by: Baptiste Lepers <baptiste.lepers@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210906015310.12802-1-baptiste.lepers@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 70f437fb43 upstream.
Dispatching requests inline with the .queue_rq() call may block while
holding the send_mutex. If the tcp io_work also happens to schedule, it
may see the req_list is non-empty, leaving "pending" true and remaining
in TASK_RUNNING. Since io_work is of higher scheduling priority, the
.queue_rq task may not get a chance to run, blocking forward progress
and leading to io timeouts.
Instead of checking for pending requests within io_work, let the queueing
restart io_work outside the send_mutex lock if there is more work to be
done.
Fixes: a0fdd14180 ("nvme-tcp: rerun io_work if req_list is not empty")
Reported-by: Samuel Jones <sjones@kalrayinc.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ee27e330a9 upstream.
Fixes the below flow of sleeping in atomic context by releasing
the RCU lock before calling to free_match_list.
build_match_list() <- disables preempt
-> free_match_list()
-> tree_put_node()
-> down_write_ref_node() <- take write lock
Fixes: 693c6883bb ("net/mlx5: Add hash table for flow groups in flow table")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 57f0ff059e upstream.
It's later supposed to be either a correct address or NULL. Without the
initialization, it may contain an undefined value which results in the
following segmentation fault:
# perf top --sort comm -g --ignore-callees=do_idle
terminates with:
#0 0x00007ffff56b7685 in __strlen_avx2 () from /lib64/libc.so.6
#1 0x00007ffff55e3802 in strdup () from /lib64/libc.so.6
#2 0x00005555558cb139 in hist_entry__init (callchain_size=<optimized out>, sample_self=true, template=0x7fffde7fb110, he=0x7fffd801c250) at util/hist.c:489
#3 hist_entry__new (template=template@entry=0x7fffde7fb110, sample_self=sample_self@entry=true) at util/hist.c:564
#4 0x00005555558cb4ba in hists__findnew_entry (hists=hists@entry=0x5555561d9e38, entry=entry@entry=0x7fffde7fb110, al=al@entry=0x7fffde7fb420,
sample_self=sample_self@entry=true) at util/hist.c:657
#5 0x00005555558cba1b in __hists__add_entry (hists=hists@entry=0x5555561d9e38, al=0x7fffde7fb420, sym_parent=<optimized out>, bi=bi@entry=0x0, mi=mi@entry=0x0,
sample=sample@entry=0x7fffde7fb4b0, sample_self=true, ops=0x0, block_info=0x0) at util/hist.c:288
#6 0x00005555558cbb70 in hists__add_entry (sample_self=true, sample=0x7fffde7fb4b0, mi=0x0, bi=0x0, sym_parent=<optimized out>, al=<optimized out>, hists=0x5555561d9e38)
at util/hist.c:1056
#7 iter_add_single_cumulative_entry (iter=0x7fffde7fb460, al=<optimized out>) at util/hist.c:1056
#8 0x00005555558cc8a4 in hist_entry_iter__add (iter=iter@entry=0x7fffde7fb460, al=al@entry=0x7fffde7fb420, max_stack_depth=<optimized out>, arg=arg@entry=0x7fffffff7db0)
at util/hist.c:1231
#9 0x00005555557cdc9a in perf_event__process_sample (machine=<optimized out>, sample=0x7fffde7fb4b0, evsel=<optimized out>, event=<optimized out>, tool=0x7fffffff7db0)
at builtin-top.c:842
#10 deliver_event (qe=<optimized out>, qevent=<optimized out>) at builtin-top.c:1202
#11 0x00005555558a9318 in do_flush (show_progress=false, oe=0x7fffffff80e0) at util/ordered-events.c:244
#12 __ordered_events__flush (oe=oe@entry=0x7fffffff80e0, how=how@entry=OE_FLUSH__TOP, timestamp=timestamp@entry=0) at util/ordered-events.c:323
#13 0x00005555558a9789 in __ordered_events__flush (timestamp=<optimized out>, how=<optimized out>, oe=<optimized out>) at util/ordered-events.c:339
#14 ordered_events__flush (how=OE_FLUSH__TOP, oe=0x7fffffff80e0) at util/ordered-events.c:341
#15 ordered_events__flush (oe=oe@entry=0x7fffffff80e0, how=how@entry=OE_FLUSH__TOP) at util/ordered-events.c:339
#16 0x00005555557cd631 in process_thread (arg=0x7fffffff7db0) at builtin-top.c:1114
#17 0x00007ffff7bb817a in start_thread () from /lib64/libpthread.so.0
#18 0x00007ffff5656dc3 in clone () from /lib64/libc.so.6
If you look at the frame #2, the code is:
488 if (he->srcline) {
489 he->srcline = strdup(he->srcline);
490 if (he->srcline == NULL)
491 goto err_rawdata;
492 }
If he->srcline is not NULL (it is not NULL if it is uninitialized rubbish),
it gets strdupped and strdupping a rubbish random string causes the problem.
Also, if you look at the commit 1fb7d06a50, it adds the srcline property
into the struct, but not initializing it everywhere needed.
Committer notes:
Now I see, when using --ignore-callees=do_idle we end up here at line
2189 in add_callchain_ip():
2181 if (al.sym != NULL) {
2182 if (perf_hpp_list.parent && !*parent &&
2183 symbol__match_regex(al.sym, &parent_regex))
2184 *parent = al.sym;
2185 else if (have_ignore_callees && root_al &&
2186 symbol__match_regex(al.sym, &ignore_callees_regex)) {
2187 /* Treat this symbol as the root,
2188 forgetting its callees. */
2189 *root_al = al;
2190 callchain_cursor_reset(cursor);
2191 }
2192 }
And the al that doesn't have the ->srcline field initialized will be
copied to the root_al, so then, back to:
1211 int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
1212 int max_stack_depth, void *arg)
1213 {
1214 int err, err2;
1215 struct map *alm = NULL;
1216
1217 if (al)
1218 alm = map__get(al->map);
1219
1220 err = sample__resolve_callchain(iter->sample, &callchain_cursor, &iter->parent,
1221 iter->evsel, al, max_stack_depth);
1222 if (err) {
1223 map__put(alm);
1224 return err;
1225 }
1226
1227 err = iter->ops->prepare_entry(iter, al);
1228 if (err)
1229 goto out;
1230
1231 err = iter->ops->add_single_entry(iter, al);
1232 if (err)
1233 goto out;
1234
That al at line 1221 is what hist_entry_iter__add() (called from
sample__resolve_callchain()) saw as 'root_al', and then:
iter->ops->add_single_entry(iter, al);
will go on with al->srcline with a bogus value, I'll add the above
sequence to the cset and apply, thanks!
Signed-off-by: Michael Petlan <mpetlan@redhat.com>
CC: Milian Wolff <milian.wolff@kdab.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Fixes: 1fb7d06a50 ("perf report Use srcline from callchain for hist entries")
Link: https //lore.kernel.org/r/20210719145332.29747-1-mpetlan@redhat.com
Reported-by: Juri Lelli <jlelli@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 040b8907cc upstream.
With the new static annotation, the compiler warns when the functions
are actually unused:
drivers/gpu/drm/rockchip/cdn-dp-core.c:1123:12: error: 'cdn_dp_resume' defined but not used [-Werror=unused-function]
1123 | static int cdn_dp_resume(struct device *dev)
| ^~~~~~~~~~~~~
Mark them __maybe_unused to suppress that warning as well.
[ Not so 'new' static annotations any more, and I removed the part of
the patch that added __maybe_unused to cdn_dp_suspend(), because it's
used by the shutdown/remove code.
So only the resume function ends up possibly unused if CONFIG_PM isn't
set - Linus ]
Fixes: 7c49abb4c2 ("drm/rockchip: cdn-dp-core: Make cdn_dp_core_suspend/resume static")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f4bb62e64c upstream.
In tipc_sk_enqueue() we use hardcoded 2 jiffies to extract
socket buffer from generic queue to particular socket.
The 2 jiffies is too short in case there are other high priority
tasks get CPU cycles for multiple jiffies update. As result, no
buffer could be enqueued to particular socket.
To solve this, we switch to use constant timeout 20msecs.
Then, the function will be expired between 2 jiffies (CONFIG_100HZ)
and 20 jiffies (CONFIG_1000HZ).
Fixes: c637c10355 ("tipc: resolve race problem at unicast message reception")
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e3f0cc1a94 upstream.
A number of users have reported that they were not able to get the PHY
to successfully link up, especially after commit c36757eb9d ("net:
phy: consider AN_RESTART status when reading link status") where we
stopped reading just BMSR, but we also read BMCR to determine the link
status.
Andrius at NetBSD did a wonderful job at debugging the problem
and found out that the MDIO bus clock frequency would be incorrectly set
back to its default value which would prevent the MDIO bus controller
from reading PHY registers properly. Back when we only read BMSR, if we
read all 1s, we could falsely indicate a link status, though in general
there is a cable plugged in, so this went unnoticed. After a second read
of BMCR was added, a wrong read will lead to the inability to determine
a link UP condition which is when it started to be visibly broken, even
if it was long before that.
The fix consists in restoring the value of the MD_CSR register that was
set prior to the MAC reset.
Link: http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=53494
Fixes: 90f750a81a ("r6040: consolidate MAC reset to its own function")
Reported-by: Andrius V <vezhlys@gmail.com>
Reported-by: Darek Strugacz <darek.strugacz@op.pl>
Tested-by: Darek Strugacz <darek.strugacz@op.pl>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9b6ff7eb66 upstream.
The reference count leak issue may take place in an error handling
path. If both conditions of tunnel->version == L2TP_HDR_VER_3 and the
return value of l2tp_v3_ensure_opt_in_linear is nonzero, the function
would directly jump to label invalid, without decrementing the reference
count of the l2tp_session object session increased earlier by
l2tp_tunnel_get_session(). This may result in refcount leaks.
Fix this issue by decrease the reference count before jumping to the
label invalid.
Fixes: 4522a70db7 ("l2tp: fix reading optional fields of L2TPv3")
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn>
Signed-off-by: Xin Xiong <xiongx18@fudan.edu.cn>
Signed-off-by: Xin Tan <tanxin.ctf@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d9ea761fdd upstream.
Commit 2677d20677 ("dccp: don't free ccid2_hc_tx_sock ...") fixed
a UAF but reintroduced CVE-2017-6074.
When the sock is cloned, two dccps_hc_tx_ccid will reference to the
same ccid. So one can free the ccid object twice from two socks after
cloning.
This issue was found by "Hadar Manor" as well and assigned with
CVE-2020-16119, which was fixed in Ubuntu's kernel. So here I port
the patch from Ubuntu to fix it.
The patch prevents cloned socks from referencing the same ccid.
Fixes: 2677d20677 ("dccp: don't free ccid2_hc_tx_sock ...")
Signed-off-by: Zhenpeng Lin <zplin@psu.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 7366c23ff4 upstream.
Building dp83640.c on arch/parisc/ produces a build warning for
PAGE0 being redefined. Since the macro is not used in the dp83640
driver, just make it a comment for documentation purposes.
In file included from ../drivers/net/phy/dp83640.c:23:
../drivers/net/phy/dp83640_reg.h:8: warning: "PAGE0" redefined
8 | #define PAGE0 0x0000
from ../drivers/net/phy/dp83640.c:11:
../arch/parisc/include/asm/page.h:187: note: this is the location of the previous definition
187 | #define PAGE0 ((struct zeropage *)__PAGE_OFFSET)
Fixes: cb646e2b02 ("ptp: Added a clock driver for the National Semiconductor PHYTER.")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Richard Cochran <richard.cochran@omicron.at>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Heiner Kallweit <hkallweit1@gmail.com>
Cc: Russell King <linux@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20210913220605.19682-1-rdunlap@infradead.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 7c3a0a018e upstream.
Remove the assert from the callback priv lookup function since it does
not require RTNL lock and is already protected by flow_indr_block_lock.
This will avoid warnings from being emitted to dmesg if the driver
registers its callback after an ingress qdisc was created for a
netdevice.
The warnings started after the following patch was merged:
commit 74fc4f8287 ("net: Fix offloading indirect devices dependency on qdisc order creation")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit cc19862ffe upstream.
syzbot reported an use-after-free crash:
BUG: KASAN: use-after-free in tipc_recvmsg+0xf77/0xf90 net/tipc/socket.c:1979
Call Trace:
tipc_recvmsg+0xf77/0xf90 net/tipc/socket.c:1979
sock_recvmsg_nosec net/socket.c:943 [inline]
sock_recvmsg net/socket.c:961 [inline]
sock_recvmsg+0xca/0x110 net/socket.c:957
tipc_conn_rcv_from_sock+0x162/0x2f0 net/tipc/topsrv.c:398
tipc_conn_recv_work+0xeb/0x190 net/tipc/topsrv.c:421
process_one_work+0x98d/0x1630 kernel/workqueue.c:2276
worker_thread+0x658/0x11f0 kernel/workqueue.c:2422
As Hoang pointed out, it was caused by skb_cb->bytes_read still accessed
after calling tsk_advance_rx_queue() to free the skb in tipc_recvmsg().
This patch is to fix it by accessing skb_cb->bytes_read earlier than
calling tsk_advance_rx_queue().
Fixes: f4919ff59c ("tipc: keep the skb in rcv queue until the whole data is read")
Reported-by: syzbot+e6741b97d5552f97c24d@syzkaller.appspotmail.com
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 34b1999da9 upstream.
Jiri Olsa reported a fault when running:
# cat /proc/kallsyms | grep ksys_read
ffffffff8136d580 T ksys_read
# objdump -d --start-address=0xffffffff8136d580 --stop-address=0xffffffff8136d590 /proc/kcore
/proc/kcore: file format elf64-x86-64
Segmentation fault
general protection fault, probably for non-canonical address 0xf887ffcbff000: 0000 [#1] SMP PTI
CPU: 12 PID: 1079 Comm: objdump Not tainted 5.14.0-rc5qemu+ #508
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.14.0-4.fc34 04/01/2014
RIP: 0010:kern_addr_valid
Call Trace:
read_kcore
? rcu_read_lock_sched_held
? rcu_read_lock_sched_held
? rcu_read_lock_sched_held
? trace_hardirqs_on
? rcu_read_lock_sched_held
? lock_acquire
? lock_acquire
? rcu_read_lock_sched_held
? lock_acquire
? rcu_read_lock_sched_held
? rcu_read_lock_sched_held
? rcu_read_lock_sched_held
? lock_release
? _raw_spin_unlock
? __handle_mm_fault
? rcu_read_lock_sched_held
? lock_acquire
? rcu_read_lock_sched_held
? lock_release
proc_reg_read
? vfs_read
vfs_read
ksys_read
do_syscall_64
entry_SYSCALL_64_after_hwframe
The fault happens because kern_addr_valid() dereferences existent but not
present PMD in the high kernel mappings.
Such PMDs are created when free_kernel_image_pages() frees regions larger
than 2Mb. In this case, a part of the freed memory is mapped with PMDs and
the set_memory_np_noalias() -> ... -> __change_page_attr() sequence will
mark the PMD as not present rather than wipe it completely.
Have kern_addr_valid() check whether higher level page table entries are
present before trying to dereference them to fix this issue and to avoid
similar issues in the future.
Stable backporting note:
------------------------
Note that the stable marking is for all active stable branches because
there could be cases where pagetable entries exist but are not valid -
see 9a14aefc1d ("x86: cpa, fix lookup_address"), for example. So make
sure to be on the safe side here and use pXY_present() accessors rather
than pXY_none() which could #GP when accessing pages in the direct map.
Also see:
c40a56a781 ("x86/mm/init: Remove freed kernel image areas from alias mapping")
for more info.
Reported-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Dave Hansen <dave.hansen@intel.com>
Tested-by: Jiri Olsa <jolsa@redhat.com>
Cc: <stable@vger.kernel.org> # 4.4+
Link: https://lkml.kernel.org/r/20210819132717.19358-1-rppt@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit aeef8b5089 upstream.
The end address passed to memtype_reserve() is handed directly to
sanitize_phys(). However, end is exclusive and sanitize_phys() expects
an inclusive address. If end falls at the end of the physical address
space, sanitize_phys() will return 0. This can result in drivers
failing to load, and the following warning:
WARNING: CPU: 26 PID: 749 at arch/x86/mm/pat.c:354 reserve_memtype+0x262/0x450
reserve_memtype failed: [mem 0x3ffffff00000-0xffffffffffffffff], req uncached-minus
Call Trace:
[<ffffffffa427b1f2>] reserve_memtype+0x262/0x450
[<ffffffffa42764aa>] ioremap_nocache+0x1a/0x20
[<ffffffffc04620a1>] mpt3sas_base_map_resources+0x151/0xa60 [mpt3sas]
[<ffffffffc0465555>] mpt3sas_base_attach+0xf5/0xa50 [mpt3sas]
---[ end trace 6d6eea4438db89ef ]---
ioremap reserve_memtype failed -22
mpt3sas_cm0: unable to map adapter memory! or resource not found
mpt3sas_cm0: failure at drivers/scsi/mpt3sas/mpt3sas_scsih.c:10597/_scsih_probe()!
Fix this by passing the inclusive end address to sanitize_phys().
Fixes: 510ee090ab ("x86/mm/pat: Prepare {reserve, free}_memtype() for "decoy" addresses")
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/x49o8a3pu5i.fsf@segfault.boston.devel.redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f2faea8b64 upstream.
When we forcefully evict a mapping from the the address space and thus the
MMU context, the MMU context is leaked, as the mapping no longer points to
it, so it doesn't get freed when the GEM object is destroyed. Add the
mssing context put to fix the leak.
Cc: stable@vger.kernel.org # 5.4
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Michael Walle <michael@walle.cc>
Tested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d6408538f0 upstream.
Move the refcount manipulation of the MMU context to the point where the
hardware state is programmed. At that point it is also known if a previous
MMU state is still there, or the state needs to be reprogrammed with a
potentially different context.
Cc: stable@vger.kernel.org # 5.4
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Michael Walle <michael@walle.cc>
Tested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8f3eea9d01 upstream.
The MMU state may be kept across a runtime suspend/resume cycle, as we
avoid a full hardware reset to keep the latency of the runtime PM small.
Don't pretend that the MMU state is lost in driver state. The MMU
context is pushed out when new HW jobs with a different context are
coming in. The only exception to this is when the GPU is unbound, in
which case we need to make sure to also free the last active context.
Cc: stable@vger.kernel.org # 5.4
Reported-by: Michael Walle <michael@walle.cc>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Michael Walle <michael@walle.cc>
Tested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 23e0f5a57d upstream.
While the DMA frontend can only be active when the MMU context is set, the
reverse isn't necessarily true, as the frontend can be stopped while the
MMU state is kept. Stop treating mmu_context being set as a indication that
the frontend is running and instead add a explicit property.
Cc: stable@vger.kernel.org # 5.4
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Michael Walle <michael@walle.cc>
Tested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit cda7532916 upstream.
The prev context is the MMU context at the time of the job
queueing in hardware. As a job might be queued multiple times
due to recovery after a GPU hang, we need to make sure to put
the stale prev MMU context from a prior queuing, to avoid the
reference and thus the MMU context leaking.
Cc: stable@vger.kernel.org # 5.4
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Michael Walle <michael@walle.cc>
Tested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e35ac9d0b5 upstream.
When we need a buffer for SVE register state we call sve_alloc() to make
sure that one is there. In order to avoid repeated allocations and frees
we keep the buffer around unless we change vector length and just memset()
it to ensure a clean register state. The function that deals with this
takes the task to operate on as an argument, however in the case where we
do a memset() we initialise using the SVE state size for the current task
rather than the task passed as an argument.
This is only an issue in the case where we are setting the register state
for a task via ptrace and the task being configured has a different vector
length to the task tracing it. In the case where the buffer is larger in
the traced process we will leak old state from the traced process to
itself, in the case where the buffer is smaller in the traced process we
will overflow the buffer and corrupt memory.
Fixes: bc0ee47603 ("arm64/sve: Core task context handling")
Cc: <stable@vger.kernel.org> # 4.15.x
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20210909165356.10675-1-broonie@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 16c8d2df7e upstream.
When setting up the next segment, we check what type the iter is and
handle it accordingly. However, when incrementing and processed amount
we do not, and both iter advance and addr/len are adjusted, regardless
of type. Split the increment side just like we do on the setup side.
Fixes: 4017eb91a9 ("io_uring: make loop_rw_iter() use original user supplied pointers")
Cc: stable@vger.kernel.org
Reported-by: Valentina Palmiotti <vpalmiotti@gmail.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6f93e834fa upstream.
The mount option max_inline ranges from 0 to the sectorsize (which is
now equal to page size). But we parse the mount options too early and
before the actual sectorsize is read from the superblock. So the upper
limit of max_inline is unaware of the actual sectorsize and is limited
by the temporary sectorsize 4096, even on a system where the default
sectorsize is 64K.
Fix this by reading the superblock sectorsize before the mount option
parse.
Reported-by: Alexander Tsvetkov <alexander.tsvetkov@oracle.com>
CC: stable@vger.kernel.org # 5.4+
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The fw_node pointer has already been retrieved, and using it allows
us to remove a downstream patch to the firmware driver.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This is a copy of the code and approach from our DAC+ driver.
It allows to probe (and activate) an optional TPA6130A2 headphone
amplifier. Updated email address.
Signed-off-by: Joerg Schambacher <joerg@hifiberry.com>
This is a copy of the approach from our DAC+ driver.
It allows to probe (and activate) an optional TPA6130A2 headphone
amplifier.
Signed-off-by: Joerg Schambacher <joerg@hifiberry.com>
A regression introduced in https://github.com/raspberrypi/linux/pull/3887
meant that if the newly scheduled transfer immediately returned data, and
the driver resubmitted a single URB after every transfer, then the effective
polling interval would end up being approx 1ms.
Use the larger of SCHEDULE_SLOP or the configured endpoint interval.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
Adds an overlay to configure the TinyDRM driver for ST7735R
based 160x128 and 128x128 (untested) displays such as the
Adafruit 1.8" display.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
This commit updates the imx519 driver to adverise support for embedded
data streams.
The imx519 sensor subdevice overloads the media pad to differentiate
between image stream (pad 0) and embedded data stream (pad 1) when
performing the v4l2_subdev_pad_ops functions.
Signed-off-by: Lee Jackson <info@arducam.com>
Adds a driver for the 16MPix IMX519 CSI2 sensor.
Whilst the sensor supports 2 or 4 CSI2 data lanes, this driver
currently only supports 2 lanes.
The following Bayer modes are currently available:
4656x3496 10-bit @ 10fps
3840x2160 10-bit (cropped) @ 21fps
2328x1748 10-bit (binned) @ 30fps
1920x1080 10-bit (cropped/binned) @ 60fps
1280x720 10-bit (cropped/binned) @ 120fps
Signed-off-by: Lee Jackson <info@arducam.com>
The Linux support for controlling card power via regulators appears to
be contentious. I would argue that the default behaviour is contrary to
the SDHCI spec - turning off the power writes a reserved value to the
SD Bus Voltage Select field of the Power Control Register, which
seems to kill the Arasan/iProc controller - but fortunately there is a
hook in sdhci_ops to override the behaviour.
Signed-off-by: Juerg Haefliger <juergh@canonical.com>
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This reverts commit aed19399a0.
Commit 6c92ae1e45 ("mmc: sdhci: Introduce sdhci_set_power_and_bus_voltage()")
introduced a generic helper that does the same thing so use that instead in
the following commit.
Signed-off-by: Juerg Haefliger <juergh@canonical.com>
Check for errors only if we actually tried to enable the bvb clock.
Fixes: 01a6d727b4 ("vc4/drm: hdmi: Handle case when bvb clock is null")
Signed-off-by: Juerg Haefliger <juergh@canonical.com>
When doing an asynchronous page flip (PAGE_FLIP ioctl with the
DRM_MODE_PAGE_FLIP_ASYNC flag set), the current code waits for the
possible GPU buffer being rendered through a call to
vc4_queue_seqno_cb().
On the BCM2835-37, the GPU driver is part of the vc4 driver and that
function is defined in vc4_gem.c to wait for the buffer to be rendered,
and once it's done, call a callback.
However, on the BCM2711 used on the RaspberryPi4, the GPU driver is
separate (v3d) and that function won't do anything. This was working
because we were going into a path, due to uninitialized variables, that
was always scheduling the callback.
However, we were never actually waiting for the buffer to be rendered
which was resulting in frames being displayed out of order.
The generic API to signal those kind of completion in the kernel are the
DMA fences, and fortunately the v3d drivers supports them and signal
when its job is done. That API also provides an equivalent function that
allows to have a callback being executed when the fence is signalled as
done.
Let's change our driver a bit to rely on the previous function for the
older SoCs, and on DMA fences for the BCM2711.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The read and write functions did not use the correct pointer offset
when dealing with an odd number of bytes after a DMA transfer. Also,
only handle the remaining odd bytes if the DMA transfer completed
successfully.
Submitted-by: @madimario (GitHub)
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The driver was implementing a get_brightness function that
tried to read back the PWM setting of the display to report
as the current brightness.
The controller on the display does not support that, therefore
we end up reporting a brightness of 0, and that confuses
systemd's backlight service.
Remove the hook so that the framework returns the current
brightness automatically.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
This allows using the video-i2c camera driver with MLX90640 thermal
infrared sensors, connected to Raspberry Pi. CONFIG_VIDEO_V4L2_I2C
has to be selected to use the camera.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
The ISP can do H & V flips whilst resizing or converting
the image, so expose that via V4L2_CID_[H|V]FLIP.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Currently the code was only setting some controls from
bcm2835_codec_set_ctrls, but it's simpler to use
v4l2_ctrl_handler_setup to avoid forgetting to adding new
controls to the list.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Completions don't reference count, so setting the completion
on the first buffer returned and then not reinitialising it
means that the flush function doesn't behave as intended.
Signal the completion when the last buffer is returned.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Should we go through the timeout failure case with port_disable
not returning all buffers for whatever reason, the
buffers_with_vpu counter gets left at a non-zero value, which
will cause reference counting issues should the instance be
reused.
Reset the count when the port is enabled again, but before
any buffers have been sent to the VPU.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
When a buffer is returned on a port that is disabled, return it
to the videobuf2 QUEUED state instead of DONE which returns it
to the client.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The firmware defaults to not stopping video decode if only the
pixel aspect ratio or colourspace change. V4L2 requires us
to stop decoding on any change, therefore tell the firmware
of the desire for this alternate behaviour.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
When a format changed event occurs, the spec says that it
triggers an implicit drain, and that needs to be signalled
via -EPIPE.
For BCM2835, the format changed event happens at the point
the format change occurs, so no further buffers exist from
before the resolution changed point. We therefore signal the
last buffer immediately.
We don't have a V4L2 available to us at this point, so set
the videobuf2 queue last_buffer_dequeued flag directly.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The list of includes was slightly over generic, and wasn't
in alphabetical order. Clean it up.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
With video decode we now enable both input and output ports on
the component. This means that buffers will get passed to the VPU
earlier than desired if they are queued befoer STREAMON.
Check that the queue is streaming before sending buffers to the VPU.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The V4L2 stateful video decoder API requires that you can STREAMON
on only the OUTPUT queue, feed in buffers, and wait for the
SOURCE_CHANGE event.
This requires that we enable the MMAL output port at the same time
as the input port, because the output port is the one that creates
the SOURCE_CHANGED event.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Before uniniting the decode context sync with the IRQ queues to ensure
that decode no longer has any buffers in use. This fixes a problem that
manifested as ffmpeg leaking CMA buffers when it did a stream off on
OUTPUT before CAPTURE, though in reality it was probably much more
dangerous than that.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Remove unused ctx state tracking variable and associated defines.
Their presence implies they might be used, but they aren't.
Signed-off-by: John Cox <jc@kynesim.co.uk>
With EDPWRDOWN set in idle, it must be cleared before checking for
ENERGYON going high, indicating that a link is being established.
The existing code allows 640ms for ENERGYON to go high, but on
Raspberry Pis that appears not to be enough, causing link detection
to fail.
Increase the polling timeout to 1500ms - with a polling interval of
10ms it shouldn't cause unnecessary delays.
See: https://github.com/raspberrypi/linux/issues/4393
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
V4L2 spec says that G/S/TRY_FMT IOCTLs should never return errors for
anything other than wrong buffer types. Improve the capture format
function such that this is so and unsupported values get converted
to supported ones properly.
Signed-off-by: John Cox <jc@kynesim.co.uk>
A failure of some CPU cores to come online has been traced to the
failure of a stm instruction while the cache is disabled. The symptom
is that the saved values read back as zeroes, a catastrophic error since
one of the values is a return address.
This patch forces a readback and retry until the correct value is
returned,
Notes:
At this stage in the boot process the core is running with its cache
disabled. Before enabling the cache its contents must be explicitly
invalidated, a process that requires quite a few registers that the
caller must preserve. Evidence suggests that something is writing a
block of zeroes over that space at a time when all other cores should
be idle, possibly some kind of write-combiner, and retrying is an
attempt to avoid the problem.
The previous attempted fix (forcing the accesses to only be 4-byte
aligned) appears to have only worked for a while and likely for less
obvious reasons such as a change in code alignment.
See: https://github.com/Hexxeh/rpi-firmware/issues/232
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This reverts commit fe4cc0ed59.
The speculative patch that this commit reverts is proving to not be
effective any more, so revert it and try a new approach.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
As we're wanting to wrap the image_fx component for deinterlacing,
add the deinterlace algorithm values to enum mmal_parameter_imagefx
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If the client provides a bytesperline value in try_fmt/s_fmt then
validate it and correct if necessary.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Should start_streaming fail, or buffers be queued during
stop_streaming, they should be returned to the core as QUEUED
and not (as currently) as ERROR.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The video decoder can support decoding interlaced streams, so add
the required plumbing to signal this correctly.
The encoder and ISP do NOT support interlaced data, so trying to
configure an interlaced format on those nodes will be rejected.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Adds enum mmal_interlace_type and struct
mmal_parameter_video_interlace_type to allow for querying the
interlacing mode on decoders.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
In order to effectively guarantee that a V4L2_EVENT_SOURCE_CHANGE
event occurs, adopt a default resolution of 32x32 so that it
is incredibly unlikely to be decoding a stream of that resolution
and therefore failing to note a "change" requiring the event.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The exit path of vc4_hdmi_encoder_pre_crtc_configure() is fairly hard to
maintain given its numerous error conditions.
Switch to a goto based approach to simplify it.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Unlike pm_runtime_get_sync(), pm_runtime_resume_and_get() doesn't take a
reference on failure, so we don't need to call pm_runtime_put() on
failure.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Our detect function calls pm_runtime_resume_and_get() and
pm_runtime_put() to make sure the device is properly powered before
trying to access the controller.
However, it also makes sure the HSM clock is properly enabled (and
disabled), which is redundant with what runtime_pm is doing already.
Let's just remove it.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The older panel-raspberrypi-touchscreen driver had issues in
that it also controlled the power for the touchscreen without
having an appropriate hook for the touchscreen driver to control
that.
Mainline has now added a Toshiba TC358762 bridge driver, and
a regulator/backlight driver for the ATTiny microcontroller on
the board. That allows clean integration with the touchscreen
driver.
Switch the overlays over to using newer drivers.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
We need independent control of the resets for the panel&bridge,
vs the touch controller.
Expose the reset lines that are on the Atmel's port C via the GPIO
API so that they can be controlled appropriately.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The Atmel was doing a load of automatic sequencing of
control lines, however it was combining the touch controller's
reset with the bridge/panel control.
Change to control the control signals directly rather than
through the automatic POWERON control.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The initial state of the Atmel is not defined, so ensure the
backlight PWM is set to 0 by default.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The driver was using the regmap lock to serialise the
individual accesses, but we really need to protect the
timings of enabling the regulators, including any communication
with the Atmel.
Use a mutex within the driver to control overall accesses to
the Atmel, instead of the regmap lock.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The Atmel is doing some things in the I2C ISR, during which
period it will not respond to further commands. This is
particularly true of the POWERON command.
Increase delays appropriately, and retry should I2C errors be
reported.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
There's no reason why 2 Raspberry Pi DSI displays can't be
attached to a Pi Compute Module, so the backlight names need to
be unique.
Use the parent dev_name. It's not as readable, but is unique.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If no interrupt is defined then a timer and workqueue are used
to poll the controller.
On remove these were not being cleaned up correctly.
Fixes: ca61fdaba7 "Input: edt-ft5x06: Poll the device if no interrupt is
configured."
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The Raspberry Pi 7" 800x480 panel uses a Toshiba TC358762 DSI
to DPI bridge chip, so there is a requirement for the timings
to be specified for the end panel. Add such a definition.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
rpi_touchscreen_i2c_read returns any errors from i2c_transfer,
or the 8 bit received value.
Check for error values before trying to process the data as
valid.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The panel has a prepare call which is before video starts, and an
enable call which is after.
The Toshiba bridge should be configured before video, so move
the relevant power and initialisation calls to prepare.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If a call to rpi_touchscreen_i2c_write from rpi_touchscreen_probe
fails before mipi_dsi_device_register_full is called, then
in trying to log the error message if uses ts->dsi->dev when
it is still NULL.
Use ts->i2c->dev instead, which is initialised earlier in probe.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The divider calculations tried to find the divider
just faster than the clock requested. However if
it required a divider of 7 then the for loop
aborted without handling the "error" case, and could
end up with a clock lower than requested.
Correct the loop so that we always have a clock greater
than requested.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
On Pi0-3 the driver allocates a buffer and requests a DMA channel
because the ARM can't write to DSI1's registers directly.
However unbind and the error paths in bind don't release the buffer or
the DMA channel.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The HDMI block can repeat pixels for double clocked modes,
and the firmware is now configuring the block to do this as
the PV is doing it incorrectly when at 2pixels/clock.
If the kernel doesn't reset it then we end up with strange
modes.
Reset MISC_CONTROL.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The sysfs delay_ms value is calculated live, and it is possible for
the time left to appear to be negative briefly if the timer handling
hasn't completed. Ensure the displayed value never goes below zero,
for the sake of appearances.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
A module parameter "dpc_enable" is added to allow the control of the
sensor's on-board DPC (Defective Pixel Correction) function.
This is a global setting to be configured before using the sensor;
there is no intention that this would ever be changed on-the-fly.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
drm_connector_helper_hpd_irq_event() calls
drm_kms_helper_hotplug_event() with the mode-setting lock taken while
it's supposed to be called without that lock taken.
This results in a lockdep warning, and a deadlock if we were to wake up
a TV through CEC (and possibly other cases).
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
This adds an overlay to connect the QCA7000 in UART mode via UART0.
The qcauart driver uses the serial device bus instead of deprecated
line disciplines.
Signed-off-by: Stefan Wahren <stefan.wahren@in-tech.com>
When the firmware doesn't setup the HSM rate (such as when booting
without an HDMI cable plugged in), its rate is 0 and thus any register
access results in a CPU stall, even though HSM is enabled.
Let's enforce a minimum rate at boot to avoid this issue.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Commit 9d44abbbb8 ("drm/vc4: Fall back to using an EDID probe in the
absence of a GPIO.") added some code to read the EDID through DDC in the
HDMI driver detect hook since the Pi3 had no HPD GPIO back then.
However, commit b1b8f45b31 ("ARM: dts: bcm2837: Add missing GPIOs of
Expander") changed that a couple of years later.
This causes an issue though since some TV (like the LG 55C8) when it
comes out of standy will deassert the HPD line, but the EDID will
remain readable.
It causes an issues nn platforms without an HPD GPIO, like the Pi4,
where the DDC probing will be our primary mean to detect a display, and
thus we will never detect the HPD pulse. This was fine before since the
pulse was small enough that we would never detect it, and we also didn't
have anything (like the scrambler) that needed to be set up in the
display.
However, now that we have both, the display during the HPD pulse will
clear its scrambler status, and since we won't detect the
disconnect/reconnect cycle we will never enable the scrambler back.
As our main reason for that DDC probing is gone, let's just remove it.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The drm_helper_hpd_irq_event() documentation states that this function
is "useful for drivers which can't or don't track hotplug interrupts for
each connector." and that "Drivers which support hotplug interrupts for
each connector individually and which have a more fine-grained detect
logic should bypass this code and directly call
drm_kms_helper_hotplug_event()". This is thus what we ended-up doing.
However, what this actually means, and is further explained in the
drm_kms_helper_hotplug_event() documentation, is that
drm_kms_helper_hotplug_event() should be called by drivers that can
track the connection status change, and if it has changed we should call
that function.
This underlying expectation we failed to provide is that the caller of
drm_kms_helper_hotplug_event() should call drm_helper_probe_detect() to
probe the new status of the connector.
Since we didn't do it, it meant that even though we were sending the
notification to user-space and the DRM clients that something changed we
never probed or updated our internal connector status ourselves.
This went mostly unnoticed since the detect callback usually doesn't
have any side-effect. Also, if we were using the DRM fbdev emulation
(which is a DRM client), or any user-space application that can deal
with hotplug events, chances are they would react to the hotplug event
by probing the connector status eventually.
However, now that we have to enable the scrambler in detect() if it was
enabled it has a side effect, and an application such as Kodi or
modetest doesn't deal with hotplug events. This resulted with a black
screen when Kodi or modetest was running when a screen was disconnected
and then reconnected, or switched off and on.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The drm_helper_hpd_irq_event() function is iterating over all the
connectors when an hotplug event is detected.
During that iteration, it will call each connector detect function and
figure out if its status changed.
Finally, if any connector changed, it will notify the user-space and the
clients that something changed on the DRM device.
This is supposed to be used for drivers that don't have a hotplug
interrupt for individual connectors. However, drivers that can use an
interrupt for a single connector are left in the dust and can either
reimplement the logic used during the iteration for each connector or
use that helper and iterate over all connectors all the time.
Since both are suboptimal, let's create a helper that will only perform
the status detection on a single connector.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The PoE HAT cooling levels are not well suited for the PoE+ HAT - the
fan fails to come on until the temperature reaches the third trip point
(50C).
Give the rpi-poe-plus overlay a different set of cooling levels -
0 32 64 128 255, as suggested by @chris-kai-in.
