[ Upstream commit d2bd39c045 ]
The PCIe bandwidth controller added by a subsequent commit will require
selecting PCIe Link Speeds that are lower than the Maximum Link Speed.
The struct pci_bus only stores max_bus_speed. Even if PCIe r6.1 sec 8.2.1
currently disallows gaps in supported Link Speeds, the Implementation Note
in PCIe r6.1 sec 7.5.3.18, recommends determining supported Link Speeds
using the Supported Link Speeds Vector in the Link Capabilities 2 Register
(when available) to "avoid software being confused if a future
specification defines Links that do not require support for all slower
speeds."
Reuse code in pcie_get_speed_cap() to add pcie_get_supported_speeds() to
query the Supported Link Speeds Vector of a PCIe device. The value is taken
directly from the Supported Link Speeds Vector or synthesized from the Max
Link Speed in the Link Capabilities Register when the Link Capabilities 2
Register is not available.
The Supported Link Speeds Vector in the Link Capabilities Register 2
corresponds to the bus below on Root Ports and Downstream Ports, whereas it
corresponds to the bus above on Upstream Ports and Endpoints (PCIe r6.1 sec
7.5.3.18):
Supported Link Speeds Vector - This field indicates the supported Link
speed(s) of the associated Port.
Add supported_speeds into the struct pci_dev that caches the
Supported Link Speeds Vector.
supported_speeds contains a set of Link Speeds only in the case where PCIe
Link Speed can be determined. Root Complex Integrated Endpoints do not have
a well-defined Link Speed because they do not implement either of the Link
Capabilities Registers, which is allowed by PCIe r6.1 sec 7.5.3 (the same
limitation applies to determining cur_bus_speed and max_bus_speed that are
PCI_SPEED_UNKNOWN in such case). This is of no concern from PCIe bandwidth
controller point of view because such devices are not attached into a PCIe
Root Port that could be controlled.
The supported_speeds field keeps the extra reserved zero at the least
significant bit to match the Link Capabilities 2 Register layout.
An attempt was made to store supported_speeds field into the struct pci_bus
as an intersection of both ends of the Link, however, the subordinate
struct pci_bus is not available early enough. The Target Speed quirk (in
pcie_failed_link_retrain()) can run either during initial scan or later,
requiring it to use the API provided by the PCIe bandwidth controller to
set the Target Link Speed in order to co-exist with the bandwidth
controller. When the Target Speed quirk is calling the bandwidth controller
during initial scan, the struct pci_bus is not yet initialized. As such,
storing supported_speeds into the struct pci_bus is not viable.
Suggested-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/20241018144755.7875-4-ilpo.jarvinen@linux.intel.com
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
[bhelgaas: move pcie_get_supported_speeds() decl to drivers/pci/pci.h]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Stable-dep-of: 6cff20ce3b ("PCI/ACPI: Fix runtime PM ref imbalance on Hot-Plug Capable ports")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b1a7f99967 ]
Many functions in PCI use accessor macros such as pci_resource_len(),
which take a BAR index. That index, however, is never checked for
validity, potentially resulting in undefined behavior by overflowing the
array pci_dev.resource in the macro pci_resource_n().
Since many users of those macros directly assign the accessed value to
an unsigned integer, the macros cannot be changed easily anymore to
return -EINVAL for invalid indexes. Consequently, the problem has to be
mitigated in higher layers.
Add pci_bar_index_valid(). Use it where appropriate.
Link: https://lore.kernel.org/r/20250312080634.13731-4-phasta@kernel.org
Closes: https://lore.kernel.org/all/adb53b1f-29e1-3d14-0e61-351fd2d3ff0d@linux.intel.com/
Reported-by: Bingbu Cao <bingbu.cao@linux.intel.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
[kwilczynski: correct if-statement condition the pci_bar_index_is_valid()
helper function uses, tidy up code comments]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
[bhelgaas: fix typo]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 2fa046449a ]
The "bus" and "cxl_bus" reset methods reset a device by asserting Secondary
Bus Reset on the bridge leading to the device. These only work if the
device is the only device below the bridge.
Add a sysfs 'reset_subordinate' attribute on bridges that can assert
Secondary Bus Reset regardless of how many devices are below the bridge.
This resets all the devices below a bridge in a single command, including
the locking and config space save/restore that reset methods normally do.
This may be the only way to reset devices that don't support other reset
methods (ACPI, FLR, PM reset, etc).
