| Age | Commit message (Collapse) | Author |
|---|
|
Simplify the JIT code by replacing the custom expolines with the ones
defined in the kernel text.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/r/20250519223646.66382-4-iii@linux.ibm.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
After the V!=R rework (commit c98d2ecae08f ("s390/mm: Uncouple physical
vs virtual address spaces")), kernel and BPF programs are allocated
within a 4G region, making it possible to use relative addressing to
directly use kernel functions from BPF code.
Add two new macros for calling kernel functions from BPF code:
EMIT6_PCREL_RILB_PTR() and EMIT6_PCREL_RILC_PTR(). Factor out parts
of the existing macros that are helpful for implementing the new ones.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/r/20250519223646.66382-3-iii@linux.ibm.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
It would be convenient to use the following pattern in the BPF JIT:
if (nospec_uses_trampoline())
emit_call(__s390_indirect_jump_r1);
Unfortunately with CONFIG_EXPOLINE=n the compiler complains about the
missing prototype of __s390_indirect_jump_r1(). One could wrap the
whole "if" statement in an #ifdef, but this clutters the code.
Instead, declare expoline thunk prototypes even when compiling without
expolines. When using the above code structure and compiling without
expolines, references to them are optimized away, and there are no
linker errors.
Acked-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/r/20250519223646.66382-2-iii@linux.ibm.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When xdp is attached or detached, dev->ndo_bpf() is called by
do_setlink(), and it acquires netdev_lock() if needed.
Unlike other drivers, the bnxt driver is protected by netdev_lock while
xdp is attached/detached because it sets dev->request_ops_lock to true.
So, the bnxt_xdp(), that is callback of ->ndo_bpf should not acquire
netdev_lock().
But the xdp_features_{set | clear}_redirect_target() was changed to
acquire netdev_lock() internally.
It causes a deadlock.
To fix this problem, bnxt driver should use
xdp_features_{set | clear}_redirect_target_locked() instead.
Splat looks like:
============================================
WARNING: possible recursive locking detected
6.15.0-rc6+ #1 Not tainted
--------------------------------------------
bpftool/1745 is trying to acquire lock:
ffff888131b85038 (&dev->lock){+.+.}-{4:4}, at: xdp_features_set_redirect_target+0x1f/0x80
but task is already holding lock:
ffff888131b85038 (&dev->lock){+.+.}-{4:4}, at: do_setlink.constprop.0+0x24e/0x35d0
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(&dev->lock);
lock(&dev->lock);
*** DEADLOCK ***
May be due to missing lock nesting notation
3 locks held by bpftool/1745:
#0: ffffffffa56131c8 (rtnl_mutex){+.+.}-{4:4}, at: rtnl_setlink+0x1fe/0x570
#1: ffffffffaafa75a0 (&net->rtnl_mutex){+.+.}-{4:4}, at: rtnl_setlink+0x236/0x570
#2: ffff888131b85038 (&dev->lock){+.+.}-{4:4}, at: do_setlink.constprop.0+0x24e/0x35d0
stack backtrace:
CPU: 1 UID: 0 PID: 1745 Comm: bpftool Not tainted 6.15.0-rc6+ #1 PREEMPT(undef)
Hardware name: ASUS System Product Name/PRIME Z690-P D4, BIOS 0603 11/01/2021
Call Trace:
<TASK>
dump_stack_lvl+0x7a/0xd0
print_deadlock_bug+0x294/0x3d0
__lock_acquire+0x153b/0x28f0
lock_acquire+0x184/0x340
? xdp_features_set_redirect_target+0x1f/0x80
__mutex_lock+0x1ac/0x18a0
? xdp_features_set_redirect_target+0x1f/0x80
? xdp_features_set_redirect_target+0x1f/0x80
? __pfx_bnxt_rx_page_skb+0x10/0x10 [bnxt_en
? __pfx___mutex_lock+0x10/0x10
? __pfx_netdev_update_features+0x10/0x10
? bnxt_set_rx_skb_mode+0x284/0x540 [bnxt_en
? __pfx_bnxt_set_rx_skb_mode+0x10/0x10 [bnxt_en
? xdp_features_set_redirect_target+0x1f/0x80
xdp_features_set_redirect_target+0x1f/0x80
bnxt_xdp+0x34e/0x730 [bnxt_en 11cbcce8fa11cff1dddd7ef358d6219e4ca9add3]
dev_xdp_install+0x3f4/0x830
? __pfx_bnxt_xdp+0x10/0x10 [bnxt_en 11cbcce8fa11cff1dddd7ef358d6219e4ca9add3]
? __pfx_dev_xdp_install+0x10/0x10
dev_xdp_attach+0x560/0xf70
dev_change_xdp_fd+0x22d/0x280
do_setlink.constprop.0+0x2989/0x35d0
? __pfx_do_setlink.constprop.0+0x10/0x10
? lock_acquire+0x184/0x340
? find_held_lock+0x32/0x90
? rtnl_setlink+0x236/0x570
? rcu_is_watching+0x11/0xb0
? trace_contention_end+0xdc/0x120
? __mutex_lock+0x946/0x18a0
? __pfx___mutex_lock+0x10/0x10
? __lock_acquire+0xa95/0x28f0
? rcu_is_watching+0x11/0xb0
? rcu_is_watching+0x11/0xb0
? cap_capable+0x172/0x350
rtnl_setlink+0x2cd/0x570
Fixes: 03df156dd3a6 ("xdp: double protect netdev->xdp_flags with netdev->lock")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20250520071155.2462843-1-ap420073@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Pull NVMe fix from Christoph:
"nvme fixes for Linux 6.15
- do not create the newly added multipath sysfs group for
non-multipath nodes (Nilay Shroff)"
* tag 'nvme-6.15-2025-05-22' of git://git.infradead.org/nvme:
nvme: avoid creating multipath sysfs group under namespace path devices
|
|
With a VRF, ipv4 and ipv6 FIB expression behave differently.
fib daddr . iif oif
Will return the input interface name for ipv4, but the real device
for ipv6. Example:
If VRF device name is tvrf and real (incoming) device is veth0.