See: https://github.com/raspberrypi/firmware/issues/1607
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
See https://github.com/raspberrypi/linux/issues/3981
An unknown unsafe memory access can result in the ep_state variable
in xhci_virt_ep being trampled with a stuck SET_DEQ_PENDING state
despite successful completion of a Set TR Deq Pointer command.
All URB enqueue/dequeue calls for the endpoint will fail in this state
so no transfers are possible until the device is reconnected.
As a workaround, clear the flag if we see it set and issue a new Set
TR Deq command anyway - this should be harmless, as a prior Set TR Deq
command will only have been issued in the Stopped state, and if the
endpoint is Running then the controller is required to ignore it and
respond with a Context State Error event TRB.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
Reduce the default CMA allocation requested by the vc4-kms-v3d-pi4 and
vc4-fkms-v3d-pi4 overlays to 320MB.
Use magic values of the form (<n>*64 - 4)MB to encode default values
of <n>*64MB, allowing these defaults to be distinguished from values
set explicitly by the user with the usual overlay parameters (e.g.
"cma-384"). Only default values will be capped if the Pi RAM is too
small or the gpu_mem setting too large for it to be viable.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Add call to v4l2_ctrl_new_fwnode_properties to read and
create the fwnode based controls.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add call to v4l2_ctrl_new_fwnode_properties to read and
create the fwnode based controls.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add call to v4l2_ctrl_new_fwnode_properties to read and
create the fwnode based controls.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add the orientation parameter to all the camera sensor overlays to
avoid libcamera complaining, and add the rotation parameter where
it hadn't been added before.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
There are no error exists once device_create has succeeded, and
therefore no need to call device_destroy from vc_mem_init.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The firmware patches the PHY MDIO address in the DTB to cope with
variations between board revisions, but the default for the CM4 PHY
is currently 1 when it should be 0.
Reported-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Pi 4s have at least 1GB, and there are advantages to having more CMA
available (HEVC works out of the box, support for more complex video
setups, etc.) without significant disadvantages.
Can be overridden by appending a parameter to the dtoverlay line, e.g.
dtoverlay=vc4-fkms-v3d,cma-256
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
In the bind hook, we actually need the device to have the HSM clock
running during the final part of the display initialisation where we
reset the controller and initialise the CEC component.
Failing to do so will result in a complete, silent, hang of the CPU.
Fixes: 411efa18e4 ("drm/vc4: hdmi: Move the HSM clock enable to runtime_pm")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
[ drm-misc commit 2eecd93b74 ]
There'a limit to how big a kmalloc buffer can be, and as memory gets
fragmented it becomes more difficult to get big buffers. The downside of
smaller buffers is that the driver has to split the transfer up which
hampers performance. Compression might also take a hit because of the
splitting.
Solve this by allocating the transfer buffer using vmalloc and create a
SG table to be passed on to the USB subsystem. vmalloc_32() is used to
avoid DMA bounce buffers on USB controllers that can only access 32-bit
addresses.
This also solves the problem that split transfers can give host side
tearing since flushing is decoupled from rendering.
usb_sg_wait() doesn't have timeout handling builtin, so it is wrapped in
a timer like 4 out of 6 users in the kernel have done.
v2:
- Use DIV_ROUND_UP (Linus)
- Add timeout note to the commit log (Linus)
- Expand note about upper buffer limit (Linus)
- Change var name s/timer/ctx/ in gud_usb_bulk_timeout()
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210701170748.58009-2-noralf@tronnes.org
[ drm-misc commit f8ac863b6a ]
Free transfer and compression buffers on device removal instead of at
DRM device removal time. This ensures that the usual 2x8MB buffers are
released when the device is unplugged and not kept around should
userspace keep the DRM device fd open.
At least Ubuntu 20.04 doesn't release the DRM device on unplug.
The damage_lock mutex is not destroyed because it is used outside the
drm_dev_enter/exit block in gud_pipe_update(). AFAICT it's possible for
an open fbdev descriptor to trigger a commit after the USB device is gone.
v2: Don't destroy damage_lock
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210701170748.58009-1-noralf@tronnes.org
[ Upstream commit 40e1a70b4a ]
This adds a USB display driver with the intention that it can be
used with future USB interfaced low end displays/adapters. The Linux
gadget device driver will serve as the canonical device implementation.
The following DRM properties are supported:
- Plane rotation
- Connector TV properties
There is also support for backlight brightness exposed as a backlight
device.
Display modes can be made available to the host driver either as DRM
display modes or through EDID. If both are present, EDID is just passed
on to userspace.
Performance is preferred over color depth, so if the device supports
RGB565, DRM_CAP_DUMB_PREFERRED_DEPTH will return 16.
If the device transfer buffer can't fit an uncompressed framebuffer
update, the update is split up into parts that do fit.
Optimal user experience is achieved by providing damage reports either by
setting FB_DAMAGE_CLIPS on pageflips or calling DRM_IOCTL_MODE_DIRTYFB.
LZ4 compression is used if the device supports it.
The driver supports a one bit monochrome transfer format: R1. This is not
implemented in the gadget driver. It is added in preparation for future
monochrome e-ink displays.
The driver is MIT licensed to smooth the path for any BSD port of the
driver.
v2:
- Use devm_drm_dev_alloc() and drmm_mode_config_init()
- drm_fbdev_generic_setup: Use preferred_bpp=0, 16 was a copy paste error
- The drm_backlight_helper is dropped, copy in the code
- Support protocol version backwards compatibility for device
v3:
- Use donated Openmoko USB pid
- Use direct compression from framebuffer when pitch matches, not only on
full frames, so split updates can benefit
- Use __le16 in struct gud_drm_req_get_connector_status
- Set edid property when the device only provides edid
- Clear compression fields in struct gud_drm_req_set_buffer
- Fix protocol version negotiation
- Remove mode->vrefresh, it's calculated
v4:
- Drop the status req polling which was a workaround for something that
turned out to be a dwc2 udc driver problem
- Add a flag for the Linux gadget to require a status request on
SET operations. Other devices will only get status req on STALL errors
- Use protocol specific error codes (Peter)
- Add a flag for devices that want to receive the entire framebuffer on
each flush (Lubomir)
- Retry a failed framebuffer flush
- If mode has changed wait for worker and clear pending damage before
queuing up new damage, fb width/height might have changed
- Increase error counter on bulk transfer failures
- Use DRM_MODE_CONNECTOR_USB
- Handle R1 kmalloc error (Peter)
- Don't try and replicate the USB get descriptor request standard for the
display descriptor (Peter)
- Make max_buffer_size optional (Peter), drop the pow2 requirement since
it's not necessary anymore.
- Don't pre-alloc a control request buffer, it was only 4k
- Let gud.h describe the whole protocol explicitly and don't let DRM
leak into it (Peter)
- Drop display mode .hskew and .vscan from the protocol
- Shorten names: s/GUD_DRM_/GUD_/ s/gud_drm_/gud_/ (Peter)
- Fix gud_pipe_check() connector picking when switching connector
- Drop gud_drm_driver_gem_create_object() cached is default now
- Retrieve USB device from struct drm_device.dev instead of keeping a
pointer
- Honour fb->offsets[0]
- Fix mode fetching when connector status is forced
- Check EDID length reported by the device
- Use drm_do_get_edid() so userspace can overrride EDID
- Set epoch counter to signal connector status change
- gud_drm_driver can be const now
v5:
- GUD_DRM_FORMAT_R1: Use non-human ascii values (Daniel)
- Change name to: GUD USB Display (Thomas, Simon)
- Change one __u32 -> __le32 in protocol header
- Always log fb flush errors, unless the previous one failed
- Run backlight update in a worker to avoid upsetting lockdep (Daniel)
- Drop backlight_ops.get_brightness, there's no readback from the device
so it doesn't really add anything.
- Set dma mask, needed by dma-buf importers
v6:
- Use obj-y in Makefile (Peter)
- Fix missing le32_to_cpu() when using GUD_DISPLAY_MAGIC (Peter)
- Set initial brightness on backlight device
v7:
- LZ4_compress_default() can return zero, check for that
- Fix memory leak in gud_pipe_check() error path (Peter)
- Improve debug and error messages (Peter)
- Don't pass length in protocol structs (Peter)
- Pass USB interface to gud_usb_control_msg() et al. (Peter)
- Improve gud_connector_fill_properties() (Peter)
- Add GUD_PIXEL_FORMAT_RGB111 (Peter)
- Remove GUD_REQ_SET_VERSION (Peter)
- Fix DRM_IOCTL_MODE_OBJ_SETPROPERTY and the rotation property
- Fix dma-buf import (Thomas)
v8:
- Forgot to filter RGB111 from reaching userspace
- Handle a device that only returns unknown device properties (Peter)
- s/GUD_PIXEL_FORMAT_RGB111/GUD_PIXEL_FORMAT_XRGB1111/ (Peter)
- Fix R1 and XRGB1111 format conversion
- Add FIXME about Big Endian being broken (Peter, Ilia)
Cc: Lubomir Rintel <lkundrak@v3.sk>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Peter Stuge <peter@stuge.se>
Tested-by: Peter Stuge <peter@stuge.se>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210313112545.37527-4-noralf@tronnes.org
[ backport changes:
- Remove include drm_gem_atomic_helper.h
- s/drm_gem_simple_display_pipe_prepare_fb/drm_gem_fb_simple_display_pipe_prepare_fb/
- Remove const from gud_drm_driver
- Change drm_gem_shmem_{vmap,vunmap} signatures
- Add gud_gem_create_object() to get cached memory mapping.
]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
The imx477 driver's line length for this mode had not been updated to
the value supplied to us by the sensor manufacturer. With this
correction the sensor delivers the framerates that are expected.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
The bcm2711 vc3-kms-v3d overlay enables the I2C instances used for
EDID data. Give these distinct I2C interface numbers (20 & 21) to
clearly separate them from other regular I2C blocks (1, 3-6) and the
mux on I2C0 (10+).
The 2711 DTS tree no longer includes i2c2, so the explicit deletion can
be removed.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Replace drm_encoder_helper_funcs::atomic_check with
drm_connector_helper_funcs::atomic_check - the former is not called
during drm_mode_obj_set_property_ioctl(). Set crtc_state->mode_changed
if TV norm changes even without explicit mode change. This makes things
like "xrandr --output Composite-1 --set mode PAL-M" work properly.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Similar to the ch7006 and nouveau drivers, introduce a "tv_mode" module
parameter that allow setting the TV norm by specifying vc4.tv_norm= on
the kernel command line.
If that is not specified, try inferring one of the most popular norms
(PAL or NTSC) from the video mode specified on the command line. On
Raspberry Pis, this causes the most common cases of the sdtv_mode
setting in config.txt to be respected.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Add support for the following composite output modes (all of them are
somewhat more obscure than the previously defined ones):
- NTSC_443 - NTSC-style signal with the chroma subcarrier shifted to
4.43361875 MHz (the PAL subcarrier frequency). Never used for
broadcasting, but sometimes used as a hack to play NTSC content in PAL
regions (e.g. on VCRs).
- PAL_N - PAL with alternative chroma subcarrier frequency,
3.58205625 MHz. Used as a broadcast standard in Argentina, Paraguay
and Uruguay to fit 576i50 with colour in 6 MHz channel raster.
- PAL60 - 480i60 signal with PAL-style color at normal European PAL
frequency. Another non-standard, non-broadcast mode, used in similar
contexts as NTSC_443. Some displays support one but not the other.
- SECAM - French frequency-modulated analog color standard; also have
been broadcast in Eastern Europe and various parts of Africa and Asia.
Uses the same 576i50 timings as PAL.
Also added some comments explaining color subcarrier frequency
registers.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
PAL-M is a Brazilian analog TV standard that uses a PAL-style chroma
subcarrier at 3.575611[888111] MHz on top of 525-line (480i60) timings.
This commit makes the driver actually use the proper VEC preset for this
mode instead of just changing PAL subcarrier frequency.
DRM mode constant names have also been changed, as they no longer
correspond to the "NTSC" or "PAL" terms.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Change the mode_set function pointer logic to declarative config0,
config1 and custom_freq fields, to make TV mode setting logic more
concise and uniform.
Additionally, remove the superfluous tv_mode field, which was redundant
with the mode field in struct drm_tv_connector_state.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
This commit fixes vertical timings of the VEC (composite output) modes
to accurately represent the 525-line ("NTSC") and 625-line ("PAL") ITU-R
standards.
Previous timings were actually defined as 502 and 601 lines, resulting
in non-standard 62.69 Hz and 52 Hz signals being generated,
respectively.
Changes to vc4_crtc.c have also been made, to make the PixelValve
vertical timings accurately correspond to the DRM modeline in interlaced
modes. The resulting VERTA/VERTB register values have been verified
against the reference values set by the Raspberry Pi firmware.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
The new vec node in bcm2711.dtsi should have the correct interrupt
number to start with, rather than include the bcm283x version and
patch it later.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The calculations clipped the right/bottom edge of the clipped
range based on the left/top margins.
Fixes: 666e73587f ("drm/vc4: Take margin setup into account when updating planes")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The alternative is to move them into the source file that uses then,
but they are large and intrusive, so that strategy is being avoided.
Fixes the following W=1 kernel build warning(s):
drivers/gpu/drm/vc4/vc4_hdmi_regs.h:282:39: warning: ‘vc5_hdmi_hdmi1_fields’ defined but not used [-Wunused-const-variable=]
drivers/gpu/drm/vc4/vc4_hdmi_regs.h:206:39: warning: ‘vc5_hdmi_hdmi0_fields’ defined but not used [-Wunused-const-variable=]
drivers/gpu/drm/vc4/vc4_hdmi_regs.h:145:39: warning: ‘vc4_hdmi_fields’ defined but not used [-Wunused-const-variable=]
Cc: Eric Anholt <eric@anholt.net>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20201116174112.1833368-38-lee.jones@linaro.org
idr_init() uses base 0 which is an invalid identifier for this driver.
The idr_alloc for this driver uses VC4_PERFMONID_MIN as start value for
ID range and it is #defined to 1. The new function idr_init_base allows
IDR to set the ID lookup from base 1. This avoids all lookups that
otherwise starts from 0 since 0 is always unused / available.
References: commit 6ce711f275 ("idr: Make 1-based IDRs more efficient")
Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20201105202135.GA145111@localhost
To support legacy gamma ioctls the drivers need to set
drm_crtc_funcs.gamma_set either to a custom implementation or to
drm_atomic_helper_legacy_gamma_set. Most of the atomic drivers do the
latter.
We can simplify this by making the core handle it automatically.
Move the drm_atomic_helper_legacy_gamma_set() functionality into
drm_color_mgmt.c to make drm_mode_gamma_set_ioctl() use
drm_crtc_funcs.gamma_set if set or GAMMA_LUT property if not.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Philippe Cornu <philippe.cornu@st.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201211114237.213288-2-tomi.valkeinen@ti.com
Now that atomic_check takes the global atomic state as a parameter, we
don't need to go through the pointer in the CRTC state.
This was done using the following coccinelle script:
@ crtc_atomic_func @
identifier helpers;
identifier func;
@@
static struct drm_crtc_helper_funcs helpers = {
...,
.atomic_check = func,
...,
};
@@
identifier crtc_atomic_func.func;
identifier crtc, state;
@@
func(struct drm_crtc *crtc, struct drm_atomic_state *state) {
...
- struct drm_crtc_state *crtc_state = drm_atomic_get_new_crtc_state(state, crtc);
... when != crtc_state
- crtc_state->state
+ state
...
}
@@
struct drm_crtc_state *crtc_state;
identifier crtc_atomic_func.func;
identifier crtc, state;
@@
func(struct drm_crtc *crtc, struct drm_atomic_state *state) {
...
- crtc_state->state
+ state
...
}
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201102133834.1176740-3-maxime@cerno.tech
The core clock needs to be raised temporarily during a modeset to
500MHz. However, the HVS core clock requirement might be higher than
500MHz. This rate will be enforced at the end of the mode setting,
meaning that might might end up with a core clock rate lower than
planned on the first mode set.
Use the maximum value of 500MHz and the HVS core clock rate for our
temporary boost to fix this issue.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Depending on a given HVS output (HVS to PixelValves) and input (planes
attached to a channel) load, the HVS needs for the core clock to be
raised above its boot time default.
Failing to do so will result in a vblank timeout and a stalled display
pipeline.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Now that the semaphore is gone, our atomic_commit implementation is
basically drm_atomic_helper_commit with a somewhat custom commit_tail,
the main difference being that we're using wait_for_flip_done instead of
wait_for_vblanks used in the drm_atomic_helper_commit_tail helper.
Let's switch to using drm_atomic_helper_commit.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Now that we have proper ordering guaranteed by the previous patch, the
semaphore is redundant and can be removed.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The HVS state now has both unassigned_channels that reflects the
channels that are not used in the associated state, and the in_use
boolean for each channel that says whether or not a particular channel
is in use.
Both express pretty much the same thing, and we need the in_use variable
to properly track the commits, so let's get rid of unassigned_channels.
Suggested-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20201204151138.1739736-6-maxime@cerno.tech
If we're having two subsequent, non-blocking, commits on two different
CRTCs that share no resources, there's no guarantee on the order of
execution of both commits.
However, the second one will consider the first one as the old state,
and will be in charge of freeing it once that second commit is done.
If the first commit happens after that second commit, it might access
some resources related to its state that has been freed, resulting in a
use-after-free bug.
The standard DRM objects are protected against this, but our HVS private
state isn't so let's make sure we wait for all the previous FIFO users
to finish their commit before going with our own.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
When we can't allocate a new channel, we can simply return instead of
having to handle both cases, and that simplifies a bit the code.
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The private objects have a gotcha that could result in a use-after-free,
make sure it's properly documented.
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Private objects storing a state shared across all CRTCs need to be
carefully handled to avoid a use-after-free issue.
The proper way to do this to track all the commits using that shared
state and wait for the previous commits to be done before going on with
the current one to avoid the reordering of commits that could occur.
However, this commit setup needs to be done after
drm_atomic_helper_setup_commit(), because before the CRTC commit
structure hasn't been allocated before, and before the workqueue is
scheduled, because we would be potentially reordered already otherwise.
That means that drivers currently have to roll their own
drm_atomic_helper_commit() function, even though it would be identical
if not for the commit setup.
Let's introduce a hook to do so that would be called as part of
drm_atomic_helper_commit, allowing us to reuse the atomic helpers.
Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The vidioc_enum_input() v4l2 ioctl is capable of returning
sensor/input status as well. This is used in current
GStreamer HEAD for signal detection [1].
bcm2835-unicam does handle this syscall, but it didn't ask
the subdevice driver about the input status. The input then
appeared as always present.
This commit adds the necessary query. There is a precedent for
this - the R-Car VIN V4L2 driver does a similar call [2].
[1]: ce0be27caf/sys/v4l2/gstv4l2src.c (L553)
[2]: 7fb9d006d3/drivers/media/platform/rcar-vin/rcar-v4l2.c (L548)
Signed-off-by: Jakub Vaněk <linuxtardis@gmail.com>
Our hotplug handler will currently call the drm_kms_helper_hotplug_event
every time a hotplug interrupt is called.
However, since the device is registered after all the drivers have
finished their bind callback, we have a window between when we install
our interrupt handler and when drm_dev_register() is eventually called
where our handler can run and call drm_kms_helper_hotplug_event but the
device hasn't been registered yet, causing a null pointer dereference.
Fix this by making sure we only call drm_kms_helper_hotplug_event if our
device has been properly registered.
Fixes: f4790083c7 ("drm/vc4: hdmi: Rely on interrupts to handle hotplug")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The hotplugs interrupt handlers are registered through the
devm_request_threaded_irq function. However, while free_irq is indeed
called properly when the device is unbound or bind fails, it's called
after unbind or bind is done.
In our particular case, it means that on failure it creates a window
where our interrupt handler can be called, but we're freeing every
resource (CEC adapter, DRM objects, etc.) it might need.
In order to address this, let's switch to the non-devm variant to
control better when the handler will be unregistered and allow us to
make it safe.
Fixes: f4790083c7 ("drm/vc4: hdmi: Rely on interrupts to handle hotplug")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The CEC interrupt handlers are registered through the
devm_request_threaded_irq function. However, while free_irq is indeed
called properly when the device is unbound or bind fails, it's called
after unbind or bind is done.
In our particular case, it means that on failure it creates a window
where our interrupt handler can be called, but we're freeing every
resource (CEC adapter, DRM objects, etc.) it might need.
In order to address this, let's switch to the non-devm variant to
control better when the handler will be unregistered and allow us to
make it safe.
Fixes: 15b4511a4a ("drm/vc4: add HDMI CEC support")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
NetBSD have changed their licensing requirements such that the 2-clause
licence is preferred. Update usb.h in the downstream dwc_otg code
accordingly.
See https://www.netbsd.org/about/redistribution.html for more
information.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Commit ecdd08fd9b ("drm/vc4: hdmi: Make sure the device is powered
with CEC") made sure that the device is powered while there is
CEC-related accesses but missed one register read in the variable
declaration.
Move the variable assignment after the pm_runtime_resume_and_get.
Fixes: ecdd08fd9b ("drm/vc4: hdmi: Make sure the device is powered with CEC")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
We've had many silent hangs where the kernel would look like it just
stalled due to the access to one of the HDMI registers while the
controller was disabled.
Add a warning if we're about to do that so that it's at least not silent
anymore.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In vc4_hdmi_encoder_pre_crtc_configure, if clk_request_start for the HSM
clock fails, we don't call clk_disable_unprepare on the pixel clock even
though it's enabled by now.
Make sure it's there to avoid leaking that reference.
Fixes: cd4cb49dc5 ("drm/vc4: hdmi: Adjust HSM clock rate depending on pixel rate")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Similarly to what we encountered with the detect hook with DRM, nothing
actually prevents any of the CEC callback from being run while the HDMI
output is disabled.
However, this is an issue since any register access to the controller
when it's powered down will result in a silent hang.
Let's make sure we run the runtime_pm hooks when the CEC adapter is
opened and closed by the userspace to avoid that issue.
Fixes: 15b4511a4a ("drm/vc4: add HDMI CEC support")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In order to ease further additions to the CEC enable and disable, let's
split the function into two functions, one to enable and the other to
disable.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In the vc4_hdmi_encoder_pre_crtc_configure() function error path we
never actually call pm_runtime_put() even though
pm_runtime_resume_and_get() is our very first call.
Fixes: 4f6e3d66ac ("drm/vc4: Add runtime PM support to the HDMI encoder driver")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In the bind hook, we actually need the device to have the HSM clock
running during the final part of the display initialisation where we
reset the controller and initialise the CEC component.
Failing to do so will result in a complete, silent, hang of the CPU.
Fixes: 411efa18e4 ("drm/vc4: hdmi: Move the HSM clock enable to runtime_pm")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The HPD GPIO retrieval code on failure will jump to the
err_unprepare_hsm label that calls pm_runtime_disable.
However at that point we haven't called pm_runtime_enable, so we end up
with an unbalanced call.
The next error than can occur (and therefore the next label) needs both
pm_runtime_disable and drm_encoder_cleanup though, so let's rearrange
the labels to match what we expect.
Fixes: cd4cb49dc5 ("drm/vc4: hdmi: Adjust HSM clock rate depending on pixel rate")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In order to access the HDMI controller, we need to make sure the HSM
clock is enabled. If we were to access it with the clock disabled, the
CPU would completely hang, resulting in an hard crash.
Since we have different code path that would require it, let's move that
clock enable / disable to runtime_pm that will take care of the
reference counting for us.
Fixes: 4f6e3d66ac ("drm/vc4: Add runtime PM support to the HDMI encoder driver")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210525091059.234116-3-maxime@cerno.tech
Add the firmware phandle to the vc4 node so that we can send it the
message that we're done with the firmware display.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Once the call to drm_fb_helper_remove_conflicting_framebuffers() has
been made, simplefb has been unregistered and the KMS driver is entirely
in charge of the display.
Thus, we can notify the firmware it can free whatever resource it was
using to maintain simplefb functional.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The bind hooks will modify their controller registers, so simplefb is
going to be unusable anyway. Let's avoid any transient state where it
could still be in the system but no longer functionnal.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The RPI_FIRMWARE_NOTIFY_DISPLAY_DONE firmware call allows to tell the
firmware the kernel is in charge of the display now and the firmware can
free whatever resources it was using.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The vc4 driver will need to tell the firmware that it takes over the
display for the firmware to free its resources (lower the clock, free
some memory, etc.)
Let's add an optional phandle to our firmware node.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The raspberrypi,firmware property has been documented as required in the
binding but was never actually used in the final version of the binding.
Remove it.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Lift the set of RTCs out of i2c-rtc and i2c-rtc-gpio to update
i2c-rtc-gpio and to reduce duplication.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The imx378 sensor is almost identical to the imx477 and can be
supported as a "compatible" sensor with just a few extra register
writes.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
The bcm2835 ISP requires the base address of all input/output planes to have 32
byte alignment. Using a Y stride of 32 bytes would not guarantee that the V
plane would fulfil this, e.g. a height of 650 lines would mean the V plane
buffer is not 32 byte aligned for YUV420 formats.
Having a Y stride of 64 bytes would ensure both U and V planes have a 32 byte
alignment, as the luma height will always be an even number of lines.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
From the requesting issue:
"This option is necessary for having bridge-like networking on KVM VMs
with the host root filesystem on NFS (since regular bridge networking
doesn't work in that case),"
See: https://github.com/raspberrypi/linux/issues/4413
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
If we have a state already and disconnect/reconnect the display, the
SCDC messages won't be sent again since we didn't go through a disable /
enable cycle.
In order to fix this, let's call the vc4_hdmi_enable_scrambling function
in the detect callback if there is a mode and it needs the scrambler to
be enabled.
Fixes: 74465b84fa ("drm/vc4: hdmi: Enable the scrambler")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Before the introduction of the BCM2711 support, the HSM clock rate was
fixed, and was the CEC and audio clock source on the SoCs previously
supported.
The HSM clock is also the source of the internal state machine of the
controller and needs to run faster than the pixel clock. All these
requirements were met by running at 101% of the maximum pixel rate,
meeting the fixed clock requirement for audio and CEC, while remaining
faster than any pixel clock we might need.
However, the BCM2711 brought support for 4k and therefore increased
significantly the rate needed for the HSM, and new, independant, clocks
to feed the audio and CEC clocks. Since the HSM clock can also run much
higher, we also need to lower its rate if possible to reduce its power
consumption.
The CEC support code changes its clock divider when the HSM clock rate
is changed, but the audio support never had a similar feature and will
glitch out if audio is played back during a mode set.
Since the HSM rate was meant to be fixed on the SoCs prior to the
BCM2711 anyway, let's introduce back a fixed HSM rate and fix audio.
Fixes: cd4cb49dc5 ("drm/vc4: hdmi: Adjust HSM clock rate depending on pixel rate")
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
From the original Rockchip driver, the subdev was renamed
from the default to being "mov9281 <dev_name>" whereas the
default would have been "ov9281 <dev_name>".
Remove the override to drop back to the default rather than
a vendor custom string.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The core clock needs to be raised temporarily during a modeset to
500MHz. However, the HVS core clock requirement might be higher than
500MHz. This rate will be enforced at the end of the mode setting,
meaning that might might end up with a core clock rate lower than
planned on the first mode set.
Use the maximum value of 500MHz and the HVS core clock rate for our
temporary boost to fix this issue.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Adjust the DVP enable/disable sequence to avoid a pixel getting stuck
in an internal, non resettable FIFO within PixelValve when changing
HDMI resolution.
The blank pixels features of the DVP can prevent signals back to
pixelvalve causing it to not clear the FIFO. Adjust the ordering
and timing of operations to ensure the clear signal makes it through to
pixelvalve.
Signed-off-by: Tim Gover <tim.gover@raspberrypi.com>
A matching media bus format was added and an overlay for using it,
both with FB and VC4 was added as well.
Signed-off-by: Joerg Quinten <aBUGSworstnightmare@gmail.com>
It is legitimate, though unusual, for an aux ent associated with a slot
to be selected in phase 0 before a previous selection has been used and
released in phase 2. Fix such that if the slot is found to be in use
that the aux ent associated with it is reused rather than an new aux
ent being created. This fixes a problem where when the first aux ent
was released the second was lost track of.
This bug spotted in Nick's testing. It may explain some other occasional,
unreliable decode error reports where dmesg included "Missing DPB AUX
ent" logging.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Increase the number of post-sync blanking lines on odd fields instead of
decreasing it on even fields. This makes the total number of lines
properly match the modelines.
Additionally fix the value of PV_VCONTROL_ODD_DELAY, which did not take
pixels_per_clock into account, causing some displays to invert the
fields when driven by bcm2711.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Depending on a given HVS output (HVS to PixelValves) and input (planes
attached to a channel) load, the HVS needs for the core clock to be
raised above its boot time default.
Failing to do so will result in a vblank timeout and a stalled display
pipeline.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The load tracker was initially designed to report and warn about a load
too high for the HVS. To do so, it computes for each plane the impact
it's going to have on the HVS, and will warn (if it's enabled) if we go
over what the hardware can process.
While the limits being used are a bit irrelevant to the BCM2711, the
algorithm to compute the HVS load will be one component used in order to
compute the core clock rate on the BCM2711.
Let's remove the hooks to prevent the load tracker to do its
computation, but since we don't have the same limits, don't check them
against them, and prevent the debugfs file to enable it from being
created.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The encoder retrieval code has been a source of bugs and glitches in the
past and the crtc <-> encoder association been wrong in a number of
different ways.
Add some logging to quickly spot issues if they occur.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
It turns out the encoder retrieval code, in addition to being
unnecessarily complicated, has a bug when only the planes and crtcs are
affected by a given atomic commit.
Indeed, in such a case, either drm_atomic_get_old_connector_state or
drm_atomic_get_new_connector_state will return NULL and thus our encoder
retrieval code will not match on anything.
We can however simplify the code by using drm_for_each_encoder_mask, the
drm_crtc_state storing the encoders a given CRTC is connected to
directly and without relying on any other state.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
vc4_crtc_config_pv() retrieves the encoder again, even though its only
caller, vc4_crtc_atomic_enable(), already did.
Pass the encoder pointer as an argument instead of going through all the
connectors to retrieve it again.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
When the clock setups were added for the alternate external clocks,
the settings for 2 lane 720p and 4 lane 1080p were transposed.
2 lane 720p still worked, but 4 lane 1080p didn't.
Correct the assignments.
Fixes: 6b0c094a5b (media: i2c: imx290: Add support for 74.25MHz clock")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Whilst the datasheet lists the link frequency changing between
1080p and 720p modes, reality is that with the default blanking
we have
(1920 + 280) * (1080 + 45) * 60fps = 148.5MPix/s
and
(1280 + 2020) * (720 + 30) * 60fps = 148.5MPix/s
and this reflects reality whether in 10 or 12 bit readout modes.
How this relates to link frequency is unclear as it differs
from the datasheet, but all exposure and frame rate calcs need
the pixel rate to be correct, so make it so.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Commit "97589ad61c73 media: i2c: imx290: Add support for 2 data lanes"
added support for running in two lane mode (instead of 4), but
without changing the link frequency that resulted in a max of 30fps.
Commit "98e0500eadb7 media: i2c: imx290: Add configurable link frequency
and pixel rate" then doubled the link frequency when in 2 lane mode,
but didn't undo the correction for running at only 30fps, just extending
horizontal blanking instead.
It also didn't update the CSI timing registers in accordance with the
datasheet.
Remove the 30fps limit on 2 lane by correcting the register config
in accordance with the datasheet for 60fps operation over 2 lanes.
Frame rate control (via V4L2_CID_VBLANK or HBLANK) can still reduce
the frame rate on 2 lanes back to 30fps.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
In order to allow containerised Android apps to run, add the following
config settings on 64-bit builds:
CONFIG_ANDROID=y
CONFIG_ANDROID_BINDER_IPC=y
CONFIG_ANDROID_BINDERFS=y
CONFIG_ASHMEM=y
This has a memory overhead of about 130kB and no other performance
penalty.
See: https://github.com/raspberrypi/linux/issues/4162
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Enable NO_WAIT_RESP, DMA_WIDE_SOURCE, DMA_WIDE_DEST, and bump the DMA
panic and AXI priorities to avoid any DMA transfer error with HBR audio
(8 channel, 192Hz).
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The hdmi-codec brings a lot of advanced features, including the HDMI
channel mapping. Let's use it in our driver instead of our own codec.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The IEC958 status bit is usually set by the userspace after hw_params
has been called, so in order to use whatever is set by the userspace, we
need to implement the prepare hook. Let's add it to the hdmi_codec_ops,
and mandate that either prepare or hw_params is implemented.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In some situations, like a codec probe, we need to provide an IEC status
default but don't have access to the sampling rate and width yet since
no stream has been configured yet.
Each and every driver has its own default, whereas the core iec958 code
also has some buried in the snd_pcm_create_iec958_consumer functions.
Let's split these functions in two to provide a default that doesn't
rely on the sampling rate and width, and another function to fill them
when available.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
We're going to add more controls to support the IEC958 output, so let's
rework the control registration a bit to support more of them.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Depending on a given HVS output (HVS to PixelValves) and input (planes
attached to a channel) load, the HVS needs for the core clock to be
raised above its boot time default.
Failing to do so will result in a vblank timeout and a stalled display
pipeline.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
drm_atomic_get_crtc_state relies on drm_atomic_get_existing_crtc_state
that is deprecated and isn't really clear on which state it provides.
Even worse, after the states have been swapped, if the CRTC is present
in the state drm_atomic_get_existing_crtc_state will return the old
state, whereas if the CRTC is not present, we'll use a copy of
crtc->state. crtc->state for the CRTC that is there at that point is the new
state though, which leads to confusion.
Let's provide two new helpers that make it clear what our expectations
are.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In the event that the ENABLE signal from the codec goes low before
RELAY2 has been enabled, wait until the full 1000ms has elapsed then
enable RELAY2 and jump to amp_on_wait, i.e. as if output had been
fully enabled then disabled, rather than returning to the amp_off
idle state.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Do not scale IMX477_EXPOSURE_OFFSET with the long exposure factor during
the limit calculations. This allows larger exposure times, and does seem to be
what the sensor is doing internally.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Increasing the sleep time after clock selection to 3-4ms
allows the correct detection of all combinations of DAC+ Pro
and DAC+ADC Pro sound cards and the various PI revisions.
Signed-off-by: Joerg Schambacher <joerg@hifiberry.com>
There is a CAN protocol named J1939 which is supported by Linux Kernel
since v5.4. Recently J1939 header file turned up in
raspberrypi-kernel-headers but the module is not built for kernel
v5.10.y due to the missing flag. Therefore set it to build as module:
CONFIG_CAN_J1939=m
Signed-off-by: Patrick Menschel <menschel.p@posteo.de>
Fix stream on & off such that failures leave the driver in the correct
state. Ensure that the clock is on when we are streaming and off when
all contexts attached to this device have stopped streaming.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Due to overheads in assembling controls and requests it is worth having
the slice assembly phase separate from the h/w pass1 processing. Create
a queue to service pass1 rather than have the pass1 finished callback
trigger the next slice job.
This requires a rework of the logic that splits up the buffer and
request done events. This code contains two ways of doing that, we use
Ezequiel Garcias <ezequiel@collabora.com> solution, but expect that
in the future this will be handled by the framework in a cleaner manner.
Fix up the handling of some of the memory exhaustion crashes uncovered
in the process of writing this code.
Signed-off-by: John Cox <jc@kynesim.co.uk>
This is probably not the API we will want to add, but it
should show what semantics are needed by drivers.
The goal is to allow the OUTPUT (aka source) buffer and the
controls associated to a request to be released from the request,
and in particular return the OUTPUT buffer back to userspace,
without signalling the media request fd.
This is useful for devices that are able to pre-process
the OUTPUT buffer, therefore able to release it before
the decoding is finished. These drivers should signal
the media request fd only after the CAPTURE buffer is done.
Tested-by: John Cox <jc@kynesim.co.uk>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Add an enable count to the irq Q structures to allow the irq logic to
block further callbacks if resources associated with the irq are not
yet available.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Use multi-planar interface rather than single plane interface. This
allows dmabufs holding compressed data to be resized.
Signed-off-by: John Cox <jc@kynesim.co.uk>
VAAPI H265 has num entry points but never sets it. Allow a VAAPI
shim to work without requiring rewriting the VAAPI driver.
num_entry_points can be calculated from the slice_segment_addr
of the next slice so delay processing until we have that.
Also includes some minor cosmetics.
Signed-off-by: John Cox <jc@kynesim.co.uk>
When a request is re-inited it will release all control handler
objects that are still in the request. It does that by unbinding
and putting all those objects. When the object is unbound the
obj->req pointer is set to NULL, and the object's unbind op is
called. When the object it put the object's release op is called
to free the memory.
For a request object that contains a control handler that means
that v4l2_ctrl_handler_free() is called in the release op.
A control handler used in a request has a pointer to the main
control handler that is created by the driver and contains the
current state of all controls. If the device is unbound (due to
rmmod or a forced unbind), then that main handler is freed, again
by calling v4l2_ctrl_handler_free(), and any outstanding request
objects that refer to that main handler have to be unbound and put
as well.
It does that by this test:
if (!hdl->req_obj.req && !list_empty(&hdl->requests)) {
I.e. the handler has no pointer to a request, so is the main
handler, and one or more request objects refer to this main
handler.
However, this test is wrong since hdl->req_obj.req is actually
NULL when re-initing a request (the object unbind will set req to
NULL), and the only reason this seemingly worked is that the
requests list is typically empty since the request's unbind op
will remove the handler from the requests list.
But if another thread is at the same time adding a new control
to a request, then there is a race condition where one thread
is removing a control handler object from the requests list and
another thread is adding one. The result is that hdl->requests
is no longer empty and the code thinks that a main handler is
being freed instead of a control handler that is part of a request.
There are two bugs here: first the test for hdl->req_obj.req: this
should be hdl->req_obj.ops since only the main control handler will
have a NULL pointer there.