Link: https://lore.kernel.org/r/20241025222755.3756162-1-kbusch@meta.com
Signed-off-by: Keith Busch <kbusch@kernel.org>
[bhelgaas: commit log, add capable(CAP_SYS_ADMIN) check]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Amey Narkhede <ameynarkhede03@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
- Add DT "ti,syscon-acspcie-proxy-ctrl" and driver support to enable the
ACSPCIE module to drive Refclk for the Endpoint (Siddharth Vadapalli)
- Extract the cadence link setup from cdns_pcie_host_setup() so link setup
can be done separately during resume (Thomas Richard)
- Use dev_err_probe() to simplify j721e probe (Thomas Richard)
- Add T_PERST_CLK_US definition for the mandatory delay between Refclk
becoming stable and PERST# being deasserted (Thomas Richard)
- Add j721e suspend and resume support (Théo Lebrun)
* pci/controller/j721e:
PCI: j721e: Add suspend and resume support
PCI: j721e: Use T_PERST_CLK_US macro
PCI: Add T_PERST_CLK_US macro
PCI: j721e: Add reset GPIO to struct j721e_pcie
PCI: j721e: Use dev_err_probe() in the probe() function
PCI: cadence: Set cdns_pcie_host_init() global
PCI: cadence: Extract link setup sequence from cdns_pcie_host_setup()
PCI: j721e: Enable ACSPCIE Refclk if "ti,syscon-acspcie-proxy-ctrl" exists
dt-bindings: PCI: ti,j721e-pci-host: Add ACSPCIE proxy control property
- Initialize leds class earlier (with an unfortunate Makefile ordering
change) so the PCI NPEM driver can use it (Mariusz Tkaczyk)
- Add Native PCIe Enclosure Management (NPEM) support for sysfs control of
NVMe RAID storage indicators (ok/fail/locate/rebuild/etc) (Mariusz
Tkaczyk)
- Add support for the ACPI _DSM PCIe SSD status LED management, which is
functionally similar to NPEM but mediated by platform firmware (Mariusz
Tkaczyk)
* pci/npem:
PCI/NPEM: Add _DSM PCIe SSD status LED management
PCI/NPEM: Add Native PCIe Enclosure Management support
leds: Init leds class earlier
- Clear LBMS bit after a manual link retrain so we don't try to retrain a
link when there's no downstream device anymore (Maciej W. Rozycki)
- Revert to the original link speed after retraining fails instead of
leaving it restricted to 2.5GT/s, so a future device has a chance to use
higher speeds (Maciej W. Rozycki)
- Correct interpretation of pcie_retrain_link() return status and update it
to return 0/errno instead of true/false (Maciej W. Rozycki)
* pci/enumeration:
PCI: Use an error code with PCIe failed link retraining
PCI: Correct error reporting with PCIe failed link retraining
PCI: Revert to the original speed after PCIe failed link retraining
PCI: Clear the LBMS bit after a link retrain
- Export pcim_request_region(), a managed counterpart of
pci_request_region(), for use by drivers (Philipp Stanner)
- Request the PCI BAR used by xboxvideo (Philipp Stanner)
- Export pcim_iomap_region() and deprecate pcim_iomap_regions() (Philipp
Stanner)
- Request and map drm/ast BARs with pcim_iomap_region() (Philipp Stanner)
* pci/devres:
drm/ast: Request PCI BAR with devres
PCI: Deprecate pcim_iomap_regions() in favor of pcim_iomap_region()
drm/vboxvideo: Add PCI region request
PCI: Make pcim_request_region() a public function
PCIe r6.0 changed the abbreviation for "Configuration Request Retry Status"
Completion Status from "CRS" to "RRS" and uses the terminology of
"Configuration RRS Software Visibility" instead of "CRS Software
Visibility".
Align the Linux usage with the r6.0 spec language. No functional change
intended.
It's confusing to make this change, but I think "RRS" *is* a better
abbreviation because it was easy to interpret "CRS" as "Completion Retry
Status", which really didn't make any sense.
Link: https://lore.kernel.org/r/20240827234848.4429-4-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
After a device reset, delays are required before the device can
successfully complete config accesses. PCIe r6.0, sec 6.6, specifies some
delays required before software can perform config accesses. Devices that
require more time after those delays may respond to config accesses with
Configuration Request Retry Status (RRS) completions.
Callers of pci_dev_wait() are responsible for delays until the device can
respond to config accesses. pci_dev_wait() waits any additional time until
the device can successfully complete config accesses.
Reading config space of devices that are not present or not ready typically
returns ~0 (PCI_ERROR_RESPONSE). Previously we polled the Command register
until we got a value other than ~0. This is sometimes a problem because
Root Complex handling of RRS completions may include several retries and
implementation-specific behavior that is invisible to software (see sec
2.3.2), so the exponential backoff in pci_dev_wait() may not work as
intended.
Linux enables Configuration RRS Software Visibility on all Root Ports that
support it. If it is enabled, read the Vendor ID instead of the Command
register. RRS completions cause immediate return of the 0x0001 reserved
Vendor ID value, so the pci_dev_wait() backoff works correctly.
When a read of Vendor ID eventually completes successfully by returning a
non-0x0001 value (the Vendor ID or 0xffff for VFs), the device should be
initialized and ready to respond to config requests.
For conventional PCI devices or devices below Root Ports that don't support
Configuration RRS Software Visibility, poll the Command register as before.
This was developed independently, but is very similar to Stanislav
Spassov's previous work at
https://lore.kernel.org/linux-pci/20200223122057.6504-1-stanspas@amazon.com
Link: https://lore.kernel.org/r/20240827234848.4429-2-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Tested-by: Duc Dang <ducdang@google.com>
Native PCIe Enclosure Management (NPEM, PCIe r6.1 sec 6.28) allows managing
LEDs in storage enclosures. NPEM is indication oriented and it does not
give direct access to LEDs. Although each indication *could* represent an
individual LED, multiple indications could also be represented as a single,
multi-color LED or a single LED blinking in a specific interval. The
specification leaves that open.
Each enabled indication (capability register bit on) is represented as a
ledclass_dev which can be controlled through sysfs. For every ledclass
device only 2 brightness states are allowed: LED_ON (1) or LED_OFF (0).