First round is ok, both ipv4 and ipv6 will yield 'veth0'.
But in the second round (incoming device will be set to "tvrf"), ipv4
will yield "tvrf" whereas ipv6 returns "veth0" for the second round too.
This makes ipv6 behave like ipv4.
A followup patch will add a test case for this, without this change
it will fail with:
get element inet t fibif6iif { tvrf . dead:1::99 . tvrf }
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FAIL: did not find tvrf . dead:1::99 . tvrf in fibif6iif
Alternatively we could either not do anything at all or change
ipv4 to also return the lower/real device, however, nft (userspace)
doc says "iif: if fib lookup provides a route then check its output
interface is identical to the packets input interface." which is what
the nft fib ipv4 behaviour is.
Fixes: f6d0cbcf09c5 ("netfilter: nf_tables: add fib expression")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
It was located in conntrack_vrf.sh because that already had the VRF bits.
Lets not add to this and move it to nft_fib.sh where this belongs.
No functional changes for the subtest intended.
The subtest is limited, it only covered 'fib oif'
(route output interface query) when the incoming interface is part
of a VRF.
Next we can extend it to cover 'fib type' for VRFs and also check fib
results when there is an unrelated VRF in same netns.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
fib can either lookup the interface id/name of the output interface that
would be used for the given address, or it can check for the type of the
address according to the fib, e.g. local, unicast, multicast and so on.
This can be used to e.g. make a locally configured address only reachable
through its interface.
Example: given eth0:10.1.1.1 and eth1:10.1.2.1 then 'fib daddr type' for
10.1.1.1 arriving on eth1 will be 'local', but 'fib daddr . iif type' is
expected to return 'unicast', whereas 'fib daddr' and 'fib daddr . iif'
are expected to indicate 'local' if such a packet arrives on eth0.
So far nft_fib.sh only covered oif/oifname, not type.
Repeat tests both with default and a policy (ip rule) based setup.
Also try to run all remaining tests even if a subtest has failed.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
only builds
Its now possible to build a kernel that has no support for the classic
xtables get/setsockopt interfaces and builtin tables.
In this case, we have CONFIG_IP6_NF_MANGLE=n and
CONFIG_IP_NF_ARPTABLES=n.
For optstript, the ipv6 code is so small that we can enable it if
netfilter ipv6 support exists. For mark, check if either classic
arptables or NFT_ARP_COMPAT is set.
Fixes: a9525c7f6219 ("netfilter: xtables: allow xtables-nft only builds")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
representation
Pipapo supports a more compact '4 bit group' format that is chosen when
the memory needed for the default exceeds a threshold (2mb).
Add coverage for those code paths, the existing tests use small sets that
are handled by the default representation.
This comes with a test script run-time increase, but I think its ok:
normal: 2m35s -> 3m9s
debug: 3m24s -> 5m29s (with KSFT_MACHINE_SLOW=yes).
Cc: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
If, in a previous transfer, the controller sends more data than expected
by the DSPI target, SR.RFDF (RX FIFO is not empty) will remain asserted.
When flushing the FIFOs at the beginning of a new transfer (writing 1
into MCR.CLR_TXF and MCR.CLR_RXF), SR.RFDF should also be cleared.
Otherwise, when running in target mode with DMA, if SR.RFDF remains
asserted, the DMA callback will be fired before the controller sends any
data.
Take this opportunity to reset all Status Register fields.
Fixes: 5ce3cc567471 ("spi: spi-fsl-dspi: Provide support for DSPI slave mode operation (Vybryd vf610)")
Signed-off-by: Larisa Grigore <larisa.grigore@nxp.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Link: https://patch.msgid.link/20250522-james-nxp-spi-v2-3-bea884630cfb@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
The XSPI mode implementation in this driver still uses the EOQ flag to
signal the last word in a transmission and deassert the PCS signal.
However, at speeds lower than ~200kHZ, the PCS signal seems to remain
asserted even when SR[EOQF] = 1 indicates the end of a transmission.
This is a problem for target devices which require the deassertation of
the PCS signal between transfers.
Hence, this commit 'forces' the deassertation of the PCS by stopping the
module through MCR[HALT] after completing a new transfer. According to
the reference manual, the module stops or transitions from the Running
state to the Stopped state after the current frame, when any one of the
following conditions exist:
- The value of SR[EOQF] = 1.
- The chip is in Debug mode and the value of MCR[FRZ] = 1.
- The value of MCR[HALT] = 1.
This shouldn't be done if the last transfer in the message has cs_change
set.
Fixes: ea93ed4c181b ("spi: spi-fsl-dspi: Use EOQ for last word in buffer even for XSPI mode")
Signed-off-by: Bogdan-Gabriel Roman <bogdan-gabriel.roman@nxp.com>
Signed-off-by: Larisa Grigore <larisa.grigore@nxp.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Link: https://patch.msgid.link/20250522-james-nxp-spi-v2-2-bea884630cfb@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
DSPI registers are NOT continuous, some registers are reserved and
accessing them from userspace will trigger external abort, add regmap
register access table to avoid below abort.
