Florian Fainelli says:
====================
Rework GENET MDIO controller clocking
This patch series reworks the way that we manage the GENET MDIO
controller clocks around I/O accesses. During testing with a fully
modular build where bcmgenet, mdio-bcm-unimac, and the Broadcom PHY
driver (broadcom) are all loaded as modules, with no particular care
being taken to order them to mimize deferred probing the following bus
error was obtained:
[ 4.344831] printk: console [ttyS0] enabled
[ 4.351102] 840d000.serial: ttyS1 at MMIO 0x840d000 (irq = 29, base_baud = 5062500) is a Broadcom BCM7271 UART
[ 4.363110] 840e000.serial: ttyS2 at MMIO 0x840e000 (irq = 30, base_baud = 5062500) is a Broadcom BCM7271 UART
[ 4.387392] iproc-rng200 8402000.rng: hwrng registered
[ 4.398012] Consider using thermal netlink events interface
[ 4.403717] brcmstb_thermal a581500.thermal: registered AVS TMON of-sensor driver
[ 4.440085] bcmgenet 8f00000.ethernet: GENET 5.0 EPHY: 0x0000
[ 4.482526] unimac-mdio unimac-mdio.0: Broadcom UniMAC MDIO bus
[ 4.514019] bridge: filtering via arp/ip/ip6tables is no longer available by default. Update your scripts to load br_netfilter if you need this.
[ 4.551304] SError Interrupt on CPU2, code 0x00000000bf000002 -- SError
[ 4.551324] CPU: 2 PID: 8 Comm: kworker/u8:0 Not tainted 6.1.53-0.1pre-g5a26d98e908c #2
[ 4.551330] Hardware name: BCM972180HB_V20 (DT)
[ 4.551336] Workqueue: events_unbound deferred_probe_work_func
[ 4.551363] pstate: 00000005 (nzcv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 4.551368] pc : el1_abort+0x2c/0x58
[ 4.551376] lr : el1_abort+0x20/0x58
[ 4.551379] sp : ffffffc00a383960
[ 4.551380] x29: ffffffc00a383960 x28: ffffff80029fd780 x27: 0000000000000000
[ 4.551385] x26: 0000000000000000 x25: ffffff8002839005 x24: ffffffc00a1f9bd0
[ 4.551390] x23: 0000000040000005 x22: ffffffc000a48084 x21: ffffffc00a3dde14
[ 4.551394] x20: 0000000096000210 x19: ffffffc00a3839a0 x18: 0000000000000579
[ 4.551399] x17: 0000000000000000 x16: 0000000100000000 x15: ffffffc00a3838c0
[ 4.551403] x14: 000000000000000a x13: 6e69622f7273752f x12: 3a6e6962732f7273
[ 4.551408] x11: 752f3a6e69622f3a x10: 6e6962732f3d4854 x9 : ffffffc0086466a8
[ 4.551412] x8 : ffffff80049ee100 x7 : ffffff8003231938 x6 : 0000000000000000
[ 4.551416] x5 : 0000002200000000 x4 : ffffffc00a3839a0 x3 : 0000002000000000
[ 4.551420] x2 : 0000000000000025 x1 : 0000000096000210 x0 : 0000000000000000
[ 4.551429] Kernel panic - not syncing: Asynchronous SError Interrupt
[ 4.551432] CPU: 2 PID: 8 Comm: kworker/u8:0 Not tainted 6.1.53-0.1pre-g5a26d98e908c #2
[ 4.551435] Hardware name: BCM972180HB_V20 (DT)
[ 4.551437] Workqueue: events_unbound deferred_probe_work_func
[ 4.551443] Call trace:
[ 4.551445] dump_backtrace+0xe4/0x124
[ 4.551452] show_stack+0x1c/0x28
[ 4.551455] dump_stack_lvl+0x60/0x78
[ 4.551462] dump_stack+0x14/0x2c
[ 4.551467] panic+0x134/0x304
[ 4.551472] nmi_panic+0x50/0x70
[ 4.551480] arm64_serror_panic+0x70/0x7c
[ 4.551484] do_serror+0x2c/0x5c
[ 4.551487] el1h_64_error_handler+0x2c/0x40
[ 4.551491] el1h_64_error+0x64/0x68
[ 4.551496] el1_abort+0x2c/0x58
[ 4.551499] el1h_64_sync_handler+0x8c/0xb4
[ 4.551502] el1h_64_sync+0x64/0x68
[ 4.551505] unimac_mdio_readl.isra.0+0x4/0xc [mdio_bcm_unimac]
[ 4.551519] __mdiobus_read+0x2c/0x88
[ 4.551526] mdiobus_read+0x40/0x60