This corresponds to the NPEM control register (Indication bit on/off).
Ledclass devices appear in sysfs as child devices (subdirectory) of PCI
device which has an NPEM Extended Capability and indication is enabled in
NPEM capability register. For example, these are LEDs created for pcieport
"10000:02:05.0" on my setup:
leds/
├── 10000:02:05.0:enclosure:fail
├── 10000:02:05.0:enclosure:locate
├── 10000:02:05.0:enclosure:ok
└── 10000:02:05.0:enclosure:rebuild
They can be also found in "/sys/class/leds" directory. The parent PCIe
device domain/bus/device/function address is used to guarantee uniqueness
across leds subsystem.
To enable/disable a "fail" indication, the "brightness" file can be edited:
echo 1 > ./leds/10000:02:05.0:enclosure:fail/brightness
echo 0 > ./leds/10000:02:05.0:enclosure:fail/brightness
PCIe r6.1, sec 7.9.19.2 defines the possible indications.
Multiple indications for same parent PCIe device can conflict and hardware
may update them when processing new request. To avoid issues, driver
refresh all indications by reading back control register.
This driver expects to be the exclusive NPEM extended capability manager.
It waits up to 1 second after imposing new request, it doesn't verify if
controller is busy before write, and it assumes the mutex lock gives
protection from concurrent updates.
If _DSM LED management is available, we assume the platform may be using
NPEM for its own purposes (see PCI Firmware Spec r3.3 sec 4.7), so the
driver does not use NPEM. A future patch will add _DSM support; an info
message notes whether NPEM or _DSM is being used.
NPEM is a PCIe extended capability so it should be registered in
pcie_init_capabilities() but it is not possible due to LED dependency. The
parent pci_device must be added earlier for led_classdev_register() to be
successful. NPEM does not require configuration on kernel side, so it is
safe to register LED devices later.
Link: https://lore.kernel.org/r/20240904104848.23480-3-mariusz.tkaczyk@linux.intel.com
Suggested-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Tested-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
The "Power Sequencing and Reset Signal Timings" table of the PCI
Express Card Electromechanical Specification, Revision 5.1, Section
2.9.2, indicates PERST# should be deasserted after minimum of 100us
once REFCLK is stable (symbol T_PERST-CLK).
Add a macro so that PCIe controller drivers can use it.
Link: https://lore.kernel.org/linux-pci/20240102-j7200-pcie-s2r-v7-5-a2f9156da6c3@bootlin.com
Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
[kwilczynski: commit log, update sleep interval macros code comments]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
- Use dev_err_probe() in dw-rockchip probe error path so the failures
aren't silent (Uwe Kleine-König)
- Sleep PCIE_T_PVPERL_MS (100ms) before deasserting PERST# (Damien Le Moal)
- Sleep PCIE_T_RRS_READY_MS (100ms) after conventional reset, before a
config access (Damien Le Moal)
- Request the PERST# GPIO with GPIOD_OUT_LOW so it matches the POR value,
which avoids a spurious PERST# assertion and fixes a Qcom modem firmware
crash and issues with WLAN controllers, e.g., RTL8822CE (Manivannan
Sadhasivam for rockchip, Niklas Cassel for dw-rockchip)
- Refactor dw-rockchip and add support for Endpoint mode for rk3568 and
rk3588 (Niklas Cassel)
* pci/controller/rockchip:
PCI: dw-rockchip: Use pci_epc_init_notify() directly
PCI: dw-rockchip: Add endpoint mode support
PCI: dw-rockchip: Refactor the driver to prepare for EP mode
PCI: dw-rockchip: Add rockchip_pcie_get_ltssm() helper
PCI: dw-rockchip: Fix weird indentation
PCI: dw-rockchip: Fix initial PERST# GPIO value
PCI: dw-rockchip: Add error messages in .probe() error paths
PCI: rockchip: Use GPIOD_OUT_LOW flag while requesting ep_gpio
PCI: rockchip-host: Wait 100ms after reset before starting configuration
PCI: rockchip-host: Fix rockchip_pcie_host_init_port() PERST# handling
- Use devm_clk_bulk_get_all() to get all the clocks from DT to avoid
writing out all the clock names (Manivannan Sadhasivam)
- Add DT binding and driver support for the SA8775P SoC (Mrinmay Sarkar)
- Refactor dw_pcie_edma_find_chip() to enable adding support for Hyper DMA
(HDMA) (Manivannan Sadhasivam)
- Enable drivers to supply the eDMA channel count since some can't auto
detect this (Manivannan Sadhasivam)
- Add HDMA support for the SA8775P SoC (Mrinmay Sarkar)
- Override the SA8775P NO_SNOOP default to avoid possible memory corruption
(Mrinmay Sarkar)
- Make sure resources are disabled during PERST# assertion, even if the
link is already disabled (Manivannan Sadhasivam)
- Vote for the CPU-PCIe ICC (interconnect) path to ensure it stays active
even if other drivers don't vote for it (Krishna chaitanya chundru)
- Add Operating Performance Points (OPP) to scale performance state based
on aggregate link bandwidth to improve SoC power efficiency (Krishna
chaitanya chundru)
- Return failure instead of success if dev_pm_opp_find_freq_floor() fails