For example on S32G:
# cat /sys/kernel/debug/regmap/401d8000.spi/registers
Internal error: synchronous external abort: 96000210 1 PREEMPT SMP
...
Call trace:
regmap_mmio_read32le+0x24/0x48
regmap_mmio_read+0x48/0x70
_regmap_bus_reg_read+0x38/0x48
_regmap_read+0x68/0x1b0
regmap_read+0x50/0x78
regmap_read_debugfs+0x120/0x338
Fixes: 1acbdeb92c87 ("spi/fsl-dspi: Convert to use regmap and add big-endian support")
Co-developed-by: Xulin Sun <xulin.sun@windriver.com>
Signed-off-by: Xulin Sun <xulin.sun@windriver.com>
Signed-off-by: Larisa Grigore <larisa.grigore@nxp.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Link: https://patch.msgid.link/20250522-james-nxp-spi-v2-1-bea884630cfb@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
When an entity from application B is killed, drm_sched_entity_kill()
removes all jobs belonging to that entity through
drm_sched_entity_kill_jobs_work(). If application A's job depends on a
scheduled fence from application B's job, and that fence is not properly
signaled during the killing process, application A's dependency cannot be
cleared.
This leads to application A hanging indefinitely while waiting for a
dependency that will never be resolved. Fix this issue by ensuring that
scheduled fences are properly signaled when an entity is killed, allowing
dependent applications to continue execution.
Signed-off-by: Lin.Cao <lincao12@amd.com>
Reviewed-by: Philipp Stanner <phasta@kernel.org>
Signed-off-by: Christian König <christian.koenig@amd.com>
Link: https://lore.kernel.org/r/20250515020713.1110476-1-lincao12@amd.com
|
|
Commit 835a0c10d33b ("leds: Rename simple directory to simatic") renames
the driver's directory to drivers/leds/simatic, but misses to adjust the
SIEMENS IPC LED DRIVERS section in MAINTAINERS.
Hence, ./scripts/get_maintainer.pl --self-test=patterns complains about a
broken reference. Adjust the file entry to this renaming.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@redhat.com>
Link: https://lore.kernel.org/r/20250325092159.213463-1-lukas.bulwahn@redhat.com
Signed-off-by: Lee Jones <lee@kernel.org>
|
|
The I2C core now provides a debugfs entry for each client. Let this
driver use it instead of the root directory. Further improvements by
this change: automatic support of multiple instances.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20250318091234.22170-1-wsa+renesas@sang-engineering.com
Signed-off-by: Lee Jones <lee@kernel.org>
|
|
The I2C core now provides a debugfs entry for each client. Let this
driver use it instead of the custom directory in debugfs root. Further
improvements by this change: automatic clean up on removal, support of
multiple instances.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20250318091426.22258-2-wsa+renesas@sang-engineering.com
Signed-off-by: Lee Jones <lee@kernel.org>
|
|
'ib-mfd-gpio-nvmem-6.16', 'ib-mfd-regulator-6.16' and 'ib-mfd-regulator-6.16-1' into ibs-for-mfd-merged
|
|
multi-uprobe attach logic"
This reverts commit 4a8f635a6054.
Althought get_pid_task() internally already calls rcu_read_lock() and
rcu_read_unlock(), the find_vpid() was not.
The documentation for find_vpid() clearly states:
"Must be called with the tasklist_lock or rcu_read_lock() held."
Add proper rcu_read_lock/unlock() to protect the find_vpid().
Fixes: 4a8f635a6054 ("bpf: remove unnecessary rcu_read_{lock,unlock}() in multi-uprobe attach logic")
Reported-by: Xuewen Yan <xuewen.yan@unisoc.com>
Signed-off-by: Di Shen <di.shen@unisoc.com>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20250520054943.5002-1-xuewen.yan@unisoc.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
to_cxlfs() was declared 'inline' in the header but only defined in
drivers/cxl/core/features.c. This has worked because features.c was
the only file using the function and the definition happened to be
available in the same compilation unit.
However, in preparation for a second .c file using the header and
needing to call the function, the inline specifier became an issue.
Sparse flagged the declaration as invalid since 'inline' requires a
visible definition at the point of use.
Defining the function in the header was considered but rejected, as
it depends on internal symbols not visible at that level.
Remove the inline specifier to correct the linkage violation.
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Shiju Jose <shiju.jose@huawei.com>
Link: https://patch.msgid.link/20250521233625.1745849-1-alison.schofield@intel.com
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
|
|
Prior changes ensured that when zpci_release_device() is called and it
removed the zdev from the zpci_list this instance can not be found via
the zpci_list anymore even while allowing re-add of reserved devices.
This only accounts for the overall lifetime and zpci_list addition and
removal, it does not yet prevent concurrent add of a new instance for
the same underlying device. Such concurrent add would subsequently cause
issues such as attempted re-use of the same IOMMU sysfs directory and is
generally undesired.
Introduce a new zpci_add_remove_lock mutex to serialize adding a new
device with removal. Together this ensures that if a struct zpci_dev is
not found in the zpci_list it was either already removed and torn down,
or its removal and tear down is in progress with the
zpci_add_remove_lock held.