The second is that adding or deleting request objects from the
requests list of the main handler isn't protected by taking the
main handler's lock.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reported-by: John Cox <jc@kynesim.co.uk>
Fixes: 6fa6f831f0 ("media: v4l2-ctrls: add core request support")
Tested-by: John Cox <jc@kynesim.co.uk>
Reported-by: John Cox <jc@kynesim.co.uk>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
(cherry picked from commit be7e8af98f)
Update the symbol as well as the alias so that other overlays will
then be applied against the right node.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Fixes the following v4l2-compliance failure:
fail: v4l2-test-controls.cpp(871): subscribe event for control 'User Controls' failed test
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Trial and error reveals that the minimum vblank value appears to be 24
(the OV5647 data sheet does not give any clues). This fixes streaming
lock-ups in full resolution mode.
Fixes: 9b5a5ebedc ("media: i2c: ov5647: Add support for V4L2_CID_VBLANK")
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
The top offset in the pixel array is actually 6 (see page 3-1 of the
OV5647 data sheet).
Fixes: f2f7ad5ce5 ("media: i2c: ov5647: Selection compliance fixes")
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
It looks like some displays (like the LG 27UL850-W) don't enable the
scrambling when the HDMI driver enables it. However, if we set later the
scrambler enable bit, the display will work as expected.
Let's create delayed work queue to periodically look at the display
scrambling status, and if it's not set yet try to enable it again.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
DRM currently polls for the HDMI connector status every 10s, which can
be an issue when we connect/disconnect a display quickly or the device
on the other end only issues a hotplug pulse (for example on EDID
change).
Switch the driver to rely on the internal controller logic for the
BCM2711/RPi4.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
When we have the entire DRM state, retrieving the connector state only
requires the drm_connector pointer. Fortunately for us, we have
allocated it as a part of the vc4_hdmi structure, so we can retrieve get
a pointer by simply accessing our field in that structure.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
If the HPD GPIO is not available and drm_probe_ddc fails, we end up
reading the HDMI_HOTPLUG register, but the controller might be powered
off resulting in a CPU hang. Make sure we have the power domain and the
HSM clock powered during the detect cycle to prevent the hang from
happening.
Fixes: 4f6e3d66ac ("drm/vc4: Add runtime PM support to the HDMI encoder driver")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
pm_runtime_get_sync increases the PM usage counter even if it fails, and
forgetting to do so will result in a reference leak. We can't really do
anything in atomic_enable in case of a failure though, and we probably
can't recover either, but at least switching to
pm_runtime_resume_and_get makes us play nice with the PM subsystem.
Fixes: 4f6e3d66ac ("drm/vc4: Add runtime PM support to the HDMI encoder driver")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
At boot, we can't rely on the vc4_get_crtc_encoder since we don't have a
state yet and thus will not be able to figure out which connector is
attached to our CRTC.
However, we have a muxing bit in the CRTC register we can use to get the
encoder currently connected to the pixelvalve. We can thus read that
register, lookup the associated register through the vc4_pv_data
structure, and then pass it to vc4_crtc_disable so that we can perform
the proper operations.
Fixes: 875a4d5368 ("drm/vc4: drv: Disable the CRTC at boot time")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The vc4_get_crtc_encoder function currently only works when the
connector->state->crtc pointer is set, which is only true when the
connector is currently enabled.
However, we use it as part of the disable path as well, and our lookup
will fail in that case, resulting in it returning a null pointer we
can't act on.
We can access the connector that used to be connected to that crtc
though using the old connector state in the disable path.
Since we want to support both the enable and disable path, we can
support it by passing the state accessor variant as a function pointer,
together with the atomic state.
Fixes: 792c3132bc ("drm/vc4: encoder: Add finer-grained encoder callbacks")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The vc4_crtc_config_pv will need to access the drm_atomic_state
structure and its only parent function, vc4_crtc_atomic_enable already
has access to it. Let's pass it as a parameter.
Fixes: 792c3132bc ("drm/vc4: encoder: Add finer-grained encoder callbacks")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
This reverts commit 3cf3d39b75.
This commit was making the assumption that we had a 1:1 mapping between
the encoders and their CRTC. While this is true for the HDMI controllers
on the BCM2711, this isn't true for the other encoders (DSI0 and DPI
share the PixelValve 0, and DSI1 and SMI share the PixelValve1), and
this isn't true at all on the older SoCs, effectively breaking the
encoder retrieval logic.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The current core while setting the min and max rate properly in the
clk_request structure will not make sure that the requested rate is
within these boundaries, leaving it to each and every driver to make
sure it is.
Add a clamp call to make sure it's always done.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The dwc2 gadget support maps and unmaps DMA buffers as necessary. When
mapping and unmapping it uses the direction of the endpoint to select
the direction of the DMA transfer, but this fails for Control OUT
transfers because the unmap occurs after the endpoint direction has
been reversed for the status phase.
A possible solution would be to unmap the buffer before the direction
is changed, but a safer, less invasive fix is to remember the buffer
direction independently of the endpoint direction.
See: https://github.com/raspberrypi/Raspberry-Pi-OS-64bit/issues/127
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This is a simple overlay based on the simple-audio-card and the dmic
codec. It has the speciality that it is configured to use the codec
as a master I2S device. It works for example with the Si468x DAB
receiver on the uGreen DABBoard.
See: https://github.com/raspberrypi/linux/issues/4304
Signed-off-by: Christoph Orth <c.orth@ugreen.eu>
The 2711 pixel valve can't produce odd horizontal timings, and
checks were added to vc4_hdmi_encoder_atomic_check and
vc4_hdmi_encoder_mode_valid to filter out/block selection of
such modes.
Modes with DRM_MODE_FLAG_DBLCLK double all the horizontal timing
values before programming them into the PV. The PV values,
therefore, can not be odd, and so the modes can be supported.
Amend the filtering appropriately.
See https://github.com/raspberrypi/linux/issues/4307
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Keeping a copy of the old poweroff handler allows it to be restored
should this module be unloaded, but also provides a fallback if the
power hasn't been removed when the timeout elapses.
See: https://github.com/raspberrypi/rpi-eeprom/issues/330
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
There is no reason not to support multiple gpio-shutdown signals,
so add the necessary __override__ magic.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The result of dividing a u32 by a size_t is an unsigned int on arm32
and a long unsigned int on arm64. Use "%zu" (the size_t format) to
remove the build warning for 64-bit builds.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
During decode, setting the CAPTURE queue format was setting the crop
rectangle to the requested height before aligning up the format to
cater for simple clients that weren't expecting to deal with cropping
and the SELECTION API.
This caused problems on some resolution change events if the client
didn't also then use the selection API.
Disable the crop update after a resolution change.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Whilst the hardware can't achieve the limits of level 4.2 under
all situations, it can exceed level 4.0.
Allow selection of levels 4.1 and 4.2.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
MMAL has the flag MMAL_BUFFER_HEADER_FLAG_CORRUPTED but that
wasn't being passed through, so add it.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Video decode supports YUV and RGB formats. YUV needs to report SMPTE170M
or REC709 appropriately, whilst RGB should report SRGB.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The driver said it supported H264 levels 4.1 and 4.2, but
was missing the V4L2 to MMAL mappings.
Add in those mappings.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The current code will first dereference the req pointer and then test if
it's NULL, resulting in a NULL pointer dereference if req is indeed
NULL. Reorder the test and derefence to avoid the issue
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The VEC's DAC on BCM2711 is slightly different compared to the one on
BCM283x and needs different configuration. In particular, bit 3
(mask 0x8) switches the BCM2711 DAC input to "self-test input data",
which makes the output unusable. Separating two compatible variants in
devicetrees and the DRM driver was therefore necessary.
The configurations used for both variants have been borrowed from
Raspberry Pi (model 3B for BCM283x, 4B for BCM2711) firmware defaults.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
On the BCM2711 (Raspberry Pi 4), the VEC is actually connected to
output 2 of pixelvalve3.
NOTE: This contradicts the Broadcom docs, but has been empirically
tested and confirmed by Raspberry Pi firmware devs.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
The VEC has a different address (0x7ec13000) on the BCM2711 (used in
e.g. Raspberry Pi 4) compared to BCM238x (e.g. Pi 3 and earlier). This
was erroneously not taken account for.
Definition of the VEC in the devicetrees had to be moved from
bcm283x.dtsi to bcm2711.dtsi and bcm2835-common.dtsi to allow for this
differentiation.
Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Extend the spi-rtc overlay to support the ds3232 and ds3234 RTCs, as
well as adding parameters to select difference SPI controllers and
chip selects.
N.B. The default CS is now active-low - use the "cs_high" parameter to
override this.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Now that we have the infrastructure in place, we can raise the maximum
pixel rate we can reach for HDMI0 on the BCM2711.
HDMI1 is left untouched since its pixelvalve has a smaller FIFO and
would need a clock faster than what we can provide to support the same
modes.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The HDMI controller on the BCM2711 includes a scrambler in order to
reach the HDMI 2.0 modes that require it. Let's add the support for it.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
In order to reach the frequencies needed to output at 594MHz, the
firmware needs to be configured with the appropriate parameters in the
config.txt file (enable_hdmi_4kp60 and force_turbo).
Let's detect it at bind time, warn the user if we can't, and filter out
the relevant modes.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The BVB clock rate computation doesn't take into account a mode clock of
594MHz that we're going to need to support 4k60.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
We'll need to have the HVS binding before the HDMI controllers so that
we can check whether the firmware allows to run in 4kp60. Reorder a bit
the component list, and document the current constraints we're aware of.
Acked-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Since we fixed the hooks to disable the encoder at boot, we now have an
unbalanced clk_disable call at boot since we never enabled them in the
first place.
Let's mimic the state of the hardware and enable the clocks at boot if
the controller is enabled to get the use-count right.
Cc: <stable@vger.kernel.org> # v5.10+
Fixes: 09c438139b ("drm/vc4: hdmi: Implement finer-grained hooks")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Due to a FIFO that cannot be flushed between the pixelvalve and the HDMI
controllers on BCM2711, we need to carefully disable both at boot time
if they were left enabled by the firmware.
However, at the time we're running that code, the struct drm_connector
encoder pointer isn't set yet, and thus we cannot retrieve the encoder
associated to our CRTC.
We can however make use of the fact that we have a less flexible setup
than what DRM allows where we have a 1:1 relationship between our CRTCs
and encoders (and connectors), and thus store the crtc associated to our
encoder at boot time.
We cannot do that at the time the encoders are probed though, since the
CRTCs won't be probed yet and thus we don't know at that time which CRTC
index we're going to get, so let's do this in two passes: we can first
bind all the components and then once they all are bound, we can iterate
over all the encoders to find their associated CRTC and set the pointer.
This is similar to what we're doing to set the possible_crtcs field.
Fixes: 875a4d5368 ("drm/vc4: drv: Disable the CRTC at boot time")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The vc4_set_crtc_possible_masks is meant to run over all the encoders
and then set their possible_crtcs mask to their associated pixelvalve.
However, since the commit 39fcb28083 ("drm/vc4: txp: Turn the TXP into
a CRTC of its own"), the TXP has been turned to a CRTC and encoder of
its own, and while it does indeed register an encoder, it no longer has
an associated pixelvalve. The code will thus run over the TXP encoder
and set a bogus possible_crtcs mask, overriding the one set in the TXP
bind function.
In order to fix this, let's skip any virtual encoder.
Cc: <stable@vger.kernel.org> # v5.9+
Fixes: 39fcb28083 ("drm/vc4: txp: Turn the TXP into a CRTC of its own")
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The current code does a binary OR on the possible_crtcs variable of the
TXP encoder, while we want to set it to that value instead.
Cc: <stable@vger.kernel.org> # v5.9+
Fixes: 39fcb28083 ("drm/vc4: txp: Turn the TXP into a CRTC of its own")
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Credit to forum member gizmomouse on
https://www.raspberrypi.org/forums/viewtopic.php?f=98&t=253912 and
Andrey Vostrukhin of Harlab for these overlays.
See https://github.com/harlab/CM4_LCD_LT070ME05000 for
schematics and docs for the adapter board to connect this panel which
is found in the Asus/Google 2013 Nexus 7" tablet and therefore
relatively easily available.
Note that this uses 4 DSI data lanes, and therefore MUST be used
with DISP1 on a Compute Module. It can not be used on a standard
Pi.
There are two versions of the adapter board. V1 connects the
display controls to Pi GPIOs, whilst v2 uses an I2C GPIO expander
so needs no additional connections beyond the FFC and power.
The touchscreen overlay for these panels varies, so that part
is not configured.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
There is no reason why the control GPIOs for the panel can not
be connected to I2C or similar GPIO interfaces that may need to
sleep, therefore switch from gpiod_set_value to
gpiod_set_value_cansleep calls to configure them.
Without that you get complaints from gpiolib every time the state
is changed.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
vc4_dsi_encoder_disable is partially an open coded version of
drm_bridge_chain_disable, but it missed a termination condition
in the loop for ->disable which meant that no post_disable
calls were made.
Add in the termination clause.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
DSI0 seemingly had very little or no testing as a load of
the register mappings were incorrect/missing, so host
transfers always timed out due to enabling/checking incorrect
bits in the interrupt enable and status registers.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
vc4_dsi was registering both dsi0 and dsi1 as VC4_ENCODER_TYPE_DSI1
which seemed to work OK for a single DSI display, but fails
if there are two DSI displays connected.
Update to register the correct type.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
For slightly unknown reasons, dsi0 takes a different pixel format
to dsi1, and that has to be set in the pixel valve.
Amend the setup accordingly.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Replace an undeclared variable used by DWC_DEBUGPL with the real endpoint address. DWC_DEBUGPL does nothing with DEBUG undefined so it did not go wrong before.
Signed-off-by: Zixuan Wang <wangzixuan@sjtu.edu.cn>
commit c2198943e3 upstream.
The vendor mode is not always at config #1, so it is necessary to
set the correct configuration number.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 195aae321c upstream.
Support RTL8153C, RTL8153D, RTL8156A, and RTL8156B. The RTL8156A
and RTL8156B are the 2.5G ethernet.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 67ce1a806f upstream.
The different chips may have different requests when changing mtu.
Therefore, add a new help function of rtl_ops to change mtu. Besides,
reset the tx/rx after changing mtu.
Additionally, add mtu_to_size() and size_to_mtu() macros to simplify
the code.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 5133bcc748 upstream.
Set the maximum inter frame gap time (144ns) for speed 10M/half and
100M/half. It improves the performance for those speeds. And, there
is no effect for the other speeds.
For 10M/half and 100M/half, the fast inter frame gap time let the
device couldn't use the feature of the aggregation effectively,
because the transfer would be completed fastly. Therefore, use the
maximum value to improve the effect of the aggregation. However, you
may not feel the improvement for fast CPUs, because they compensate
for the effect of the aggregation.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 156c320761 upstream.
Some messages are before calling register_netdev(), so replace
netif_err() with dev_err().
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
commit c79515e479 upstream.
Return error code if autosuspend_en, eee_get, or eee_set don't exist.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
commit 7a0ae61acd upstream.
U1/U2 shoued be enabled for USB 3.0 or later. The USB 2.0 doesn't
support it.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
commit 80fd850b31 upstream.
For runtime resuming, the RTL8153B may be resumed from the state
of power cut, when enabling the feature of UPS. Then, the PHY
would be reset, so it is necessary to be initailized again.
Besides, the USB_U1U2_TIMER also has to be set again, so I move
it from r8153b_init() to r8153b_hw_phy_cfg().
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
commit a08c0d309d upstream.
Replace r8153_patch_request() with rtl_phy_patch_request().
Replace r8153_pre_ram_code() with rtl_pre_ram_code().
Replace r8153_post_ram_code() with rtl_post_ram_code().
Add rtl_patch_key_set().
The new functions have an additional parameter. It is used to wait
the patch request command finished. When the PHY is resumed from
the state of power cut, the PHY is at a safe mode and the
OCP_PHY_PATCH_STAT wouldn't be updated. For this situation, it is
safe to set patch request command without waiting OCP_PHY_PATCH_STAT.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
commit f3163f1cb8 upstream.
This converts the driver to use the new tasklet API introduced in
commit 12cc923f1c ("tasklet: Introduce new initialization API")
Signed-off-by: Emil Renner Berthing <kernel@esmil.dk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
commit c1aedf015e upstream.
Support ECM mode based on cdc_ether with relative mii functions,
when CONFIG_USB_RTL8152 is not set, or the device is not supported
by r8152 driver.
Both r8152 and r8153_ecm would check the return value of
rtl8152_get_version() in porbe(). If rtl8152_get_version()
return none zero value, the r8152 is used for the device
with vendor mode. Otherwise, the r8153_ecm is used for the
device with ECM mode.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Link: https://lore.kernel.org/r/1394712342-15778-392-Taiwan-albertk@realtek.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
commit 34e653efb6 upstream.
Fixes the following W=1 kernel build warning(s):
drivers/net/usb/r8152.c:934: warning: Function parameter or member 'blk_hdr' not described in 'fw_mac'
drivers/net/usb/r8152.c:934: warning: Function parameter or member 'reserved' not described in 'fw_mac'
drivers/net/usb/r8152.c:947: warning: Function parameter or member 'blk_hdr' not described in 'fw_phy_patch_key'
drivers/net/usb/r8152.c:947: warning: Function parameter or member 'reserved' not described in 'fw_phy_patch_key'
drivers/net/usb/r8152.c:986: warning: Function parameter or member 'blk_hdr' not described in 'fw_phy_nc'
drivers/net/usb/r8152.c:986: warning: Function parameter or member 'mode_pre' not described in 'fw_phy_nc'
drivers/net/usb/r8152.c:986: warning: Function parameter or member 'mode_post' not described in 'fw_phy_nc'
drivers/net/usb/r8152.c:986: warning: Function parameter or member 'reserved' not described in 'fw_phy_nc'
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Acked-by: Hayes Wang <hayeswang@realtek.com>
Link: https://lore.kernel.org/r/20201102114512.1062724-23-lee.jones@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The new clock request API allows us to increase the rate of the
core clock as required during mode set while decreasing it when
we're done, resulting in a better power-efficiency.
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
The new clock request API allows us to increase the rate of the HSM
clock to match our pixel rate requirements while decreasing it when
we're done, resulting in a better power-efficiency.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
It's not unusual to find clocks being shared across multiple devices
that need to change the rate depending on what the device is doing at a
given time.
The SoC found on the RaspberryPi4 (BCM2711) is in such a situation
between its two HDMI controllers that share a clock that needs to be
raised depending on the output resolution of each controller.
The current clk_set_rate API doesn't really allow to support that case
since there's really no synchronisation between multiple users, it's
essentially a fire-and-forget solution.
clk_set_min_rate does allow for such a synchronisation, but has another
drawback: it doesn't allow to reduce the clock rate once the work is
over.
In our previous example, this means that if we were to raise the
resolution of one HDMI controller to the largest resolution and then
changing for a smaller one, we would still have the clock running at the
largest resolution rate resulting in a poor power-efficiency.
In order to address both issues, let's create an API that allows user to
create temporary requests to increase the rate to a minimum, before
going back to the initial rate once the request is done.
This introduces mainly two side-effects:
* There's an interaction between clk_set_rate and requests. This has
been addressed by having clk_set_rate increasing the rate if it's
greater than what the requests asked for, and in any case changing
the rate the clock will return to once all the requests are done.
* Similarly, clk_round_rate has been adjusted to take the requests
into account and return a rate that will be greater or equal to the
requested rates.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
be suspended
Webcams with microphones are composite devices, and autosuspend is set
at the device level. If uvcvideo is probed after snd-usb-audio, the effect
of the quirk applied by snd-usb-audio is undone by uvcvideo's global
application of autosuspend.
Incrementing the interface's PM refcount in such cases prevents runtime PM
from happening, thus the device is left active.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
These devices use a type of Sonix chipset that produces broken microphone
data if suspended/resumed.
They also don't support readback of the sample rate.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
https://gist.github.com/popcornmix/6b3e23103c60170b02b148e0ba5d6ed7
is the script used to generate the 601, 709 and 2020 colourspaces.
I've regenetated the existing ones using script so it is reprocable
but there are lsb dfferences compared to values here (copied from spec)
whose origin is now lost.
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
When importing there was a missing call to detach the buffer,
so each import leaked the sg table entry.
Actually the release process for both locally allocated and
imported buffers is identical, so fix them to both use the same
function.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Our driver while supporting HDR didn't send the proper colorimetry info
in the AVI infoframe.
Let's add the property needed so that the userspace can let us know what
the colorspace is supposed to be.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The intel driver uses the same logic to attach the Colorspace property
in multiple places and we'll need it in vc4 too. Let's move that common
code in a helper.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
All the drivers that support the HDR metadata property have a similar
function to compare the metadata from one connector state to the next,
and force a mode change if they differ.
All these functions run pretty much the same code, so let's turn it into
an helper that can be shared across those drivers.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
All the drivers that implement HDR output call pretty much the same
function to initialise the hdr_output_metadata property, and while the
creation of that property is in a helper, every driver uses the same
code to attach it.
Provide a helper for it as well
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Commit 158e800e0f upstream
A test was added to the probe function to ensure the device was
actually connected and working before successfully completing a
probe. If the device was actually there, but the I2C bus was not
ready yet for whatever reason, the probe fails permanently.
Change the probe so that we defer the probe on a regmap read
failure so that we try the probe again when the dependent drivers
are potentially loaded. This should not affect the case where the
device truly isn't present because the probe will never successfully
complete.
Fixes: 2aa916e67d ("sc16is7xx: Read the LSR register for basic device presence check")
Cc: stable@vger.kernel.org
Signed-off-by: Annaliese McDermond <nh6z@nh6z.net>
Link: https://lore.kernel.org/r/010101787f9c3fd8-c1815c00-2d6b-4c85-a96a-a13e68597fda-000000@us-west-2.amazonses.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add unique names to the individual dac coded drivers
Remove some of the codec controls that are not used.
Signed-off-by: Paul Hermann <paul@picoreplayer.org>
# Conflicts:
# sound/soc/bcm/allo-piano-dac-plus.c
warning: address of array 'desc->wMaxPacketSize' will always evaluate to 'true'
The wMaxPacketSize field is actually a two element array which content should
be accessed via the UGETW macro.
x_off should only be within current stripe
The stripe number is accounted for elsewhere
Fixes: 9b1b1beb1c ("vc4/drm: Fix source offsets with DRM_FORMAT_P030")
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Neither CM4 nor Pi 400 have appeared upstream yet, and as a result
they have missed out on improvements to the Pi 4B platform.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
ARM: dts: Bring bcm2711-rpi-400.dts up-to-date
Pi 400 support has not appeared upstream yet, and as a result it has
missed out on improvements to the other Pi 4 platforms.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
1. Reduce the delay between RELAY1 and RELAY2 to 1000ms.
2. Rename the states to simplify LED control by an external script.
3. Claim all the required GPIOs, enabling pull-ups on the input.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The Adafruit Mini-PiTFT13 display needs offsets applying when rotated,
so use the "variant" mechanism to select a custom set_addr_win method
using a dedicated compatible string of "fbtft,minipitft13".
See: https://github.com/raspberrypi/firmware/issues/1524
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
DMABUFs are all handled by videobuf2, so there is no reason not
to enable support for them.
Note that this driver is still using the vmalloc allocator, so
the buffers it allocates will not be compatible with the codec
or ISP driver that require contiguous buffers. However this
driver should be able to import the buffers allocated by them.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
There is little harm in ignoring fractional coordinates
(they just get truncated).
Without this:
modetest -M vc4 -F tiles,gradient -s 32:1920x1080-60 -P89@74:1920x1080*.1.1@XR24
is rejected. We have the same issue in Kodi when trying to
use zoom options on video.
Note: even if all coordinates are fully integer. e.g.
src:[0,0,1920,1080] dest:[-10,-10,1940,1100]
it will still get rejected as drm_atomic_helper_check_plane_state
uses drm_rect_clip_scaled which transforms this to fractional src coords
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Spec says: bits [31:4] of the given address should point to
the 128-bit word containing the desired starting pixel,
and bits[3:0] should be between 0 and 11, indicating which
of the 12-pixels in that 128-bit word is the first pixel to be used
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Run the PoE hat fan with reduced noise when possible. The PoE hat fan will spin even at PWM=1 and spins almost silent. Tested at 17ºC, the fan PWM alternates between 1 and 10, which is a lot less noisy then alternating between PWM levels 0 and 31.
Without aliases for the new SPI interfaces in BCM2711, spidev instances
will be allocated sequential numbers that may not match the number of
the physical interface.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
In order to accommodate full PCI DMA access to memory on newer BCM2711
revisions, we're forced to map PCIe's view of physical memory with an
offset. This offset makes DMA addressing dependent on having 64bit
support on the PCI device's side. Which isn't always the case.
In order to mitigate this, introduce the pcie-32bit-dma overlay which
will forbid firmware from updating the default inbound memory window.
The default setting, albeit limited to accessing the lower 3GB of
memory, will allow for 32bit DMA addresses at the expense of having to
bounce buffers.
Link: https://github.com/raspberrypi/linux/issues/4197
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Experimentally have found PV on hvs4 reports fifo full
error with expected settings and does not with one less
This appears as:
[drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:82:crtc-3] flip_done timed out
with bit 10 of PV_STAT set "HVS driving pixels when the PV FIFO is full"
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
With HBR audio (8 channel 192kHz) we get occasional VC4_HD_MAI_CTL_DLATE error flags in
HDMI_MAI_CTL which seem to correspond with audio dropouts.
Increasing the normal AXI priority for dma is needed to avoid these
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
We are getting occasional VC4_HD_MAI_CTL_ERRORF in
HDMI_MAI_CTL which seem to correspond with audio dropouts.
Reduce the threshold where we deassert DREQ to avoid the fifo overfilling
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
v4l_cropcap calls our vidioc_g_pixelaspect function to get the pixel
aspect ratio, but also calls g_selection for V4L2_SEL_TGT_CROP_BOUNDS
and V4L2_SEL_TGT_CROP_DEFAULT. Whilst it allows for vidioc_g_pixelaspect
not to be implemented, it doesn't allow for either of the other two.
Add in support for the additional selection targets.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If the format is detected by the driver and a V4L2_EVENT_SOURCE_CHANGE
event is generated, then pass on the pixel aspect ratio as well.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Fixes: "staging/bcm2835-codec: Log the number of excess supported formats"
Which used %u for printing a size_t, and 64bit builds then log a warning.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
In order to get the intended behaviour of the stateful video
decoder API where only the OUTPUT queue needs to be enabled and fed
buffers in order to get the SOURCE_CHANGED event that configures the
CAPTURE queue, we want the device to run should either queue be
streaming.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The recently improved RTC_HCTOSYS option now works with RTC drivers in
modules, making it much more useful in that it removes the need to run
hwclock -s in order to synchronise the system clock.
See: https://github.com/raspberrypi/linux/issues/4205
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The i2c_csi_dsi parameter of the i2c-rtc overlay (added for the CM4IO
board) causes the RTC devices to be probed on the I2C0 bus appearing
on GPIOs 44 and 45. However, it didn't enable the other nodes necessary
for it to work - "dtparam=i2c_vc=on" was also required.
Fix that.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The kernel modules aes-neon-blk and aes-neon-bs perform poorly, at least on
Cortex-A72 without crypto extensions. In fact, aes-arm64 outperforms them
on benchmarks, despite it being a simpler implementation (only accelerating
the single-block AES cipher).
For modes of operation where multiple cipher blocks can be processed in
parallel, aes-neon-bs outperforms aes-neon-blk by around 60-70% and aes-arm64
is another 10-20% faster still. But the difference is even more marked with
modes of operation with dependencies between neighbouring blocks, such as
CBC encryption, which defeat parallelism: in these cases, aes-arm64 is
typically around 250% faster than either aes-neon-blk or aes-neon-bs.
The key trade-off with aes-arm64 is that the look-up tables are situated in
RAM. This leaves them potentially open to cache timing attacks. The two other
modules, by contrast, load the look-up tables into NEON registers and so are
able to perform in constant time.
This patch aims to load aes-arm64 more often.
If none of the currently-loaded crypto modules implement a given algorithm,
a new one is typically selected for loading using a platform-neutral alias
describing the required algorithm. To enable users to still
load aes-neon-blk or aes-neon-bs if they really want them, while still
ensuring that aes-arm64 is usually selected, remove the aliases from
aes-neonbs-glue.c and aes-glue.c and apply them to aes-cipher-glue.c, but
still build the two NEON modules.
Since aes-glue.c can also be used to build aes-ce-blk, leave them enabled
if USE_V8_CRYPTO_EXTENSIONS is defined, to ensure they are selected if we
in future use a CPU which has the crypto extensions enabled.
Note that the algorithm priority specifiers are unchanged, so if
aes-neon-bs is loaded at the same time as aes-arm64, the former will be
used in preference. However, aes-neon-blk and aes-arm64 have tied priority,
so whichever module was loaded first will be used (assuming aes-neon-bs is
not loaded).
Signed-off-by: Ben Avison <bavison@riscosopen.org>
If multiple sets of interrupts occur simultaneously, it may be unsafe
to swap buffers, as the hardware may already be re-using the current
buffers. In such cases, avoid swapping buffers, and wait for the next
opportunity at the Frame End interrupt to signal completion.
Additionally, check the packet compare status when watching for frame
end for buffers swaps, as this could also signify a frame end event.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
This reverts commit 92516cd97f.
Thi commit "Bluetooth: Always request for user confirmation for Just
Works" prevents BLE devices pairing in (at least) the Raspberry Pi OS
GUI. After reverting it, pairing works again.
If another solution to the problem is found then this reversion will
be removed.
See: https://github.com/raspberrypi/linux/issues/4139
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This reverts commit ffee202a78.
The commit "Bluetooth: Always request for user confirmation for Just
Works" prevents BLE devices pairing in (at least) the Raspberry Pi OS
GUI. After reverting it, pairing works again. Although this companion
commit ("... (LE SC)") has not been demonstrated to be problematic,
it follows the same logic and therefore could affect some use cases.
If another solution to the problem is found then this reversion will
be removed.
See: https://github.com/raspberrypi/linux/issues/4139
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The struct imx477 *ctrl parameter is not used in the function
imx477_adjust_exposure_range(), so remove it.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The only field in this struct that is used is the format code, so
replace the struct with this single field.
Save the format code in imx477_set_pad_format() when setting up a new
mode so that imx477_get_pad_format() performs the right lookup.
Otherwise, this caused a bug where the mode lookup occurred on the
12-bit table rather than the 10-bit table.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The existing 1012x760 120 fps mode has significant IQ problem using
the internal sensor scaler. Replace this mode with a 1332x990 120 fps
mode instead. This new mode has a smaller field of view, but does not
suffer from the bad IQ of the original mode.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
When vblank changes we must modify the exposure range. Also, with this
sensor, the effective exposure time implicitly changes when vblank
does, so we have to reset it after every vblank update.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Add support for very long exposures by using the exposure multiplier
register. Userland does not need to pass any additional controls to
enable long exposures, it simply requests a larger vblank to extend the
exposure control range appropriately.
Currently, since hblank is fixed, a maximum of approximately 124 seconds
of exposure time can be used. In a future change, hblank could also be
controlled in userland to give over 200 seconds of exposure time.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The V4L2_CID_EXPOSURE_AUTO_PRIORITY was used to let the sensor control
frame length (effectively framerate) based on the requested exposure
time requested. Remove this feature as it is never used, and goes
against how V4L2 likes to handle exposure and vblank controls.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
If the shutdown process is delayed enough to trigger the shutdown
timeout then one or more states in the shutdown sequence might be
skipped. Ensure that all LEDs are turned off regardless by explicitly
doing so in the shutdown state, as an example of good practices.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The driver is intended to jump directly to a shutdown state in the
event of a timeout during shutdown, but the sense of the test was
inverted.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Add gpio-fsm sysfs entries under /sys/class/gpio-fsm. For each state
machine show the current state, which state (if any) will be entered
after a delay, and the current value of that delay.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
DPI hasn't really been used up until now, so the default has
been meaningless.
In theory we should be able to pass the desired format for the
adjacent bridge chip through, but framework seems to be missing
for that.
As the main device to use DPI is the VGA666 or Adafruit Kippah,
both of which use RGB666, change the default to being RGB666 instead
of RGB888.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
VGA666 uses "vga-connector" from DRM_DISPLAY_CONNECTOR, and
"dumb-vga-dac" from DRM_SIMPLE_BRIDGE to connect up, so add them
to the defconfigs.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
VGA666 doesn't use the DE or PCLK signals, therefore there is
no point in claiming their use. It's also then possible to
use GPIOs 0&1 for DDC to read the EDID from the display.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If realloc to increase coeff size fails then attempt to re-allocate
the original size. If that also fails then flag a fatal error to abort
all further decode.
Signed-off-by: John Cox <jc@kynesim.co.uk>
According to CTA-861 specification, HDR static metadata data block allows a
sink to indicate which HDR metadata types it supports by setting the SM_0 to
SM_7 bits. Currently, only Static Metadata Type 1 is supported and this is
indicated by setting the SM_0 bit to 1.
However, the connector->hdr_sink_metadata.hdmi_type1.metadata_type is always
0, because hdr_metadata_type() in drm_edid.c checks the wrong bit.
This patch corrects the HDMI_STATIC_METADATA_TYPE1 bit position.
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@nxp.com>
The gpio-fsm property "num-soft-gpios" triggers a kernel DT checker
that warns about the lack of #gpio-cells on a random node with the
phandle that just happens to match the number of soft GPIOs. Rename
the property to "num-swgpios" to avoid the warning.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
As of 5.10, the Device Tree parser warns about properties that look
like references to "suppliers" of various services. "num-soft-gpios"
resembles a declaration of a GPIO called "num-soft", causing the value
to be interpreted as a phandle, the owner of which is checked for a
"#gpio-cells" property.
To avoid this warning, rename the gpio-fsm property to "num-swgpios".
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
This reverts commit 93a3b097c4.
GIC interrupt 117 is shared with the HDMI interrupt controller
and all the standard BSC I2C controllers. Whilst the BSC driver
flags the interrupt as shared, there doesn't appear to be an
easy way to flag the controller as being on a shared parent
interrupt.
Revert to polling the I2C controllers.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Adds 2 msecs delay when switching between oscillators to allow
correct PLL settling.
Thanks to Clive Messer for the support!
Signed-off-by: Joerg Schambacher <joerg@hifiberry.com>
Each supported format now includes a mask showing the allowed colour
spaces, as well as a default colour space for when one was not
specified.
Additionally we translate the colour space to mmal format and pass it
over to the VideoCore.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
The cherry-pick of the patch that added the greyworld AWB mode
was incomplete. Fix it up.
Fixes: b3ef481fe2 "staging: bcm2835-camera: Add greyworld AWB mode"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Unlike the previous generations, the HSM clock limitation is way above
what we can reach without scrambling, so let's move the maximum
frequency we support to the maximum clock frequency without scrambling.
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The CEC and hotplug interrupts go through an interrupt controller shared
between the two HDMI controllers.
Let's add that interrupt controller and the interrupts for both HDMI
controllers
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The BSC controllers used for the HDMI DDC have an interrupt controller
shared between both instances. Let's add it to avoid polling.
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The CEC and hotplug interrupts were missing when that binding was
introduced, let's add them in now that we've figured out how it works.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
We introduced the BCM2711 support to the vc4 HDMI controller with 5.10,
but this was lacking any of the interrupts of the CEC controller so we
have to deal with the backward compatibility.
Do so by simply ignoring the CEC setup if the DT doesn't have the
interrupts property.
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The HDMI controller found in the BCM2711 has an external interrupt
controller for the CEC and hotplug interrupt shared between the two
instances.
Let's add a variant flag to register a single interrupt handler and
deals with the interrupt handler setup, or two interrupt handlers
relying on an external irqchip.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The BCM2711 has two different interrupt sources to transmit and receive
CEC messages, provided through an external interrupt chip shared between
the two HDMI interrupt controllers.
The rest of the CEC controller is identical though so we need to change
a bit the code organisation to share the code as much as possible, yet
still allowing to register independant handlers.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
While the BCM2835 had the CEC clock derived from the HSM clock, the
BCM2711 has a dedicated parent clock for it.
Let's introduce a separate clock for it so that we can handle both
cases.
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
As part of the enable sequence we might change the HSM clock rate if the
pixel rate is different than the one we were already dealing with.
On the BCM2835 however, the CEC clock derives from the HSM clock so any
rate change will need to be reflected in the CEC clock divider to output
40kHz.
Fixes: cd4cb49dc5 ("drm/vc4: hdmi: Adjust HSM clock rate depending on pixel rate")
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The CEC clock divider needs to output a frequency of 40kHz from the HSM
rate on the BCM2835. The driver used to have a fixed frequency for it,
but that changed for the BCM2711 and we now need to compute it
dynamically to maintain the proper rate.
Fixes: cd4cb49dc5 ("drm/vc4: hdmi: Adjust HSM clock rate depending on pixel rate")
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Currently we call cec_phys_addr_invalidate on a hotplug deassert.
That may be due to a TV power cycling, or an AVR being switched
on (and switching edid).
This makes CEC unusable since our controller wouldn't have a physical
address anymore.
Set it back up again on the hotplug assert.
Fixes: 15b4511a4a ("drm/vc4: add HDMI CEC support")
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The commit 311e305fdb ("drm/vc4: hdmi: Implement a register layout
abstraction") forgot one CEC register, and made a copy and paste mistake
for another one. Fix those mistakes.
Fixes: 311e305fdb ("drm/vc4: hdmi: Implement a register layout abstraction")
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The code prior to 311e305fdb ("drm/vc4: hdmi: Implement a register
layout abstraction") was relying on the fact that the register offset
was incremented by 4 for each readl call. That worked since the register
width is 4 bytes.
However, since that commit the HDMI_READ macro is now taking an enum,
and the offset doesn't increment by 4 but 1 now. Divide the index by 4
to fix this.
Fixes: 311e305fdb ("drm/vc4: hdmi: Implement a register layout abstraction")
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The hdmi reset got moved to a later point in the commit 9045e91a47
("drm/vc4: hdmi: Add reset callback").
However, the reset now occurs after vc4_hdmi_cec_init and so tramples
the setup of registers like HDMI_CEC_CNTRL_1
This only affects pi0-3 as on pi4 the cec registers are in a separate
block
Fixes: 9045e91a47 ("drm/vc4: hdmi: Add reset callback")
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The BCM2711 has a number of instances of interrupt controllers handled
by the driver behind the BRCMSTB_L2_IRQ Kconfig option (irq-brcmstb-l2).