(Dan Carpenter)
- Avoid an error pointer dereference if dev_pm_opp_find_freq_exact() fails
(Dan Carpenter)
- Prevent use of uninitialized data in qcom_pcie_suspend_noirq() (Dan
Carpenter)
* pci/controller/qcom:
PCI: qcom: Prevent use of uninitialized data in qcom_pcie_suspend_noirq()
PCI: qcom: Prevent potential error pointer dereference
PCI: qcom: Fix missing error code in qcom_pcie_probe()
PCI: qcom: Add OPP support to scale performance
PCI: Bring the PCIe speed to MBps logic to new pcie_dev_speed_mbps()
PCI: qcom: Add ICC bandwidth vote for CPU to PCIe path
PCI: qcom-ep: Disable resources unconditionally during PERST# assert
PCI: qcom-ep: Override NO_SNOOP attribute for SA8775P EP
PCI: qcom: Override NO_SNOOP attribute for SA8775P RC
PCI: epf-mhi: Enable HDMA for SA8775P SoC
PCI: qcom-ep: Add HDMA support for SA8775P SoC
PCI: dwc: Pass the eDMA mapping format flag directly from glue drivers
PCI: dwc: Skip finding eDMA channels count for HDMA platforms
PCI: dwc: Refactor dw_pcie_edma_find_chip() API
PCI: qcom-ep: Add support for SA8775P SOC
dt-bindings: PCI: qcom-ep: Add support for SA8775P SoC
PCI: qcom: Use devm_clk_bulk_get_all() API
- Move PLDA XpressRICH generic DT binding properties to
plda,xpressrich3-axi-common.yaml where they can be shared across
PLDA-based drivers (Minda Chen)
- Create a drivers/pci/controller/plda/ directory for PLDA-based drivers
and move pcie-microchip-host.c there (Minda Chen)
- Move PLDA generic macros to pcie-plda.h where they can be shared across
drivers (Minda Chen)
- Extract PLDA generic structures from pcie-microchip-host.c, rename them
to be generic, and move them to pcie-plda-host.c where they can be shared
across drivers (Minda Chen)
- Add a .request_event_irq() callback for requesting device-specific
interrupts in addition to PLDA-generic interrupts (Minda Chen)
- Add DT binding and driver for the StarFive JH7110 SoC, based on PLDA IP
(Minda Chen)
* pci/controller/microchip:
PCI: starfive: Add JH7110 PCIe controller
dt-bindings: PCI: Add StarFive JH7110 PCIe controller
PCI: Add PCIE_RESET_CONFIG_DEVICE_WAIT_MS waiting time value
PCI: plda: Pass pci_host_bridge to plda_pcie_setup_iomems()
PCI: plda: Add host init/deinit and map bus functions
PCI: plda: Add event bitmap field to struct plda_pcie_rp
PCI: microchip: Move IRQ functions to pcie-plda-host.c
PCI: microchip: Add event irqchip field to host port and add PLDA irqchip
PCI: microchip: Add get_events() callback and PLDA get_event()
PCI: microchip: Add INTx and MSI event num to struct plda_event
PCI: microchip: Add request_event_irq() callback function
PCI: microchip: Add num_events field to struct plda_pcie_rp
PCI: microchip: Rename interrupt related functions
PCI: microchip: Move PLDA functions to pcie-plda-host.c
PCI: microchip: Rename PLDA functions to be generic
PCI: microchip: Move PLDA structures to plda-pcie.h
PCI: microchip: Rename PLDA structures to be generic
PCI: microchip: Add bridge_addr field to struct mc_pcie
PCI: microchip: Move PLDA IP register macros to pcie-plda.h
PCI: microchip: Move pcie-microchip-host.c to PLDA directory
dt-bindings: PCI: Add PLDA XpressRICH PCIe host common properties
# Conflicts:
# drivers/pci/pci.h
- Use msleep() in DWC core instead of usleep_range() for ~100 ms sleep
(Konrad Dybcio)
- Fix iATU slot management to avoid using the wrong slot after PERST#
assert/deassert, which could potentially cause DMA to go the wrong place
(Frank Li)
- Consolidate dw_pcie_prog_outbound_atu() arguments into a struct to ease
adding new functionality like initiating Message TLPs (Yoshihiro Shimoda)
- Add support for endpoints to initiate PCIe messages (Yoshihiro Shimoda)
- Add #defines for PCIe INTx messages (Yoshihiro Shimoda)
- Add support for endpoints to initiate PCIe PME_Turn_Off messages for
system suspend (Frank Li)
- Add dw_pcie_ep_linkdown() to reinitialize registers that are lost when
the link goes down (Manivannan Sadhasivam)
- Use dw_pcie_ep_linkdown() to reinitialize qcom non-sticky registers that
are lost when the link goes down (Manivannan Sadhasivam)
- Enforce DWC limitation that 64-bit BARs must start with the even numbered
BAR (Niklas Cassel)
* pci/controller/dwc:
PCI: dwc: ep: Enforce DWC specific 64-bit BAR limitation
PCI: layerscape-ep: Use the generic dw_pcie_ep_linkdown() API to handle Link Down event
PCI: qcom-ep: Use the generic dw_pcie_ep_linkdown() API to handle Link Down event
PCI: dwc: ep: Remove dw_pcie_ep_init_notify() wrapper
PCI: dwc: ep: Add a generic dw_pcie_ep_linkdown() API to handle Link Down event
PCI: dwc: Add generic MSG TLP support for sending PME_Turn_Off when system suspend
PCI: Add PCIE_MSG_CODE_PME_TURN_OFF message macro
PCI: Add PCIE_MSG_CODE_ASSERT_INTx message macros
PCI: dwc: Add outbound MSG TLPs support
PCI: dwc: Consolidate args of dw_pcie_prog_outbound_atu() into a structure
PCI: dwc: Fix index 0 incorrectly being interpreted as a free ATU slot
PCI: dwc: Use msleep() in dw_pcie_wait_for_link()