Cc: stable@vger.kernel.org
Fixes: a46044a92add ("s390/pci: fix zpci_zdev_put() on reserve")
Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Tested-by: Gerd Bayer <gbayer@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
|
|
The architecture assumes that PCI functions can be removed synchronously
as PCI events are processed. This however clashes with the reference
counting of struct pci_dev which allows device drivers to hold on to a
struct pci_dev reference even as the underlying device is removed. To
bridge this gap commit 2a671f77ee49 ("s390/pci: fix use after free of
zpci_dev") keeps the struct zpci_dev in ZPCI_FN_STATE_RESERVED state
until common code releases the struct pci_dev. Only when all references
are dropped, the struct zpci_dev can be removed and freed.
Later commit a46044a92add ("s390/pci: fix zpci_zdev_put() on reserve")
moved the deletion of the struct zpci_dev from the zpci_list in
zpci_release_device() to the point where the device is reserved. This
was done to prevent handling events for a device that is already being
removed, e.g. when the platform generates both PCI event codes 0x304
and 0x308. In retrospect, deletion from the zpci_list in the release
function without holding the zpci_list_lock was also racy.
A side effect of this handling is that if the underlying device
re-appears while the struct zpci_dev is in the ZPCI_FN_STATE_RESERVED
state, the new and old instances of the struct zpci_dev and/or struct
pci_dev may clash. For example when trying to create the IOMMU sysfs
files for the new instance. In this case, re-adding the new instance is
aborted. The old instance is removed, and the device will remain absent
until the platform issues another event.
Fix this by allowing the struct zpci_dev to be brought back up right
until it is finally removed. To this end also keep the struct zpci_dev
in the zpci_list until it is finally released when all references have
been dropped.
Deletion from the zpci_list from within the release function is made
safe by using kref_put_lock() with the zpci_list_lock. This ensures that
the releasing code holds the last reference.
Cc: stable@vger.kernel.org
Fixes: a46044a92add ("s390/pci: fix zpci_zdev_put() on reserve")
Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Tested-by: Gerd Bayer <gbayer@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
|
|
As disable_slot() takes a struct zpci_dev from the Configured to the
Standby state. In Standby there is still a hotplug slot so this is not
usually a case of sysfs self deletion. This is important because self
deletion gets very hairy in terms of locking (see for example
recover_store() in arch/s390/pci/pci_sysfs.c).
Because the pci_dev_put() is not within the critical section of the
zdev->state_lock however, disable_slot() can turn into a case of self
deletion if zPCI device event handling slips between the mutex_unlock()
and the pci_dev_put(). If the latter is the last put and
zpci_release_device() is called this then tries to remove the hotplug
slot via zpci_exit_slot() which will try to remove the hotplug slot
directory the disable_slot() is part of i.e. self deletion.
Prevent this by widening the zdev->state_lock critical section to
include the pci_dev_put() which is then guaranteed to happen with the
struct zpci_dev still in Standby state ensuring it will not lead to
a zpci_release_device() call as at least the zPCI event handling code
still holds a reference.
Cc: stable@vger.kernel.org
Fixes: a46044a92add ("s390/pci: fix zpci_zdev_put() on reserve")
Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Tested-by: Gerd Bayer <gbayer@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
|
|
Remove zpci_bus_remove_device() and zpci_disable_device() calls from
zpci_release_device(). These calls were done when the device
transitioned into the ZPCI_FN_STATE_STANDBY state which is guaranteed to
happen before it enters the ZPCI_FN_STATE_RESERVED state. When
zpci_release_device() is called the device is known to be in the
ZPCI_FN_STATE_RESERVED state which is also checked by a WARN_ON().
Cc: stable@vger.kernel.org
Fixes: a46044a92add ("s390/pci: fix zpci_zdev_put() on reserve")
Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Reviewed-by: Julian Ruess <julianr@linux.ibm.com>
Tested-by: Gerd Bayer <gbayer@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
|
|
Replace spin_lock_irqsave/spin_unlock_irqrestore with
spin_lock_irq/spin_unlock_irq at places where it is known that interrupts
are enabled.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
|
|
ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/jic23/iio into char-misc-next
Jonathan writes:
IIO: New device support, features and cleanup for 6.16 - take 2
Note - last minute rebase was to drop a typo patch that I'd accidentally
picked up (in the microblaze arch Kconfig)
Take 2 is due to that rebase messing up some fixes tags that were
referring to patches after that point.
There is a known merge conflict due to changes in neighbouring lines.
Stephen's resolution in linux-next is:
https://lore.kernel.org/linux-next/20250506155728.65605bae@canb.auug.org.au/
Added 3 named IIO reviewers to MAINTAINERS. This is a reflection of those
who have been doing much of this work for some time. Lars-Peter is
removed from the entry having moved on to other topics. Thanks
Nuno, David and Andy for stepping up and Lars-Peter for all your
hard work in the past!
Includes the usual mix of new device support, features and general
cleanup.
This time we also have some tree wide changes.
- Rip out the iio_device_claim_direct_scoped() as it proved hard to work
with. This series includes quite a few related cleanups such as use
of guard or factoring code out to allow direct returns.
- Switch from iio_device_claim/release_direct_mode() to new
iio_device_claim/release_direct() which is structured so that sparse
can warn on failed releases. There were a few false positives but
those were mostly in code that benefited from being cleaned up as part
of this process.
- Introduce iio_push_to_buffers_with_ts() to replace the _timestamp()
version over time. This version takes the size of the supplied buffer
which the core checks is at least as big as expected by calculation
from channel descriptions of those channels enabled. Use this in
an initial set of drivers.