Let's select that driver as part of the ARCH_BCM2835 Kconfig option.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Under certain circumstance the DONE flag can appear to be set early.
Fortunately the TA flag is often still set at that time, and it can be
used as an indication that DONE is invalid.
Handle the other cases - when TA is not set but not all data is
available - by silently accepting it and hoping for a second DONE.
See: https://github.com/raspberrypi/linux/issues/3064
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The number is only a hint, but may as well be correct.
Fixes: 471e0029e9 ("media: i2c: imx290: Convert HMAX setting into V4L2_CID_HBLANK")
Fixes: be0b9b7ad1 ("media: i2c: imx290: Add support for V4L2_CID_VBLANK")
Fixes: 8483f0d759 ("media: i2c: imx290: Add exposure control to the driver.")
Fixes: 9764f3459c ("media: i2c: imx290: Add H and V flip controls")
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Most software (including libcamera) requires V4L2_CID_ANALOGUE_GAIN,
not V4L2_CID_GAIN.
The range for the control is 0 to 100 for which the sensor uses only
analogue gain; higher values would involve digital gain which this
control should not apply.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Adds the necessary GPIO handling and ALSA mixer extensions.
Also fixes a problem with the PLL/CLK control when switching sample rates.
Thanks to Clive Messer for the support!
Signed-off-by: Joerg Schambacher <joerg@hifiberry.com>
Commit 0a038c1c29 ("drm/vc4: Move LBM creation out of
vc4_plane_mode_set()") changed the LBM allocation logic from first
allocating the LBM memory for the plane to running mode_set,
adding a gap in the LBM, and then running the dlist allocation filling
that gap.
The gap was introduced by incrementing the dlist array index, but was
never checking whether or not we were over the array length, leading
eventually to memory corruptions if we ever crossed this limit.
vc4_dlist_write had that logic though, and was reallocating a larger
dlist array when reaching the end of the buffer. Let's share the logic
between both functions.
Cc: Boris Brezillon <boris.brezillon@collabora.com>
Cc: Eric Anholt <eric@anholt.net>
Fixes: 0a038c1c29 ("drm/vc4: Move LBM creation out of vc4_plane_mode_set()")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Much effort has been put into finding ways to avoid warnings from dtc
about overlays, usually to do with the presence of #address-cells and
size-cells, but not exclusively so. Since the issues being warned about
are harmless, suppress the warnings to declutter the build output and
to avoid alarming users.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
A relatively recent commit ([1]) contained optimisation for the PIO
SPI FIFO-filling functions. The commit message includes the phrase
"[t]he blind and counted loops are always called with nonzero count".
This is technically true, but it is still possible for count to become
zero before the loop is entered - if tfr->len is zero. Moving the loop
exit condition to the end of the loop saves a few cycles, but results
in a near-infinite loop should the revised count be zero on entry.
Strangely, zero-lengthed transfers aren't filtered by the SPI framework
and, even more strangely, the Python3 spidev library is triggering them
for no obvious reason.
Avoid the problem completely by bailing out of the main transfer
function early if trf->len is zero, although there may be a case for
moving the mitigation into the framework.
See: https://github.com/raspberrypi/linux/issues/4100
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
[1] 26751de25d ("spi: bcm2835: Micro-optimise FIFO loops")
RPi4's co-processor will copy the board's bootloader[1] configuration
into memory for the OS to consume. Specifically, for the bootloader
configuration and upgrade user-space routines to query it through
nvmem's sysfs interface.
Introduce a reserved-memory area template for the co-processor to edit
before booting the system so as for Linux not to overwrite that memory
and to expose it as an nvmem device.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
[1] https://www.raspberrypi.org/documentation/hardware/raspberrypi/bcm2711_bootloader_config.md
---
Changes since v1:
- Introduce compatible string
- Change alias name to something more explicit
Firmware/co-processors might use reserved memory areas in order to pass
data stemming from an nvmem device otherwise non accessible to Linux.
For example an EEPROM memory only physically accessible to firmware, or
data only accessible early at boot time.
In order to expose this data to other drivers and user-space, the driver
models the reserved memory area as an nvmem device.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
---
Changes since v1:
- Remove reserved memory phandle indirection by directly creating a
platform device from the reserved memory DT node
- Only map memory upon reading it to avoid corruption
- Small cosmetic cleanups
Firmware/co-processors might use reserved memory areas in order to pass
data stemming from an nvmem device otherwise non accessible to Linux.
For example an EEPROM memory only physically accessible to firmware, or
data only accessible early at boot time.
Introduce the dt-bindings to nvmem's rmem.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
---
Changes since v1:
- Update schema to new driver design
Add colour denoise control to the bcm2835 driver through a new v4l2
control: V4L2_CID_USER_BCM2835_ISP_CDN.
Add the accompanying MMAL configuration structure definitions as well.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The arm64 version of bcm2711_defconfig is intended for Pi 4, but the
Raspberry Pi OS kernel8.img built from it is also used for Pi 3. It is
therefore necessary to include some legacy drivers such as the BCM2835
thermal driver, otherwise there is no support for reading the CPU
temperature on a Pi 3 in 64-bit mode.
See: https://github.com/raspberrypi/linux/issues/4077
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
LBM base address is measured in units of pixels per cycle.
That is 4 for 2711 (hvs5) and 2 for 2708.
We are wasting 75% of lbm by indexing without the scaling.
But we were also using too high a size for the lbm resulting
in partial corruption (right hand side) of vertically
scaled images, usually at 4K or lower resolutions with more layers.
The physical RAM of LBM on 2711 is 8 * 1920 * 16 * 12-bit
(pixels are stored 12-bits per component regardless of format).
The LBM adress indexes work in units of pixels per clock,
so for 4 pixels per clock that means we have 32 * 1920 = 60K
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
This reverts commit 1c15edc0dca002c8536e9f1f5e1ec43017815018.
This revert and its neighbour are opposites. When squashing, delete
the original commits as well.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
It turns out the used CAN SPI chip is not a good way to identify the version of
the CAN HAT.
There are two different board layouts of the Seeed Studio CAN BUS FD HAT. The
v1 board doesn't have a battery holder, while the v2 board has. Update the
overlay README accordinly.
Link: https://github.com/Seeed-Studio/seeed-linux-dtoverlays/issues/13
Cc: Patrick Menschel <menschel.p@posteo.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Update those overlays that use the regulator framework to use the
new cam1_reg node to control the camera shutdown line, and remove
the firmware workaround nodes.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The current firmware fixup of camera sensor overlays is not
particularly nice, and it stops you being able to load them
dynamically.
It's also incompatible with creating a simple DT that can be
loaded for both CAM1 and CAM0 on a CM as they would both
try to claim the one GPIO.
Almost all sensors have a hook of some form for a regulator, so
it's relatively straightforward to convert them all to use a
fixed regulator with GPIO control.
Add a fixed regulator node for each platform with the GPIO
correctly configured for the camera shutdown line. (The LED line
is ignored).
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
This patch adds the overlay for the Seeed Studio CAN BUS FD HAT v1 with two CAN
FD Channels (based on mcp2517fd).
https://www.seeedstudio.com/2-Channel-CAN-BUS-FD-Shield-for-Raspberry-Pi-p-4072.html
The overlay was generated by:
ovmerge -c spi1-1cs-overlay.dts,cs0_pin=18,cs0_spidev=false \
mcp251xfd-overlay.dts,spi0-0,interrupt=25 \
mcp251xfd-overlay.dts,spi1-0,interrupt=24
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
There are several versions of the Seeed Studio CAN BUS FD HAT. This is the
second version, based on the mcp2518fd, so give it a -v2 postfix.
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
To comply with the intended usage of the V4L2 selection target when
used to retrieve a sensor image properties, adjust the rectangles
returned by the ov5647 driver.
The top/left crop coordinates of the TGT_CROP rectangle were set to
(0, 0) instead of (16, 16) which is the offset from the larger physical
pixel array rectangle. This was also a mismatch with the default values
crop rectangle value, so this is corrected. Found with v4l2-compliance.
While at it, add V4L2_SEL_TGT_CROP_BOUNDS support: CROP_DEFAULT and
CROP_BOUNDS have the same size as the non-active pixels are not readable
using the selection API. Found with v4l2-compliance.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Use (reserved) bits 24 and 25 of the dreq value
(the second cell of the DT DMA descriptor) to request
that wide source reads or wide dest writes are required
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
When logging that the firmware has provided more supported formats
than we had allocated storage for, log the number allocated and
returned.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Now that the firmware supports the unpacked (16bpp) variants
of the MIPI raw formats, add the mappings.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
When logging that the firmware has provided more supported formats
than we had allocated storage for, log the number allocated and
returned.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Now that the firmware supports the unpacked (16bpp) variants
of the MIPI raw formats, add the mappings.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Support has been added for the unpacked (16bpp) versions of
the MIPI raw 10, 12, and 14 formats, so add the 4CCs for them.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
For some reason lost in history function vchiq_mmal_init used
a static variable for storing the vchiq_instance.
This value is retrieved from vchiq per instance, so worked fine
until you try to call vchiq_mmal_init multiple times concurrently
when things then go wrong. This seemed to happen quite frequently
if using the cutdown firmware (no MMAL or VCSM services running)
as the vchiq_connect then failed, and one or other vchiq_shutdown
was working on an invalid handle.
Remove the static so that each caller gets a unique vchiq_instance.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The recent change to the bulk transfer compat function missed the fact
the relevant ioctl command is VCHIQ_IOC_QUEUE_BULK_TRANSMIT32, not
VCHIQ_IOC_QUEUE_BULK_TRANSMIT, as any attempt to send a bulk block
to the VPU would have shown.
Fixes: a4367cd2b2 ("staging: vchiq: convert compat bulk transfer")
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
error logs:
drivers/staging/vc04_services/vc-sm-cma/Kconfig:1:error: recursive dependency detected!
drivers/staging/vc04_services/vc-sm-cma/Kconfig:1: symbol BCM_VC_SM_CMA is selected by BCM2835_VCHIQ_MMAL
drivers/staging/vc04_services/vchiq-mmal/Kconfig:1: symbol BCM2835_VCHIQ_MMAL depends on BCM2835_VCHIQ
drivers/staging/vc04_services/Kconfig:14: symbol BCM2835_VCHIQ is selected by BCM_VC_SM_CMA
For a resolution refer to Documentation/kbuild/kconfig-language.rst
subsection "Kconfig recursive dependency limitations"
Tested-by: make ARCH=arm64 bcm2711_defconfig
Test platform: fedora 33
Branch: rpi-5.10.y
This patch adds the overlay for the Seeed Studio CAN BUS FD HAT
with two CAN FD Channels and an RTC.
https://www.seeedstudio.com/CAN-BUS-FD-HAT-for-Raspberry-Pi-p-4742.html
The overlay was generated by
ovmerge -c mcp251xfd-overlay.dts,spi0-0,interrupt=25 \
mcp251xfd-overlay.dts,spi0-1,interrupt=24 \
i2c-rtc-overlay.dts,pcf85063
Also, add description on how to generate overlays
Signed-off-by: Patrick Menschel <menschel.p@posteo.de>
The addition of the local 'userdata' pointer to
vchiq_irq_queue_bulk_tx_rx omitted the case where neither BLOCKING nor
WAITING modes are used, in which case the value provided by the
caller is replaced with a NULL.
Fixes: 4184da4f31 ("staging: vchiq: fix __user annotations")
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The BCM2711 supports higher bpc count than just 8, so let's support it in
our driver.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Unlike the previous generations, the HSM clock limitation is way above
what we can reach without scrambling, so let's move the maximum
frequency we support to the maximum clock frequency without scrambling.
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The PHY initialisation parameters are not based on the pixel clock but
the TMDS clock rate which can be the pixel clock in the standard case,
but could be adjusted based on some parameters like the bits per color.
Since the TMDS clock rate is stored in our custom connector state
already, let's reuse it from there instead of computing it again.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
The pixel rate is for now quite simple to compute, but with more features
(30 and 36 bits output, YUV output, etc.) will depend on a bunch of
connectors properties.
Let's store the rate we have to run the pixel clock at in our custom
connector state, and compute it in atomic_check.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
When run with a higher bpc than 8, the clock of the HDMI controller needs
to be adjusted. Let's create a connector state that will be used at
atomic_check and atomic_enable to compute and store the clock rate
associated to the state.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
drm_atomic_helper_connector_reset uses kmalloc which, from an API
standpoint, can fail, and thus setting connector->state to NULL.
However, our reset hook then calls drm_atomic_helper_connector_tv_reset
that will access connector->state without checking if it's a valid
pointer or not.
Make sure we don't end up accessing a NULL pointer.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Suggested-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Commit 63495f6b4a ("drm/vc4: hdmi: Make sure our clock rate is within
limits") was intended to compute the pixel rate to make sure we remain
within the boundaries of what the hardware can provide.
However, unlike what mode_valid was checking for, we forgot to take
into account the clock doubling flag that can be set for modes. Let's
honor that flag if it's there.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Reported-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Fixes: 63495f6b4a ("drm/vc4: hdmi: Make sure our clock rate is within limits")
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
We'll need to access the connector state in our encoder setup, so let's
just pass the whole DRM state to our private encoder hooks.
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Since the CRTC setup in vc4 is split between the PixelValves/TXP and the
HVS, only the PV/TXP atomic hooks were updated in the previous commits, but
it makes sense to update the HVS ones too.
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
It appears that the V4L2 M2M framework requires the driver to manually
call vb2_clear_last_buffer_dequeued on the CAPTURE queue during a
V4L2_DEC_CMD_START.
Add such a call.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Control V4L2_CID_MPEG_VIDEO_HEADER_MODE controls whether the encoder
is meant to emit the header bytes as a separate packet or with the
first encoded frame.
Add support for it.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
To comply with the intended usage of the V4L2 selection target when
used to retrieve a sensor image properties, adjust the rectangles
returned by the imx477 driver.
The top/left crop coordinates of the TGT_CROP rectangle were set to
(0, 0) instead of (8, 16) which is the offset from the larger physical
pixel array rectangle. This was also a mismatch with the default values
crop rectangle value, so this is corrected. Found with v4l2-compliance.
While at it, add V4L2_SEL_TGT_CROP_BOUNDS support: CROP_DEFAULT and
CROP_BOUNDS have the same size as the non-active pixels are not readable
using the selection API. Found with v4l2-compliance.
This commit mirrors 543790f777 done for
the imx219 sensor.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The overlays for the ads1015 and ads1115 I2C ADCs omitted the addresses
in the main device node names. As well as breaking the conventions for
I2C devices, this prevents the firmware from renaming them when the
"reg" property is modified, which in turn stops the overlays from being
instantiated multiple times.
See: https://www.raspberrypi.org/forums/viewtopic.php?f=107&t=294465
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Eric's view was that there was no point in having zpos
support on vc4 as all the planes had the same functionality.
Can be later squashed into (and fixes):
drm/vc4: Add firmware-kms mode
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
The mpu6050 starts up at address 0x68 by default, but can be set to
0x69 if the ADO pin is pulled high. Give the overlay an addr parameter
to allow devices at the alternate address to be used.
See: https://github.com/Hexxeh/rpi-firmware/issues/252
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
lan78xx_link_reset explicitly clears the MAC's view of the PHY's IRQ
status. In doing so it potentially leaves the PHY with a pending
interrupt that will never be acknowledged, at which point no further
interrupts will be generated.
Avoid the problem by acknowledging any pending PHY interrupt after
clearing the MAC's status bit.
See: https://github.com/raspberrypi/linux/issues/2937
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The firmware by default tries to ensure that decoded frame
timestamps always increment. This is counter to the V4L2 API
which wants exactly the OUTPUT queue timestamps passed to the
CAPTURE queue buffers.
Disable the firmware option.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add support for the PCF85063 and PCF85063A RTC devices to the
i2c-rtc overlay.
Also enable the device to be used on i2c0 (i2c_vc) on GPIOs 0&1 (use
parameter "i2c0") and GPIOs 44 & 45 (use parameter "i2c_csi_dsi").
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Although the BRCMSTB PCIe interface doesn't technically support the
MSI-X spec, in practise it seems to work provided no more than 32
MSI-Xs are required. Add the required flag to the driver to allow
experimentation with devices that demand MSI-X support.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Commit 65e08c4650 failed to clear the
clock state when the device stopped streaming. Fix this, as it might
again cause the same problems when doing an unprepare.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
clk_disable_unprepare() is called unconditionally in stop_streaming().
This is incorrect in the cases where start_streaming() fails, and
unprepares all clocks as part of the failure cleanup. To avoid this,
ensure that clk_disable_unprepare() is only called in stop_streaming()
if the clocks are in a prepared state.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
On a failure in start_streaming(), the error code would not propagate to
the calling function on all conditions. This would cause the userland
caller to not know of the failure.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The DSI1_PHY_AFEC0_PD_DLANE1 and DSI1_PHY_AFEC0_PD_DLANE3 register
definitions were swapped, so trying to use more than a single data
lane failed as lane 1 would get powered down.
(In theory a 4 lane device would work as all lanes would remain
powered).
Correct the definitions.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The hdmi reset got moved to a later point in
"drm/vc4: hdmi: Add reset callback"
which now occurs after vc4_hdmi_cec_init
and so tramples the setup of registers like
HDMI_CEC_CNTRL_1
This only affects pi0-3 as on pi4 the cec
resgisters are in a separate block
Fixes: ed9a1f6eb4
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
BCM2711 DSI1 doesn't have the issue with the ARM not being
able to write to the registers, therefore remove the DMA
workaround for that compatible string.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
DSI1 on BCM2711 doesn't require the DMA workaround that is used
on BCM2835/6/7, therefore it needs a new compatible string.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
DSI0 was partially supported, but didn't register with the main
driver, and the code was inconsistent as to whether it checked
port == 0 or port == 1.
Add compatible string and other support to make it consistent.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Breaks out common register set and adds the different registers
for 1280x720 (cropped) and 640x480 (skipped) modes
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
CONFIG_BRIDGE_NETFILTER=m used to be the default when CONFIG_NETFILTER
was enabled, but that was removed in 5.9. The way that defconfigs work
caused this wanted setting to be lost in rpi-5.9.y and rpi-5.10.y -
restore it now.
See: https://github.com/Hexxeh/rpi-firmware/issues/248
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The CAN_ISOTP setting was in the wrong position, and it's better for
bisecting and reverting if this doesn't get rolled into the next
config-changing commit.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
In switching to the hardware I2C controller there is an issue
where we seem to not get back the correct state from the Pi
touchscreen.
Insert a delay before polling to avoid this condition.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
We now have the hardware I2C controller pinmuxed to the drive the
display I2C, but this controller does not support clock stretching.
The Atmel micro-controller in the panel requires clock stretching
to allow it to prepare any data to be read.
Split the rpi_touchscreen_i2c_read into two independent transactions with
a delay between them for the Atmel to prepare the data.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
This touchscreen controller is used by the 7" DSI panel, and
this overlay configures it for when it is NOT being polled by
the firmware.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Not all systems have the interrupt line wired up, so switch to
polling the touchscreen off a timer if no interrupt line is
configured.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Using a hdmi analyser the bytes in packet ram
registers beyond the length were visible in the
infoframes and it flagged the checksum as invalid.
Zeroing unused words of packet RAM avoids this
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
[1] replaced a single reset function with a pointer to one of two
implementations, but also removed the call asserting the reset
at the start of brcm_pcie_setup. Doing so breaks Raspberry Pis with
VL805 XHCI controllers lacking dedicated SPI EEPROMs, which have been
used for USB booting but then need to be reset so that the kernel
can reconfigure them. The lack of a reset causes the firmware's loading
of the EEPROM image to RAM to fail, breaking USB for the kernel.
See: https://www.raspberrypi.org/forums/viewtopic.php?p=1758157#p1758157
Fixes: 04356ac307 ("PCI: brcmstb: Add bcm7278 PERST# support")
[1] 04356ac307 ("PCI: brcmstb: Add bcm7278 PERST# support")
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
* Add mcp251xfd driver module to the RPi kernel builds.
* Add isotp can protocol module to the RPi kernel builds.
Signed-off-by: Patrick Menschel <menschel.p@posteo.de>
The enable_headphones parameter of the snd_bcm2835 module is forced
to 1 if enable_compat_alsa is 0, so setting them both on the kernel
command line is pointless (and, in the case of Pi 400 and Pi Zeroes,
confusing).
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The sdtweak overlay has been superseded by the board-specific
sd_* parameters such as sd_poll_once, sd_overclock etc.
For example, replace:
dtoverlay=sdtweak,poll_once
with:
dtparam=sd_poll_once
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The last nibble is a revision ID, and the 54213pe is a later rev
than the 54210e. Running the 54210e setup code on a 54213pe results
in a broken RGMII interface.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
Define a new mailbox (SET_REBOOT_FLAGS) which may be used to
pass optional flags to the Raspberry Pi firmware that changes
the behaviour of the bootloader and firmware during a reboot.
Currently this just defines the 'tryboot' flag which causes
the firmware to load tryboot.txt instead config.txt. This
alternate configuration file can be used to specify the
path of an alternate firmware and kernels allowing a fallback
mechanism to be implemented for OS upgrades.
Delete a large amount of unused declaration from "usb.h", some of which
were causing build warnings, and get the module building cleanly.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
RASPBERRYPI_FIRMWARE was missing from the arm64 bcm2711_defconfig.
bcmrpi3_defconfig was also missing a few options. Add the misssing
settings and regenerate all the Pi defconfigs.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Attempts to connect external GPUs to Compute Module 4's PCIe bus have
highlighted that the existing "outbound window" - the fraction of the
PCI address base that is appears in the host's memory map - is
restrictively small. Expand the window to a full 1GB.
See: https://www.raspberrypi.org/forums/viewtopic.php?f=98&t=288902
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Add a property to allow the headphone output to be disabled. Use an
integer property rather than a boolean so that an overlay can clear it.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The VL805 controller can't cope with the TR Dequeue Pointer for an endpoint
being set to a Link TRB. The hardware-maintained endpoint context ends up
stuck at the address of the Link TRB, leading to erroneous ring expansion
events whenever the enqueue pointer wraps to the dequeue position.
If the search for the end of the current TD and ring cycle state lands on
a Link TRB, move to the next segment.
See: https://github.com/raspberrypi/linux/issues/3919
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.com>
With vc4-drv node not being under /soc on Pi4, we need to
adopt the correct DMA parameters from a suitable sub-component.
Add "brcm,bcm2711-hvs" to that list of components.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The upstream driver is checking the link frequency parameter, and
the overlay had the wrong value.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Composite gets enabled automatically if HDMI isn't detected,
which can cause some grief in X should it be not connected
and touchscreens are in use.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The upstream bcm2837.dtsi uses cortex-a53-pmu, so we can do the same
but with a fallback to the cortex-a7-pmu which is supported by the
32-bit kernel.
Now that we're using the natural fallback mechanism of compatible
strings, the RPI364 macro no longer serves any purpose - remove it.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The vc5 HDMI registers hadn't been added into the debugfs
register sets, therefore weren't dumped on request.
Add them in.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The order of precedence should be:
- hotplug GPIO being defined
- DDC probe
- hotplug register
In particular the hotplug register is not valid if a GPIO is defined
(eg on Pi0-3), but was being checked.
Fixes "943f078 vc4: cec: Restore cec physical address on reconnect"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The gpio-fsm driver implements simple state machines that allow GPIOs
to be controlled in response to inputs from other GPIOs - real and
soft/virtual - and time delays. It can:
+ create dummy GPIOs for drivers that demand them,
+ drive multiple GPIOs from a single input, with optional delays,
+ add a debounce circuit to an input,
+ drive pattern sequences onto LEDs
etc.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Overlays are unable to remove properties in the base DTB, but they
can overwrite them. Allow a present but empty 'dmas' property
to also disable the HDMI audio interface.
See: https://github.com/raspberrypi/linux/issues/2489
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Driver for the BCM2835 ISP hardware block. This driver uses the MMAL
component to program the ISP hardware through the VC firmware.
The ISP component can produce two video stream outputs, and Bayer
image statistics. This can't be encompassed in a simple V4L2
M2M device, so create a new device that registers 4 video nodes.
This patch squashes all the development patches from the earlier
rpi-5.4.y branch into one
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
This file defines the userland interface to the bcm2835-isp driver
that will follow in a separate commit.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If CONFIG_DMA_BCM2708 isn't enabled there's no need to mask out
one of the already scarce DMA channels.
Signed-off-by: Matthias Reichl <hias@horus.com>
This adds a V4L2 memory to memory device that wraps the MMAL
video decode and video_encode components for H264 and MJPEG encode
and decode, MPEG4, H263, and VP8 decode (and MPEG2 decode
if the appropriate licence has been purchased).
This patch squashes all the work done in developing the driver
on the Raspberry Pi rpi-5.4.y kernel branch.
Thanks to Kieran Bingham, Aman Gupta, Chen-Yu Tsai, and
Marek Behún for their contributions. Please refer to the
rpi-5.4.y branch for the full history.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
With the vc-sm-cma driver we can support zero copy of buffers between
the kernel and VPU. Add this support to mmal-vchiq.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add Broadcom VideoCore Shared Memory support.
This new driver allows contiguous memory blocks to be imported
into the VideoCore VPU memory map, and manages the lifetime of
those objects, only releasing the source dmabuf once the VPU has
confirmed it has finished with it.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If the RBUF logic is not reset when the kernel starts then there
may be some data left over from any network boot loader. If the
64-byte packet headers are enabled then this can be fatal.
Extend bcmgenet_dma_disable to do perform the reset, but not when
called from bcmgenet_resume in order to preserve a wake packet.
N.B. This different handling of resume is just based on a hunch -
why else wouldn't one reset the RBUF as well as the TBUF? If this
isn't the case then it's easy to change the patch to make the RBUF
reset unconditional.
See: https://github.com/raspberrypi/linux/issues/3850
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Increase the delay before entering the lower power state to 2 seconds
(the maximum allowed) in order to reduce the packet latencies,
particularly for inbound packets.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Display variants are intended as a replacement for the now-deleted
fbtft_device drivers. Drivers can register additional compatible
strings with a custom callback that can make the required changes
to the fbtft_display structure.
Start the ball rolling by adding adafruit18, adafruit18_green and
sainsmart18 displays.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Since the unicam driver was modified to write to a dummy buffer when no
user-supplied buffer is available, it can now write to and return a
buffer even when there's only a single one. Enable this by changing the
min_buffers_needed in the vb2_queue; it will be useful for enabling
still captures without allocating more memory than absolutely necessary.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
The change to retrieve the pixel format always on g_fmt didn't
check whether the native or unpacked version of the format
had been requested, and always returned the packed one.
Correct this so that the packing setting is retained whereever
possible.
Fixes "9d59e89 media: bcm2835-unicam: Re-fetch mbus code from subdev
on a g_fmt call"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
From when bringing up the driver, there was a check in the isr
to ignore interrupts (claiming them handled) should the driver
not be streaming.
The VPU now will not register a camera driver if it finds a
CSI2 node enabled in device tree, therefore this flawed check is
redundant.
https://github.com/raspberrypi/linux/issues/3602
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Parse device properties and register controls for them using the V4L2
fwnode properties helpers.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Use V4L2_CID_EXPOSURE_AUTO_PRIORITY to control if the driver should
automatically adjust the sensor frame length based on exposure time,
allowing variable frame rates and longer exposures.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Adds a driver for the 12MPix Sony IMX477 CSI2 sensor.
Whilst the sensor supports 2 or 4 CSI2 data lanes, this driver
currently only supports 2 lanes.
The following Bayer modes are currently available:
4056x3040 12-bit @ 10fps
2028x1520 12-bit (binned) @ 40fps
2028x1050 12-bit (cropped/binned) @ 50fps
1012x760 10-bit (scaled) @ 120 fps
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Under FKMS, the firmware (via FKMS) also requires the VideoCore cache
aliases for image planes, as defined by the dma-ranges under /soc.
Add rpi-firmware-kms to the list of acceptable nodes to look for
to copy dma config from.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
This is a squash of all firmware-kms related patches from previous
branches, up to and including
"drm/vc4: Set the possible crtcs mask correctly for planes with FKMS"
plus a couple of minor fixups for the 5.9 branch.
Please refer to earlier branches for full history.
This patch includes work by Eric Anholt, James Hughes, Phil Elwell,
Dave Stevenson, Dom Cobley, and Jonathon Bell.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
drm/vc4: Fixup firmware-kms after "drm/atomic: Pass the full state to CRTC atomic enable/disable"
Prototype for those calls changed, so amend fkms (which isn't
upstream) to match.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The BT601/BT709 color encoding and limited vs full
range properties were not being exposed, defaulting
always to BT601 limited range.
Expose the parameters and set the registers appropriately.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
This currently doesn't handle non-zero source rectangles correctly,
but add support for DRM_FORMAT_P030 with DRM_FORMAT_MOD_BROADCOM_SAND128
modifier to planes when running on HVS5.
WIP still for source cropping SAND/P030 formats
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Currently we call cec_phys_addr_invalidate on a hotplug deassert.
That may be due to a TV power cycling, or an AVR being switched
on (and switching edid). This makes CEC unusable.
Set it back up again on the hotplug assert.
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
This follows logic in hdmi-codec.c to use speaker layout
from ELD to choose a suitable speaker mapping based on
number of channels requested and signal that in audio
infoframe and report this back to userspace.
This allows apps like speaker-test and kodi to get the
output to the right speakers.
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Without this bit set, HDMI_MAI_FORMAT doesn't pick up
the format and samplerate from DVP_CFG_MAI0_FMT and you
can't get HDMI_HDMI_13_AUDIO_STATUS_1 to indicate HBR mode
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
This was a workaround for bugs in hardware on earlier Pi models
and wasn't totally successful.
It makes audio quality worse on a Pi4 at the higher sample rates
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Configuring HDMI audio registers in prepare allows us to take
IEC958 bits into account which are set by the alsa hook after
the hw_params call.
Signed-off-by: Matthias Reichl <hias@horus.com>
Although vc4 get an IEC958 formatted stream passed in from userspace
the driver needs the info from the channel status bits to properly
set up the hardware, eg for HBR passthrough.
Add iec958 controls so the channel status bits can be passed in
from userspace.
Signed-off-by: Matthias Reichl <hias@horus.com>
The hardware uses this for generating the right audio
data island packets when using formats other than PCM
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Symptom is random switching of speakers when using multichannel.
Repeatedly running speakertest -c8 occasionally starts with
channels jumbled. This is fixed with HD_CTL_WHOLSMP.
The other bit looks beneficial and apears harmless in testing so
I'd suggest adding it too.
Documentation says: HD_CTL_WHILSMP_SET
Wait for whole sample. When this bit is set MAI transmit will start
only when there is at least one whole sample available in the fifo.
Documentation says: HD_CTL_CHALIGN_SET
Channel Align When Overflow. This bit is used to realign the audio
channels in case of an overflow.
If this bit is set, after the detection of an overflow, equal
amount of dummy words to the missing words will be written to fifo,
filling up the broken sample and maintaining alignment.
Signed-off-by: Dom Cobley <popcornmix@gmail.com>
vc4_drv isn't necessarily under the /soc node in DT as it is a
virtual device, but it is the one that does the allocations.
The DMA addresses are consumed by primarily the HVS or V3D, and
those require VideoCore cache alias address mapping, and so will be
under /soc.
During probe find the a suitable device node for HVS or V3D,
and adopt the DMA configuration of that node.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
During a bulk transfer we request a DMA allocation to hold the
scatter-gather list. Most of the time, this allocation is small
(<< PAGE_SIZE), however it can be requested at a high enough frequency
to cause fragmentation and/or stress the CMA allocator (think time
spent in compaction here, or during allocations elsewhere).
Implement a pool to serve up small DMA allocations, falling back
to a coherent allocation if the request is greater than
VCHIQ_DMA_POOL_SIZE.
Signed-off-by: Oliver Gjoneski <ogjoneski@gmail.com>
The sensor supports 8 bit mode as well as 10bit, so add the
relevant code to allow selection of this.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Fix commit "media: tc358743: Return an appropriate colorspace from
tc358743_set_fmt" to ensure that the format passed in to set_fmt
is checked to be valid, and reset to the current format if not.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Pi 0&1 pass all ARM accesses through the VPU L2 cache, therefore
the dma-ranges property sets the cache alias bits to other
than the direct alias, hence this WARN was firing.
It was overprotective coding, so assume that everything is OK
with the dma-ranges, and remove the WARN.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
MEDIA_CONTROLLER_REQUEST_API is a hidden option. If rpivid depends on it,
the user would need to first enable another driver that selects
MEDIA_CONTROLLER_REQUEST_API, and only then rpivid would become available.
By selecting it instead of depending on it, it becomes possible to enable
rpivid without having to enable other potentially unnecessary drivers.
Signed-off-by: Hristo Venev <hristo@venev.name>
In an attempt to prevent the problem of CPUn not starting, explicitly
misalign the scratch space used to save registers acros the cache
invalidation.
Notes:
At this stage in the boot process the core is running with its cache
disabled. Before enabling the cache its contents must be explicitly
invalidated, a process that requires quite a few registers that the
caller must preserve. Evidence suggests that something is writing a
block of zeroes over that space at a time when all other cores should
be idle, possibly some kind of write-combiner, and the misalignment is
designed to disrupt any write-coalescing.
In truth, I don't understand why this patch works, and when the failure
is so random it is hard to be certain that this isn't just rolling the
dice again. One interesting test would be to change the "addeq r12, #4"s
to "addeq r12, #0"s determine see if the offset itself is significant or
just the additional code.
See: https://github.com/Hexxeh/rpi-firmware/issues/232
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Although it is no longer necessary for vchiq's children to have a
different DMA configuration to the parent, they do still need to
explicitly to have their DMA configuration set - to be that of the
parent.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The actpwr trigger is a meta trigger that cycles between an inverted
mmc0 and default-on. It is written in a way that could fairly easily
be generalised to support alternative sets of source triggers.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Parse device properties and register controls for them using the V4L2
fwnode properties helpers.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
When streaming with Unicam, the VPU must have a clock frequency of at
least 250Mhz. Otherwise, the input fifos could overrun, causing
image corruption.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
[g|s]_selection pass in a buffer type that needs to be validated
before passing on to the sensor subdev.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
v4l2-compliance throws a failure if the device doesn't advertise
V4L2_CAP_READWRITE but allows read or write operations.
We do support read, so reinstate the flag.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The colorspace fields were left untouched in imx290_set_fmt
which lead to a v4l2-compliance failure.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Userspace needs to know the cropping arrangements for each mode,
so expose this through g_selection.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
__v4l2_ctrl_modify_range only updates the current value should
it be invalid within the new range. That can leave modes producing
odd frame rates.
Explicitly update the HBLANK and VBLANK values so that on mode
change we revert to the default frame rate for the mode.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Vision Components have made an OV9281 module which blocks reading
back the majority of registers to comply with NDAs, and in doing
so doesn't allow auto-increment register reading as used when
reading the chip ID.
Use two reads and manually combine the results.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The Rockchip driver was based on a 4.4 kernel, and had several custom
Rockchip parts.
Update to 5.4 kernel APIs, with the relevant controls required by
libcamera, and remove custom Rockchip parts.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Adds the ov9281 parts of the Rockchip patch adding enum_frame_interval to
a large number of drivers.
Change-Id: I03344cd6cf278dd7c18fce8e97479089ef185a5c
Signed-off-by: Zefa Chen <zefa.chen@rock-chips.com>
Takes the ov9281 part only from the Rockchip's patch.
Change-Id: I30e833baf2c1bb07d6d87ddb3b00759ab45a90e4
Signed-off-by: Zefa Chen <zefa.chen@rock-chips.com>
Use bit 27 of the dreq value (the second cell of the DT DMA descriptor)
to request that the WAIT_RESP bit is not set.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Now that the 14bit non-packed Bayer formats are defined, add them
into the supported formats lookup table.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Now that V4L2_PIX_FMT_Y14 and V4L2_PIX_FMT_Y14P are defined,
allow passing 14bit mono data through the peripheral.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Now that V4L2_PIX_FMT_Y12P is defined, allow passing raw 12bit
mono packed data through the peripheral.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
imx290_set_hmax was using two independent writes to set up hmax,
when all other multi-register writes were using imx290_write_buffered_reg
which claims the group hold first.
Switch imx290_set_hmax to using imx290_write_buffered_reg too.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The IMX290 module is available as either mono or colour (Bayer).
Update the driver so that it can advertise the correct mono
formats instead of the colour ones.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The IMX290 module is available as either monochrome or colour and
the variant is not detectable at runtime.
Add a new compatible string for the monochrome version.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The sensor supports horizontal and vertical flips, so support them
through V4L2_CID_HFLIP and V4L2_CID_VFLIP.
This sensor does NOT change the Bayer order when changing the
direction of readout, therefore no special handling is required for
that.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Adds support for V4L2_CID_EXPOSURE so that userspace can control
the sensor exposure time.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
In order to calculate framerate and durations userspace needs
the vertical blanking information. This can be configurable,
and indeed the datasheet lists different values for VBLANK for
the 1080p and 720p modes.
Add the new control, and adopt the datasheet values for each mode.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Userspace needs to know HBLANK if it is to work out exposure times
and frame rates, therefore convert it to map onto V4L2_CID_HBLANK
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The datasheet lists the gain as being 0.0 to 72.0dB in 0.3dB steps, which
makes 238 steps total.
Correct the 0-72 range defined in the driver.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The existing driver only supported a clock of 37.125MHz, but the
sensor also supports 74.25MHz.
Add the relevant register modifications to support this alternate
clock frequency.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Commit d46cfdc86c upstream.
With the current driver 'media-ctl -p' issued right after the imx290 driver
is loaded prints:
pad0: Source
[fmt:unknown/0x0]
The format value of zero is due to the current_format field of the imx290
struct not being initialized yet.
As imx290_entity_init_cfg() calls imx290_set_fmt(), the current_mode field
is also initialized, so the line which set current_mode to a default value
in driver's probe() function is no longer needed.
Signed-off-by: Andrey Konovalov <andrey.konovalov@linaro.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Older gcc versions object to = { 0 } initialisation if the first
elemtn in the structure is a substructure.
Use = { } to avoid this compiler warning.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Use the get_mbus_config pad subdev call to allow a source to use
fewer than the number of CSI2 lanes defined in device tree.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Add a driver for the Unicam camera receiver block on BCM283x processors.