- Move the PRESERVE_BOOT_CONFIG ACPI _DSM evaluation from drivers/acpi to
drivers/pci so we can unify with similar DT functionality (Vidya Sagar)
- Add of_pci_preserve_config() to check for a DT "linux,pci-probe-only"
property on a per-host bridge basis in addition to a global basis (Vidya
Sagar)
- Unify ACPI PRESERVE_BOOT_CONFIG _DSM and DT "linux,pci-probe-only" in a
generic pci_preserve_config() path (Vidya Sagar)
* pci/enumeration:
PCI: Use preserve_config in place of pci_flags
PCI: Unify ACPI and DT 'preserve config' support
PCI: of: Add of_pci_preserve_config() for per-host bridge support
PCI: Move PRESERVE_BOOT_CONFIG _DSM evaluation to pci_register_host_bridge()
Thanks to preceding cleanup steps, pcim_release() is now not needed
anymore and can be replaced by pcim_disable_device(), which is the exact
counterpart to pcim_enable_device().
This permits removing further parts of the old PCI devres implementation.
Replace pcim_release() with pcim_disable_device(). Remove the now unused
function get_pci_dr(). Remove the struct pci_devres from pci.h.
Link: https://lore.kernel.org/r/20240613115032.29098-12-pstanner@redhat.com
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The bit describing whether the PCI device is currently pinned is stored
in struct pci_devres. To clean up and simplify the PCI devres API, it's
better if this information is stored in struct pci_dev.
This will later permit simplifying pcim_enable_device().
Move the 'pinned' boolean bit to struct pci_dev.
Restructure bits in struct pci_dev so the pm / pme fields are next to
each other.
Link: https://lore.kernel.org/r/20240613115032.29098-9-pstanner@redhat.com
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The struct pci_devres has a separate boolean to track whether a device is
enabled. That, however, can easily be tracked in an agnostic manner through
the function pci_is_enabled().
Using it allows for simplifying the PCI devres implementation.
Replace the separate 'enabled' status bit from struct pci_devres with
calls to pci_is_enabled() at the appropriate places.
Link: https://lore.kernel.org/r/20240613115032.29098-8-pstanner@redhat.com
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
These existing functions:
pci_request_region()
pci_request_selected_regions()
pci_request_selected_regions_exclusive()
are "hybrid" functions built on __pci_request_region() and are managed if
pcim_enable_device() has been called, but unmanaged otherwise.
Add these new functions:
pcim_request_region()
pcim_request_region_exclusive()
These are *always* managed and use the new pcim_addr_devres tracking
infrastructure instead of find_pci_dr() and struct pci_devres.region_mask.
Implement the hybrid functions using the new "pure" functions and remove
struct pci_devres.region_mask, which is no longer needed.
Link: https://lore.kernel.org/r/20240613115032.29098-6-pstanner@redhat.com
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
[bhelgaas: commit log]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The pcim_iomap_devres table tracks entire-BAR mappings, so we can't use it
to build a managed version of pci_iomap_range(), which maps partial BARs.
Add struct pcim_addr_devres, which can track request and mapping of both
entire BARs and partial BARs.
Add the following internal devres functions based on struct
pcim_addr_devres:
pcim_iomap_region() # request & map entire BAR
pcim_iounmap_region() # unmap & release entire BAR
pcim_request_region() # request entire BAR
pcim_release_region() # release entire BAR
pcim_request_all_regions() # request all entire BARs
pcim_release_all_regions() # release all entire BARs
Rework the following public interfaces using the new infrastructure
listed above:
pcim_iomap() # map partial BAR
pcim_iounmap() # unmap partial BAR
pcim_iomap_regions() # request & map specified BARs
pcim_iomap_regions_request_all() # request all BARs, map specified BARs
pcim_iounmap_regions() # unmap & release specified BARs
Link: https://lore.kernel.org/r/20240613115032.29098-4-pstanner@redhat.com
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
[bhelgaas: commit log]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
PCIe r6.0, sec 6.6.1, states that the host should wait for at least 100
msec from the end of a conventional reset (PERST# is de-asserted) before
sending a configuration request to ensure that the device is able to
respond with a "Request Retry Status" completion.
Add the PCIE_T_RRS_READY_MS macro to define this wait time and modify
rockchip_pcie_host_init_port() to add this 100ms sleep after deasserting
PERST# using the ep_gpio GPIO.
Link: https://lore.kernel.org/linux-pci/20240413004120.1099089-3-dlemoal@kernel.org
Suggested-by: Bjorn Helgaas <helgaas@kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Add of_pci_preserve_config() to look for the "linux,pci-probe-only"
property under a specified node. If it's not found there, look under
"of_chosen" in addition.