- Add macros for IIO_DECLARE_BUFFER_WITH_TS() and
IIO_DECLARE_DMA_BUFFER_WITH_TS() to avoid lots of fiddly code to ensure
correctly aligned buffers for timestamps being added onto the end of
channel data.
New device support
------------------
adi,ad3530r
- New driver for AD3530, AD3530R, AD3531 and AD3531R DACs with
programmable gain controls. R variants have internal references.
adi,ad7476
- Add support (dt compatible only) for the Rohm BU79100G ADC which is
fully compatible with the ti,ads7866.
adi,ad7606
- Support ad7606c-16 and ad7606c-18 devices. Includes switch to dynamic
channel information allocation.
adi,ad7380
- Add support for the AD7389-4
dfrobot,sen0322
- New driver for this oxygen sensor.
mediatek,mt2701-auxadc
- Add binding for MT6893 which is fully compatible with already supported
MT8173.
meson-saradc
- Support the GXLX SoCs. Mostly this is a workaround for some unrelated
clock control bits found in the ADC register map.
nuvoton,nct7201
- New driver for NCT7201 and NCT7202 I2C ADCs.
rohm,bd79124
- New driver for this 12-bit, 8-channel SAR ADC.
- Switch to new set_rv etc gpio callbacks that were added in 6.15.
rohm,bd79703
- Add support for BD79700, BD79701 and BD79702 DACs that have subsets of
functionality of the already supported bd79703. Included making this
driver suitable for support device variants.
st,stm32-lptimer
- Add support for stm32pm25 to this trigger.
Features
--------
Beyond IIO
- Property iterator for named children.
core
- Enable writes for 64 bit integers used for standard IIO ABI elements.
Previously these could be read only.
- Helper library that should avoid code duplication for simpler ADC
bindings that have a child node per channel.
- Enforce that IIO_DMA_MINALIGN is always at least 8 (almost always true
and simplifies code on all significant architectures)
core/backend
- Add support to control source of data - useful when the HDL includes
things like generated ramps for testing purposes. Enable this for
adi-axi-dac
adi,ad3552-hs
- Add debugfs related callbacks to allow debug access to register contents.
adi,ad4000
- Support SPI offload with appropriate FPGA firmware along with improving
documentation.
adi,ad7293
- Add support for external reference voltage.
adi,ad7606
- Support SPI offload.
adi,ad7768-1
- Support reset GPIO.
adi,admv8818
- Support filter frequencies beyond 2^32.
adi,adxl345
- Add single and double tap events.
hid-sensor-prox
- Support 16-bit report sizes as seen on some Intel platforms.
invensense,icm42600
- Enable use of named interrupts to avoid problems with some wiring choices.
Get the interrupt by name, but fallback to previous assumption on the first
being INT1 if no names are supplied.
microchip,mcp3911
- Add reset gpio support.
rohm,bh7150
- Add reset gpio support.
st,stm32
- Add support to control oversampling.
ti,adc128s052
- Add support for ROHM BD79104 which is early compatible with the TI
parts already supported by this driver. Includes some general driver
cleanup and a separate dt binding.
- Simplify reference voltage handling by assuming it is fixed after enabling
the supply.
winsen,mhz19b
- New driver for this C02 sensor.
Cleanup and minor fixes
-----------------------
dt-bindings
- Correct indentation and style for DTS examples.
- Use unevalutateProperties for SPI devices instead of additionalProperties
to allow generic SPI properties from spi-peripheral-props.yaml
ABI Docs
- Add missing docs for sampling_frequency when it applies only to events.
Treewide
- Various minor tweaks, comment fixes and similar.
- Sort TI ADCs in Kconfig that had gotten out of order.
- Switch various drives that provide GPIO chip functionality to the new
callbacks with return values.
- Standardize on { } formatting for all array sentinels.
- Make use of aligned_s64 in a few places to replace either wrong types
or manually defined equivalents.
- Drop places where spi bits_per_word is set to 8 because that is the
default anyway.
adi,ad_sigma_delta library
- Avoid a potential use of uninitialized data if reg_size has a value
that is not supported (no drivers hit this but it is reasonable hardening)
adi,ad4030
- Add error checking for scan types and no longer store it in state.
- Rework code to reduce duplication.
- Move setting the mode from buffer preenable() to update_scan_mode(),
better matching expected semantics of the two different callbacks.
- Improve data marshalling comments.
adi,ad4695
- Use u16 for buffer elements as oversampling is not yet supported except
with SPI offload (which doesn't use this path).
adi,ad5592r
- Clean up destruction of mutexes.
- Use lock guards to simplify code (later patch fixes a missed unlock)
adi,ad5933
- Correct some incorrect settling times.
adi,ad7091
- Deduplicate handling of writable vs volatile registers as they are the
inverse of each other for this device.
adi,ad7124
- Fix 3db Filter frequency.
- Remove ability to directly write the filter frequency (which was broken)
- Register naming improvements.
adi,ad7606
- Add a missing return value check.
- Fill in max sampling rates for all chips.
- Use devm_mutex_init()
- Fix up some kernel-doc formatting issues.
- Remove some camel case that snuck in.
- Drop setting address field in channels as easily established from other
fields.
- Drop unnecessary parameter to ad76060_scale_setup_cb_t.
adi,ad7768-1
- Convert to regmap.
- Factor out buffer allocation.