[ 4.551530] phy_read+0x18/0x20
[ 4.551534] bcm_phy_config_intr+0x20/0x84
[ 4.551537] phy_disable_interrupts+0x2c/0x3c
[ 4.551543] phy_probe+0x80/0x1b0
[ 4.551545] really_probe+0x1b8/0x390
[ 4.551550] __driver_probe_device+0x134/0x14c
[ 4.551554] driver_probe_device+0x40/0xf8
[ 4.551559] __device_attach_driver+0x108/0x11c
[ 4.551563] bus_for_each_drv+0xa4/0xcc
[ 4.551567] __device_attach+0xdc/0x190
[ 4.551571] device_initial_probe+0x18/0x20
[ 4.551575] bus_probe_device+0x34/0x94
[ 4.551579] deferred_probe_work_func+0xd4/0xe8
[ 4.551583] process_one_work+0x1ac/0x25c
[ 4.551590] worker_thread+0x1f4/0x260
[ 4.551595] kthread+0xc0/0xd0
[ 4.551600] ret_from_fork+0x10/0x20
[ 4.551608] SMP: stopping secondary CPUs
[ 4.551617] Kernel Offset: disabled
[ 4.551619] CPU features: 0x00000,00c00080,0000420b
[ 4.551622] Memory Limit: none
[ 4.833838] ---[ end Kernel panic - not syncing: Asynchronous SError Interrupt ]---
The issue here is that we managed to probe the GENET controller, the
mdio-bcm-unimac MDIO controller, but the PHY was still being held in a
probe deferral state because it depended upon a GPIO controller provider
not loaded yet. As soon as that provider is loaded however, the PHY
continues to probe, tries to disable the interrupts, and this causes a
MDIO transaction. That MDIO transaction requires I/O register accesses
within the GENET's larger block, and since its clocks are turned off,
the CPU gets a bus error signaled as a System Error.
The patch series takes the simplest approach of keeping the clocks
enabled just for the duration of the I/O accesses. This is also
beneficial to other drivers like bcmasp2 which make use of the same MDIO
controller driver.
Changes in v2:
- added missing ret assignment in the if (IS_ERR(priv->clk)) branch
- added Jacob's R-by tags
- corrected the commit ID being reverted in patch #3
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 1b5ea7ffb7 ("net:
bcmgenet: Ensure MDIO unregistration has clocks enabled"). This is no
longer necessary now that the MDIO bus controller has a clock that it
can manage around the I/O accesses.
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
GENET has historically had to create a MDIO platform device for its
controller and pass some auxiliary data to it, like a MDIO completion
callback. Now we also pass the "main" clock to allow for the MDIO bus
controller to manage that clock adequately around I/O accesses.
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Up until now we have managed not to have the mdio-bcm-unimac manage its
clock except during probe and suspend/resume. This works most of the
time, except where it does not.
With a fully modular build, we can get into a situation whereby the
GENET driver is fully registered, and so is the mdio-bcm-unimac driver,
however the Ethernet PHY driver is not yet, because it depends on a
resource that is not yet available (e.g.: GPIO provider). In that state,
the network device is not usable yet, and so to conserve power, the
GENET driver will have turned off its "main" clock which feeds its MDIO
controller.
When the PHY driver finally probes however, we make an access to the PHY
registers to e.g.: disable interrupts, and this causes a bus error
within the MDIO controller space because the MDIO controller clock(s)
are turned off.