If the caller didn't specify a node, look under "of_chosen".
With a future patch, this will support "linux,pci-probe-only" on a per host
bridge basis based on the presence of the property in the respective PCI
host bridge DT node.
Implement of_pci_check_probe_only() using of_pci_preserve_config().
Link: https://lore.kernel.org/r/20240508174138.3630283-3-vidyas@nvidia.com
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Move the PRESERVE_BOOT_CONFIG _DSM evaluation from acpi_pci_root_create()
to pci_register_host_bridge().
This will help unify the ACPI _DSM path and the DT-based
"linux,pci-probe-only" paths.
This should be safe because it happens earlier than it used to:
acpi_pci_root_create
pci_create_root_bus
pci_register_host_bridge
+ bridge->preserve_config = pci_preserve_config(bridge)
pci_acpi_preserve_config
+ acpi_evaluate_dsm_typed(DSM_PCI_PRESERVE_BOOT_CONFIG)
- acpi_evaluate_dsm_typed(DSM_PCI_PRESERVE_BOOT_CONFIG)
No functional change intended.
Link: https://lore.kernel.org/r/20240508174138.3630283-2-vidyas@nvidia.com
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Add the PCIE_RESET_CONFIG_DEVICE_WAIT_MS macro to define the minimum
waiting time between exit from a conventional reset and sending the
first configuration request to the device.
As described in PCIe r6.0, sec 6.6.1 <Conventional Reset>, there are two
different use cases of the value:
- "With a Downstream Port that does not support Link speeds greater
than 5.0 GT/s, software must wait a minimum of 100 ms following exit
from a Conventional Reset before sending a Configuration Request to
the device immediately below that Port."
- "With a Downstream Port that supports Link speeds greater than
5.0 GT/s, software must wait a minimum of 100 ms after Link training
completes before sending a Configuration Request to the device
immediately below that Port."
[kwilczynski: commit log]
Link: https://lore.kernel.org/linux-pci/20240328091835.14797-21-minda.chen@starfivetech.com
Signed-off-by: Kevin Xie <kevin.xie@starfivetech.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mason Huo <mason.huo@starfivetech.com>
Pull PCI updates from Bjorn Helgaas:
"Enumeration:
- Consolidate interrupt related code in irq.c (Ilpo Järvinen)
- Reduce kernel size by replacing sysfs resource macros with
functions (Ilpo Järvinen)
- Reduce kernel size by compiling sysfs support only when
CONFIG_SYSFS=y (Lukas Wunner)
- Avoid using Extended Tags on 3ware-9650SE Root Port to work around
an apparent hardware defect (Jörg Wedekind)
Resource management:
- Fix an MMIO mapping leak in pci_iounmap() (Philipp Stanner)
- Move pci_iomap.c and other PCI-specific devres code to drivers/pci
(Philipp Stanner)
- Consolidate PCI devres code in devres.c (Philipp Stanner)
Power management:
- Avoid D3cold on Asus B1400 PCI-NVMe bridge, where firmware doesn't
know how to return correctly to D0, and remove previous quirk that
wasn't as specific (Daniel Drake)
- Allow runtime PM when the driver enables it but doesn't need any
runtime PM callbacks (Raag Jadav)
- Drain runtime-idle callbacks before driver removal to avoid races
between .remove() and .runtime_idle(), which caused intermittent
page faults when the rtsx .runtime_idle() accessed registers that
its .remove() had already unmapped (Rafael J. Wysocki)
Virtualization:
- Avoid Secondary Bus Reset on LSI FW643 so it can be assigned to VMs
with VFIO, e.g., for professional audio software on many Apple
machines, at the cost of leaking state between VMs (Edmund Raile)
Error handling:
- Print all logged TLP Prefixes, not just the first, after AER or DPC
errors (Ilpo Järvinen)
- Quirk the DPC PIO log size for Intel Raptor Lake Root Ports, which
still don't advertise a legal size (Paul Menzel)
- Ignore expected DPC Surprise Down errors on hot removal (Smita
Koralahalli)
- Block runtime suspend while handling AER errors to avoid races that
prevent the device form being resumed from D3hot (Stanislaw
Gruszka)
Peer-to-peer DMA:
- Use atomic XA allocation in RCU read section (Christophe JAILLET)
ASPM:
- Collect bits of ASPM-related code that we need even without
CONFIG_PCIEASPM into aspm.c (David E. Box)
- Save/restore L1 PM Substates config for suspend/resume (David E.