- Tidy up headers.
adi,ad7944
- Stop setting bits_per_word in SPI xfers with no data.
adi,ad9832
- Add of_device_id table rather than just relying on fallbacks.
- Use FIELD_PREP() to set values of fields.
adi,admv1013
- Cleanup a pointless ternary.
adi,admv8818
- Fix up LPF Band 5 frequency which was slightly wrong.
- Fix an integer overflow.
- Fix range calculation
adi,adt7316
- Replace irqd_get_trigger_type(irq_get_irq_data()) with simpler
irq_get_trigger_type()
adi,adxl345
- Use regmap cache instead of various state variables that were there to
reduce bus accesses.
- Make regmap return value checking consistent across all call sites.
adi,axi-dac
- Add a check on number of channels (0 to 15 valid)
allwinner,sun20i
- Use new adc-helpers to replace local parsing code for channel nodes.
bosch,bmp290
- Move to local variables for sensor data marshalling removing the need
for a messy definition that has to work for all supported parts.
Follow up fix adds a missing initialization.
dynaimage,al3010 and dynaimage,al3320a
- Various minor cleanup to bring these drivers inline with reviewed feedback
given on a new driver.
- Fix an error path in which power down is not called when it should be.
- Switch to regmap.
google,cros_ec
- Fix up a flexible array in middle of structure warning.
- Flush fifo when changing the timeout to avoid potential long wait
for samples.
hid-sensor-rotation
- Remove an __aligned(16) marking that doesn't seem to be justified.
kionix,kxcjk-1013
- Deduplicate code for setting up interrupts.
microchip,mcp3911
- Fix handling of conversion results register which differs across supported
devices.
idt,zopt2201
- Avoid duplicating register lists as all volatile registers are the
inverse of writeable registers on this device.
renesas,rzg2l
- Use new adc-helpers to replace local parsing code for channel nodes.
ti,ads1298
- Fix a missing Kconfig dependency.
* tag 'iio-for-6.16a-take2' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/jic23/iio: (260 commits)
dt-bindings: iio: adc: Add ROHM BD79100G
iio: adc: add support for Nuvoton NCT7201
dt-bindings: iio: adc: add NCT7201 ADCs
iio: chemical: Add driver for SEN0322
dt-bindings: trivial-devices: Document SEN0322
iio: adc: ad7768-1: reorganize driver headers
iio: bmp280: zero-init buffer
iio: ssp_sensors: optimalize -> optimize
HID: sensor-hub: Fix typo and improve documentation
iio: admv1013: replace redundant ternary operator with just len
iio: chemical: mhz19b: Fix error code in probe()
iio: adc: at91-sama5d2: use IIO_DECLARE_BUFFER_WITH_TS
iio: accel: sca3300: use IIO_DECLARE_BUFFER_WITH_TS
iio: adc: ad7380: use IIO_DECLARE_DMA_BUFFER_WITH_TS
iio: adc: ad4695: rename AD4695_MAX_VIN_CHANNELS
iio: adc: ad4695: use IIO_DECLARE_DMA_BUFFER_WITH_TS
iio: introduce IIO_DECLARE_BUFFER_WITH_TS macros
iio: make IIO_DMA_MINALIGN minimum of 8 bytes
iio: pressure: zpa2326_spi: remove bits_per_word = 8
iio: pressure: ms5611_spi: remove bits_per_word = 8
...
|
|
`bindgen` currently generates the wrong type for an `enum` when there
is a forward reference to it. For instance:
enum E;
enum E { A };
generates:
pub const E_A: E = 0;
pub type E = i32;
instead of the expected:
pub const E_A: E = 0;
pub type E = ffi::c_uint;
The issue was reported to upstream `bindgen` [1].
Now, both GCC and Clang support silently these forward references to
`enum` types, unless `-Wpedantic` is passed, and it turns out that some
headers in the kernel depend on them.
Thus, depending on how the headers are included, which in turn may depend
on the kernel configuration or the architecture, we may get a different
type on the Rust side for a given C `enum`.
That can be quite confusing, to say the least, especially since
developers may only notice issues when building for other architectures
like in [2]. In particular, they may end up forcing a cast and adding
an `#[allow(clippy::unnecessary_cast)]` like it was done in commit
94e05a66ea3e ("rust: hrtimer: allow timer restart from timer handler"),
which isn't great.
Instead, let's have a section at the top of our `bindings_helper.h` that
`#include`s the headers with the affected types -- hopefully there are
not many cases and there is a single ordering that covers all cases.
This allows us to remove the cast and the `#[allow]`, thus keeping the
correct code in the source files. When the issue gets resolved in upstream
`bindgen` (and we update our minimum `bindgen` version), we can easily
remove this section at the top.
Link: https://github.com/rust-lang/rust-bindgen/issues/3179 [1]
Link: https://lore.kernel.org/rust-for-linux/87tt7md1s6.fsf@kernel.org/ [2]
Acked-by: Andreas Hindborg <a.hindborg@kernel.org>
Link: https://lore.kernel.org/r/20250325184309.97170-1-ojeda@kernel.org
[ Added extra paragraph on the comment to clarify that the workaround may
not be possible in some cases. - Miguel ]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
|
|
Merge series from "Sheetal ." <sheetal@nvidia.com>:
The patch series includes the necessary changes to enable
support for the Tegra264 platforms in AHUB drivers.
|
|
Multi-PTP source support within a network topology has been merged,
but the hardware timestamp source is not yet exposed to users.
Currently, users only see the PTP index, which does not indicate
whether the timestamp comes from a PHY or a MAC.