Compared to the bcm2835-camera driver present in staging, this driver
handles the Unicam block only (CSI-2 receiver), and doesn't depend on
the VC4 firmware running on the VPU.
The commit is made up of a series of changes cherry-picked from the
rpi-5.4.y branch of https://github.com/raspberrypi/linux/ with
additional enhancements, forward-ported to the mainline kernel.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reported-by: kbuild test robot <lkp@intel.com>
Commit 8353fe6f1e ("Revert "staging: bcm2835-audio: Drop DT
dependency"") reverts the upstream change and makes bcm2835-audio use
device tree again, however, it also removes the MODULE_ALIAS for the
platform device. This MODULE_ALIAS is needed, because VCHIQ registers
bcm2835-audio as a child platform device since commit 25c7597af2
("staging: vchiq_arm: Register a platform device for audio"), and this
mechanism is adopted also in the downstream kernel.
This commit puts back that MODULE_ALIAS to make bcm2835-audio
autoprobing work again. The rest of VCHIQ children have their
MODULE_ALIASes in place.
Fixes: 8353fe6f1e ("Revert "staging: bcm2835-audio: Drop DT dependency"")
Signed-off-by: Maxim Mikityanskiy <maxtram95@gmail.com>
When closing the video device, the irs1125 is put in power down state.
To keep V4L2 ctrls and the HW in sync, v4l2_ctrl_handler_setup is
called after power up.
The compound ctrl IRS1125_CID_MOD_PLL however has a default value
of all zeros, which puts the imager into a non responding state.
Thus, this ctrl is not written by the driver into HW after power up.
The userspace has to take care to write senseful data.
Signed-off-by: Markus Proeller <markus.proeller@pieye.org>
Instead of changing the exposure and framerate settings for all sequences,
they can be changed for every sequence individually now. Therefore the
IRS1125_CID_SAFE_RECONFIG ctrl has been removed and replaced by
IRS1125_CID_SAFE_RECONFIG_S<seq_num>_EXPO and *_FRAME ctrls.
The consistency check in the sequence ctrl IRS1125_CID_SEQ_CONFIG
is removed.
Signed-off-by: Markus Proeller <markus.proeller@pieye.org>
Reading data over i2c is done by using i2c_transfer to ensure that this
operation can't be interrupted.
Signed-off-by: Markus Proeller <markus.proeller@pieye.org>
The BRCM PCIe block has controls to enable control of the CLKREQ#
signal by the L1SS, and to gate the refclk with the CLKREQ# input.
These controls are mutually exclusive - the upstream code sets the
latter, but some use cases require the former.
Add a Device Tree property - brcm,enable-l1ss - to switch to the
L1SS configuration.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Upstream Linux deems using output GPIOs to generate IRQs as a bogus
use case, even though the BCM2835 GPIO controller is capable of doing
so. A number of users would like to make use of this facility, so
disable the checks.
See: https://github.com/raspberrypi/linux/issues/2527
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Things don't work too well when both the vc4 driver and the firmware
driver are trying to control the same audio output:
[ 763.569406] bcm2835_audio bcm2835_audio: vchi message timeout, msg=5
Hence, when the vc4 HDMI driver is used, let it control audio. This is done
by introducing a new device tree property to the audio node, and
extending the vc4-kms-v3d overlays to set it appropriately.
Signed-off-by: Hristo Venev <hristo@venev.name>
Since the unicam driver was modified to write to a dummy buffer when no
user-supplied buffer is available, it can now write to and return a
buffer even when there's only a single one. Enable this by changing the
min_buffers_needed in the vb2_queue; it will be useful for enabling
still captures without allocating more memory than absolutely necessary.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Manage the split between addresses for the VPU and addresses for the
40-bit DMA controller with a dedicated DMA device pointer that on non-
BCM2711 platforms is the same as the main VCHIQ device. This allows
the VCHIQ node to stay in the usual place in the DT, and removes the
ugly VC_SAFE macros.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
staging: vchiq_arm: Use g_dma_dev for dma_unmap_sg
Commit "staging: vchiq_arm: Clean up 40-bit DMA support" failed to
change one of the calls to dma_unmap_sg to pass in g_dma_dev (rather
than g_dev). Correct that oversight.
See: https://github.com/raspberrypi/linux/issues/3647
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Without this patch GPIOs don't seem to work properly, primarily
noticeable as broken LEDs.
Squash with "pinctrl-bcm2835: Set base to 0 give expected gpio numbering"
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Enabling zswap support in the kernel configuration costs about 1.5MB
of RAM, even when zswap is not enabled at runtime. This cost can be
reduced significantly by deferring initialisation (including pool
creation) until the "enabled" parameter is set to true. There is a
small cost to this in that some initialisation code has to remain in
memory after the init phase, just in case they are needed later,
but the total size increase is negligible.
See: https://github.com/raspberrypi/linux/pull/3432
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The change to retrieve the pixel format always on g_fmt didn't
check whether the native or unpacked version of the format
had been requested, and always returned the packed one.
Correct this so that the packing setting is retained whereever
possible.
Fixes "9d59e89 media: bcm2835-unicam: Re-fetch mbus code from subdev
on a g_fmt call"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The GCC plugin feature leads to different kernel configurations on what
ought to be equivalent build systems because they depend on the build
hosts native compilers rather than the cross compilers needed for the
target. This causes problems with module symbol version mismatches.
Disable GCC plugins for all build hosts.
Advanced build script hackery borrowed from a patch by milhouse.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
V4L2 wishes to have the codec header bytes in the same buffer as the
first encoded frame, so it does become 1-in 1-out for encoding.
The firmware now has an option to do this, so enable it.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The SC16IS7XX hardware flow control is mishandled by the driver in
a number of ways:
1. The set_baud method accidentally clears it when setting EFR bit.
2. Even though hardware flow control is enabled, it isn't indicated
back to the serial framework.
3. Applying the flow control clears the EFR bit.
4. The CTS support is not indicated in the return from
sc16is7xx_get_mctrl.
Address all of those issues using a mixture of patches found on the
linked pages.
See: https://github.com/raspberrypi/linux/issues/2542
See: https://www.spinics.net/lists/linux-serial/msg21794.html
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
From when bringing up the driver, there was a check in the isr
to ignore interrupts (claiming them handled) should the driver
not be streaming.
The VPU now will not register a camera driver if it finds a
CSI2 node enabled in device tree, therefore this flawed check is
redundant.
https://github.com/raspberrypi/linux/issues/3602
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
If the firmware hasn't detected a display, the driver would assume
one display was available, but because it had failed to retrieve the
display size it would try to allocate a zero-sized buffer.
Avoid the allocation failure by bailing out early if no display is
found.
See: https://github.com/raspberrypi/linux/issues/3598
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The reference counting of node->open was only incremented after
a check that the node was v4l2_fh_is_singular_file, which resulted
in the counting going wrong and s_power not being called at an
appropriate time.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
unicam_release calls _vb2_fop_release, which will call stop_streaming
if that particular node was streaming. Calling it unconditionally (as
the code was) means that if a second handle was opened eg to alter
a setting, on closing that connection it also stopped Unicam.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Sensors are now reflecting cropping and scaling parameters through
the selection API, therefore Unicam needs to forward the requests
through to the subdev.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Exposure is clipped by the VTS of the mode, so needs to be updated as
and when this is changed.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
To make adding new members to the mode structures easier, use
the member names in the initialisers.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The controls for analogue gain and exposure were defined with
V4L2_CTRL_FLAG_EXECUTE_ON_WRITE. This is not required as we only need
to send changes to the sensor.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The ov5647 subdev can generate control events, therefore set
the V4L2_SUBDEV_FL_HAS_EVENTS flag.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Clients need to know the pixel rate in order to compute exposure
and frame rate values.
Advertise it.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
"87f3ab9 media: ov5647: Add basic support for multiple sensor modes."
added a return path ov5647_set_fmt that didn't release the device
mutex that it had claimed.
Release the mutex.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
In order to apply lens shading correctly the client needs to know how
each mode crops or scales the image compared to the full sensor array.
Implement this (based on the imx219 equivalent).
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Specifically:
* AWB is now off by default.
* AEC/AGC is also off by default.
* The default mode is changed to the 10-bit 2x2 binned mode.
AWB and AEC/AGC can be re-enabled using the usual V4L2 controls. The
original 8-bit mode will be respected if an application requests the
8-bit format.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The 8-bit VGA mode remains, we add the following 10-bit modes:
Mode 0: 2592x1944 full resolution.
Mode 1: 1920x1080 full resolution, but centre-cropped.
(This mode achieves 30fps, mode 0 does not.)
Mode 2: 1296x972 full field-of-view 2x2 binned mode.
Mode 3: VGA full field of view mode.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Added basic v4l2_ctrl_handler infrastructure (there was none
previously).
Added controls to let AWB/AEC/AGC run in the sensor's auto mode or
manually. Also controls to set exposure (in lines) and analogue gain
(as a register code) from user code.
Also delete registers (just the one) from the VGA mode register set
that are now controlled by the new V4L2 controls.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Specifically:
Added a structure ov5647_mode and a list of supported_modes (though no
actual new modes as yet). The state object points to the "current mode".
ov5647_enum_mbus_code, ov5647_enum_frame_size, ov5647_set_fmt and
ov5647_get_fmt all needed upgrading to cope with multiple modes.
__sensor_init (which writes all the registers) is now called by
ov5647_stream_on (once the mode is known) rather than by
ov5647_sensor_power.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Previously they were returning positive non-zero codes for success,
which were getting passed up the call stack. Since release 4.19,
do_dentry_open (fs/open.c) has been catching these and flagging an
error. (So this driver has been broken since that date.)
Fixes: 3c2472a [media] media: i2c: Add support for OV5647 sensor
Signed-off-by: David Plowman <david.plowman@raspberrypi.org>
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The flag V4L2_SUBDEV_FL_HAS_EVENTS is required if the subdev can
generate events. It can create events from the ctrl handler, therefore
this is required.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Like many overlays, the i2c-gpio overlay goes to efforts to avoid
generating warnings about #address-cells and #size-cells not
being defined, which it does by defining them. Unfortunately this
is fatal if they don't match what the system requires, and the
recent switch to #size-cells = 2 on 2711 made i2c-gpio very
dangerous.
In the absence of the knowledge of a clean way to fix this, just delete
the declarations and suffer the warnings.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
BCM2711 has 4 DMA channels with a 40-bit address range, allowing them
to access the full 4GB of memory on a Pi 4.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcmn2835_isp is a platform driver dependent on vchiq,
therefore add the load/unload functions for it to vchiq.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Add V4L2_META_FMT_BCM2835_ISP_STATS V4L2 format type.
This new format will be used by the BCM2835 ISP device to return
out ISP statistics for 3A.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The sensor subdevice may change the Bayer order if a H/V flip is
requested after a s_fmt call. Unicam g_fmt must call the subdev get_fmt
in case this has happened and return out the correct format 4cc.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
This commit updates the imx219 driver to adverise support for embedded
data streams. This can then be used by the bcm2835-unicam driver, which
has recently been updated to expose the embedded data stream to
userland.
The imx219 sensor subdevice overloads the media pad to differentiate
between image stream (pad 0) and embedded data stream (pad 1) when
performing the v4l2_subdev_pad_ops functions.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
The unicam driver supports both the SOURCE_CHANGE and CTRL events. Both
events are only generated on the image video node, so the event-related
ioctls are useless on the medatada node. Disable them.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
If no buffer has been queued by a userland application, we use an
internal dummy buffer for the hardware to spin in. This will allow
the driver to release the existing userland buffer back to the
application for processing.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
This patch adds a new node in the bcm2835-unicam driver to support
CSI-2 embedded data streams. The subdevice is queried to see if
embedded data is available from the sensor.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Move device node specific state out of the device state structure and
into a new node structure. This separation will be needed for future
changes where we will add an embedded data node to the driver to work
alongside the existing image data node.
Currently only use a single image node, so this commit does not add
any functional changes.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
This patch adds MEDIA_BUS_FMT_SENSOR_DATA used by the bcm2835-unicam
driver to support CSI-2 embedded data streams from camera sensors.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Add V4L2_META_FMT_SENSOR_DATA format 4CC.
This new format will be used by the BCM2835 Unicam device to return
out camera sensor embedded data.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Add driver for the Unicam camera receiver block on
BCM283x processors.
This commit is made up of a series of changes cherry-picked from the
rpi-4.19.y branch.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Limit mappings to the permitted range, but don't map more than asked
for otherwise we walk off the end of the allocated VMA.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Commit f3186dd876 ("spi: Optionally use GPIO descriptors for CS GPIOs")
amended of_spi_parse_dt() to always set SPI_CS_HIGH for SPI slaves whose
Chip Select is defined by a "cs-gpios" devicetree property.
This change breaks drivers whose probe functions set the mode field of
the spi_device because in doing so they clear the SPI_CS_HIGH flag.
Fix by setting SPI_CS_HIGH in spi_setup (under the same conditions as
in of_spi_parse_dt()).
See also: 83b2a8fe43 ("spi: spidev: Fix CS polarity if GPIO descriptors are used")
Fixes: f3186dd876 ("spi: Optionally use GPIO descriptors for CS GPIOs")
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
SQUASH: spi: Demote SPI_CS_HIGH warning to KERN_DEBUG
This warning is unavoidable from a client's perspective and
doesn't indicate anything wrong (just surprising).
SQUASH with "spi: use_gpio_descriptor fixup moved to spi_setup"
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
framebuffer_check was computing a minimum pitch value and ensuring
that the provided value was greater than this.
That check is only valid if the format is linear.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The HDMI controllers found in the BCM2711 SoC need some adjustments to the
bindings, especially since the registers have been shuffled around in more
register ranges.
Cc: Rob Herring <robh+dt@kernel.org>
Cc: devicetree@vger.kernel.org
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
When the MMC isn't plugged in, the driver will spam the console which is
pretty annoying when using NFS.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
This driver is for the HEVC/H265 decoder block on the Raspberry
Pi 4, and conforms to the V4L2 stateless decoder API.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Allow the capture buffer to be detached from a v4l2 request job such
that another job can start before the capture buffer is returned. This
allows h/w codecs that can process multiple requests at the same time
to operate more efficiently.
Signed-off-by: John Cox <jc@kynesim.co.uk>
Some of the Broadcom codec blocks use a column based YUV4:2:0 image
format, so add the documentation and defines for both 8 and 10 bit
versions.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
WPP can allow greater parallelism within the decode, but needs
offset information to be passed in.
Adds num_entry_point_offsets and entry_point_offset_minus1 to
v4l2_ctrl_hevc_slice_params.
This is based on Jernej Skrabec's patches for cedrus which
implement the same feature.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
From https://patchwork.linuxtv.org/patch/60725/
Changes requested, but mainly docs.
If HEVC frame consists of multiple slices, segment address has to be
known in order to properly decode it.
Add segment address field to slice parameters.
Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net>
Adds a format that is 3 10bit YUV 4:2:0 samples packed into
a 32bit work (with 2 spare bits).
Supported on Broadcom BCM2711 chips.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
This is adjustment to commit
d368ceaacd "kbuild: Allow .dtbo overlays to be built piecemeal"
prepare3 target has gone from mainline tree in branch 5.4.y
Signed-off-by: Nataliya Korovkina <malus.brandywine@gmail.com>
The DT bindings description of the Brcmstb PCIe device is described. This
node can be used by almost all Broadcom settop box chips, using
ARM, ARM64, or MIPS CPU architectures.
Signed-off-by: Jim Quinlan <jim2101024@gmail.com>
When symbols from overlays are added to the live tree their paths must
be rebased. The translated symbol is normally the result of joining
the fragment-relative path (with a leading "/") to the target path
(either copied directly from the "target-path" property or resolved
from the phandle). This translation fails when the target is the root
node (a common case for Raspberry Pi overlays) because the resulting
path starts with a double slash. For example, if target-path is "/" and
the fragment adds a node called "newnode", the label associated with
that node will be assigned the path "//newnode", which can't be found
in the tree.
Fix the failure case by explicitly replacing a target path of "/" with
an empty string.
Fixes: d1651b03c2 ("of: overlay: add overlay symbols to live device tree")
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The definition of compat_ptr is now common for most platforms, but
requires the inclusion of <linux/compat.h>.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The definition of compat_ptr is now common for most platforms, but
requires the inclusion of <linux/compat.h>.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The downstream .dts files only request two GPIO IRQs. Truncate the
array of parent IRQs when irq_of_parse_and_map returns 0.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
pinctrl-bcm2835 is a combined pinctrl/gpio driver. Currently the gpio
side is registered first, but this breaks gpio hogs (which are
configured during gpiochip_add_data). Part of the hog initialisation
is a call to pinctrl_gpio_request, and since the pinctrl driver hasn't
yet been registered this results in an -EPROBE_DEFER from which it can
never recover.
Change the initialisation sequence to register the pinctrl driver
first.
See: https://www.raspberrypi.org/forums/viewtopic.php?f=107&t=260600
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
A failure in gpiochip_irqchip_add leads to a leak of a gpiochip. Fix
the leak with the use of devm_gpiochip_add_data.
Fixes: 85ae9e512f ("pinctrl: bcm2835: switch to GPIOLIB_IRQCHIP")
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
vchiq kernel clients are now instantiated as platform drivers rather
than using DT, but the children of the vchiq interface may still
benefit from access to DT properties. Give them the option of a
a sub-node of the vchiq parent for configuration and to allow
them to be disabled.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The IMA (Integrity Measurement Architecture) looks for a TPM (Trusted
Platform Module) having been registered when it initialises; otherwise
it assumes there is no TPM. It has been observed on BCM2835 that IMA
is initialised before TPM, and that initialising the BCM2835 clock
driver before the firmware driver has the effect of reversing this
order.
Change the firmware driver to initialise at core_initcall, delaying the
BCM2835 clock driver to postcore_initcall.
See: https://github.com/raspberrypi/linux/issues/3291https://github.com/raspberrypi/linux/pull/3297
Signed-off-by: Luke Hinds <lhinds@redhat.com>
Co-authored-by: Phil Elwell <phil@raspberrypi.org>
vchiq on Pi4 is no longer under the soc node, therefore it
doesn't get the dma-ranges for the VPU.
Switch to using the configuration of the old dma controller as
that will set the dma-ranges correctly.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The VCHIQ driver now loads the audio, camera, codec, and vc-sm
drivers as platform drivers. However they were not being given
the correct DMA configuration.
Call of_dma_configure with the parent (VCHIQ) parameters to be
inherited by the child.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Both coherent_dma_mask and dma_mask act as constraints on allocations
and bounce buffer usage, so be sure to set dma_mask to the appropriate
value otherwise the effective mask could be incorrect.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
For performance/power it is beneficial to adjust gpu clocks with arm clock.
This is how the downstream cpufreq driver works
Signed-off-by: popcornmix <popcornmix@gmail.com>
Setting the v3d clock to low value allows firmware to handle dvfs in case
where v3d hardware is not being actively used (e.g. console use).
Signed-off-by: popcornmix <popcornmix@gmail.com>
Add device tree entries and code to allow the specification of
the lighting modes for the LED's on the ethernet connector.
Signed-off-by: James Hughes <james.hughes@raspberrypi.org>
Following the same pattern as bcm2835-camera and bcm2835-audio,
register the V4L2 codec driver as a platform driver
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Following the same pattern as bcm2835-camera and bcm2835-audio,
register the vcsm-cma driver as a platform driver
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
The irq_fence and done_fence are given a reference that is never
released. The necessary dma_fence_put()s seem to have been
deleted in error in an earlier commit.
Fixes: 0b73676836b2 ("drm/v3d: Clock V3D down when not in use.")
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The v3d driver currently encounters a lot of MMU PTE exceptions, so
only log the first to avoid swamping the kernel log.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
MMU exception conditions are reported in the V3D_MMU_CTRL register as
write-1-to-clear (W1C) bits. The MMU interrupt handling code clears any
exceptions, but does so by masking out any other bits and writing the
result back. There are some important control bits in that register,
including MMU_ENABLE, so a safer approach is to simply write back the
value just read unaltered.
This patch doesn't remove the cause of the apparent PTE errors, but it
does reduce the impact to just an error in the kernel log.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The Infineon IRS1125 is a time of flight depth sensor that
has a CSI-2 interface.
Add a V4L2 subdevice driver for this device.
Signed-off-by: Markus Proeller <markus.proeller@pieye.org>
HDMI Alsa devices renamed to match names used by DRM, to
HDMI 1 and HDMI 2
Check for which HDMI devices are connected and only create
devices for those that are present.
The rename of the devices might cause some backwards compatibility
issues, but since this particular part of the driver needs to be
specifically enabled, I suspect the number of people who will see
the problem will be very small.
Signed-off-by: James Hughes <james.hughes@raspberrypi.org>
After the decision to use bcm2711 compatible for upstream, we should
switch all accepted compatibles to bcm2711. So we can boot with
one DTB the down- and the upstream kernel.
Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
Before 4.20, it was possible to build an arbitrary overlay by copying
it to arm/boot/dts/overlays/mytest-overlay.dts and running:
make ARCH=arm overlays/mytest.dtbo
In 4.20 the .dtb build rules were centralised, requiring the dowstream
.dtbo build rules to be changed. They were, enough to support "make ...
dtbs", but not sufficiently to allow this ad-hoc, one-off building of
individual files.
Add the missing makefile rule to support this way of building.
See: https://github.com/raspberrypi/linux/issues/3250
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
This is mainly used for the NoIR camera which has no IR
filter and can completely confuse normal AWB presets.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Adds a simple greyworld white balance preset, mainly for use
with cameras without an IR filter (eg Raspberry Pi NoIR)
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The IMX219 is an 8MPix CSI2 sensor, supporting 2 or 4 data lanes.
Document the binding for this device.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Users have reported log spam created by "Event Ring Full" xHC event
TRBs. These are caused by interrupt latency in conjunction with a very
busy set of devices on the bus. The errors are benign, but throughput
will suffer as the xHC will pause processing of transfers until the
event ring is drained by the kernel. Expand the number of event TRB slots
available by increasing the number of event ring segments in the ERST.
Controllers have a hardware-defined limit as to the number of ERST
entries they can process, so make the actual number in use
min(ERST_MAX_SEGS, hw_max).
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
Some combinations of Pi 4Bs and Ethernet switches don't reliably get a
DCHP-assigned IP address, leaving the unit with a self=assigned 169.254
address. In the failure case, the Pi is left able to receive packets
but not send them, suggesting that the MAC<->PHY link is getting into
a bad state.
It has been found empirically that skipping a reset step by the genet
driver prevents the failures. No downsides have been discovered yet,
and unlike the forced renegotiation it doesn't increase the time to
get an IP address, so the workaround is enabled by default; add
genet.skip_umac_reset=n
to the command line to disable it.
See: https://github.com/raspberrypi/linux/issues/3108
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
These wireless mouse/keyboard combo remote control devices specify
multiple "wheel" events in their report descriptors. The wheel events
are incorrectly defined and apparently map to accelerometer data, leading
to spurious mouse scroll events being generated at an extreme rate when
the device is moved.
As a workaround, use HID_QUIRK_INCREMENT_USAGE_ON_DUPLICATE to mask
feeding the extra wheel events to the input subsystem.
See: https://github.com/raspberrypi/firmware/issues/1189
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
Based on the gpiomem driver, allow mapping of the decoder register
spaces such that userspace can access control/status registers.
This driver is intended for use with a custom ffmpeg backend accelerator
prior to a v4l2 driver being written.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
driver: char: rpivid: Destroy the legacy device on remove
The legacy name support created a new device that was never destroyed.
If the driver was unloaded and reloaded, it failed due to the
device already existing.
Fixes: "75f1d14 driver: char: rpivid - also support legacy name"
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
driver: char: rpivid: Clean up error handling use of ERR_PTR/IS_ERR
The driver used an unnecessary intermediate void* variable so it
only called ERR_PTR once to convert to the error value.
Switch to converting as the error arises to remove these intermediate
variables.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
driver: char: rpivid: Add error handling to the legacy device load
The return value from device_create for the legacy device was never
checked or handled. Add the required error handling.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
driver: char: rpivid: Fix coding style whitespace issues.
Makes checkpatch happier.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
driver: char: rpimem: Add SPDX licence header.
Stops checkpatch complaining.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
driver: char: rpivid: Fix access to freed memory
The error path during probe frees the private memory block, and
then promptly dereferences it to log an error message.
Use the base device instead of the pointer to it in the private
structure.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
My various attempts at re-enabling runtime PM have failed, so just
crank the clock down when V3D is idle to reduce power consumption.
Signed-off-by: Eric Anholt <eric@anholt.net>
Something is still unstable -- on starting a new glxgears from an idle
X11, I get an MMU violation in high addresses. The CTS also failed
quite quickly. With this, CTS progresses for an hour before OOMing
(allocating some big buffers when my board only has 600MB available to
Linux)
Signed-off-by: Eric Anholt <eric@anholt.net>
The BCM2835 I2C blocks have a register to set the clock-stretch
timeout - how long the device is allowed to hold SCL low - in bus
cycles. The current driver doesn't write to the register, therefore
the default value of 64 cycles is being used for all devices.
Set the timeout to the value recommended for SMBus - 35ms.
See: https://github.com/raspberrypi/linux/issues/3064
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Seen on a VLI VL805 PCIe to USB controller. For non-stream endpoints
at least, if the xHC halts on a particular TRB due to an error then
the DCS field in the Out Endpoint Context maintained by the hardware
is not updated with the current cycle state.
Using the quirk XHCI_EP_CTX_BROKEN_DCS and instead fetch the DCS bit
from the TRB that the xHC stopped on.
See: https://github.com/raspberrypi/linux/issues/3060
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
In translating the runtime PM code from vc4, I missed the ".pm"
assignment to actually connect them up. Fixes missing MMU setup if
runtime PM resets V3D.
Signed-off-by: Eric Anholt <eric@anholt.net>
(cherry picked from commit ca197699af29baa8236c74c53d4904ca8957ee06)
If it's off, we know it will be reset on poweron, so the MMU won't
have any TLB cached from before this point. Avoids failed waits for
MMU flush to reply.
Signed-off-by: Eric Anholt <eric@anholt.net>
(cherry picked from commit 3ee4e2e0a9e9587eacbb69b067bbc72ab2cdc47b)
Must be called in a non-atomic context, after the endpoint
has been registered with the hardware via xhci_add_endpoint
and before the first URB is submitted for the endpoint.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
xHCI caches device and endpoint data after the interface is configured,
so an explicit command needs to be issued for any device driver wanting
to alter the polling interval of an endpoint.
Add usb_fixup_endpoint() to allow drivers to do this. The fixup must be
called after calculating endpoint bandwidth requirements but before any
URBs are submitted.
If polling intervals are shortened, any bandwidth reservations are no
longer valid but in practice polling intervals are only ever relaxed.
Limit the scope to interrupt transfers for now.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
This falls under the same "we can reprogram glitch-free as long as we
pause generation" rule as updating the div/frac fields. This can be
used for runtime reclocking of V3D to manage power leakage.
Signed-off-by: Eric Anholt <eric@anholt.net>
Without the actual power management part any more, there's a lot less
to set up for V3D. We just need to clear the RSTN field for the power
domain, and expose the reset controller for toggling it again.
This is definitely incomplete -- the old ISP and H264 is in the old
bridge, but since we have no consumers of it I've just done the
minimum to get V3D working.
Signed-off-by: Eric Anholt <eric@anholt.net>
If the VL805 EEPROM has not been programmed then boot will hang for five
seconds. The timeout seems to be arbitrary and is an unecessary
delay on the first boot. Remove the timeout.
This is common code and probably can't be upstreamed unless the timeout
can be configurable somehow or perhaps the XHCI driver can be skipped
on the first boot.
There are several warts surrounding bcmgenet_mii_probe() as this
function is called from ndo_open, but it's calling registration-type
functions. The probe should be called at probe time and refactored
such that the PHY device data can be extracted to limit the scope
of this flag to Broadcom PHYs.
For now, pass this flag in as it puts our attached PHY into a low-power
state when disconnected.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
Set defaults for TX and RX packet coalescing to be equivalent to:
# ethtool -C eth0 tx-frames 10
# ethtool -C eth0 rx-usecs 50
This may be something we want to set via DT parameters in the
future.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Conditional on a new compatible string, change the pagelist encoding
such that the top 24 bits are the pfn, leaving 8 bits for run length
(-1).
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
staging/vchiq_arm: Fix bcm2711 compatible string
Fixes: "vchiq: Add 36-bit address support"
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The Linux support for controlling card power via regulators appears to
be contentious. I would argue that the default behaviour is contrary to
the SDHCI spec - turning off the power writes a reserved value to the
SD Bus Voltage Select field of the Power Control Register, which
seems to kill the Arasan/iProc controller - but fortunately there is a
hook in sdhci_ops to override the behaviour. Borrow the implementation
from sdhci_arasan_set_power.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The HWRNG on the BCM2838 is compatible to iproc-rng200, so add the
support to this driver instead of bcm2835-rng.
Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
hwrng: iproc-rng200: Correct SoC name
The Pi 4 SoC is called BCM2711, not BCM2838.
Fixes: "hwrng: iproc-rng200: Add BCM2838 support"
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
The legacy peripherals can only address the first gigabyte of RAM, so
ensure that DMA allocations are restricted to that region.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The ioremapping creates mappings within the vmalloc area. The
equivalent early function, create_mapping, now checks that the
requested explicit virtual address is between VMALLOC_START and
VMALLOC_END. As there is no reason to have any correlation between
the physical and virtual addresses, put the required mappings at
VMALLOC_START and above.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
On error, vchiq_mmal_component_init could leave the
event context allocated for ports.
Clean them up in the error path.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
vchiq_mmal_component_init calls init_event_context for the
control port, but vchiq_mmal_component_finalise didn't free
it, causing a memory leak..
Add the free call.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
mmal_parameters.h hasn't been updated to reflect additions made
over the last few years. Update it to reflect the currently
supported parameters.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The list of formats was copied before Bayer support was added.
The ISP supports Bayer and is being supported by the bcm2835_codec
driver, so add in the encodings for them.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
All calls to the gpio library are in contexts that can sleep,
therefore there is no issue with having those GPIOs controlled
by controllers which require sleeping (eg I2C GPIO expanders).
Switch to using gpiod_set_value_cansleep instead of gpiod_set_value
to avoid triggering the warning in gpiolib should the GPIO
controller need to sleep.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The debug text for how many clocks have been registered
uses "%d" with a size_t. Correct it to "%zd".
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The MMAL client_component field is used with the event
mechanism to allow the client to identify the component for
which the event is generated.
The field is only 32bits in size, therefore we can't use a
pointer to the component in a 64 bit kernel.
Component handles are already held in an array per VCHI
instance, so use the array index as the client_component handle
to avoid having to create a new IDR for this purpose.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
videobuf2 only allowed exporting a dmabuf as a file descriptor,
but there are instances where having the struct dma_buf is
useful within the kernel.
Split the current implementation into two, one step which
exports a struct dma_buf, and the second which converts that
into an fd.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Add the ability to send data to ports. This only supports
zero copy mode as the required bulk transfer setup calls
are not done.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
(Preparation for the codec driver).
The codec uses the event mechanism to report things such as
resolution changes. It is signalled by the cmd field of the buffer
being non-zero.
Add support for passing this information out to the client.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Fixes up a checkpatch error "Avoid using bool structure members
because of possible alignment issues".
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
When calling tc358743_set_fmt, the code was calling tc358743_get_fmt
to choose a valid format. However that sets the colorspace
based on what was read back from the chip. When you set the format,
then the driver would choose and program the colorspace based
on the format code.
The result was that if you called try or set format for UYVY
when the current format was RGB3 then you would get told sRGB,
and try RGB3 when current was UYVY and you would get told
SMPTE170M.
The value programmed into the chip is determined by this driver,
therefore there is no need to read back the value. Return the
colorspace based on the format set/tried instead.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Document the DT bindings for the CSI2/CCP2 receiver peripheral
(known as Unicam) on BCM283x SoCs.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Acked-by: Rob Herring <robh@kernel.org>
New helper defines that allow printing of a FOURCC using
printf(V4L2_FOURCC_CONV, V4L2_FOURCC_CONV_ARGS(fourcc));
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The ADV7282M can support YPbPr on AIN1-3, but this was
not selectable from the driver. Add it to the list of
supported input modes.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The hardware default is differential CVBS on AIN1 & 2, which
isn't very useful.
Select the first input that is defined as valid for the
chip variant (typically CVBS_AIN1).
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The probe for the TC358743 reads the CHIPID register from
the device and compares it to the expected value of 0.
If the I2C request fails then that also returns 0, so
the driver loads thinking that the device is there.
Generally I2C communications are reliable so there is
limited need to check the return value on every transfer,
therefore only amend the one read during probe to check
for I2C errors.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Adds register setups for running the CSI lanes at 972Mbit/s,
which allows 1080P50 UYVY down 2 lanes.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
g_mbus_config was supposed to indicate all supported lane numbers, not
only the number of those currently in active use. Since the TC358743
can dynamically reduce the number of active lanes if the required
bandwidth allows for it, report all lane numbers up to the connected
number of lanes as supported in pdata mode.
In device tree mode, do not report lane count and clock mode at all, as
the receiver driver can determine these from the device tree.
To allow communicating the number of currently active lanes, add a new
bitfield to the v4l2_mbus_config flags. This is a temporary fix, to be
used only until a better solution is found.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
The existing fixed value of 16 worked for UYVY 720P60 over
2 lanes at 594MHz, or UYVY 1080P60 over 4 lanes. (RGB888
1080P60 needs 6 lanes at 594MHz).
It doesn't allow for lower resolutions to work as the FIFO
underflows.
374 is required for 1080P24-30 UYVY over 2 lanes @ 972Mbit/s, but
>374 means that the FIFO underflows on 1080P50 UYVY over 2 lanes
@ 972Mbit/s.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The driver was only supporting continuous clock mode
although this was not stated anywhere.
Non-continuous clock saves a small amount of power and
on some SoCs is easier to interface with.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
There's no way to query the subdevice for the supported
resolutions.
Add set_fmt and get_fmt implementations.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
The lan78xx uses a 12-byte hardware rx header, so there is no need
to allocate SKBs with NET_IP_ALIGN set. Removes alignment faults
in both dwc_otg and in ipv6 processing.
Add the ability to interpret the high bits of the dreq specifier as
flags to be included in the DMA_CS register. The motivation for this
change is the ability to set the DISDEBUG flag for SD card transfers
to avoid corruption when using the VPU debugger.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The bInterval is set to 4 (i.e. 8 microframes => 1ms) and the only bit
that the driver pays attention to is "link was reset". If there's a
flapping status bit in that endpoint data, (such as if PHY negotiation
needs a few tries to get a stable link) then polling at a slower rate
would act as a de-bounce.
See: https://github.com/raspberrypi/linux/issues/2447
The driver already reported the firmware build date during probe.
The mailbox calls have been extended to also report the variant
1 = standard start.elf
2 = start_x.elf (includes camera stack)
3 = start_db.elf (includes assert logging)
4 = start_cd.elf (cutdown version for smallest memory footprint).
Log the variant during probe.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
firmware: raspberrypi: Report the fw git hash during probe
The firmware can now report the git hash from which it was built
via the mailbox, so report it during probe.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Ethernet cables with faulty or missing pairs (specifically pairs C and
D) allow auto-negotiation to 1000Mbs, but do not support the successful
establishment of a link. Add a DT property, "microchip,downshift-after",
to configure the number of auto-negotiation failures after which it
falls back to 100Mbs. Valid values are 2, 3, 4, 5 and 0, where 0 means
never downshift.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Step wise governor increases the mitigation level when the temperature
goes above a threshold and will decrease the mitigation when the
temperature falls below the threshold. If it were a case, where the
temperature hovers around a threshold, the mitigation will be applied
and removed at every iteration. This reaction to the temperature is
inefficient for performance.
The use of hysteresis temperature could avoid this ping-pong of
mitigation by relaxing the mitigation to happen only when the
temperature goes below this lower hysteresis value.
Signed-off-by: Ram Chandrasekar <rkumbako@codeaurora.org>
Signed-off-by: Lina Iyer <ilina@codeaurora.org>
Avoid a hard userspace ABI change by adding a compatible get_throttled
sysfs entry. Its value is now feed by the GET_THROTTLED requests of the
new hwmon driver. The first access to get_throttled will generate
a warning.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Although the correct fix for low voltage warnings is to
improve the power supply, the current implementation
of the detection can fill the log if the warning
happens freqently. This replaces the logging with
slightly custom ratelimited logging.
Signed-off-by: James Hughes <james.hughes@raspberrypi.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Enable EEE mode as soon as possible after connecting to the PHY, and
before phy_start. This avoids a second link negotiation, which speeds
up booting and stops the interface failing to become ready.
See: https://github.com/raspberrypi/linux/issues/2437
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
As of 4.18, a firmware that implements the update_connect_params
method but doesn't claim to support roaming causes an error. We
disabled firmware roaming in 4.4 [1] because it appeared to
prevent disconnects, but let's try with the current firmware to see
if things have improved.
[1] dd91880117
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The published API to the dynamic overlay application mechanism now
takes a Flattened Device Tree blob as input so that it can manage the
lifetime of the unflattened tree. Conveniently, the new API call -
of_overlay_fdt_apply - is virtually a drop-in replacement for
create_overlay, which can now be deleted.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
ad83c7cb2f ("irqchip/irq-bcm2836: Add support for DT interrupt polarity")
changed the way that the BCM2836/7 local interrupts are mapped; instead
of being pre-mapped they are now mapped on-demand. A side effect of this
change is that the call to irq_of_parse_and_map from armctrl_of_init
creates a new mapping, forming a gap between the IRQs and the FIQs. This
gap breaks the FIQ<->IRQ mapping which up to now has been done by assuming:
1) that the value of FIQ_START is the same as the number of normal IRQs
that will be mapped (still true), and
2) that this value is also the offset between an IRQ and its equivalent
FIQ (which is no longer the case).
Remove both assumptions by measuring the interval between the last IRQ
and the last FIQ, passing it as the parameter to init_FIQ().