Box)
- Update save_save when ASPM config is changed, so a .slot_reset()
during error recovery restores the changed config, not the
.probe()-time config (Vidya Sagar)
Endpoint framework:
- Refactor and improve pci_epf_alloc_space() API (Niklas Cassel)
- Clean up endpoint BAR descriptions (Niklas Cassel)
- Fix ntb_register_device() name leak in error path (Yang Yingliang)
- Return actual error code for pci_vntb_probe() failure (Yang
Yingliang)
Broadcom STB PCIe controller driver:
- Fix MDIO write polling, which previously never waited for
completion (Jonathan Bell)
Cadence PCIe endpoint driver:
- Clear the ARI "Next Function Number" of last function (Jasko-EXT
Wojciech)
Freescale i.MX6 PCIe controller driver:
- Simplify by replacing switch statements with function pointers for
different hardware variants (Frank Li)
- Simplify by using clk_bulk*() API (Frank Li)
- Remove redundant DT clock and reg/reg-name details (Frank Li)
- Add i.MX95 DT and driver support for both Root Complex and Endpoint
mode (Frank Li)
Microsoft Hyper-V host bridge driver:
- Reduce memory usage by limiting ring buffer size to 16KB instead of
4 pages (Michael Kelley)
Qualcomm PCIe controller driver:
- Add X1E80100 DT and driver support (Abel Vesa)
- Add DT 'required-opps' for SoCs that require a minimum performance
level (Johan Hovold)
- Make DT 'msi-map-mask' optional, depending on how MSI interrupts
are mapped (Johan Hovold)
- Disable ASPM L0s for sc8280xp, sa8540p and sa8295p because the PHY
configuration isn't tuned correctly for L0s (Johan Hovold)
- Split dt-binding qcom,pcie.yaml into qcom,pcie-common.yaml and
separate files for SA8775p, SC7280, SC8180X, SC8280XP, SM8150,
SM8250, SM8350, SM8450, SM8550 for easier reviewing (Krzysztof
Kozlowski)
- Enable BDF to SID translation by disabling bypass mode (Manivannan
Sadhasivam)
- Add endpoint MHI support for Snapdragon SA8775P SoC (Mrinmay
Sarkar)
Synopsys DesignWare PCIe controller driver:
- Allocate 64-bit MSI address if no 32-bit address is available (Ajay
Agarwal)
- Fix endpoint Resizable BAR to actually advertise the required 1MB
size (Niklas Cassel)
MicroSemi Switchtec management driver:
- Release resources if the .probe() fails (Christophe JAILLET)
Miscellaneous:
- Make pcie_port_bus_type const (Ricardo B. Marliere)"
* tag 'pci-v6.9-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci: (77 commits)
PCI/ASPM: Update save_state when configuration changes
PCI/ASPM: Disable L1 before configuring L1 Substates
PCI/ASPM: Call pci_save_ltr_state() from pci_save_pcie_state()
PCI/ASPM: Save L1 PM Substates Capability for suspend/resume
PCI: hv: Fix ring buffer size calculation
PCI: dwc: endpoint: Fix advertised resizable BAR size
PCI: cadence: Clear the ARI Capability Next Function Number of the last function
PCI: dwc: Strengthen the MSI address allocation logic
PCI: brcmstb: Fix broken brcm_pcie_mdio_write() polling
PCI: qcom: Add X1E80100 PCIe support
dt-bindings: PCI: qcom: Document the X1E80100 PCIe Controller
PCI: qcom: Enable BDF to SID translation properly
PCI/AER: Generalize TLP Header Log reading
PCI/AER: Use explicit register size for PCI_ERR_CAP
PCI: qcom: Disable ASPM L0s for sc8280xp, sa8540p and sa8295p
dt-bindings: PCI: qcom: Do not require 'msi-map-mask'
dt-bindings: PCI: qcom: Allow 'required-opps'
PCI/AER: Block runtime suspend when handling errors
PCI/ASPM: Move pci_save_ltr_state() to aspm.c
PCI/ASPM: Always build aspm.c
...
- Compile pci-sysfs.c only if CONFIG_SYSFS=y, which reduces kernel size by
~120KB when it's disabled (Lukas Wunner)
- Remove obsolete pci_cleanup_rom() declaration (Lukas Wunner)
- Rework pci_dev_resource_resize_attr(n) macros to call a function instead
of duplicating most of the body, which saves about 2.5KB of text (Ilpo
Järvinen)
* pci/sysfs:
PCI/sysfs: Demacrofy pci_dev_resource_resize_attr(n) functions
PCI: Remove obsolete pci_cleanup_rom() declaration
PCI/sysfs: Compile pci-sysfs.c only if CONFIG_SYSFS=y
# Conflicts:
# drivers/pci/Makefile
- Unmap MMIO mappings in pci_iounmap() to avoid a leak when
ARCH_HAS_GENERIC_IOPORT_MAP is defined (Philipp Stanner)
- Move pci_iomap.c to drivers/pci/ since it's all PCI-related (Philipp
Stanner)
- Move other PCI-related devres code from lib/devres.c to drivers/pci/
(Philipp Stanner)
- Move other devres code from pci.c to devres.c (Philipp Stanner)
* pci/devres:
PCI: Move devres code from pci.c to devres.c
PCI: Move PCI-specific devres code to drivers/pci/
PCI: Move pci_iomap.c to drivers/pci/
pci_iounmap(): Fix MMIO mapping leak
- Collect ASPM-related code into aspm.c (David E. Box)
- Save and restore ASPM L1 PM Substates configuration so these states
continue working after suspend/resume (David E. Box)
- Move the ASPM L1.2-related LTR save/restore next to the ASPM save/restore
(David E. Box)
- Move the required L1 disable before L1 Substate configuration into
pci_restore_aspm_l1ss_state() (Bjorn Helgaas)
- Update save_save when ASPM config is changed, so a .slot_reset() during
error recovery restores the changed config, not the .probe()-time config
(Vidya Sagar)
* pci/aspm:
PCI/ASPM: Update save_state when configuration changes
PCI/ASPM: Disable L1 before configuring L1 Substates
PCI/ASPM: Call pci_save_ltr_state() from pci_save_pcie_state()
PCI/ASPM: Save L1 PM Substates Capability for suspend/resume
PCI/ASPM: Move pci_save_ltr_state() to aspm.c
PCI/ASPM: Always build aspm.c
PCI/ASPM: Move pci_configure_ltr() to aspm.c
Both AER and DPC RP PIO provide TLP Header Log registers (PCIe r6.1 secs
7.8.4 & 7.9.14) to convey error diagnostics but the struct is named after
AER as the struct aer_header_log_regs. Also, not all places that handle TLP
Header Log use the struct and the struct members are named individually.