Add support for reporting the hwtstamp source using a
hwtstamp-source field, alongside hwtstamp-phyindex, to describe
the origin of the hardware timestamp.
Remove HWTSTAMP_SOURCE_UNSPEC enum value as it is not used at all.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Link: https://patch.msgid.link/20250519-feature_ptp_source-v4-1-5d10e19a0265@bootlin.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Convert cdn-dp-rockchip.txt to yaml.
Add new "port@1" property which represents the CDN DP output to keep
the same style as the other display interfaces.
This patch also changes the constraints for "phys" and "extcon". For
the original binding, only one phy and the corresponding extcon can
be specified. In the new binding, one or two phys can be specified.
Since the RK3399 has two DP-USB PHYs, specifying one allows output
via the corresponding PHY, while specifying two lets the driver
choose one PHY for output. This rule also applies to extcon, which
provides the cable state for the corresponding PHY.
Signed-off-by: Chaoyi Chen <chaoyi.chen@rock-chips.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20250520024718.142-3-kernel@airkyi.com
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
|
|
Some GPIO chips allow to rise an IRQ on GPIO level changes but do not
provide an IRQ status for each separate line: only the current gpio
level can be retrieved.
Add support for these chips, emulating IRQ status by comparing GPIO
levels with the levels during the previous interrupt.
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20250522-mdb-max7360-support-v9-5-74fc03517e41@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
The cirrus_dir pointer should be const data but was missing the second
const needed to achieve this.
I haven't marked this as a 'Fixes' because it isn't causing any bugs,
it's only a code improvement.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20250522103816.543919-1-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
FSL SOC Changes for 6.16:
- Various fixes and cleanups from Ioana on fsl-mc bus driver
- Other misc fixes on fsl-mc bus driver
- A build fix on DPAA2 driver
- Some preparation work from Kees on qbman driver
- Some cleanup on FSL qe interrupt driver init
* tag 'soc_fsl-6.16-1' of https://github.com/chleroy/linux:
soc: fsl: qe: Consolidate chained IRQ handler install/remove
bus: fsl-mc: Use strscpy() instead of strscpy_pad()
soc: fsl: qbman: Remove const from portal->cgrs allocation type
bus: fsl_mc: Fix driver_managed_dma check
bus: fsl-mc: increase MC_CMD_COMPLETION_TIMEOUT_MS value
bus: fsl-mc: drop useless cleanup
bus: fsl-mc: add the dprc_get_mem() command to the whitelist
bus: fsl-mc: fix GET/SET_TAILDROP command ids
bus: fsl-mc: do not add a device-link for the UAPI used DPMCP device
bus: fsl-mc: fix double-free on mc_dev
soc: fsl: Do not enable DPAA2_CONSOLE by default during compile testing
Link: https://lore.kernel.org/r/9e5b3e8e-3280-4ff9-915b-9a8b934bac22@csgroup.eu
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
Felix Fietkau says:
===================
mt76 patches for 6.16
- fixes
- EHT improvements
- new device ids
===================
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/ath/ath
Jeff Johnson says:
==================
ath.git patches for v6.16
ath12k:
Add monitor mode support for WCN7850.
Enhance regulatory support including 6 GHz power modes.
In addition, perform the usual set of bug fixes and cleanups across
all supported drivers.
==================
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
https://github.com/Broadcom/stblinux into soc/arm
This pull request contains MAINTAINERS file update for Broadcom SoCs,
please pull the following for 6.16:
- Krzysztof updates the BCM2711/BCM283X entry to also match on the
bcm2712 pattern to cover the Raspberry Pi 5 files
* tag 'arm-soc/for-6.16/maintainers' of https://github.com/Broadcom/stblinux:
ARM: broadcom: MAINTAINERS: Cover bcm2712 files
Link: https://lore.kernel.org/r/20250513140808.2047729-1-florian.fainelli@broadcom.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux into soc/arm
TI DaVinci updates for v6.16-rc1
- remove support for the da830 SoC
- increase the panel FIFO threshold in da850-evm
* tag 'davinci-updates-for-v6.16-rc1' of https://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
ARM: dts: davinci: da850-evm: Increase fifo threshold
ARM: davinci: remove support for da830
Link: https://lore.kernel.org/r/20250519090232.13147-1-brgl@bgdev.pl
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
When a list head is a const pointer, the list entry for that head also
must remain a const pointer, otherwise we are just "throwing it away"
for no good reason. Fix this up by properly marking these structures as
const.
Cc: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://patch.msgid.link/2025052213-semifinal-sublevel-d631@gregkh
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
While the struct spi_device * passed into spi_to_spi_gpio() is a const
one, the struct spi_bitbang * that is retrieved from the controller
field in the spi_device is NOT a const pointer, as it is coming from the
spi_controller_get_devdata() call, and then passed to container_of()
which would strip off the const attribute for no good reason (i.e. if a
const pointer is passed to container_of() it still is const coming out).
Fix this all up by properly declaring the struct spi_bitbang * as not
const.
Cc: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://patch.msgid.link/2025052225-scallion-ritzy-dbbd@gregkh
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Some places in the spi core pass in a const pointer to a device and the
default container_of() casts that away, which is not a good idea.
Preserve the proper const attribute by using container_of_const() for
to_spi_device() instead, which is what it was designed for.
Note, this removes the NULL check for a device pointer in the call, but
no one was ever checking for that return value, and a device pointer
should never be NULL overall anyway, so this should be a safe change.