To remedy that, we manage the clock around all of the I/O accesses to
the hardware which are done exclusively during read, write and clock
divider configuration.
This ensures that the register space is accessible, and this also
ensures that there are not unnecessarily elevated reference counts
keeping the clocks active when the network device is administratively
turned off. It would be the case with the previous way of managing the
clock.
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove documentation of non-existent children field
from the Kernel doc for struct framer_ops.
Introduced by 82c944d05b ("net: wan: Add framer framework support")
Signed-off-by: Simon Horman <horms@kernel.org>
Acked-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Kalle Valo says:
====================
wireless-next patches for v6.9
The second "new features" pull request for v6.9. Lots of iwlwifi and
stack changes this time. And naturally smaller changes to other drivers.
We also twice merged wireless into wireless-next to avoid conflicts
between the trees.
Major changes:
stack
* mac80211: negotiated TTLM request support
* SPP A-MSDU support
* mac80211: wider bandwidth OFDMA config support
iwlwifi
* kunit tests
* bump FW API to 89 for AX/BZ/SC devices
* enable SPP A-MSDUs
* support for new devices
ath12k
* refactoring in preparation for Multi-Link Operation (MLO) support
* 1024 Block Ack window size support
* provide firmware wmi logs via a trace event
ath11k
* 36 bit DMA mask support
* support 6 GHz station power modes: Low Power Indoor (LPI), Standard
Power) SP and Very Low Power (VLP)
rtl8xxxu
* TP-Link TL-WN823N V2 support
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Kunwu Chan says:
====================
net: Use KMEM_CACHE instead of kmem_cache_create
As Jiri Pirko suggests,
I'm using a patchset to cleanup the same issues in the 'net' module.
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create
to simplify the creation of SLAB caches.
Some cache names are changed to be the same as struct names.
This change is recorded in the changelog for easy reference.
It's harmless cause it's used in /proc/slabinfo to identify this cache.
---
Changes in v2:
- Delete a patch as Eric said in https://lore.kernel.org/all/CANn89iLkWvum6wSqSya_K+1eqnFvp=L2WLW=kAYrZTF8Ei4b7g@mail.gmail.com/
- No code changes,only add Reviewed-by tag
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create
to simplify the creation of SLAB caches.
And change cache name from 'ip_dst_cache' to 'rtable'.
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create
to simplify the creation of SLAB caches.
And change cache name from 'ip_mrt_cache' to 'mfc_cache'.
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create
to simplify the creation of SLAB caches.
And change cache name from 'ip6_mrt_cache' to 'mfc6_cache'.
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create
to simplify the creation of SLAB caches.
And change cache name from 'kcm_mux_cache' to 'kcm_mux',
'kcm_psock_cache' to 'kcm_psock'.
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With commit 34d21de99c ("net: Move {l,t,d}stats allocation to core and
convert veth & vrf"), stats allocation could be done on net core instead
of this driver.
With this new approach, the driver doesn't have to bother with error
handling (allocation failure checking, making sure free happens in the
right spot, etc). This is core responsibility now.
Move dummy driver to leverage the core allocation.
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The current code overwrites fields in tp->eee with unchecked data from
edata, e.g. the bitmap with supported modes. ethtool properly returns
the received data from get_eee() call, but we have no guarantee that
other users of the ioctl set_eee() interface behave properly too.
Therefore copy only fields which are actually needed.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ricardo B. Marliere says:
====================
net: constify struct device_type usage
This is a simple and straight forward cleanup series that makes all device
types in the net subsystem constants. This has been possible since 2011 [1]
but not all occurrences were cleaned. I have been sweeping the tree to fix
them all.
I was not sure if I should send these squashed, but there are quite a few
changes so I decided to send them separately. Please let me know if that is
not desirable.