Fixes: https://github.com/raspberrypi/linux/issues/2432
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Register for reboot notifications, sending RPI_FIRMWARE_NOTIFY_REBOOT
over the mailbox interface on reception.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Add two new DT properties:
* microchip,eee-enabled - a boolean to enable EEE
* microchip,tx-lpi-timer - time in microseconds to wait before entering
low power state
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
I2C busses can be assigned specific bus numbers using aliases in
Device Tree - string properties where the name is the alias and the
value is the path to the node. The current DT parameter mechanism
does not allow property names to be derived from a parameter value
in any way, so it isn't possible to generate unique or matching
aliases for nodes from an overlay that can generate multiple
instances, e.g. i2c-gpio.
Work around this limitation (at least temporarily) by allowing
the i2c adapter number to be initialised from the "reg" property
if present.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
There is a new test in __irq_startup that the IRQ is activated, which
hasn't been the case for FIQs since they bypass some of the usual setup.
Augment enable_fiq to include a call to irq_activate to avoid the
warning.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Create a semi-static mapping for the USB registers early in the boot
process, before additional kernel threads are started, so all threads
will have the mappings from the start. This avoids the need for
data aborts to lazily update them.
See: https://github.com/raspberrypi/linux/issues/2450
Signed-off-by: Floris Bos <bos@je-eigen-domein.nl>
The VideoCore bootloader passes in Serial number and
Revision number through Device Tree. Make these available to
userspace through /proc/cpuinfo.
Mainline status:
There is a commit in linux-next that standardize passing the serial
number through Device Tree (string: /serial-number):
ARM: 8355/1: arch: Show the serial number from devicetree in cpuinfo
There was an attempt to do the same with the revision number, but it
didn't get in:
[PATCH v2 1/2] arm: devtree: Set system_rev from DT revision
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Uses the debugfs I/F to provide access to the AXI
bus performance monitors.
Requires the new mailbox peripheral access for access
to the VPU performance registers, system bus access
is done using direct register reads.
Signed-off-by: James Hughes <james.hughes@raspberrypi.org>
raspberrypi_axi_monitor: suppress warning
Suppress the following warning by casting the pointer to and uintptr_t
before to u32:
Signed-off-by: Matteo Croce <mcroce@redhat.com>
IRQ-CPU mapping is round robined on ARM64 to increase
concurrency and allow multiple interrupts to be serviced
at a time. This reduces the need for FIQ.
Signed-off-by: Michael Zoran <mzoran@crowfest.net>
brcmfmac: Disable power management
Disable wireless power saving in the brcmfmac WLAN driver. This is a
temporary measure until the connectivity loss resulting from power
saving is resolved.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
brcmfmac: Use original country code as a fallback
Commit 73345fd212:
brcmfmac: Configure country code using device specific settings
prevents region codes from working on devices that lack a region code
translation table. In the event of an absent table, preserve the old
behaviour of using the provided code as-is.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
brcmfmac: Plug memory leak in brcmf_fill_bss_param
See: https://github.com/raspberrypi/linux/issues/1471
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
brcmfmac: do not use internal roaming engine by default
Some evidence of curing disconnects with this disabled, so make it a default.
Can be overridden with module parameter roamoff=0
See: http://projectable.me/optimize-my-pi-wi-fi/
brcmfmac: Change stop_ap sequence
Patch from Broadcom/Cypress to resolve a customer error
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Revert "brcmfmac: Disable power management"
Shortly after the release of the Pi 3B, a loss of SSH connectivity
over WiFi was traced to the power management handling, so power
management was disabled. And so it has remained ever since.
Enabling power management saves 55mA (~270mW) on a Pi 4B, so is very
much worth the minimal effort of reverting this patch, which was
squashed and rebased many times since then to the commit hash is
meaningless.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
This is a port of Pantelis Antoniou's v3 port that makes use of the
new upstreamed configfs support for binary attributes.
Original commit message:
Add a runtime interface to using configfs for generic device tree overlay
usage. With it its possible to use device tree overlays without having
to use a per-platform overlay manager.
Please see Documentation/devicetree/configfs-overlays.txt for more info.
Changes since v2:
- Removed ifdef CONFIG_OF_OVERLAY (since for now it's required)
- Created a documentation entry
- Slight rewording in Kconfig
Changes since v1:
- of_resolve() -> of_resolve_phandles().
Originally-signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
DT configfs: Fix build errors on other platforms
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
DT configfs: fix build error
There is an error when compiling rpi-4.6.y branch:
CC drivers/of/configfs.o
drivers/of/configfs.c:291:21: error: initialization from incompatible pointer type [-Werror=incompatible-pointer-types]
.default_groups = of_cfs_def_groups,
^
drivers/of/configfs.c:291:21: note: (near initialization for 'of_cfs_subsys.su_group.default_groups.next')
The .default_groups is linked list since commit
1ae1602de0.
This commit uses configfs_add_default_group to fix this problem.
Signed-off-by: Slawomir Stepien <sst@poczta.fm>
configfs: New of_overlay API
Add a mailbox-driven backlight controller for the Raspberry Pi DSI
touchscreen display. Requires updated GPU firmware to recognise the
mailbox request.
Signed-off-by: Gordon Hollingworth <gordon@raspberrypi.org>
Add Raspberry Pi firmware driver to the dependencies of backlight driver
Otherwise the backlight driver fails to build if the firmware
loading driver is not in the kernel
Signed-off-by: Alex Riesen <alexander.riesen@cetitec.com>
The driver crashed while a NULL pointer returned by i2c_get_adapter()
has been used to access the i2c bus functions.
The headphone probing function hb_hp_probe() now returns -EPROBE_DEFER
in case the i2c module has not been loaded yet.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
ASoC: Add support for Rpi-DAC
ASoC: Add prompt for ICS43432 codec
Without a prompt string, a config setting can't be included in a
defconfig. Give CONFIG_SND_SOC_ICS43432 a prompt so that Pi soundcards
can use the driver.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Add IQaudIO Sound Card support for Raspberry Pi
Set a limit of 0dB on Digital Volume Control
The main volume control in the PCM512x DAC has a range up to
+24dB. This is dangerously loud and can potentially cause massive
clipping in the output stages. Therefore this sets a sensible
limit of 0dB for this control.
Allow up to 24dB digital gain to be applied when using IQAudIO DAC+
24db_digital_gain DT param can be used to specify that PCM512x
codec "Digital" volume control should not be limited to 0dB gain,
and if specified will allow the full 24dB gain.
Modify IQAudIO DAC+ ASoC driver to set card/dai config from dt
Add the ability to set the card name, dai name and dai stream name, from
dt config.
Signed-off-by: DigitalDreamtime <clive.messer@digitaldreamtime.co.uk>
IQaudIO: auto-mute for AMP+ and DigiAMP+
IQAudIO amplifier mute via GPIO22. Add dt params for "one-shot" unmute
and auto mute.
Revision 2, auto mute implementing HiassofT suggestion to mute/unmute
using set_bias_level, rather than startup/shutdown....
"By default DAPM waits 5 seconds (pmdown_time) before shutting down
playback streams so a close/stop immediately followed by open/start
doesn't trigger an amp mute+unmute."
Tested on both AMP+ (via DAC+) and DigiAMP+, with both options...
dtoverlay=iqaudio-dacplus,unmute_amp
"one-shot" unmute when kernel module loads.
dtoverlay=iqaudio-dacplus,auto_mute_amp
Unmute amp when ALSA device opened by a client. Mute, with 5 second delay
when ALSA device closed. (Re-opening the device within the 5 second close
window, will cancel mute.)
Revision 4, using gpiod.
Revision 5, clean-up formatting before adding mute code.
- Convert tab plus 4 space formatting to 2x tab
- Remove '// NOT USED' commented code
Revision 6, don't attempt to "one-shot" unmute amp, unless card is
successfully registered.
Signed-off-by: DigitalDreamtime <clive.messer@digitaldreamtime.co.uk>
ASoC: iqaudio-dac: fix S24_LE format
Remove set_bclk_ratio call so 24-bit data is transmitted in
24 bclk cycles.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: iqaudio-dac: use modern dai_link style
Signed-off-by: Matthias Reichl <hias@horus.com>
Added support for HiFiBerry DAC+
The driver is based on the HiFiBerry DAC driver. However HiFiBerry DAC+ uses
a different codec chip (PCM5122), therefore a new driver is necessary.
Add support for the HiFiBerry DAC+ Pro.
The HiFiBerry DAC+ and DAC+ Pro products both use the existing bcm sound driver with the DAC+ Pro having a special clock device driver representing the two high precision oscillators.
An addition bug fix is included for the PCM512x codec where by the physical size of the sample frame is used in the calculation of the LRCK divisor as it was found to be wrong when using 24-bit depth sample contained in a little endian 4-byte sample frame.
Limit PCM512x "Digital" gain to 0dB by default with HiFiBerry DAC+
24db_digital_gain DT param can be used to specify that PCM512x
codec "Digital" volume control should not be limited to 0dB gain,
and if specified will allow the full 24dB gain.
Add dt param to force HiFiBerry DAC+ Pro into slave mode
"dtoverlay=hifiberry-dacplus,slave"
Add 'slave' param to use HiFiBerry DAC+ Pro in slave mode,
with Pi as master for bit and frame clock.
Signed-off-by: DigitalDreamtime <clive.messer@digitaldreamtime.co.uk>
Fixed a bug when using 352.8kHz sample rate
Signed-off-by: Daniel Matuschek <daniel@hifiberry.com>
ASoC: pcm512x: revert downstream changes
This partially reverts commit 185ea05465
which was added by https://github.com/raspberrypi/linux/pull/1152
The downstream pcm512x changes caused a regression, it broke normal
use of the 24bit format with the codec, eg when using simple-audio-card.
The actual bug with 24bit playback is the incorrect usage
of physical_width in various drivers in the downstream tree
which causes 24bit data to be transmitted with 32 clock
cycles. So it's not the pcm512x that needs fixing, it's the
soundcard drivers.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: hifiberry_dacplus: fix S24_LE format
Remove set_bclk_ratio call so 24-bit data is transmitted in
24 bclk cycles.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: hifiberry_dacplus: transmit S24_LE with 64 BCLK cycles
Signed-off-by: Matthias Reichl <hias@horus.com>
hifiberry_dacplus: switch to snd_soc_dai_set_bclk_ratio
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: hifiberry_dacplus: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Add driver for rpi-proto
Forward port of 3.10.x driver from https://github.com/koalo
We are using a custom board and would like to use rpi 3.18.x
kernel. Patch works fine for our embedded system.
URL to the audio chip:
http://www.mikroe.com/add-on-boards/audio-voice/audio-codec-proto/
Playback tested with devicetree enabled.
Signed-off-by: Waldemar Brodkorb <wbrodkorb@conet.de>
ASoC: rpi-proto: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Add Support for JustBoom Audio boards
justboom-dac: Adjust for ALSA API change
As of 4.4, snd_soc_limit_volume now takes a struct snd_soc_card *
rather than a struct snd_soc_codec *.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
ASoC: justboom-dac: fix S24_LE format
Remove set_bclk_ratio call so 24-bit data is transmitted in
24 bclk cycles.
Also remove hw_params as it's no longer needed.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: justboom-dac: use modern dai_link style
Signed-off-by: Matthias Reichl <hias@horus.com>
New AudioInjector.net Pi soundcard with low jitter audio in and out.
Contains the sound/soc/bcm ALSA machine driver and necessary alterations to the Kconfig and Makefile.
Adds the dts overlay and updates the Makefile and README.
Updates the relevant defconfig files to enable building for the Raspberry Pi.
Thanks to Phil Elwell (pelwell) for the review, simple-card concepts and discussion. Thanks to Clive Messer for overlay naming suggestions.
Added support for headphones, microphone and bclk_ratio settings.
This patch adds headphone and microphone capability to the Audio Injector sound card. The patch also sets the bit clock ratio for use in the bcm2835-i2s driver. The bcm2835-i2s can't handle an 8 kHz sample rate when the bit clock is at 12 MHz because its register is only 10 bits wide which can't represent the ch2 offset of 1508. For that reason, the rate constraint is added.
ASoC: audioinjector-pi-soundcard: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
New driver for RRA DigiDAC1 soundcard using WM8741 + WM8804
ASoC: digidac1-soundcard: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Add support for Dion Audio LOCO DAC-AMP HAT
Using dedicated machine driver and pcm5102a codec driver.
Signed-off-by: DigitalDreamtime <clive.messer@digitaldreamtime.co.uk>
ASoC: dionaudio_loco: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Allo Piano DAC boards: Initial 2 channel (stereo) support (#1645)
Add initial 2 channel (stereo) support for Allo Piano DAC (2.0/2.1) boards,
using allo-piano-dac-pcm512x-audio overlay and allo-piano-dac ALSA ASoC
machine driver.
NB. The initial support is 2 channel (stereo) ONLY!
(The Piano DAC 2.1 will only support 2 channel (stereo) left/right output,
pending an update to the upstream pcm512x codec driver, which will have
to be submitted via upstream. With the initial downstream support,
provided by this patch, the Piano DAC 2.1 subwoofer outputs will
not function.)
Signed-off-by: Baswaraj K <jaikumar@cem-solutions.net>
Signed-off-by: Clive Messer <clive.messer@digitaldreamtime.co.uk>
Tested-by: Clive Messer <clive.messer@digitaldreamtime.co.uk>
ASoC: allo-piano-dac: fix S24_LE format
Remove set_bclk_ratio call so 24-bit data is transmitted in
24 bclk cycles.
Also remove hw_params and ops as they are no longer needed.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: allo-piano-dac: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Add support for Allo Piano DAC 2.1 plus add-on board for Raspberry Pi.
The Piano DAC 2.1 has support for 4 channels with subwoofer.
Signed-off-by: Baswaraj K <jaikumar@cem-solutions.net>
Reviewed-by: Vijay Kumar B. <vijaykumar@zilogic.com>
Reviewed-by: Raashid Muhammed <raashidmuhammed@zilogic.com>
Add clock changes and mute gpios (#1938)
Also improve code style and adhere to ALSA coding conventions.
Signed-off-by: Baswaraj K <jaikumar@cem-solutions.net>
Reviewed-by: Vijay Kumar B. <vijaykumar@zilogic.com>
Reviewed-by: Raashid Muhammed <raashidmuhammed@zilogic.com>
PianoPlus: Dual Mono & Dual Stereo features added (#2069)
allo-piano-dac-plus: Master volume added + fixes
Master volume added, which controls both DACs volumes.
See: https://github.com/raspberrypi/linux/pull/2149
Also fix initial max volume, default mode value, and unmute.
Signed-off-by: allocom <sparky-dev@allo.com>
ASoC: allo-piano-dac-plus: fix S24_LE format
Remove set_bclk_ratio call so 24-bit data is transmitted in
24 bclk cycles.
Signed-off-by: Matthias Reichl <hias@horus.com>
sound: bcm: Fix memset dereference warning
This warning appears with GCC 6.4.0 from toolchains.bootlin.com:
../sound/soc/bcm/allo-piano-dac-plus.c: In function ‘snd_allo_piano_dac_init’:
../sound/soc/bcm/allo-piano-dac-plus.c:711:30: warning: argument to ‘sizeof’ in ‘memset’ call is the same expression as the destination; did you mean to dereference it? [-Wsizeof-pointer-memaccess]
memset(glb_ptr, 0x00, sizeof(glb_ptr));
^
Suggested-by: Phil Elwell <phil@raspberrypi.org>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
ASoC: allo-piano-dac-plus: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Add support for Allo Boss DAC add-on board for Raspberry Pi. (#1924)
Signed-off-by: Baswaraj K <jaikumar@cem-solutions.net>
Reviewed-by: Deepak <deepak@zilogic.com>
Reviewed-by: BabuSubashChandar <babusubashchandar@zilogic.com>
Add support for new clock rate and mute gpios.
Signed-off-by: Baswaraj K <jaikumar@cem-solutions.net>
Reviewed-by: Deepak <deepak@zilogic.com>
Reviewed-by: BabuSubashChandar <babusubashchandar@zilogic.com>
ASoC: allo-boss-dac: fix S24_LE format
Remove set_bclk_ratio call so 24-bit data is transmitted in
24 bclk cycles.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: allo-boss-dac: transmit S24_LE with 64 BCLK cycles
Signed-off-by: Matthias Reichl <hias@horus.com>
allo-boss-dac: switch to snd_soc_dai_set_bclk_ratio
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: allo-boss-dac: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Support for Blokas Labs pisound board
Pisound dynamic overlay (#1760)
Restructuring pisound-overlay.dts, so it can be loaded and unloaded dynamically using dtoverlay.
Print a logline when the kernel module is removed.
pisound improvements:
* Added a writable sysfs object to enable scripts / user space software
to blink MIDI activity LEDs for variable duration.
* Improved hw_param constraints setting.
* Added compatibility with S16_LE sample format.
* Exposed some simple placeholder volume controls, so the card appears
in volumealsa widget.
Add missing SND_PISOUND selects dependency to SND_RAWMIDI
Without it the Pisound module fails to compile.
See https://github.com/raspberrypi/linux/issues/2366
Updates for Pisound module code:
* Merged 'Fix a warning in DEBUG builds' (1c8b82b).
* Updating some strings and copyright information.
* Fix for handling high load of MIDI input and output.
* Use dual rate oversampling ratio for 96kHz instead of single
rate one.
Signed-off-by: Giedrius Trainavicius <giedrius@blokas.io>
Fixing memset call in pisound.c
Signed-off-by: Giedrius Trainavicius <giedrius@blokas.io>
Fix for Pisound's MIDI Input getting blocked for a while in rare cases.
There was a possible race condition which could lead to Input's FIFO queue
to be underflown, causing high amount of processing in the worker thread for
some period of time.
Signed-off-by: Giedrius Trainavicius <giedrius@blokas.io>
Fix for Pisound kernel module in Real Time kernel configuration.
When handler of data_available interrupt is fired, queue_work ends up
getting called and it can block on a spin lock which is not allowed in
interrupt context. The fix was to run the handler from a thread context
instead.
Pisound: Remove spinlock usage around spi_sync
ASoC: pisound: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
ASoC: pisound: fix the parameter for spi_device_match
Signed-off-by: Hui Wang <hui.wang@canonical.com>
ASoC: Add driver for Cirrus Logic Audio Card
Note: due to problems with deferred probing of regulators
the following softdep should be added to a modprobe.d file
softdep arizona-spi pre: arizona-ldo1
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: rpi-cirrus: use modern dai_link style
Signed-off-by: Matthias Reichl <hias@horus.com>
sound: Support for Dion Audio LOCO-V2 DAC-AMP HAT
Signed-off-by: Miquel Blauw <info@dionaudio.nl>
ASoC: dionaudio_loco-v2: fix S24_LE format
Remove set_bclk_ratio call so 24-bit data is transmitted in
24 bclk cycles.
Also remove hw_params and ops as they are no longer needed.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: dionaudio_loco-v2: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Add support for Fe-Pi audio sound card. (#1867)
Fe-Pi Audio Sound Card is based on NXP SGTL5000 codec.
Mechanical specification of the board is the same the Raspberry Pi Zero.
3.5mm jacks for Headphone/Mic, Line In, and Line Out.
Signed-off-by: Henry Kupis <fe-pi@cox.net>
ASoC: fe-pi-audio: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Add support for the AudioInjector.net Octo sound card
AudioInjector Octo: sample rates, regulators, reset
This patch adds new sample rates to the Audioinjector Octo sound card. The
new supported rates are (in kHz) :
96, 48, 32, 24, 16, 8, 88.2, 44.1, 29.4, 22.05, 14.7
Reference the bcm270x DT regulators in the overlay.
This patch adds a reset GPIO for the AudioInjector.net octo sound card.
Audioinjector octo : Make the playback and capture symmetric
This patch ensures that the sample rate and channel count of the audioinjector
octo sound card are symmetric.
audioinjector-octo: Add continuous clock feature
By user request, add a switch to prevent the clocks being stopped when
the stream is paused, stopped or shutdown. Provide access to the switch
by adding a 'non-stop-clocks' parameter to the audioinjector-addons
overlay.
See: https://github.com/raspberrypi/linux/issues/2409
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
sound: Fixes for audioinjector-octo under 4.19
1. Move the DT alias declaration to the I2C shim in the cases
where the shim is enabled. This works around a problem caused by a
4.19 commit [1] that generates DT/OF uevents for I2C drivers.
2. Fix the diagnostics in an error path of the soundcard driver to
correctly identify the reason for the failure to load.
3. Move the declaration of the clock node in the overlay outside
the I2C node to avoid warnings.
4. Sort the overlay nodes so that dependencies are only to earlier
fragments, in an attempt to get runtime dtoverlay application to
work (it still doesn't...)
See: https://github.com/Audio-Injector/Octo/issues/14
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
[1] af503716ac ("i2c: core: report OF style module alias for devices registered via OF")
ASoC: audioinjector-octo-soundcard: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Driver support for Google voiceHAT soundcard.
ASoC: googlevoicehat-codec: Use correct device when grabbing GPIO
The fixup for the VoiceHAT in 4.18 incorrectly tried to find the
sdmode GPIO pin under the card device, not the codec device.
This failed, and therefore caused the device probe to fail.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
ASoC: googlevoicehat-codec: Reformat for kernel coding standards
Fix all whitespace, indentation, and bracing errors.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
ASoC: googlevoicehat-codec: Make driver function structure const
Make voicehat_component_driver a const structure.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
ASoC: googlevoicehat-codec: Only convert from ms to jiffies once
Minor optimisation and allows to become checkpatch clean.
A msec value is read out of DT or from a define, and convert once to
jiffies, rather than every time that it is used.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Driver and overlay for Allo Katana DAC
Allo Katana DAC: Updated default values
Signed-off-by: Jaikumar <jaikumar@cem-solutions.com>
Added mute stream func
Signed-off-by: Jaikumar <jaikumar@cem-solutions.net>
codecs: Correct Katana minimum volume
Update Katana minimum volume to get the exact 0.5 dB value in each step.
Signed-off-by: Sudeep Kumar <sudeepkumar@cem-solutions.net>
ASoC: Add generic RPI driver for simple soundcards.
The RPI simple sound card driver provides a generic ALSA SOC card driver
supporting a variety of Pi HAT soundcards. The intention is to avoid
the duplication of code for cards that can't be fully supported by
the soc simple/graph cards but are otherwise almost identical.
This initial commit adds support for the ADAU1977 ADC, Google VoiceHat,
HifiBerry AMP, HifiBerry DAC and RPI DAC.
Signed-off-by: Tim Gover <tim.gover@raspberrypi.org>
ASoC: Use correct card name in rpi-simple driver
Use the specific card name from drvdata instead of the snd_rpi_simple
rpi-simple-soundcard: Use nicer driver name "RPi-simple"
Rename the driver from "RPI simple soundcard" to "RPi-simple" so that
the driver name won't be mangled allowing to be used unaltered as the
card conf filename.
ASoC: rpi-simple-soundcard: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
ASoC: Add Kconfig and Makefile for sound/soc/bcm
Signed-off-by: popcornmix <popcornmix@gmail.com>
ASoC: Create a generic Pi Hat WM8804 driver
Reduce the amount of duplicated code by creating a generic driver for
Pi Hat digi cards using the WM8804 codec.
This replaces the
Allo DigiOne, Hifiberry Digi/Pro, JustBoom Digi and IQAudIO Digi
dedicate soundcard drivers with a generic driver.
There are no significant changes to the runtime behavior of the drivers
and end users should not have to change any configuration settings
after upgrading.
Minor changes
* Check the return value of snd_soc_component_update_bits
* Added some pr_debug tracing
* Various checkpatch tidyups
* Updated allodigi-one to use use 128FS at > 96 Khz. This appears to
be an omission in the original driver code so followed the Hifiberry
DAC driver approach.
ASoC: rpi-wm8804-soundcard: use modern dai_link style
Signed-off-by: Matthias Reichl <hias@horus.com>
rpi-wm8804-soundcard: drop PWRDN register writes
Since kernel 4.0 the PWRDN register bits are under DAPM
control from the wm8804 driver.
Drop code that modifies that register to avoid interfering
with DAPM.
Signed-off-by: Matthias Reichl <hias@horus.com>
rpi-wm8804-soundcard: configure wm8804 clocks only on rate change
This should avoid clicks when stopping and immediately afterwards
starting a stream with the same samplerate as before.
Signed-off-by: Matthias Reichl <hias@horus.com>
rpi-wm8804-soundcard: Fixed MCLKDIV for Allo Digione
The Allo Digione board wants a fixed MCLKDIV of 256.
See: https://github.com/raspberrypi/linux/issues/3296
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
ASoC: Add support for AudioSense-Pi add-on soundcard
AudioSense-Pi is a RPi HAT based on a TI's TLV320AIC32x4 stereo codec
This hardware provides multiple audio I/O capabilities to the RPi.
The codec connects to the RPi's SoC through the I2S Bus.
The following devices can be connected through a 3.5mm jack
1. Line-In: Plain old audio in from mobile phones, PCs, etc.,
2. Mic-In: Connect a microphone
3. Line-Out: Connect the output to a speaker
4. Headphones: Connect a Headphone w or w/o microphones
Multiple Inputs:
It supports the following combinations
1. Two stereo Line-Inputs and a microphone
2. One stereo Line-Input and two microphones
3. Two stereo Line-Inputs, a microphone and
one mono line-input (with h/w hack)
4. One stereo Line-Input, two microphones and
one mono line-input (with h/w hack)
Multiple Outputs:
Audio output can be routed to the headphones or
speakers (with additional hardware)
Signed-off-by: b-ak <anur.bhargav@gmail.com>
ASoC: audiosense-pi: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Added driver for the HiFiBerry DAC+ ADC (#2694)
Signed-off-by: Daniel Matuschek <daniel@hifiberry.com>
hifiberry_dacplusadc: switch to snd_soc_dai_set_bclk_ratio
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: hifiberry_dacplusadc: fix DAI link setup
The driver only defines a single DAI link and the code that tries
to setup the second (non-existent) DAI link looks wrong - using dmic
as a CPU/platform driver doesn't make any sense.
The DT overlay doesn't define a dmic property, so the code was never
executed (otherwise it would have resulted in a memory corruption).
So drop the offending code to prevent issues if a dmic property
should be added to the DT overlay.
Signed-off-by: Matthias Reichl <hias@horus.com>
ASoC: hifiberry_dacplusadc: use modern dai_link style
Signed-off-by: Matthias Reichl <hias@horus.com>
Audiophonics I-Sabre 9038Q2M DAC driver
Signed-off-by: Audiophonics <contact@audiophonics.fr>
ASoC: i-sabre-q2m: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Added IQaudIO Pi-Codec board support (#2969)
Add support for the IQaudIO Pi-Codec board.
Signed-off-by: Gordon <gordon@iqaudio.com>
Fixed 48k timing issue
ASoC: iqaudio-codec: use modern dai_link style
Signed-off-by: Hui Wang <hui.wang@canonical.com>
adds the Hifiberry DAC+ADC PRO version
This adds the driver for the DAC+ADC PRO version of the Hifiberry soundcard with software controlled PCM1863 ADC
Signed-off-by: Joerg Schambacher joerg@i2audio.com
Add Hifiberry DAC+DSP soundcard driver (#3224)
Adds the driver for the Hifiberry DAC+DSP. It supports capture and
playback depending on the DSP firmware.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
Allow simultaneous use of JustBoom DAC and Digi
Signed-off-by: Johannes Krude <johannes@krude.de>
Pisound: MIDI communication fixes for scaled down CPU.
* Increased maximum SPI communication speed to avoid running too slow
when the CPU is scaled down and losing MIDI data.
* Keep track of buffer usage in millibytes for higher precision.
Signed-off-by: Giedrius Trainavičius <giedrius@blokas.io>
sound: Add the HiFiBerry DAC+HD version
This adds the driver for the DAC+HD version supporting HiFiBerry's
PCM179x based DACs. It also adds PLL control for clock generation.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
Fix master mode settings of HiFiBerry DAC+ADC PRO card (#3424)
This patch fixes the board DAI setting when in master-mode.
Wrong setting could have caused random pop noise.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
adds LED OFF feature to HiFiBerry DAC+ADC PRO sound card
This adds a DT overlay parameter 'leds_off' which allows
to switch off the onboard activity LEDs at all times
which has been requested by some users.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
adds LED OFF feature to HiFiBerry DAC+ADC sound card
This adds a DT overlay parameter 'leds_off' which allows
to switch off the onboard activity LEDs at all times
which has been requested by some users.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
adds LED OFF feature to HiFiBerry DAC+/DAC+PRO sound cards
This adds a DT overlay parameter 'leds_off' which allows
to switch off the onboard activity LEDs at all times
which has been requested by some users.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
pisound: Added reading Pisound board hardware revision and exposing it (#3425)
pisound: Added reading Pisound board hardware revision and exposing it in kernel log and sysfs file:
/sys/kernel/pisound/hw_version
Signed-off-by: Giedrius <giedrius@blokas.io>
Added driver for HiFiBerry Amp amplifier add-on board
The driver contains a low-level hardware driver for the TAS5713 and the
drivers for the Raspberry Pi I2S subsystem.
TAS5713: return error if initialisation fails
Existing TAS5713 driver logs errors during initialisation, but does not return
an error code. Therefore even if initialisation fails, the driver will still be
loaded, but won't work. This patch fixes this. I2C communication error will now
reported correctly by a non-zero return code.
HiFiBerry Amp: fix device-tree problems
Some code to load the driver based on device-tree-overlays was missing. This is added by this patch.
According to 5713 pdf doc CLOCK_CTRL is a readonly status register, and it behaves so. Remove useless setting
sound: pcm512x-codec: Adding 352.8kHz samplerate support
sound/soc: only first codec is master in multicodec setup
When using multiple codecs, at most one codec should generate the master
clock. All codecs except the first are therefore configured for slave
mode.
Signed-off-by: Johannes Krude <johannes@krude.de>
ASoC: Fix snd_soc_get_pcm_runtime usage
Commit [1] changed the snd_soc_get_pcm_runtime to take a dai_link
pointer instead of a string. Patch up the downstream drivers to use
the modified API.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
[1] 4468189ff3 ("ASoC: soc-core: find rtd via dai_link pointer at snd_soc_get_pcm_runtime()")
Add support for the AudioInjector.net Isolated sound card
This patch adds support for the Audio Injector Isolated sound card.
Signed-off-by: Matt Flax <flatmax@flatmax.org>
Add support for merus-amp soundcard and ma120x0p codec
Add 96KHz rate support to MA120X0P codec and make enable and mute gpio
pins optional.
Signed-off-by: AMuszkat <ariel.muszkat@gmail.com>
Fixes a problem with clock settings of HiFiBerry DAC+ADC PRO (#3545)
This patch fixes a problem of the re-calculation of
i2s-clock and -parameter settings when only the ADC is activated.
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
configs: Enable the AD193x codecs
See: https://github.com/raspberrypi/linux/issues/2850
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Switch to snd_soc_dai_set_bclk_ratio
Replaces obsolete function snd_soc_dai_set_tdm_slot
Signed-off-by: Joerg Schambacher <joerg@i2audio.com>
Enhances the DAC+ driver to control the optional headphone amplifier
Probes on the I2C bus for TPA6130A2, if successful, it sets DT-parameter
'status' from 'disabled' to 'okay' using change_sets to enable
the headphone control.
Signed-off-by: Joerg Schambacher joerg@i2audio.com
The Raspberry Pi firmware manages the power-down and reboot
process. To do this it installs a pm_power_off handler, causing
the gpio-poweroff module to abort the probe function.
This patch introduces a "force" DT property that overrides that
behaviour, and also adds a DT overlay to enable and control it.
Note that running in an active-low configuration (DT parameter
"active_low") requires a custom dt-blob.bin and probably won't
allow a reboot without switching off, so an external inversion
of the trigger signal may be preferable.
Provide a __copy_from_user that uses memcpy. On BCM2708, use
optimised memcpy/memmove/memcmp/memset implementations.
arch/arm: Add mmiocpy/set aliases for memcpy/set
See: https://github.com/raspberrypi/linux/issues/1082
copy_from_user: CPU_SW_DOMAIN_PAN compatibility
The downstream copy_from_user acceleration must also play nice with
CONFIG_CPU_SW_DOMAIN_PAN.
See: https://github.com/raspberrypi/linux/issues/1381
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Fix copy_from_user if BCM2835_FAST_MEMCPY=n
The change which introduced CONFIG_BCM2835_FAST_MEMCPY unconditionally
changed the behaviour of arm_copy_from_user. The page pinning code
is not safe on ARMv7 if LPAE & high memory is enabled and causes
crashes which look like PTE corruption.
Make __copy_from_user_memcpy conditional on CONFIG_2835_FAST_MEMCPY=y
which is really an ARMv6 / Pi1 optimization and not necessary on newer
ARM processors.
arm: fix mmap unlocks in uaccess_with_memcpy.c
This is a regression that was added with the commit 192a4e923e as of rpi-5.8.y, since that is when the move to the mmap locking API was introduced - d8ed45c5dc
The issue is that when the patch to improve performance for the __copy_to_user and __copy_from_user functions were added for the Raspberry Pi, some of the mmaps were incorrectly mapped to write instead of read. This would cause a verity of issues, and in my case, prevent the booting of a squashfs filesystem on rpi-5.8-y and above. An example of the panic you would see from this can be seen at https://pastebin.com/raw/jBz5xCzL
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Christopher Blake <chrisrblake93@gmail.com>
The "input" trigger makes the associated GPIO an input. This is to support
the Raspberry Pi PWR LED, which is driven by external hardware in normal use.
N.B. pwr_led is not available on Model A or B boards.
leds-gpio: Implement the brightness_get method
The power LED uses some clever logic that means it is driven
by a voltage measuring circuit when configured as input, otherwise
it is driven by the GPIO output value. This patch wires up the
brightness_get method for leds-gpio so that user-space can monitor
the LED value via /sys/class/gpio/led1/brightness. Using the input
trigger this returns an indication of the system power health,
otherwise it is just whatever value the trigger has written most
recently.
See: https://github.com/raspberrypi/linux/issues/1064
Add the bare minimum needed to boot BCM2708 from a Device Tree.
Signed-off-by: Noralf Tronnes <notro@tronnes.org>
BCM2708: DT: change 'axi' nodename to 'soc'
Change DT node named 'axi' to 'soc' so it matches ARCH_BCM2835.
The VC4 bootloader fills in certain properties in the 'axi' subtree,
but since this is part of an upstreaming effort, the name is changed.
Signed-off-by: Noralf Tronnes notro@tronnes.org
BCM2708_DT: Correct length of the peripheral space
Use dts-dirs feature for overlays.
The kernel makefiles have a dts-dirs target that is for vendor subdirectories.
Using this fixes the install_dtbs target, which previously did not install the overlays.
BCM270X_DT: configure I2S DMA channels
Signed-off-by: Matthias Reichl <hias@horus.com>
BCM270X_DT: switch to bcm2835-i2s
I2S soundcard drivers with proper devicetree support (i.e. not linking
to the cpu_dai/platform via name but to cpu/platform via of_node)
will work out of the box without any modifications.
When the kernel is compiled without devicetree support the platform
code will instantiate the bcm2708-i2s driver and I2S soundcard drivers
will link to it via name, as before.
Signed-off-by: Matthias Reichl <hias@horus.com>
SDIO-overlay: add poll_once-boolean parameter
Add paramter to toggle sdio-device-polling
done every second or once at boot-time.
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
BCM270X_DT: Make mmc overlay compatible with current firmware
The original DT overlay logic followed a merge-then-patch procedure,
i.e. parameters are applied to the loaded overlay before the overlay
is merged into the base DTB. This sequence has been changed to
patch-then-merge, in order to support parameterised node names, and
to protect against bad overlays. As a result, overrides (parameters)
must only target labels in the overlay, but the overlay can obviously target nodes in the base DTB.
mmc-overlay.dts (that switches back to the original mmc sdcard
driver) is the only overlay violating that rule, and this patch
fixes it.
bcm270x_dt: Use the sdhost MMC controller by default
The "mmc" overlay reverts to using the other controller.
squash: Add cprman to dt
BCM270X_DT: Use clk_core for I2C interfaces
BCM270X_DT: Use bcm283x.dtsi, bcm2835.dtsi and bcm2836.dtsi
The mainline Device Tree files are quite close to downstream now.
Let's use bcm283x.dtsi, bcm2835.dtsi and bcm2836.dtsi as base files
for our dts files.
Mainline dts files are based on these files:
bcm2835-rpi.dtsi
bcm2835.dtsi bcm2836.dtsi
bcm283x.dtsi
Current downstream are based on these:
bcm2708.dtsi bcm2709.dtsi bcm2710.dtsi
bcm2708_common.dtsi
This patch introduces this dependency:
bcm2708.dtsi bcm2709.dtsi
bcm2708-rpi.dtsi
bcm270x.dtsi
bcm2835.dtsi bcm2836.dtsi
bcm283x.dtsi
And:
bcm2710.dtsi
bcm2708-rpi.dtsi
bcm270x.dtsi
bcm283x.dtsi
bcm270x.dtsi contains the downstream bcm283x.dtsi diff.
bcm2708-rpi.dtsi is the downstream version of bcm2835-rpi.dtsi.
Other changes:
- The led node has moved from /soc/leds to /leds. This is not a problem
since the label is used to reference it.
- The clk_osc reg property changes from 6 to 3.
- The gpu nodes has their interrupt property set in the base file.
- the clocks label does not point to the /clocks node anymore, but
points to the cprman node. This is not a problem since the overlays
that use the clock node refer to it directly: target-path = "/clocks";
- some nodes now have 2 labels since mainline and downstream differs in
this respect: cprman/clocks, spi0/spi, gpu/vc4.
- some nodes doesn't have an explicit status = "okay" since they're not
disabled in the base file: watchdog and random.
- gpiomem doesn't need an explicit status = "okay".
- bcm2708-rpi-cm.dts got the hpd-gpios property from bcm2708_common.dtsi,
it's now set directly in that file.
- bcm2709-rpi-2-b.dts has the timer node moved from /soc/timer to /timer.
- Removed clock-frequency property on the bcm{2709,2710}.dtsi timer nodes.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
BCM270X_DT: Use raspberrypi-power to turn on USB power
Use the raspberrypi-power driver to turn on USB power.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
BCM270X_DT: Add a .dtbo target, use for overlays
Change the filenames and extensions to keep the pre-DDT style of
overlay (<name>-overlay.dtb) distinct from new ones that use a
different style of local fixups (<name>.dtbo), and to match other
platforms.