Generalize the struct name and members, and use it consistently where TLP
Header Log is being handled so that a pcie_read_tlp_log() helper can be
easily added.
Link: https://lore.kernel.org/r/20240206135717.8565-3-ilpo.jarvinen@linux.intel.com
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
[bhelgaas: drop ixgbe changes for now, tidy whitespace]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
It is possible to enable CONFIG_PCI but disable CONFIG_SYSFS and for
space-constrained devices such as routers, such a configuration may
actually make sense.
However pci-sysfs.c is compiled even if CONFIG_SYSFS is disabled,
unnecessarily increasing the kernel's size.
To rectify that:
* Move pci_mmap_fits() to mmap.c. It is not only needed by
pci-sysfs.c, but also proc.c.
* Move pci_dev_type to probe.c and make it private. It references
pci_dev_attr_groups in pci-sysfs.c. Make that public instead for
consistency with pci_dev_groups, pcibus_groups and pci_bus_groups,
which are likewise public and referenced by struct definitions in
pci-driver.c and probe.c.
* Define pci_dev_groups, pci_dev_attr_groups, pcibus_groups and
pci_bus_groups to NULL if CONFIG_SYSFS is disabled. Provide empty
static inlines for pci_{create,remove}_legacy_files() and
pci_{create,remove}_sysfs_dev_files().
Result:
vmlinux size is reduced by 122996 bytes in my arm 32-bit test build.
Link: https://lore.kernel.org/r/85ca95ae8e4d57ccf082c5c069b8b21eb141846e.1698668982.git.lukas@wunner.de
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
The file pci.c is very large and contains a number of devres functions.
These functions should now reside in devres.c.
Move as much devres-specific code from pci.c to devres.c as possible.
There are a few callers left in pci.c that do devres operations. These
should be ported in the future. Add corresponding TODOs.
The reason they are not moved right now in this commit is that PCI's devres
currently implements a sort of "hybrid-mode": pci_request_region(), for
instance, does not have a corresponding pcim_ equivalent, yet. Instead, the
function can be made managed by previously calling pcim_enable_device()
(instead of pci_enable_device()). This makes it unreasonable to move
pci_request_region() to devres.c. Moving the functions would require
changes to PCI's API and is, therefore, left for future work.
In summary, this commit serves as a preparation step for a following
patch series that will cleanly separate the PCI's managed and unmanaged
API.
Link: https://lore.kernel.org/r/20240131090023.12331-5-pstanner@redhat.com
Suggested-by: Danilo Krummrich <dakr@redhat.com>
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
A last minute revert in 6.7-final introduced a potential deadlock when
enabling ASPM during probe of Qualcomm PCIe controllers as reported by
lockdep:
============================================
WARNING: possible recursive locking detected
6.7.0 #40 Not tainted
--------------------------------------------
kworker/u16:5/90 is trying to acquire lock:
ffffacfa78ced000 (pci_bus_sem){++++}-{3:3}, at: pcie_aspm_pm_state_change+0x58/0xdc
but task is already holding lock:
ffffacfa78ced000 (pci_bus_sem){++++}-{3:3}, at: pci_walk_bus+0x34/0xbc
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(pci_bus_sem);
lock(pci_bus_sem);
*** DEADLOCK ***
Call trace:
print_deadlock_bug+0x25c/0x348
__lock_acquire+0x10a4/0x2064
lock_acquire+0x1e8/0x318
down_read+0x60/0x184
pcie_aspm_pm_state_change+0x58/0xdc
pci_set_full_power_state+0xa8/0x114
pci_set_power_state+0xc4/0x120
qcom_pcie_enable_aspm+0x1c/0x3c [pcie_qcom]
pci_walk_bus+0x64/0xbc
qcom_pcie_host_post_init_2_7_0+0x28/0x34 [pcie_qcom]
The deadlock can easily be reproduced on machines like the Lenovo ThinkPad
X13s by adding a delay to increase the race window during asynchronous
probe where another thread can take a write lock.
Add a new pci_set_power_state_locked() and associated helper functions that
can be called with the PCI bus semaphore held to avoid taking the read lock
twice.
Link: https://lore.kernel.org/r/ZZu0qx2cmn7IwTyQ@hovoldconsulting.com
Link: https://lore.kernel.org/r/20240130100243.11011-1-johan+linaro@kernel.org
Fixes: f93e71aea6 ("Revert "PCI/ASPM: Remove pcie_aspm_pm_state_change()"")
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: <stable@vger.kernel.org> # 6.7