[1] https://lore.kernel.org/all/1305850262-9575-5-git-send-email-gregkh@suse.de/
====================
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the hso_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the wwan_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the
nsim_bus_dev_type variable to be a constant structure as well, placing it
into read-only memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the vlan_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the l2tpeth_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the hsr_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the geneve_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the ppp_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the vxlan_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the br_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the dsa_type
variable to be a constant structure as well, placing it into read-only
memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the wlan_type
and wwan_type variables to be constant structures as well, placing it into
read-only memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Marc Kleine-Budde says:
====================
pull-request: can-next 2024-02-20
this is a pull request of 9 patches for net-next/master.
The first patch is by Francesco Dolcini and removes a redundant check
for pm_clock_support from the m_can driver.
Martin Hundebøll contributes 3 patches to the m_can/tcan4x5x driver to
allow resume upon RX of a CAN frame.
3 patches by Srinivas Goud add support for ECC statistics to the
xilinx_can driver.
The last 2 patches are by Oliver Hartkopp and me, target the CAN RAW
protocol and fix an error in the getsockopt() for CAN-XL introduced in
the previous pull request to net-next (linux-can-next-for-6.9-20240213).
linux-can-next-for-6.9-20240220
* tag 'linux-can-next-for-6.9-20240220' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next:
can: raw: raw_getsockopt(): reduce scope of err
can: raw: fix getsockopt() for new CAN_RAW_XL_VCID_OPTS
can: xilinx_can: Add ethtool stats interface for ECC errors
can: xilinx_can: Add ECC support
dt-bindings: can: xilinx_can: Add 'xlnx,has-ecc' optional property
can: tcan4x5x: support resuming from rx interrupt signal
can: m_can: allow keeping the transceiver running in suspend
dt-bindings: can: tcan4x5x: Document the wakeup-source flag
can: m_can: remove redundant check for pm_clock_support
====================
Link: https://lore.kernel.org/r/20240220085130.2936533-1-mkl@pengutronix.de
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
syzbot managed to trigger following splat:
BUG: KASAN: use-after-free in __skb_flow_dissect+0x4a3b/0x5e50
Read of size 1 at addr ffff888208a4000e by task a.out/2313
[..]
__skb_flow_dissect+0x4a3b/0x5e50
__skb_get_hash+0xb4/0x400
ip_tunnel_xmit+0x77e/0x26f0
ipip_tunnel_xmit+0x298/0x410
..
Analysis shows that the skb has a valid ->head, but bogus ->data
pointer.
skb->data gets its bogus value via the neigh layer, which does:
1556 __skb_pull(skb, skb_network_offset(skb));
... and the skb was already dodgy at this point:
skb_network_offset(skb) returns a negative value due to an
earlier overflow of skb->network_header (u16). __skb_pull thus
"adjusts" skb->data by a huge offset, pointing outside skb->head
area.
Allow debug builds to splat when we try to pull/push more than
INT_MAX bytes.
After this, the syzkaller reproducer yields a more precise splat
before the flow dissector attempts to read off skb->data memory:
WARNING: CPU: 5 PID: 2313 at include/linux/skbuff.h:2653 neigh_connected_output+0x28e/0x400
ip_finish_output2+0xb25/0xed0
iptunnel_xmit+0x4ff/0x870
ipgre_xmit+0x78e/0xbb0
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20240216113700.23013-1-fw@strlen.de
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The variable len being initialized with a value that is never read, an
if statement is initializing it in both paths of the if statement.
The initialization is redundant and can be removed.
Cleans up clang scan build warning:
net/ipv4/tcp_ao.c:512:11: warning: Value stored to 'len' during its
initialization is never read [deadcode.DeadStores]
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Link: https://lore.kernel.org/r/20240216125443.2107244-1-colin.i.king@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Mina Almasry says:
====================
Abstract page from net stack
This series is a prerequisite to the devmem TCP series. For a full
snapshot of the code which includes these changes, feel free to check:
https://github.com/mina/linux/commits/tcpdevmem-rfcv5/
Currently these components in the net stack use the struct page
directly:
1. Drivers.
2. Page pool.
3. skb_frag_t.
To add support for new (non struct page) memory types to the net stack, we
must first abstract the current memory type.