The RPi firmware uses the DDTK trailer atom to choose which type of
overlay to use for each kernel.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
BCM270X_DT: Don't generate "linux,phandle" props
The EPAPR standard says to use "phandle" properties to store phandles,
rather than the deprecated "linux,phandle" version. By default, dtc
generates both, but adding "-H epapr" causes it to only generate
"phandle"s, saving some space and clutter.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
BCM270X_DT: Add overlay for enc28j60 on SPI2
Works on SPI2 for compute module
BCM270X_DT: Add midi-uart0 overlay
MIDI requires 31.25kbaud, a baudrate unsupported by Linux. The
midi-uart0 overlay configures uart0 (ttyAMA0) to use a fake clock
so that requesting 38.4kbaud actually gets 31.25kbaud.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
BCM270X_DT: Add i2c-sensor overlay
The i2c-sensor overlay is a container for various pressure and
temperature sensors, currently bmp085 and bmp280. The standalone
bmp085_i2c-sensor overlay is now deprecated.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
BCM270X_DT: overlays/*-overlay.dtb -> overlays/*.dtbo (#1752)
We now create overlays as .dtbo files.
build: support for .dtbo files for dtb overlays
Kernel 4.4.6+ on RaspberryPi support .dtbo files for overlays, instead of .dtb.
Patch the kernel, which has faulty rules to generate .dtbo the way yocto does
Signed-off-by: Herve Jourdain <herve.jourdain@neuf.fr>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
BCM270X: Drop position requirement for CMA in VC4 overlay.
No longer necessary since 2aefcd5761,
and will probably let peeople that want to choose a larger CMA
allocation (particularly on pi0/1).
Signed-off-by: Eric Anholt <eric@anholt.net>
BCM270X_DT: RPi Device Tree tidy
Use the upstream sdhost node, add thermal-zones, and factor out some
common elements.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
kbuild: Silence unhelpful DTC warnings
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
BCM270X_DT: DT build rules no longer arch-specific
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Support booting without Device Tree.
Turn on USB power.
Load driver early because of lacking support for deferred probing
in many drivers.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
firmware: bcm2835: Don't turn on USB power
The raspberrypi-power driver is now used to turn on USB power.
This partly reverts commit:
firmware: bcm2835: Support ARCH_BCM270x
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Add module for accessing the mailbox property channel through
/dev/vcio. Was previously in bcm2708-vcio.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
char: vcio: Add compat ioctl handling
There was no compat ioctl handler, so 32 bit userspace on a
64 bit kernel failed as IOCTL_MBOX_PROPERTY used the size
of char*.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
char: vcio: Fail probe if rpi_firmware is not found.
Device Tree is now the only supported config mechanism, therefore
uncomment the block of code that fails the probe if the
firmware node can't be found.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
i2c-bcm2708: fixed baudrate
Fixed issue where the wrong CDIV value was set for baudrates below 3815 Hz (for 250MHz bus clock).
In that case the computed CDIV value was more than 0xffff. However the CDIV register width is only 16 bits.
This resulted in incorrect setting of CDIV and higher baudrate than intended.
Example: 3500Hz -> CDIV=0x11704 -> CDIV(16bit)=0x1704 -> 42430Hz
After correction: 3500Hz -> CDIV=0x11704 -> CDIV(16bit)=0xffff -> 3815Hz
The correct baudrate is shown in the log after the cdiv > 0xffff correction.
Perform I2C combined transactions when possible
Perform I2C combined transactions whenever possible, within the
restrictions of the Broadcomm Serial Controller.
Disable DONE interrupt during TA poll
Prevent interrupt from being triggered if poll is missed and transfer
starts and finishes.
i2c: Make combined transactions optional and disabled by default
i2c: bcm2708: add device tree support
Add DT support to driver and add to .dtsi file.
Setup pins in .dts file.
i2c is disabled by default.
Signed-off-by: Noralf Tronnes <notro@tronnes.org>
bcm2708: don't register i2c controllers when using DT
The devices for the i2c controllers are in the Device Tree.
Only register devices when not using DT.
Signed-off-by: Noralf Tronnes <notro@tronnes.org>
I2C: Only register the I2C device for the current board revision
i2c_bcm2708: Fix clock reference counting
Fix grabbing lock from atomic context in i2c driver
2 main changes:
- check for timeouts in the bcm2708_bsc_setup function as indicated by this comment:
/* poll for transfer start bit (should only take 1-20 polls) */
This implies that the setup function can now fail so account for this everywhere it's called
- Removed the clk_get_rate call from inside the setup function as it locks a mutex and that's not ok since we call it from under a spin lock.
i2c-bcm2708: When using DT, leave the GPIO setup to pinctrl
i2c-bcm2708: Increase timeouts to allow larger transfers
Use the timeout value provided by the I2C_TIMEOUT ioctl when waiting
for completion. The default timeout is 1 second.
See: https://github.com/raspberrypi/linux/issues/260
i2c-bcm2708/BCM270X_DT: Add support for I2C2
The third I2C bus (I2C2) is normally reserved for HDMI use. Careless
use of this bus can break an attached display - use with caution.
It is recommended to disable accesses by VideoCore by setting
hdmi_ignore_edid=1 or hdmi_edid_file=1 in config.txt.
The interface is disabled by default - enable using the
i2c2_iknowwhatimdoing DT parameter.
bcm2708-spi: Don't use static pin configuration with DT
Also remove superfluous error checking - the SPI framework ensures the
validity of the chip_select value.
i2c-bcm2708: Remove non-DT support
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Set the BSC_CLKT clock streching timeout to 35ms as per SMBus specs.
Fixes i2c_bcm2708: Write to FIFO correctly - v2 (#1574)
* i2c: fix i2c_bcm2708: Clear FIFO before sending data
Make sure FIFO gets cleared before trying to send
data in case of a repeated start (COMBINED=Y).
* i2c: fix i2c_bcm2708: Only write to FIFO when not full
Check if FIFO can accept data before writing.
To avoid a peripheral read on the last iteration of a loop,
both bcm2708_bsc_fifo_fill and ~drain are changed as well.
Signed-off-by: Luke Wren <wren6991@gmail.com>
MISC: bcm2835: smi: use clock manager and fix reload issues
Use clock manager instead of self-made clockmanager.
Also fix some error paths that showd up during development
(especially missing release of dma resources on rmmod)
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
bcm2835_smi: re-add dereference to fix DMA transfers
Signed-off-by: popcornmix <popcornmix@gmail.com>
BCM270x: Move vc_mem
Make the vc_mem module available for ARCH_BCM2835 by moving it.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
char: vc_mem: Fix up compat ioctls for 64bit kernel
compat_ioctl wasn't defined, so 32bit user/64bit kernel
always failed.
VC_MEM_IOC_MEM_PHYS_ADDR was defined with parameter size
unsigned long, so the ioctl cmd changes between sizes.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
char: vc_mem: Fix all coding style issues.
Cleans up all checkpatch errors in vc_mem.c and vc_mem.h
No functional change to the code.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
BCM2835 has two SD card interfaces. This driver uses the other one.
bcm2835-sdhost: Error handling fix, and code clarification
bcm2835-sdhost: Adding overclocking option
Allow a different clock speed to be substitued for a requested 50MHz.
This option is exposed using the "overclock_50" DT parameter.
Note that the sdhost interface is restricted to integer divisions of
core_freq, and the highest sensible option for a core_freq of 250MHz
is 84 (250/3 = 83.3MHz), the next being 125 (250/2) which is much too
high.
Use at your own risk.
bcm2835-sdhost: Round up the overclock, so 62 works for 62.5Mhz
Also only warn once for each overclock setting.
bcm2835-sdhost: Improve error handling and recovery
1) Expose the hw_reset method to the MMC framework, removing many
internal calls by the driver.
2) Reduce overclock setting on error.
3) Increase timeout to cope with high capacity cards.
4) Add properties and parameters to control pio_limit and debug.
5) Reduce messages at probe time.
bcm2835-sdhost: Further improve overclock back-off
bcm2835-sdhost: Clear HBLC for PIO mode
Also update pio_limit default in overlay README.
bcm2835-sdhost: Add the ERASE capability
See: https://github.com/raspberrypi/linux/issues/1076
bcm2835-sdhost: Ignore CRC7 for MMC CMD1
It seems that the sdhost interface returns CRC7 errors for CMD1,
which is the MMC-specific SEND_OP_COND. Returning these errors to
the MMC layer causes a downward spiral, but ignoring them seems
to be harmless.
bcm2835-mmc/sdhost: Remove ARCH_BCM2835 differences
The bcm2835-mmc driver (and -sdhost driver that copied from it)
contains code to handle SDIO interrupts in a threaded interrupt
handler rather than waking the MMC framework thread. The change
follows a patch from Russell King that adds the facility as the
preferred way of working.
However, the new code path is only present in ARCH_BCM2835
builds, which I have taken to be a way of testing the waters
rather than making the change across the board; I can't see
any technical reason why it wouldn't be enabled for MACH_BCM270X
builds. So this patch standardises on the ARCH_BCM2835 code,
removing the old code paths.
bcm2835-sdhost: Don't log timeout errors unless debug=1
The MMC card-discovery process generates timeouts. This is
expected behaviour, so reporting it to the user serves no purpose.
Suppress the reporting of timeout errors unless the debug flag
is on.
bcm2835-sdhost: Add workaround for odd behaviour on some cards
For reasons not understood, the sdhost driver fails when reading
sectors very near the end of some SD cards. The problem could
be related to the similar issue that reading the final sector
of any card as part of a multiple read never completes, and the
workaround is an extension of the mechanism introduced to solve
that problem which ensures those sectors are always read singly.
bcm2835-sdhost: Major revision
This is a significant revision of the bcm2835-sdhost driver. It
improves on the original in a number of ways:
1) Through the use of CMD23 for reads it appears to avoid problems
reading some sectors on certain high speed cards.
2) Better atomicity to prevent crashes.
3) Higher performance.
4) Activity logging included, for easier diagnosis in the event
of a problem.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Restore ATOMIC flag to PIO sg mapping
Allocation problems have been seen in a wireless driver, and
this is the only change which might have been responsible.
SQUASH: bcm2835-sdhost: Only claim one DMA channel
With both MMC controllers enabled there are few DMA channels left. The
bcm2835-sdhost driver only uses DMA in one direction at a time, so it
doesn't need to claim two channels.
See: https://github.com/raspberrypi/linux/issues/1327
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Workaround for "slow" sectors
Some cards have been seen to cause timeouts after certain sectors are
read. This workaround enforces a minimum delay between the stop after
reading one of those sectors and a subsequent data command.
Using CMD23 (SET_BLOCK_COUNT) avoids this problem, so good cards will
not be penalised by this workaround.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Firmware manages the clock divisor
The bcm2835-sdhost driver hands control of the CDIV clock divisor
register to matching firmware, allowing it to adjust to a changing
core clock. This removes the need to use the performance governor or
to enable io_is_busy on the on-demand governor in order to get the
best SD performance.
N.B. As SD clocks must be an integer divisor of the core clock, it is
possible that the SD clock for "turbo" mode can be different (even
lower) than "normal" mode.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Reset the clock in task context
Since reprogramming the clock can now involve a round-trip to the
firmware it must not be done at atomic context, and a tasklet
is not a task.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Don't exit cmd wait loop on error
The FAIL flag can be set in the CMD register before command processing
is complete, leading to spurious "failed to complete" errors. This has
the effect of promoting harmless CRC7 errors during CMD1 processing
into errors that can delay and even prevent booting.
Also:
1) Convert the last KERN_ERROR message in the register dumping to
KERN_INFO.
2) Remove an unnecessary reset call from bcm2835_sdhost_add_host.
See: https://github.com/raspberrypi/linux/pull/1492
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: mmc_card_blockaddr fix
Get the definition of mmc_card_blockaddr from drivers/mmc/core/card.h.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: New timer API
mmc: bcm2835-sdhost: Support underclocking
Support underclocking of the SD bus in two ways:
1. using the max-frequency DT property (which currently has no DT
parameter), and
2. using the exiting sd_overclock parameter.
The two methods differ slightly - in the former the MMC subsystem is
aware of the underclocking, while in the latter it isn't - but the
end results should be the same.
See: https://github.com/raspberrypi/linux/issues/2350
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
mmc: bcm2835-sdhost: Add include
highmem.h (needed for kmap_atomic) is pulled in by one of the other
include files, but only with some CONFIG settings. Make the inclusion
explicit to cater for cases where the CONFIG setting is absent.
See: https://github.com/raspberrypi/linux/issues/2366
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
mmc/bcm2835-sdhost: Recover from MMC_SEND_EXT_CSD
If the user issues an "mmc extcsd read", the SD controller receives
what it thinks is a SEND_IF_COND command with an unexpected data block.
The resulting operations leave the FSM stuck in READWAIT, a state which
persists until the MMC framework resets the controller, by which point
the root filesystem is likely to have been unmounted.
A less heavyweight solution is to detect the condition and nudge the
FSM by asserting the (self-clearing) FORCE_DATA_MODE bit.
N.B. This workaround was essentially discovered by accident and without
a full understanding the inner workings of the controller, so it is
fortunate that the "fix" only modifies error paths.
See: https://github.com/raspberrypi/linux/issues/2728
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
mmc: bcm2835-sdhost: Fix warnings on arm64
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Allow for sg entries that cross pages
The dma_complete handling code calculates a virtual address for a page
then adds an offset, but if the offset is more than a page and HIGHMEM
is in use then the summed address could be in an unmapped (or just
incorrect) page.
The upstream SDHOST driver allows for this possibility - copy the code
that does so.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Fix DMA channel leak on error/remove
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
mmc: bcm2835-sdhost: Support 64-bit physical addresses
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-sdhost: Replace obsolete struct timeval
struct timeval has been retired due to the impending linux 32-bit tv_sec
rollover (only 18 years to go) - timespec64 is the obvious replacement.
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
mmc: Disable CMD23 transfers on all cards
Pending wire-level investigation of these types of transfers
and associated errors on bcm2835-mmc, disable for now. Fallback of
CMD18/CMD25 transfers will be used automatically by the MMC layer.
Reported/Tested-by: Gellert Weisz <gellert@raspberrypi.org>
mmc: bcm2835-mmc: enable DT support for all architectures
Both ARCH_BCM2835 and ARCH_BCM270x are built with OF now.
Enable Device Tree support for all architectures.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
mmc: bcm2835-mmc: fix probe error handling
Probe error handling is broken in several places.
Simplify error handling by using device managed functions.
Replace pr_{err,info} with dev_{err,info}.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
bcm2835-mmc: Add locks when accessing sdhost registers
bcm2835-mmc: Add range of debug options for slowing things down
bcm2835-mmc: Add option to disable some delays
bcm2835-mmc: Add option to disable MMC_QUIRK_BLK_NO_CMD23
bcm2835-mmc: Default to disabling MMC_QUIRK_BLK_NO_CMD23
bcm2835-mmc: Adding overclocking option
Allow a different clock speed to be substitued for a requested 50MHz.
This option is exposed using the "overclock_50" DT parameter.
Note that the mmc interface is restricted to EVEN integer divisions of
250MHz, and the highest sensible option is 63 (250/4 = 62.5), the
next being 125 (250/2) which is much too high.
Use at your own risk.
bcm2835-mmc: Round up the overclock, so 62 works for 62.5Mhz
Also only warn once for each overclock setting.
mmc: bcm2835-mmc: Make available on ARCH_BCM2835
Make the bcm2835-mmc driver available for use on ARCH_BCM2835.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
BCM270x_DT: add bcm2835-mmc entry
Add Device Tree entry for bcm2835-mmc.
In non-DT mode, don't add the device in the board file.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
bcm2835-mmc: Don't overwrite MMC capabilities from DT
bcm2835-mmc: Don't override bus width capabilities from devicetree
Take out the force setting of the MMC_CAP_4_BIT_DATA host capability
so that the result read from devicetree via mmc_of_parse() is
preserved.
bcm2835-mmc: Only claim one DMA channel
With both MMC controllers enabled there are few DMA channels left. The
bcm2835-mmc driver only uses DMA in one direction at a time, so it
doesn't need to claim two channels.
See: https://github.com/raspberrypi/linux/issues/1327
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-mmc: New timer API
mmc: bcm2835-mmc: Support underclocking
Support underclocking of the SD bus using the max-frequency DT property
(which currently has no DT parameter). The sd_overclock parameter
already provides another way to achieve the same thing which should be
equivalent in end result, but it is a bug not to support max-frequency
as well.
See: https://github.com/raspberrypi/linux/issues/2350
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
mmc/bcm2835: Recover from MMC_SEND_EXT_CSD
If the user issues an "mmc extcsd read", the SD controller receives
what it thinks is a SEND_IF_COND command with an unexpected data block.
The resulting operations leave the FSM stuck in READWAIT, a state which
persists until the MMC framework resets the controller, by which point
the root filesystem is likely to have been unmounted.
A less heavyweight solution is to detect the condition and nudge the
FSM by asserting the (self-clearing) FORCE_DATA_MODE bit.
N.B. This workaround was essentially discovered by accident and without
a full understanding the inner workings of the controller, so it is
fortunate that the "fix" only modifies error paths.
See: https://github.com/raspberrypi/linux/issues/2728
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
bcm2835-mmc: Fix DMA channel leak
The BCM2835 MMC host driver requests a DMA channel on probe but neglects
to release the channel in the probe error path and on driver unbind.
I'm seeing this happen on every boot of the Compute Module 3: On first
driver probe, DMA channel 2 is allocated and then leaked with a "could
not get clk, deferring probe" message. On second driver probe, channel 4
is allocated.
Fix it.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Frank Pavlic <f.pavlic@kunbus.de>
bcm2835-mmc: Fix struct mmc_host leak on probe
The BCM2835 MMC host driver requests the bus address of the host's
register map on probe. If that fails, the driver leaks the struct
mmc_host allocated earlier.
Fix it.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Frank Pavlic <f.pavlic@kunbus.de>
bcm2835-mmc: Fix duplicate free_irq() on remove
The BCM2835 MMC host driver requests its interrupt as a device-managed
resource, so the interrupt is automatically freed after the driver is
unbound.
However on driver unbind, bcm2835_mmc_remove() frees the interrupt
explicitly to avoid invocation of the interrupt handler after driver
structures have been torn down.
The interrupt is thus freed twice, leading to a WARN splat in
__free_irq(). Fix by not requesting the interrupt as a device-managed
resource.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Frank Pavlic <f.pavlic@kunbus.de>
bcm2835-mmc: Handle mmc_add_host() errors
The BCM2835 MMC host driver calls mmc_add_host() but doesn't check its
return value. Errors occurring in that function are therefore not
handled. Fix it.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Frank Pavlic <f.pavlic@kunbus.de>
bcm2835-mmc: Deduplicate reset of driver data on remove
The BCM2835 MMC host driver sets the device's driver data pointer to
NULL on ->remove() even though the driver core subsequently does the
same in __device_release_driver(). Drop the duplicate assignment.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Frank Pavlic <f.pavlic@kunbus.de>
bcm2835_mmc: Remove vestigial threaded IRQ
With SDIO processing now managed by the MMC framework with a
workqueue, the bcm2835_mmc driver no longer needs a threaded
IRQ.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Add missing dma_unmap_sg calls to free relevant swiotlb bounce buffers.
This prevents DMA leaks.
Signed-off-by: Yaroslav Rosomakho <yaroslavros@gmail.com>
Limit max_req_size under arm64 (or any other platform that uses swiotlb) to prevent potential buffer overflow due to bouncing.
Signed-off-by: Yaroslav Rosomakho <yaroslavros@gmail.com>
Add support for DMA controller of BCM2708 as used in the Raspberry Pi.
Currently it only supports cyclic DMA.
Signed-off-by: Florian Meier <florian.meier@koalo.de>
dmaengine: expand functionality by supporting scatter/gather transfers sdhci-bcm2708 and dma.c: fix for LITE channels
DMA: fix cyclic LITE length overflow bug
dmaengine: bcm2708: Remove chancnt affectations
Mirror bcm2835-dma.c commit 9eba5536a7:
chancnt is already filled by dma_async_device_register, which uses the channel
list to know how much channels there is.
Since it's already filled, we can safely remove it from the drivers' probe
function.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dmaengine: bcm2708: overwrite dreq only if it is not set
dreq is set when the DMA channel is fetched from Device Tree.
slave_id is set using dmaengine_slave_config().
Only overwrite dreq with slave_id if it is not set.
dreq/slave_id in the cyclic DMA case is not touched, because I don't
have hardware to test with.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dmaengine: bcm2708: do device registration in the board file
Don't register the device in the driver. Do it in the board file.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dmaengine: bcm2708: don't restrict DT support to ARCH_BCM2835
Both ARCH_BCM2835 and ARCH_BCM270x are built with OF now.
Add Device Tree support to the non ARCH_BCM2835 case.
Use the same driver name regardless of architecture.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
BCM270x_DT: add bcm2835-dma entry
Add Device Tree entry for bcm2835-dma.
The entry doesn't contain any resources since they are handled
by the arch/arm/mach-bcm270x/dma.c driver.
In non-DT mode, don't add the device in the board file.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
bcm2708-dmaengine: Add debug options
BCM270x: Add memory and irq resources to dmaengine device and DT
Prepare for merging of the legacy DMA API arch driver dma.c
with bcm2708-dmaengine by adding memory and irq resources both
to platform file device and Device Tree node.
Don't use BCM_DMAMAN_DRIVER_NAME so we don't have to include mach/dma.h
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dmaengine: bcm2708: Merge with arch dma.c driver and disable dma.c
Merge the legacy DMA API driver with bcm2708-dmaengine.
This is done so we can use bcm2708_fb on ARCH_BCM2835 (mailbox
driver is also needed).
Changes to the dma.c code:
- Use BIT() macro.
- Cutdown some comments to one line.
- Add mutex to vc_dmaman and use this, since the dev lock is locked
during probing of the engine part.
- Add global g_dmaman variable since drvdata is used by the engine part.
- Restructure for readability:
vc_dmaman_chan_alloc()
vc_dmaman_chan_free()
bcm_dma_chan_free()
- Restructure bcm_dma_chan_alloc() to simplify error handling.
- Use device irq resources instead of hardcoded bcm_dma_irqs table.
- Remove dev_dmaman_register() and code it directly.
- Remove dev_dmaman_deregister() and code it directly.
- Simplify bcm_dmaman_probe() using devm_* functions.
- Get dmachans from DT if available.
- Keep 'dma.dmachans' module argument name for backwards compatibility.
Make it available on ARCH_BCM2835 as well.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dmaengine: bcm2708: set residue_granularity field
bcm2708-dmaengine supports residue reporting at burst level
but didn't report this via the residue_granularity field.
Without this field set properly we get playback issues with I2S cards.
dmaengine: bcm2708-dmaengine: Fix memory leak when stopping a running transfer
bcm2708-dmaengine: Use more DMA channels (but not 12)
1) Only the bcm2708_fb drivers uses the legacy DMA API, and
it requires a BULK-capable channel, so all other types
(FAST, NORMAL and LITE) can be made available to the regular
DMA API.
2) DMA channels 11-14 share an interrupt. The driver can't
handle this, so don't use channels 12-14 (12 was used, probably
because it appears to have an interrupt, but in reality that
interrupt is for activity on ANY channel). This may explain
a lockup encountered when running out of DMA channels.
The combined effect of this patch is to leave 7 DMA channels
available + channel 0 for bcm2708_fb via the legacy API.
See: https://github.com/raspberrypi/linux/issues/1110https://github.com/raspberrypi/linux/issues/1108
dmaengine: bcm2708: Make legacy API available for bcm2835-dma
bcm2708_fb uses the legacy DMA API, so in order to start using
bcm2835-dma, bcm2835-dma has to support the legacy API. Make this
possible by exporting bcm_dmaman_probe() and bcm_dmaman_remove().
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dmaengine: bcm2708: Change DT compatible string
Both bcm2835-dma and bcm2708-dmaengine have the same compatible string.
So change compatible to "brcm,bcm2708-dma".
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dmaengine: bcm2708: Remove driver but keep legacy API
Dropping non-DT support means we don't need this driver,
but we still need the legacy DMA API.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
bcm2708-dmaengine - Fix arm64 portability/build issues
dma-bcm2708: Fix module compilation of CONFIG_DMA_BCM2708
bcm2708-dmaengine.c defines functions like bcm_dma_start which are
defined as well in dma-bcm2708.h as inline versions when
CONFIG_DMA_BCM2708 is not defined. This works fine when
CONFIG_DMA_BCM2708 is built in, but when it is selected as module build
fails with redefinition errors because in the build system when
CONFIG_DMA_BCM2708 is selected as module, the macro becomes
CONFIG_DMA_BCM2708_MODULE.
This patch makes the header use CONFIG_DMA_BCM2708_MODULE too when
available.
Fixes https://github.com/raspberrypi/linux/issues/2056
Signed-off-by: Andrei Gherzan <andrei@gherzan.com>
Especially on platforms with a slower CPU but a relatively high
framebuffer fill bandwidth, like current ARM devices, the existing
console monochrome imageblit function used to draw console text is
suboptimal for common pixel depths such as 16bpp and 32bpp. The existing
code is quite general and can deal with several pixel depths. By creating
special case functions for 16bpp and 32bpp, by far the most common pixel
formats used on modern systems, a significant speed-up is attained
which can be readily felt on ARM-based devices like the Raspberry Pi
and the Allwinner platform, but should help any platform using the
fb layer.
The special case functions allow constant folding, eliminating a number
of instructions including divide operations, and allow the use of an
unrolled loop, eliminating instructions with a variable shift size,
reducing source memory access instructions, and eliminating excessive
branching. These unrolled loops also allow much better code optimization
by the C compiler. The code that selects which optimized variant is used
is also simplified, eliminating integer divide instructions.
The speed-up, measured by timing 'cat file.txt' in the console, varies
between 40% and 70%, when testing on the Raspberry Pi and Allwinner
ARM-based platforms, depending on font size and the pixel depth, with
the greater benefit for 32bpp.
Signed-off-by: Harm Hanemaaijer <fgenfb@yahoo.com>
Based on the patch authored by Ali Gholami Rudi at
https://lkml.org/lkml/2009/7/13/153
Provide an ioctl for userspace applications, but only if this operation
is hardware accelerated (otherwide it does not make any sense).
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
bcm2708_fb: Add ioctl for reading gpu memory through dma
video: bcm2708_fb: Add compat_ioctl support.
When using a 64 bit kernel with 32 bit userspace we need
compat ioctl handling for FBIODMACOPY as one of the
parameters is a pointer.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: popcornmix <popcornmix@gmail.com>
bcm2708_fb : Implement blanking support using the mailbox property interface
bcm2708_fb: Add pan and vsync controls
bcm2708_fb: DMA acceleration for fb_copyarea
Based on http://www.raspberrypi.org/phpBB3/viewtopic.php?p=62425#p62425
Also used Simon's dmaer_master module as a reference for tweaking DMA
settings for better performance.
For now busylooping only. IRQ support might be added later.
With non-overclocked Raspberry Pi, the performance is ~360 MB/s
for simple copy or ~260 MB/s for two-pass copy (used when dragging
windows to the right).
In the case of using DMA channel 0, the performance improves
to ~440 MB/s.
For comparison, VFP optimized CPU copy can only do ~114 MB/s in
the same conditions (hindered by reading uncached source buffer).
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
bcm2708_fb: report number of dma copies
Add a counter (exported via debugfs) reporting the
number of dma copies that the framebuffer driver
has done, in order to help evaluate different
optimization strategies.
Signed-off-by: Luke Diamand <luked@broadcom.com>
bcm2708_fb: use IRQ for DMA copies
The copyarea ioctl() uses DMA to speed things along. This
was busy-waiting for completion. This change supports using
an interrupt instead for larger transfers. For small
transfers, busy-waiting is still likely to be faster.
Signed-off-by: Luke Diamand <luke@diamand.org>
bcm2708: Make ioctl logging quieter
video: fbdev: bcm2708_fb: Don't panic on error
No need to panic the kernel if the video driver fails.
Just print a message and return an error.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
fbdev: bcm2708_fb: Add ARCH_BCM2835 support
Add Device Tree support.
Pass the device to dma_alloc_coherent() in order to get the
correct bus address on ARCH_BCM2835.
Use the new DMA legacy API header file.
Including <mach/platform.h> is not necessary.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
BCM270x_DT: Add bcm2708-fb device
Add bcm2708-fb to Device Tree and don't add the
platform device when booting in DT mode.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Cleanup of bcm2708_fb file to kernel coding standards
Some minor change to function - remove a use of
in_atomic, plus replacing various debug messages
that manually specify the function name with
("%s",.__func__)
Signed-off-by: James Hughes <james.hughes@raspberrypi.org>
video: bcm2708_fb: Try allocating on the ARM and passing to VPU
Currently the VPU allocates the contiguous buffer for the
framebuffer.
Try an alternate path first where we use dma_alloc_coherent
and pass the buffer to the VPU. Should the VPU firmware not
support that path, then free the buffer and revert to the
old behaviour of using the VPU allocation.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: popcornmix <popcornmix@gmail.com>
usb: dwc: fix lockdep false positive
Signed-off-by: Kari Suvanto <karis79@gmail.com>
usb: dwc: fix inconsistent lock state
Signed-off-by: Kari Suvanto <karis79@gmail.com>
Add FIQ patch to dwc_otg driver. Enable with dwc_otg.fiq_fix_enable=1. Should give about 10% more ARM performance.
Thanks to Gordon and Costas
Avoid dynamic memory allocation for channel lock in USB driver. Thanks ddv2005.
Add NAK holdoff scheme. Enabled by default, disable with dwc_otg.nak_holdoff_enable=0. Thanks gsh
Make sure we wait for the reset to finish
dwc_otg: fix bug in dwc_otg_hcd.c resulting in silent kernel
memory corruption, escalating to OOPS under high USB load.
dwc_otg: Fix unsafe access of QTD during URB enqueue
In dwc_otg_hcd_urb_enqueue during qtd creation, it was possible that the
transaction could complete almost immediately after the qtd was assigned
to a host channel during URB enqueue, which meant the qtd pointer was no
longer valid having been completed and removed. Usually, this resulted in
an OOPS during URB submission. By predetermining whether transactions
need to be queued or not, this unsafe pointer access is avoided.
This bug was only evident on the Pi model A where a device was attached
that had no periodic endpoints (e.g. USB pendrive or some wlan devices).
dwc_otg: Fix incorrect URB allocation error handling
If the memory allocation for a dwc_otg_urb failed, the kernel would OOPS
because for some reason a member of the *unallocated* struct was set to
zero. Error handling changed to fail correctly.
dwc_otg: fix potential use-after-free case in interrupt handler
If a transaction had previously aborted, certain interrupts are
enabled to track error counts and reset where necessary. On IN
endpoints the host generates an ACK interrupt near-simultaneously
with completion of transfer. In the case where this transfer had
previously had an error, this results in a use-after-free on
the QTD memory space with a 1-byte length being overwritten to
0x00.
dwc_otg: add handling of SPLIT transaction data toggle errors
Previously a data toggle error on packets from a USB1.1 device behind
a TT would result in the Pi locking up as the driver never handled
the associated interrupt. Patch adds basic retry mechanism and
interrupt acknowledgement to cater for either a chance toggle error or
for devices that have a broken initial toggle state (FT8U232/FT232BM).
dwc_otg: implement tasklet for returning URBs to usbcore hcd layer
The dwc_otg driver interrupt handler for transfer completion will spend
a very long time with interrupts disabled when a URB is completed -
this is because usb_hcd_giveback_urb is called from within the handler
which for a USB device driver with complicated processing (e.g. webcam)
will take an exorbitant amount of time to complete. This results in
missed completion interrupts for other USB packets which lead to them
being dropped due to microframe overruns.
This patch splits returning the URB to the usb hcd layer into a
high-priority tasklet. This will have most benefit for isochronous IN
transfers but will also have incidental benefit where multiple periodic
devices are active at once.
dwc_otg: fix NAK holdoff and allow on split transactions only
This corrects a bug where if a single active non-periodic endpoint
had at least one transaction in its qh, on frnum == MAX_FRNUM the qh
would get skipped and never get queued again. This would result in
a silent device until error detection (automatic or otherwise) would
either reset the device or flush and requeue the URBs.
Additionally the NAK holdoff was enabled for all transactions - this
would potentially stall a HS endpoint for 1ms if a previous error state
enabled this interrupt and the next response was a NAK. Fix so that
only split transactions get held off.
dwc_otg: Call usb_hcd_unlink_urb_from_ep with lock held in completion handler
usb_hcd_unlink_urb_from_ep must be called with the HCD lock held. Calling it
asynchronously in the tasklet was not safe (regression in
c4564d4a1a).
This change unlinks it from the endpoint prior to queueing it for handling in
the tasklet, and also adds a check to ensure the urb is OK to be unlinked
before doing so.
NULL pointer dereference kernel oopses had been observed in usb_hcd_giveback_urb
when a USB device was unplugged/replugged during data transfer. This effect
was reproduced using automated USB port power control, hundreds of replug
events were performed during active transfers to confirm that the problem was
eliminated.
USB fix using a FIQ to implement split transactions
This commit adds a FIQ implementaion that schedules
the split transactions using a FIQ so we don't get
held off by the interrupt latency of Linux
dwc_otg: fix device attributes and avoid kernel warnings on boot
dcw_otg: avoid logging function that can cause panics
See: https://github.com/raspberrypi/firmware/issues/21
Thanks to cleverca22 for fix
dwc_otg: mask correct interrupts after transaction error recovery
The dwc_otg driver will unmask certain interrupts on a transaction
that previously halted in the error state in order to reset the
QTD error count. The various fine-grained interrupt handlers do not
consider that other interrupts besides themselves were unmasked.
By disabling the two other interrupts only ever enabled in DMA mode
for this purpose, we can avoid unnecessary function calls in the
IRQ handler. This will also prevent an unneccesary FIQ interrupt
from being generated if the FIQ is enabled.
dwc_otg: fiq: prevent FIQ thrash and incorrect state passing to IRQ
In the case of a transaction to a device that had previously aborted
due to an error, several interrupts are enabled to reset the error
count when a device responds. This has the side-effect of making the
FIQ thrash because the hardware will generate multiple instances of
a NAK on an IN bulk/interrupt endpoint and multiple instances of ACK
on an OUT bulk/interrupt endpoint. Make the FIQ mask and clear the
associated interrupts.
Additionally, on non-split transactions make sure that only unmasked
interrupts are cleared. This caused a hard-to-trigger but serious
race condition when you had the combination of an endpoint awaiting
error recovery and a transaction completed on an endpoint - due to
the sequencing and timing of interrupts generated by the dwc_otg core,
it was possible to confuse the IRQ handler.
Fix function tracing
dwc_otg: whitespace cleanup in dwc_otg_urb_enqueue
dwc_otg: prevent OOPSes during device disconnects
The dwc_otg_urb_enqueue function is thread-unsafe. In particular the
access of urb->hcpriv, usb_hcd_link_urb_to_ep, dwc_otg_urb->qtd and
friends does not occur within a critical section and so if a device
was unplugged during activity there was a high chance that the
usbcore hub_thread would try to disable the endpoint with partially-
formed entries in the URB queue. This would result in BUG() or null
pointer dereferences.
Fix so that access of urb->hcpriv, enqueuing to the hardware and
adding to usbcore endpoint URB lists is contained within a single
critical section.
dwc_otg: prevent BUG() in TT allocation if hub address is > 16
A fixed-size array is used to track TT allocation. This was
previously set to 16 which caused a crash because
dwc_otg_hcd_allocate_port would read past the end of the array.
This was hit if a hub was plugged in which enumerated as addr > 16,
due to previous device resets or unplugs.
Also add #ifdef FIQ_DEBUG around hcd->hub_port_alloc[], which grows
to a large size if 128 hub addresses are supported. This field is
for debug only for tracking which frame an allocate happened in.
dwc_otg: make channel halts with unknown state less damaging
If the IRQ received a channel halt interrupt through the FIQ
with no other bits set, the IRQ would not release the host
channel and never complete the URB.
Add catchall handling to treat as a transaction error and retry.
dwc_otg: fiq_split: use TTs with more granularity
This fixes certain issues with split transaction scheduling.
- Isochronous multi-packet OUT transactions now hog the TT until
they are completed - this prevents hubs aborting transactions
if they get a periodic start-split out-of-order
- Don't perform TT allocation on non-periodic endpoints - this
allows simultaneous use of the TT's bulk/control and periodic
transaction buffers
This commit will mainly affect USB audio playback.
dwc_otg: fix potential sleep while atomic during urb enqueue
Fixes a regression introduced with eb1b482a. Kmalloc called from
dwc_otg_hcd_qtd_add / dwc_otg_hcd_qtd_create did not always have
the GPF_ATOMIC flag set. Force this flag when inside the larger
critical section.
dwc_otg: make fiq_split_enable imply fiq_fix_enable
Failing to set up the FIQ correctly would result in
"IRQ 32: nobody cared" errors in dmesg.
dwc_otg: prevent crashes on host port disconnects
Fix several issues resulting in crashes or inconsistent state
if a Model A root port was disconnected.
- Clean up queue heads properly in kill_urbs_in_qh_list by
removing the empty QHs from the schedule lists
- Set the halt status properly to prevent IRQ handlers from
using freed memory
- Add fiq_split related cleanup for saved registers
- Make microframe scheduling reclaim host channels if
active during a disconnect
- Abort URBs with -ESHUTDOWN status response, informing
device drivers so they respond in a more correct fashion
and don't try to resubmit URBs
- Prevent IRQ handlers from attempting to handle channel
interrupts if the associated URB was dequeued (and the
driver state was cleared)
dwc_otg: prevent leaking URBs during enqueue
A dwc_otg_urb would get leaked if the HCD enqueue function
failed for any reason. Free the URB at the appropriate points.
dwc_otg: Enable NAK holdoff for control split transactions
Certain low-speed devices take a very long time to complete a
data or status stage of a control transaction, producing NAK
responses until they complete internal processing - the USB2.0
spec limit is up to 500mS. This causes the same type of interrupt
storm as seen with USB-serial dongles prior to c8edb238.