Cc: Mark Brown <broonie@kernel.org>
Fixes: d69d80484598 ("driver core: have match() callback in struct bus_type take a const *")
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://patch.msgid.link/2025052230-fidgeting-stooge-66f5@gregkh
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into soc/defconfig
Enable Rockchip SAI and ES8328 modules.
* tag 'v6.16-rockchip-defconfig64-1' of https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
arm64: defconfig: Enable Rockchip SAI and ES8328
Link: https://lore.kernel.org/r/4060740.R56niFO833@phil
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
The argument has been unused since the function was added, so remove it.
Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250521160720.1893326-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/mediatek/linux into soc/defconfig
MediaTek ARM64 defconfig updates for v6.16
This enables some drivers as modules to enable fully booting
some of the supported MediaTek boards;
In particular, this enables the following drivers as module:
- PHYs for PCIe, HDMI, DSI and DisplayPort for all boards
- MediaTek UART DMA Controller, also for all boards
- MDP3 driver for MT8186/88/95 (Chromebooks and Genio variants)
- Auxiliary ADC for the MT6357/58/59 PMICs found on both Genio EVK
and Chromebooks based on designs with MT8183/86/88/92/95 SoCs
- iTE IT5205 Type-C USB Alternate Mode Passive MUX, found
on both Chromebooks and Genio EVKs with MT8188/95
- Richtek RT1715 Type-C PD Controller, found on all Genio boards
- Himax HX8279 DSI panel DriverIC and KD070FHFID015 panel, found
on various revisions of the Genio Evaluation Kit boards.
* tag 'mtk-defconfig-for-v6.16' of https://git.kernel.org/pub/scm/linux/kernel/git/mediatek/linux:
arm64: defconfig: Enable configs for MediaTek Genio EVK boards
arm64: defconfig: mediatek: enable PHY drivers
Link: https://lore.kernel.org/r/20250520114356.1194450-2-angelogioacchino.delregno@collabora.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
In the pci_acpi_scan_root() function, when creating a PCI bus fails,
we need to free up the previously allocated memory, which can avoid
invalid memory usage and save resources.
Fixes: 789befdfa389 ("arm64: PCI: Migrate ACPI related functions to pci-acpi.c")
Signed-off-by: Zhe Qiao <qiaozhe@iscas.ac.cn>
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20250430060603.381504-1-qiaozhe@iscas.ac.cn
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux into soc/defconfig
One more Qualcomm Arm64 defconfig update for v6.16
Enable global clock controller and TLMM pinctrl drivers for IPQ5424 to
make this boot.
* tag 'qcom-arm64-defconfig-for-6.16-2' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux:
arm64: defconfig: Enable IPQ5424 RDP466 base configs
Link: https://lore.kernel.org/r/20250520025119.40021-1-andersson@kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/conor/linux into soc/defconfig
RISC-V config for v6.16
PM_GENERIC_DOMAINS is enabled because T-Head SoCs need to use it and
spacemit's gpio and clock controller are enabled for more obvious
reasons.
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
* tag 'riscv-config-for-v6.16' of https://git.kernel.org/pub/scm/linux/kernel/git/conor/linux:
riscv: defconfig: spacemit: enable clock controller driver for SpacemiT K1
riscv: defconfig: spacemit: enable gpio support for K1 SoC
riscv: Enable PM_GENERIC_DOMAINS for T-Head SoCs
Link: https://lore.kernel.org/r/20250516-joystick-mortified-385820b1ad0b@spud
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
Add mac80211 rfkill_poll ops to monitor hardware rfkill state
and state change will be updated.
Signed-off-by: Allan Wang <allan.wang@mediatek.com>
Link: https://patch.msgid.link/20250507053131.4173691-1-allan.wang@mediatek.com
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
One variant of MT7992 has 5 TX paths, so extend the power delta function
to support it. Also, rename nss_delta to path_delta since the value is
based on the number of TX paths rather tha the number of spatial streams.
(path delta [0.5 dBm] = 10 * log(path number) [dBm] * 2)
Co-developed-by: StanleyYP Wang <StanleyYP.Wang@mediatek.com>
Signed-off-by: StanleyYP Wang <StanleyYP.Wang@mediatek.com>
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Link: https://patch.msgid.link/20250515032952.1653494-9-shayne.chen@mediatek.com
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Check if available_antennas_tx and available_antennas_rx are already set
during the per-chip initialization phase; otherwise, they could be
overwritten with incorrect values.
Fixes: 69d54ce7491d ("wifi: mt76: mt7996: switch to single multi-radio wiphy")
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Link: https://patch.msgid.link/20250515032952.1653494-8-shayne.chen@mediatek.com
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Some management frames are first processed by the firmware and then
passed to the driver through the MCU event rings. In CONNAC3, event rings
do not support scatter-gather and have a size limitation of 2048 bytes.
If a packet sized between 1728 and 2048 bytes arrives from an event ring,
the ring will hang because the driver attempts to use scatter-gather to
process it.
To fix this, include the size of struct skb_shared_info in the MCU RX
buffer size to prevent scatter-gather from being used for event skb in
mt76_dma_rx_fill_buf().
Fixes: 98686cd21624 ("wifi: mt76: mt7996: add driver for MediaTek Wi-Fi 7 (802.11be) devices")
Co-developed-by: Peter Chiu <chui-hao.chiu@mediatek.com>
Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com>
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Link: https://patch.msgid.link/20250515032952.1653494-7-shayne.chen@mediatek.com
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