Originally the plan was to reuse struct page* for the new memory types,
and to set the LSB on the page* to indicate it's not really a page.
However, for safe compiler type checking we need to introduce a new type.
struct netmem is introduced to abstract the underlying memory type.
Currently it's a no-op abstraction that is always a struct page underneath.
In parallel there is an undergoing effort to add support for devmem to the
net stack:
https://lore.kernel.org/netdev/20231208005250.2910004-1-almasrymina@google.com/
Cc: Jason Gunthorpe <jgg@nvidia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Yunsheng Lin <linyunsheng@huawei.com>
Cc: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
====================
Link: https://lore.kernel.org/r/20240214223405.1972973-1-almasrymina@google.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Use struct netmem* instead of page in skb_frag_t. Currently struct
netmem* is always a struct page underneath, but the abstraction
allows efforts to add support for skb frags not backed by pages.
There is unfortunately 1 instance where the skb_frag_t is assumed to be
a exactly a bio_vec in kcm. For this case, WARN_ON_ONCE and return error
before doing a cast.
Add skb[_frag]_fill_netmem_*() and skb_add_rx_frag_netmem() helpers so
that the API can be used to create netmem skbs.
Signed-off-by: Mina Almasry <almasrymina@google.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Add the netmem_ref type, an abstraction for network memory.
To add support for new memory types to the net stack, we must first
abstract the current memory type. Currently parts of the net stack
use struct page directly:
- page_pool
- drivers
- skb_frag_t
Originally the plan was to reuse struct page* for the new memory types,
and to set the LSB on the page* to indicate it's not really a page.
However, for compiler type checking we need to introduce a new type.
netmem_ref is introduced to abstract the underlying memory type.
Currently it's a no-op abstraction that is always a struct page
underneath. In parallel there is an undergoing effort to add support
for devmem to the net stack:
https://lore.kernel.org/netdev/20231208005250.2910004-1-almasrymina@google.com/
netmem_ref can be pointers to different underlying memory types, and the
low bits are set to indicate the memory type. Helpers are provided
to convert netmem pointers to the underlying memory type (currently only
struct page). In the devmem series helpers are provided so that calling
code can use netmem without worrying about the underlying memory type
unless absolutely necessary.
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Creation of sysfs entries is expensive, mainly for workloads that
constantly creates netdev and netns often.
Do not create BQL sysfs entries for devices that don't need,
basically those that do not have a real queue, i.e, devices that has
NETIF_F_LLTX and IFF_NO_QUEUE, such as `lo` interface.
This will remove the /sys/class/net/eth0/queues/tx-X/byte_queue_limits/
directory for these devices.
In the example below, eth0 has the `byte_queue_limits` directory but not
`lo`.
# ls /sys/class/net/lo/queues/tx-0/
traffic_class tx_maxrate tx_timeout xps_cpus xps_rxqs
# ls /sys/class/net/eth0/queues/tx-0/byte_queue_limits/
hold_time inflight limit limit_max limit_min
This also removes the #ifdefs, since we can also use netdev_uses_bql() to
check if the config is enabled. (as suggested by Jakub).
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://lore.kernel.org/r/20240216094154.3263843-1-leitao@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Now that direct recycling is performed basing on pool->cpuid when set,
memory leaks are possible:
1. A pool is destroyed.
2. Alloc cache is emptied (it's done only once).
3. pool->cpuid is still set.
4. napi_pp_put_page() does direct recycling basing on pool->cpuid.
5. Now alloc cache is not empty, but it won't ever be freed.
In order to avoid that, rewrite pool->cpuid to -1 when unlinking NAPI to
make sure no direct recycling will be possible after emptying the cache.
This involves a bit of overhead as pool->cpuid now must be accessed
via READ_ONCE() to avoid partial reads.
Rename page_pool_unlink_napi() -> page_pool_disable_direct_recycling()
to reflect what it actually does and unexport it.