In certain circumstances, usually while booting, this interrupt
storm could cause SD card timeouts.
dwc_otg: Fix for occasional lockup on boot when doing a USB reset
dwc_otg: Don't issue traffic to LS devices in FS mode
Issuing low-speed packets when the root port is in full-speed mode
causes the root port to stop responding. Explicitly fail when
enqueuing URBs to a LS endpoint on a FS bus.
Fix ARM architecture issue with local_irq_restore()
If local_fiq_enable() is called before a local_irq_restore(flags) where
the flags variable has the F bit set, the FIQ will be erroneously disabled.
Fixup arch_local_irq_restore to avoid trampling the F bit in CPSR.
Also fix some of the hacks previously implemented for previous dwc_otg
incarnations.
dwc_otg: fiq_fsm: Base commit for driver rewrite
This commit removes the previous FIQ fixes entirely and adds fiq_fsm.
This rewrite features much more complete support for split transactions
and takes into account several OTG hardware bugs. High-speed
isochronous transactions are also capable of being performed by fiq_fsm.
All driver options have been removed and replaced with:
- dwc_otg.fiq_enable (bool)
- dwc_otg.fiq_fsm_enable (bool)
- dwc_otg.fiq_fsm_mask (bitmask)
- dwc_otg.nak_holdoff (unsigned int)
Defaults are specified such that fiq_fsm behaves similarly to the
previously implemented FIQ fixes.
fiq_fsm: Push error recovery into the FIQ when fiq_fsm is used
If the transfer associated with a QTD failed due to a bus error, the HCD
would retry the transfer up to 3 times (implementing the USB2.0
three-strikes retry in software).
Due to the masking mechanism used by fiq_fsm, it is only possible to pass
a single interrupt through to the HCD per-transfer.
In this instance host channels would fall off the radar because the error
reset would function, but the subsequent channel halt would be lost.
Push the error count reset into the FIQ handler.
fiq_fsm: Implement timeout mechanism
For full-speed endpoints with a large packet size, interrupt latency
runs the risk of the FIQ starting a transaction too late in a full-speed
frame. If the device is still transmitting data when EOF2 for the
downstream frame occurs, the hub will disable the port. This change is
not reflected in the hub status endpoint and the device becomes
unresponsive.
Prevent high-bandwidth transactions from being started too late in a
frame. The mechanism is not guaranteed: a combination of bit stuffing
and hub latency may still result in a device overrunning.
fiq_fsm: fix bounce buffer utilisation for Isochronous OUT
Multi-packet isochronous OUT transactions were subject to a few bounday
bugs. Fix them.
Audio playback is now much more robust: however, an issue stands with
devices that have adaptive sinks - ALSA plays samples too fast.
dwc_otg: Return full-speed frame numbers in HS mode
The frame counter increments on every *microframe* in high-speed mode.
Most device drivers expect this number to be in full-speed frames - this
caused considerable confusion to e.g. snd_usb_audio which uses the
frame counter to estimate the number of samples played.
fiq_fsm: save PID on completion of interrupt OUT transfers
Also add edge case handling for interrupt transports.
Note that for periodic split IN, data toggles are unimplemented in the
OTG host hardware - it unconditionally accepts any PID.
fiq_fsm: add missing case for fiq_fsm_tt_in_use()
Certain combinations of bitrate and endpoint activity could
result in a periodic transaction erroneously getting started
while the previous Isochronous OUT was still active.
fiq_fsm: clear hcintmsk for aborted transactions
Prevents the FIQ from erroneously handling interrupts
on a timed out channel.
fiq_fsm: enable by default
fiq_fsm: fix dequeues for non-periodic split transactions
If a dequeue happened between the SSPLIT and CSPLIT phases of the
transaction, the HCD would never receive an interrupt.
fiq_fsm: Disable by default
fiq_fsm: Handle HC babble errors
The HCTSIZ transfer size field raises a babble interrupt if
the counter wraps. Handle the resulting interrupt in this case.
dwc_otg: fix interrupt registration for fiq_enable=0
Additionally make the module parameter conditional for wherever
hcd->fiq_state is touched.
fiq_fsm: Enable by default
dwc_otg: Fix various issues with root port and transaction errors
Process the host port interrupts correctly (and don't trample them).
Root port hotplug now functional again.
Fix a few thinkos with the transaction error passthrough for fiq_fsm.
fiq_fsm: Implement hack for Split Interrupt transactions
Hubs aren't too picky about which endpoint we send Control type split
transactions to. By treating Interrupt transfers as Control, it is
possible to use the non-periodic queue in the OTG core as well as the
non-periodic FIFOs in the hub itself. This massively reduces the
microframe exclusivity/contention that periodic split transactions
otherwise have to enforce.
It goes without saying that this is a fairly egregious USB specification
violation, but it works.
Original idea by Hans Petter Selasky @ FreeBSD.org.
dwc_otg: FIQ support on SMP. Set up FIQ stack and handler on Core 0 only.
dwc_otg: introduce fiq_fsm_spin(un|)lock()
SMP safety for the FIQ relies on register read-modify write cycles being
completed in the correct order. Several places in the DWC code modify
registers also touched by the FIQ. Protect these by a bare-bones lock
mechanism.
This also makes it possible to run the FIQ and IRQ handlers on different
cores.
fiq_fsm: fix build on bcm2708 and bcm2709 platforms
dwc_otg: put some barriers back where they should be for UP
bcm2709/dwc_otg: Setup FIQ on core 1 if >1 core active
dwc_otg: fixup read-modify-write in critical paths
Be more careful about read-modify-write on registers that the FIQ
also touches.
Guard fiq_fsm_spin_lock with fiq_enable check
fiq_fsm: Falling out of the state machine isn't fatal
This edge case can be hit if the port is disabled while the FIQ is
in the middle of a transaction. Make the effects less severe.
Also get rid of the useless return value.
squash: dwc_otg: Allow to build without SMP
usb: core: make overcurrent messages more prominent
Hub overcurrent messages are more serious than "debug". Increase loglevel.
usb: dwc_otg: Don't use dma_to_virt()
Commit 6ce0d20 changes dma_to_virt() which breaks this driver.
Open code the old dma_to_virt() implementation to work around this.
Limit the use of __bus_to_virt() to cases where transfer_buffer_length
is set and transfer_buffer is not set. This is done to increase the
chance that this driver will also work on ARCH_BCM2835.
transfer_buffer should not be NULL if the length is set, but the
comment in the code indicates that there are situations where this
might happen. drivers/usb/isp1760/isp1760-hcd.c also has a similar
comment pointing to a possible: 'usb storage / SCSI bug'.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dwc_otg: Fix crash when fiq_enable=0
dwc_otg: fiq_fsm: Make high-speed isochronous strided transfers work properly
Certain low-bandwidth high-speed USB devices (specialist audio devices,
compressed-frame webcams) have packet intervals > 1 microframe.
Stride these transfers in the FIQ by using the start-of-frame interrupt
to restart the channel at the right time.
dwc_otg: Force host mode to fix incorrect compute module boards
dwc_otg: Add ARCH_BCM2835 support
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dwc_otg: Simplify FIQ irq number code
Dropping ATAGS means we can simplify the FIQ irq number code.
Also add error checking on the returned irq number.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dwc_otg: Remove duplicate gadget probe/unregister function
dwc_otg: Properly set the HFIR
Douglas Anderson reported:
According to the most up to date version of the dwc2 databook, the FRINT
field of the HFIR register should be programmed to:
* 125 us * (PHY clock freq for HS) - 1
* 1000 us * (PHY clock freq for FS/LS) - 1
This is opposed to older versions of the doc that claimed it should be:
* 125 us * (PHY clock freq for HS)
* 1000 us * (PHY clock freq for FS/LS)
and reported lower timing jitter on a USB analyser
dcw_otg: trim xfer length when buffer larger than allocated size is received
dwc_otg: Don't free qh align buffers in atomic context
dwc_otg: Enable the hack for Split Interrupt transactions by default
dwc_otg.fiq_fsm_mask=0xF has long been a suggestion for users with audio stutters or other USB bandwidth issues.
So far we are aware of many success stories but no failure caused by this setting.
Make it a default to learn more.
See: https://www.raspberrypi.org/forums/viewtopic.php?f=28&t=70437
Signed-off-by: popcornmix <popcornmix@gmail.com>
dwc_otg: Use kzalloc when suitable
dwc_otg: Pass struct device to dma_alloc*()
This makes it possible to get the bus address from Device Tree.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
dwc_otg: fix summarize urb->actual_length for isochronous transfers
Kernel does not copy input data of ISO transfers to userspace
if actual_length is set only in ISO transfers and not summarized
in urb->actual_length. Fixesraspberrypi/linux#903
fiq_fsm: Use correct states when starting isoc OUT transfers
In fiq_fsm_start_next_periodic() if an isochronous OUT transfer
was selected, no regard was given as to whether this was a single-packet
transfer or a multi-packet staged transfer.
For single-packet transfers, this had the effect of repeatedly sending
OUT packets with bogus data and lengths.
Eventually if the channel was repeatedly enabled enough times, this
would lock up the OTG core and no further bus transfers would happen.
Set the FSM state up properly if we select a single-packet transfer.
Fixes https://github.com/raspberrypi/linux/issues/1842
dwc_otg: make nak_holdoff work as intended with empty queues
If URBs reading from non-periodic split endpoints were dequeued and
the last transfer from the endpoint was a NAK handshake, the resulting
qh->nak_frame value was stale which would result in unnecessarily long
polling intervals for the first subsequent transfer with a fresh URB.
Fixup qh->nak_frame in dwc_otg_hcd_urb_dequeue and also guard against
a case where a single URB is submitted to the endpoint, a NAK was
received on the transfer immediately prior to receiving data and the
device subsequently resubmits another URB past the qh->nak_frame interval.
Fixes https://github.com/raspberrypi/linux/issues/1709
dwc_otg: fix split transaction data toggle handling around dequeues
See https://github.com/raspberrypi/linux/issues/1709
Fix several issues regarding endpoint state when URBs are dequeued
- If the HCD is disconnected, flush FIQ-enabled channels properly
- Save the data toggle state for bulk endpoints if the last transfer
from an endpoint where URBs were dequeued returned a data packet
- Reset hc->start_pkt_count properly in assign_and_init_hc()
dwc_otg: fix several potential crash sources
On root port disconnect events, the host driver state is cleared and
in-progress host channels are forcibly stopped. This doesn't play
well with the FIQ running in the background, so:
- Guard the disconnect callback with both the host spinlock and FIQ
spinlock
- Move qtd dereference in dwc_otg_handle_hc_fsm() after the early-out
so we don't dereference a qtd that has gone away
- Turn catch-all BUG()s in dwc_otg_handle_hc_fsm() into warnings.
dwc_otg: delete hcd->channel_lock
The lock serves no purpose as it is only held while the HCD spinlock
is already being held.
dwc_otg: remove unnecessary dma-mode channel halts on disconnect interrupt
Host channels are already halted in kill_urbs_in_qh_list() with the
subsequent interrupt processing behaving as if the URB was dequeued
via HCD callback.
There's no need to clobber the host channel registers a second time
as this exposes races between the driver and host channel resulting
in hcd->free_hc_list becoming corrupted.
dwcotg: Allow to build without FIQ on ARM64
Signed-off-by: popcornmix <popcornmix@gmail.com>
dwc_otg: make periodic scheduling behave properly for FS buses
If the root port is in full-speed mode, transfer times at 12mbit/s
would be calculated but matched against high-speed quotas.
Reinitialise hcd->frame_usecs[i] on each port enable event so that
full-speed bandwidth can be tracked sensibly.
Also, don't bother using the FIQ for transfers when in full-speed
mode - at the slower bus speed, interrupt frequency is reduced by
an order of magnitude.
Related issue: https://github.com/raspberrypi/linux/issues/2020
dwc_otg: fiq_fsm: Make isochronous compatibility checks work properly
Get rid of the spammy printk and local pointer mangling.
Also, there is a nominal benefit for using fiq_fsm for isochronous
transfers in FS mode (~1.1k IRQs per second vs 2.1k IRQs per second)
so remove the root port speed check.
dwc_otg: add module parameter int_ep_interval_min
Add a module parameter (defaulting to ignored) that clamps the polling rate
of high-speed Interrupt endpoints to a minimum microframe interval.
The parameter is modifiable at runtime as it is used when activating new
endpoints (such as on device connect).
dwc_otg: fiq_fsm: Add non-periodic TT exclusivity constraints
Certain hub types do not discriminate between pipe direction (IN or OUT)
when considering non-periodic transfers. Therefore these hubs get confused
if multiple transfers are issued in different directions with the same
device address and endpoint number.
Constrain queuing non-periodic split transactions so they are performed
serially in such cases.
Related: https://github.com/raspberrypi/linux/issues/2024
dwc_otg: Fixup change to DRIVER_ATTR interface
dwc_otg: Fix compilation warnings
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
USB_DWCOTG: Disable building dwc_otg as a module (#2265)
When dwc_otg is built as a module, build will fail with the following
error:
ERROR: "DWC_TASK_HI_SCHEDULE" [drivers/usb/host/dwc_otg/dwc_otg.ko] undefined!
scripts/Makefile.modpost:91: recipe for target '__modpost' failed
make[1]: *** [__modpost] Error 1
Makefile:1199: recipe for target 'modules' failed
make: *** [modules] Error 2
Even if the error is solved by including the missing
DWC_TASK_HI_SCHEDULE function, the kernel will panic when loading
dwc_otg.
As a workaround, simply prevent user from building dwc_otg as a module
as the current kernel does not support it.
See: https://github.com/raspberrypi/linux/issues/2258
Signed-off-by: Malik Olivier Boussejra <malik@boussejra.com>
dwc_otg: New timer API
dwc_otg: Fix removed ACCESS_ONCE->READ_ONCE
dwc_otg: don't unconditionally force host mode in dwc_otg_cil_init()
Add the ability to disable force_host_mode for those that want to use
dwc_otg in both device and host modes.
dwc_otg: Fix a regression when dequeueing isochronous transfers
In 282bed95 (dwc_otg: make nak_holdoff work as intended with empty queues)
the dequeue mechanism was changed to leave FIQ-enabled transfers to run
to completion - to avoid leaving hub TT buffers with stale packets lying
around.
This broke FIQ-accelerated isochronous transfers, as this then meant that
dozens of transfers were performed after the dequeue function returned.
Restore the state machine fence for isochronous transfers.
fiq_fsm: rewind DMA pointer for OUT transactions that fail (#2288)
See: https://github.com/raspberrypi/linux/issues/2140
dwc_otg: add smp_mb() to prevent driver state corruption on boot
Occasional crashes have been seen where the FIQ code dereferences
invalid/random pointers immediately after being set up, leading to
panic on boot.
The crash occurs as the FIQ code races against hcd_init_fiq() and
the hcd_init_fiq() code races against the outstanding memory stores
from dwc_otg_hcd_init(). Use explicit barriers after touching
driver state.
usb: dwc_otg: fix memory corruption in dwc_otg driver
[Upstream commit 51b1b64917]
The move from the staging tree to the main tree exposed a
longstanding memory corruption bug in the dwc2 driver. The
reordering of the driver initialization caused the dwc2 driver
to corrupt the initialization data of the sdhci driver on the
Raspberry Pi platform, which made the bug show up.
The error is in calling to_usb_device(hsotg->dev), since ->dev
is not a member of struct usb_device. The easiest fix is to
just remove the offending code, since it is not really needed.
Thanks to Stephen Warren for tracking down the cause of this.
Reported-by: Andre Heider <a.heider@gmail.com>
Tested-by: Stephen Warren <swarren@wwwdotorg.org>
Signed-off-by: Paul Zimmerman <paulz@synopsys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[lukas: port from upstream dwc2 to out-of-tree dwc_otg driver]
Signed-off-by: Lukas Wunner <lukas@wunner.de>
usb: dwb_otg: Fix unreachable switch statement warning
This warning appears with GCC 7.3.0 from toolchains.bootlin.com:
../drivers/usb/host/dwc_otg/dwc_otg_fiq_fsm.c: In function ‘fiq_fsm_update_hs_isoc’:
../drivers/usb/host/dwc_otg/dwc_otg_fiq_fsm.c:595:61: warning: statement will never be executed [-Wswitch-unreachable]
st->hctsiz_copy.b.xfersize = nrpackets * st->hcchar_copy.b.mps;
~~~~~~~~~~~~~~~~~^~~~
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
dwc_otg: fiq_fsm: fix incorrect DMA register offset calculation
Rationalise the offset and update all call sites.
Fixes https://github.com/raspberrypi/linux/issues/2408
dwc_otg: fix bug with port_addr assignment for single-TT hubs
See https://github.com/raspberrypi/linux/issues/2734
The "Hub Port" field in the split transaction packet was always set
to 1 for single-TT hubs. The majority of single-TT hub products
apparently ignore this field and broadcast to all downstream enabled
ports, which masked the issue. A subset of hub devices apparently
need the port number to be exact or split transactions will fail.
usb: dwc_otg: Clean up build warnings on 64bit kernels
No functional changes. Almost all are changes to logging lines.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
usb: dwc_otg: Use dma allocation for mphi dummy_send buffer
The FIQ driver used a kzalloc'ed buffer for dummy_send,
passing a kernel virtual address to the hardware block.
The buffer is only ever used for a dummy read, so it
should be harmless, but there is the chance that it will
cause exceptions.
Use a dma allocation so that we have a genuine bus address,
and read from that.
Free the allocation when done for good measure.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
dwc_otg: only do_split when we actually need to do a split
The previous test would fail if the root port was in fullspeed mode
and there was a hub between the FS device and the root port. While
the transfer worked, the schedule mangling performed for high-speed
split transfers would break leading to an 8ms polling interval.
dwc_otg: fix locking around dequeueing and killing URBs
kill_urbs_in_qh_list() is practically only ever called with the fiq lock
already held, so don't spinlock twice in the case where we need to cancel
an isochronous transfer.
Also fix up a case where the global interrupt register could be read with
the fiq lock not held.
Fixes the deadlock seen in https://github.com/raspberrypi/linux/issues/2907
ARM64/DWC_OTG: Port dwc_otg driver to ARM64
In ARM64, the FIQ mechanism used by this driver is not current
implemented. As a workaround, reqular IRQ is used instead
of FIQ.
In a separate change, the IRQ-CPU mapping is round robined
on ARM64 to increase concurrency and allow multiple interrupts
to be serviced at a time. This reduces the need for FIQ.
Tests Run:
This mechanism is most likely to break when multiple USB devices
are attached at the same time. So the system was tested under
stress.
Devices:
1. USB Speakers playing back a FLAC audio through VLC
at 96KHz.(Higher then typically, but supported on my speakers).
2. sftp transferring large files through the buildin ethernet
connection which is connected through USB.
3. Keyboard and mouse attached and being used.
Although I do occasionally hear some glitches, the music seems to
play quite well.
Signed-off-by: Michael Zoran <mzoran@crowfest.net>
usb: dwc_otg: Clean up interrupt claiming code
The FIQ/IRQ interrupt number identification code is scattered through
the dwc_otg driver. Rationalise it, simplifying the code and solving
an existing issue.
See: https://github.com/raspberrypi/linux/issues/2612
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
dwc_otg: Choose appropriate IRQ handover strategy
2711 has no MPHI peripheral, but the ARM Control block can fake
interrupts. Use the size of the DTB "mphi" reg block to determine
which is required.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
usb: host: dwc_otg: fix compiling in separate directory
The dwc_otg Makefile does not respect the O=path argument correctly:
include paths in CFLAGS are given relatively to object path, not source
path. Compiling in a separate directory yields #include errors.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
dwc_otg: use align_buf for small IN control transfers (#3150)
The hardware will do a 4-byte write to memory on any IN packet received
that is between 1 and 3 bytes long. This tramples memory in the uvcvideo
driver, as it uses a sequence of 1- and 2-byte control transfers to
query the min/max/range/step of each individual camera control and
gives us buffers that are offsets into a struct.
Catch small control transfers in the data phase and use the align_buf
to bounce the correct number of bytes into the URB's buffer.
In general, short packets on non-control endpoints should be OK as URBs
should have enough buffer space for a wMaxPacket size transfer.
See: https://github.com/raspberrypi/linux/issues/3148
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
dwc_otg: Declare DMA capability with HCD_DMA flag
Following [1], USB controllers have to declare DMA capabilities in
order for them to be used by adding the HCD_DMA flag to their hc_driver
struct.
[1] 7b81cb6bdd ("usb: add a HCD_DMA flag instead of guestimating DMA capabilities")
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
dwc_otg: checking the urb->transfer_buffer too early (#3332)
After enable the HIGHMEM and VMSPLIT_3G, the dwc_otg driver doesn't
work well on Pi2/3 boards with 1G physical ram. Users experience
the failure when copying a file of 600M size to the USB stick. And
at the same time, the dmesg shows:
usb 1-1.1.2: reset high-speed USB device number 8 using dwc_otg
sd 0:0:0:0: [sda] tag#0 FAILED Result: hostbyte=DID_ERROR driverbyte=DRIVER_OK
blk_update_request: I/O error, dev sda, sector 3024048 op 0x1:(WRITE) flags 0x4000 phys_seg 15 prio class 0
When this happens, the sg_buf sent to the driver is located in the
highmem region, the usb_sg_init() in the core/message.c will leave
transfer_buffer to NULL if the sg_buf is in highmem, but in the
dwc_otg driver, it returns -EINVAL unconditionally if transfer_buffer
is NULL.
The driver can handle the situation of buffer to be NULL, if it is in
DMA mode, it will convert an address from transfer_dma.
But if the conversion fails or it is in the PIO mode, we should check
buffer and return -EINVAL if it is NULL.
BugLink: https://bugs.launchpad.net/bugs/1852510
Signed-off-by: Hui Wang <hui.wang@canonical.com>
dwc_otg: constrain endpoint max packet and transfer size on split IN
The hcd would unconditionally set the transfer length to the endpoint
packet size for non-isoc IN transfers. If the remaining buffer length
was less than the length of returned data, random memory would get
scribbled over, with bad effects if it crossed a page boundary.
Force a babble error if this happens by limiting the max transfer size
to the available buffer space. DMA will stop writing to memory on a
babble condition.
The hardware expects xfersize to be an integer multiple of maxpacket
size, so override hcchar.b.mps as well.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
dwc_otg: fiq_fsm: pause when cancelling split transactions
Non-periodic splits will DMA to/from the driver-provided transfer_buffer,
which may be freed immediately after the dequeue call returns. Block until
we know the transfer is complete.
A similar delay is needed when cleaning up disconnects, as the FIQ could
have started a periodic transfer in the previous microframe to the one
that triggered a disconnect.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
dwc_otg: fiq_fsm: add a barrier on entry into FIQ handler(s)
On BCM2835, there is no hardware guarantee that multiple outstanding
reads to different peripherals will complete in-order. The FIQ code
uses peripheral reads without barriers for performance, so in the case
where a read to a slow peripheral was issued immediately prior to FIQ
entry, the first peripheral read that the FIQ did could end up with
wrong read data returned.
Add dsb(sy) on entry so that all outstanding reads are retired.
The FIQ only issues reads to the dwc_otg core, so per-read barriers
in the handler itself are not required.
On BCM2836 and BCM2837 the barrier is not strictly required due to
differences in how the peripheral bus is implemented, but having
arch-specific handlers that introduce different latencies is risky.
Signed-off-by: Jonathan Bell <jonathan@raspberrypi.org>
Signed-off-by: popcornmix <popcornmix@gmail.com>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
bcm2709: Drop platform smp and timer init code
irq-bcm2836 handles this through these functions:
bcm2835_init_local_timer_frequency()
bcm2836_arm_irqchip_smp_init()
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
bcm270x: Use watchdog for reboot/poweroff
The watchdog driver already has support for reboot/poweroff.
Make use of this and remove the code from the platform files.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
board_bcm2835: Remove coherent dma pool increase - API has gone
Under some circumstances on BCM283x processors data loss can be
observed - a single byte missing from the TX output stream. These bytes
are always the last byte of a batch of 8 written from pl011_tx_chars
when from_irq is true, meaning that the FIFO full flag is not checked
before writing.
The transmit optimisation relies on the FIFO being half-empty when the
TX interrupt is raised. Instrumenting the driver further showed that
the failure case correlated with the TX FIFO full flag being set at the
point where the last byte was written to the data register, which
explains the data loss but not how the FIFO appeared to be prematurely
full. A possible explanation is that a FIFO write was in flight at the
time the interrupt was raised, but as yet there is no hypothesis as to
how this might occur.
In the absence of a clear understanding of the failure mechanism, avoid
the problem by checking the FIFO levels before writing the last byte of
the group, which will have minimal performance impact.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The PL011 driver lacks throttle and unthrottle methods. As a result,
sending more data to the Pi than it can immediately sink while CRTSCTS
is enabled causes a NULL pointer to be followed.
Add a throttle handler that disables the RX interrupts, and an
unthrottle handler that reenables them.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
pl011_tx_chars takes a "from_irq" parameter to reduce the number of
register accesses. When from_irq is true the function assumes that the
FIFO is half empty and writes up to half a FIFO's worth of bytes
without polling the FIFO status register, the reasoning being that
the function is being called as a result of the TX interrupt being
raised. This logic would work were it not for the fact that
pl011_rx_chars, called from pl011_int before pl011_tx_chars, releases
the spinlock before calling tty_flip_buffer_push.
A user thread writing to the UART claims the spinlock and ultimately
calls pl011_tx_chars with from_irq set to false. This reverts to the
older logic that polls the FIFO status register before sending every
byte. If this happen on an SMP system during the section of the IRQ
handler where the spinlock has been released, then by the time the TX
interrupt handler is called, the FIFO may already be full, and any
further writes are likely to be lost.
The fix involves adding a per-port flag that is true iff running from
within the interrupt handler and the spinlock has not yet been released.
This flag is then used as the value for the from_irq parameter of
pl011_tx_chars, causing polling to be used in the unsafe case.
Fixes: 1e84d22322 ("serial/amba-pl011: Refactor and simplify TX FIFO handling")
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The BCM2835 PL011 implementation seems to have a bug that can lead to a
transmission lockup if CTS changes frequently. A workaround was added to
the driver with a vendor-specific flag to enable it, but this flag is
currently not set for ARM implementations.
Add a "cts-event-workaround" property to Pi DTBs and use the presence
of that property to force the flag to be enabled in the driver.
See: https://github.com/raspberrypi/linux/issues/1280
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The pl011 register accessor functions use the _relaxed versions of the
standard readl() and writel() functions, meaning that there are no
automatic memory barriers. When polling a FIFO status register to check
for fullness, it is necessary to ensure that any outstanding writes have
completed; otherwise the flags are effectively stale, making it possible
that the next write is to a full FIFO.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The UART clock is initialised to be as close to the requested
frequency as possible without exceeding it. Now that there is a
clock manager that returns the actual frequencies, an expected
48MHz clock is reported as 47999625. If the requested baudrate
== requested clock/16, there is no headroom and the slight
reduction in actual clock rate results in failure.
Detect cases where it looks like a "round" clock was chosen and
adjust the reported clock to match that "round" value. As the
code comment says:
/*
* If increasing a clock by less than 0.1% changes it
* from ..999.. to ..000.., round up.
*/
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The pl011 driver looks for DT aliases of the form "serial<n>",
and if found uses <n> as the device ID. This can cause
/dev/ttyAMA0 to become /dev/ttyAMA1, which is confusing if the
other serial port is provided by the 8250 driver which doesn't
use the same logic.
For applications of the LAN78xx that don't have valid programmed
EEPROMs or OTPs, enabling both LEDs and auto-negotiation by default
seems reasonable.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The syscon node defines a register range that duplicates that used by
the local_intc node on bcm2836/7. Since irq-bcm2835 and irq-bcm2836 are
built in and always present together (both drivers are enabled by
CONFIG_ARCH_BCM2835), it is possible to replace the syscon usage with a
global variable that simplifies the code. Doing so does lose the
locking provided by regmap, but as only one side is using the regmap
interface (irq-bcm2835 uses readl and write) there is no loss of
atomicity.
See: https://github.com/raspberrypi/firmware/issues/926
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
See commit dae803e165 -- the warning is
expected sometimes when using CMA. However, that commit still spams
my kernel log with these warnings.
Signed-off-by: Eric Anholt <eric@anholt.net>
This adds a debug module parameter to aid in debugging transfer issues
by printing info to the kernel log. When enabled, status values are
collected in the interrupt routine and msg info in
bcm2835_i2c_start_transfer(). This is done in a way that tries to avoid
affecting timing. Having printk in the isr can mask issues.
debug values (additive):
1: Print info on error
2: Print info on all transfers
3: Print messages before transfer is started
The value can be changed at runtime:
/sys/module/i2c_bcm2835/parameters/debug
Example output, debug=3:
[ 747.114448] bcm2835_i2c_xfer: msg(1/2) write addr=0x54, len=2 flags= [i2c1]
[ 747.114463] bcm2835_i2c_xfer: msg(2/2) read addr=0x54, len=32 flags= [i2c1]
[ 747.117809] start_transfer: msg(1/2) write addr=0x54, len=2 flags= [i2c1]
[ 747.117825] isr: remain=2, status=0x30000055 : TA TXW TXD TXE [i2c1]
[ 747.117839] start_transfer: msg(2/2) read addr=0x54, len=32 flags= [i2c1]
[ 747.117849] isr: remain=32, status=0xd0000039 : TA RXR TXD RXD [i2c1]
[ 747.117861] isr: remain=20, status=0xd0000039 : TA RXR TXD RXD [i2c1]
[ 747.117870] isr: remain=8, status=0x32 : DONE TXD RXD [i2c1]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Christopher Alexander Tobias Schulze - May 2, 2015, 11:57 a.m.
This patch fixes a problem with VFP state save and restore related
to exception handling (panic with message "BUG: unsupported FP
instruction in kernel mode") present on VFP11 floating point units
(as used with ARM1176JZF-S CPUs, e.g. on first generation Raspberry
Pi boards). This patch was developed and discussed on
https://github.com/raspberrypi/linux/issues/859
A precondition to see the crashes is that floating point exception
traps are enabled. In this case, the VFP11 might determine that a FPU
operation needs to trap at a point in time when it is not possible to
signal this to the ARM11 core any more. The VFP11 will then set the
FPEXC.EX bit and store the trapped opcode in FPINST. (In some cases,
a second opcode might have been accepted by the VFP11 before the
exception was detected and could be reported to the ARM11 - in this
case, the VFP11 also sets FPEXC.FP2V and stores the second opcode in
FPINST2.)
If FPEXC.EX is set, the VFP11 will "bounce" the next FPU opcode issued
by the ARM11 CPU, which will be seen by the ARM11 as an undefined opcode
trap. The VFP support code examines the FPEXC.EX and FPEXC.FP2V bits
to decide what actions to take, i.e., whether to emulate the opcodes
found in FPINST and FPINST2, and whether to retry the bounced instruction.
If a user space application has left the VFP11 in this "pending trap"
state, the next FPU opcode issued to the VFP11 might actually be the
VSTMIA operation vfp_save_state() uses to store the FPU registers
to memory (in our test cases, when building the signal stack frame).
In this case, the kernel crashes as described above.
This patch fixes the problem by making sure that vfp_save_state() is
always entered with FPEXC.EX cleared. (The current value of FPEXC has
already been saved, so this does not corrupt the context. Clearing
FPEXC.EX has no effects on FPINST or FPINST2. Also note that many
callers already modify FPEXC by setting FPEXC.EN before invoking
vfp_save_state().)
This patch also addresses a second problem related to FPEXC.EX: After
returning from signal handling, the kernel reloads the VFP context
from the user mode stack. However, the current code explicitly clears
both FPEXC.EX and FPEXC.FP2V during reload. As VFP11 requires these
bits to be preserved, this patch disables clearing them for VFP
implementations belonging to architecture 1. There should be no
negative side effects: the user can set both bits by executing FPU
opcodes anyway, and while user code may now place arbitrary values
into FPINST and FPINST2 (e.g., non-VFP ARM opcodes) the VFP support
code knows which instructions can be emulated, and rejects other
opcodes with "unhandled bounce" messages, so there should be no
security impact from allowing reloading FPEXC.EX and FPEXC.FP2V.
Signed-off-by: Christopher Alexander Tobias Schulze <cat.schulze@alice-dsl.net>
At present there is no mechanism to specify driver load order,
which can lead to deferrals and repeated retries until successful.
Since this situation is expected, reduce the dmesg level to
INFO and mention that the operation will be retried.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The VPU is responsible for managing the core clock, usually under
direction from the bcm2835-cpufreq driver but not via the clk-bcm2835
driver. Since the core frequency can change without warning, it is
safer to report the maximum clock rate to users of the core clock -
I2C, SPI and the mini UART - to err on the safe side when calculating
clock divisors.
If the DT node for the clock driver includes a reference to the
firmware node, use the firmware API to query the maximum core clock
instead of reading the divider registers.
Prior to this patch, a "100KHz" I2C bus was sometimes clocked at about
160KHz. In particular, switching to the 4.9 kernel was likely to break
SenseHAT usage on a Pi3.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The claim-clocks property can be used to prevent PLLs and dividers
from being marked as critical. It contains a vector of clock IDs,
as defined by dt-bindings/clock/bcm2835.h.
Use this mechanism to claim PLLD_DSI0, PLLD_DSI1, PLLH_AUX and
PLLH_PIX for the vc4_kms_v3d driver.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The VPU configures and relies on several PLLs and dividers. Mark all
enabled dividers and their PLLs as CRITICAL to prevent the kernel from
switching them off.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
The Raspberry Pi firmware looks at the RSTS register to know which
partition to boot from. The reboot syscall command
LINUX_REBOOT_CMD_RESTART2 supports passing in a string argument.
Add support for passing in a partition number 0..63 to boot from.
Partition 63 is a special partiton indicating halt.
If the partition doesn't exist, the firmware falls back to partition 0.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Load driver early since at least bcm2708_fb doesn't support deferred
probing and even if it did, we don't want the video driver deferred.
Support the legacy DMA API which is needed by bcm2708_fb.
Don't mask out channel 2.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
An alternative strategy would be to use "rpi,spidev" instead, but that
would require many Raspberry Pi Device Tree changes.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Add a duplicate irq range with an offset on the hwirq's so the
driver can detect that enable_fiq() is used.
Tested with downstream dwc_otg USB controller driver.
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
smsc95xx is adjusting truesize when it shouldn't, and following a recent patch from Eric this is now triggering warnings.
This patch stops smsc95xx from changing truesize.
Signed-off-by: Steve Glendinning <steve.glendinning@smsc.com>
This reverts commit c7dacf5b0f.
The Pi 400 shutdown/poweroff mechanism relies on being able to set
a GPIO on the expander in the pm_power_off handler, something that
requires two mailbox calls - GET_GPIO_STATE and SET_GPIO_STATE. A
recent kernel change introduces a reasonable possibility that the
GET call doesn't completes, and bisecting led to a commit from
October that changes the timer usage of the mailbox.
My theory is that there is a race condition in the new code that breaks
the poll timer, but that it normally goes unnoticed because subsequent
mailbox activity wakes it up again. The power-off mailbox calls happen
at a time when other subsystems have been shut down, so if one of them
fails then there is nothing to allow it to recover.
See: https://github.com/raspberrypi/linux/issues/3941
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
description:Create a report to help us fix your issue
body:
- type:markdown
attributes:
value:|
**Is this the right place for my bug report?**
This repository contains the Linux kernel used on the Raspberry Pi.
If you believe that the issue you are seeing is kernel-related, this is the right place.
If not, we have other repositories for the GPU firmware at [github.com/raspberrypi/firmware](https://github.com/raspberrypi/firmware) and Raspberry Pi userland applications at [github.com/raspberrypi/userland](https://github.com/raspberrypi/userland).
If you have problems with the Raspbian distribution packages, report them in the [github.com/RPi-Distro/repo](https://github.com/RPi-Distro/repo).
If you simply have a question, then [the Raspberry Pi forums](https://www.raspberrypi.org/forums) are the best place to ask it.
- type:textarea
id:description
attributes:
label:Describe the bug
description:|
Add a clear and concise description of what you think the bug is.
validations:
required:true
- type:textarea
id:reproduce
attributes:
label:Steps to reproduce the behaviour
description:|
List the steps required to reproduce the issue.
validations:
required:true
- type:dropdown
id:model
attributes:
label:Device (s)
description:Onwhich device you are facing the bug?
multiple:true
options:
- Raspberry Pi Zero
- Raspberry Pi Zero W/WH
- Raspberry Pi Zero 2 W
- Raspberry Pi 1 Mod. A
- Raspberry Pi 1 Mod. A+
- Raspberry Pi 1 Mod. B
- Raspberry Pi 1 Mod. B+
- Raspberry Pi 2 Mod. B
- Raspberry Pi 2 Mod. B v1.2
- Raspberry Pi 3 Mod. A+
- Raspberry Pi 3 Mod. B
- Raspberry Pi 3 Mod. B+
- Raspberry Pi 4 Mod. B
- Raspberry Pi 400
- Raspberry Pi CM1
- Raspberry Pi CM3
- Raspberry Pi CM3 Lite
- Raspberry Pi CM3+
- Raspberry Pi CM3+ Lite
- Raspberry Pi CM4
- Raspberry Pi CM4 Lite
- Other
validations:
required:true
- type:textarea
id:system
attributes:
label:System
description:|
Copy and paste the results of the raspinfo command in to this section.
Alternatively, copy and paste a pastebin link, or add answers to the following questions:
* Which OS and version (`cat /etc/rpi-issue`)?
* Which firmware version (`vcgencmd version`)?
* Which kernel version (`uname -a`)?
validations:
required:true
- type:textarea
id:logs
attributes:
label:Logs
description:|
If applicable, add the relevant output from `dmesg` or similar.
about:"Please do not use GitHub for asking questions. If you simply have a question, then the Raspberry Pi forums are the best place to ask it. Thanks in advance for helping us keep the issue tracker clean!"
- name:"⛔ Problems with the Raspbian distribution packages"
url:https://github.com/RPi-Distro/repo
about:"If you have problems with the Raspbian distribution packages, please report them in the github.com/RPi-Distro/repo."
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