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/r/20240215113905.96817-1-aleksander.lobakin@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Some code manipulating the vif list is still missing some srcu_read_lock /
srcu_read_unlock, and so can trigger RCU warnings:
=============================
WARNING: suspicious RCU usage
6.8.0-rc1+ #37 Not tainted
-----------------------------
drivers/net/wireless/microchip/wilc1000/hif.c:110 RCU-list traversed without holding the required lock!!
[...]
stack backtrace:
CPU: 0 PID: 6 Comm: kworker/0:0 Not tainted 6.8.0-rc1+ #37
Hardware name: Atmel SAMA5
Workqueue: events sdio_irq_work
unwind_backtrace from show_stack+0x18/0x1c
show_stack from dump_stack_lvl+0x34/0x58
dump_stack_lvl from wilc_get_vif_from_idx+0x158/0x180
wilc_get_vif_from_idx from wilc_network_info_received+0x80/0x48c
wilc_network_info_received from wilc_handle_isr+0xa10/0xd30
wilc_handle_isr from wilc_sdio_interrupt+0x44/0x58
wilc_sdio_interrupt from process_sdio_pending_irqs+0x1c8/0x60c
process_sdio_pending_irqs from sdio_irq_work+0x6c/0x14c
sdio_irq_work from process_one_work+0x8d4/0x169c
process_one_work from worker_thread+0x8cc/0x1340
worker_thread from kthread+0x448/0x510
kthread from ret_from_fork+0x14/0x28
Fix those warnings by adding the needed lock around the corresponding
critical sections
Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Co-developed-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240215-wilc_fix_rcu_usage-v1-4-f610e46c6f82@bootlin.com
Enabling CONFIG_PROVE_RCU_LIST raises many warnings in wilc driver, even on
some places already protected by a read critical section. An example of
such case is in wilc_get_available_idx:
=============================
WARNING: suspicious RCU usage
6.8.0-rc1+ #32 Not tainted
-----------------------------
drivers/net/wireless/microchip/wilc1000/netdev.c:944 RCU-list traversed in non-reader section!!
[...]
stack backtrace:
CPU: 0 PID: 26 Comm: kworker/0:3 Not tainted 6.8.0-rc1+ #32
Hardware name: Atmel SAMA5
Workqueue: events_freezable mmc_rescan
unwind_backtrace from show_stack+0x18/0x1c
show_stack from dump_stack_lvl+0x34/0x58
dump_stack_lvl from wilc_netdev_ifc_init+0x788/0x8ec
wilc_netdev_ifc_init from wilc_cfg80211_init+0x690/0x910
wilc_cfg80211_init from wilc_sdio_probe+0x168/0x490
wilc_sdio_probe from sdio_bus_probe+0x230/0x3f4
sdio_bus_probe from really_probe+0x270/0xdf4
really_probe from __driver_probe_device+0x1dc/0x580
__driver_probe_device from driver_probe_device+0x60/0x140
driver_probe_device from __device_attach_driver+0x268/0x364
__device_attach_driver from bus_for_each_drv+0x15c/0x1cc
bus_for_each_drv from __device_attach+0x1ec/0x3e8
__device_attach from bus_probe_device+0x190/0x1c0
bus_probe_device from device_add+0x10dc/0x18e4
device_add from sdio_add_func+0x1c0/0x2c0
sdio_add_func from mmc_attach_sdio+0xa08/0xe1c
mmc_attach_sdio from mmc_rescan+0xa00/0xfe0
mmc_rescan from process_one_work+0x8d4/0x169c
process_one_work from worker_thread+0x8cc/0x1340
worker_thread from kthread+0x448/0x510
kthread from ret_from_fork+0x14/0x28
This warning is due to the section being protected by a srcu critical read
section, but the list traversal being done with classic RCU API. Fix the
warning by using corresponding SRCU read lock/unlock APIs. While doing so,
since we always manipulate the same list (managed through a pointer
embedded in struct_wilc), add a macro to reduce the corresponding
boilerplate in each call site.
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240215-wilc_fix_rcu_usage-v1-2-f610e46c6f82@bootlin.com
