module: Don't wait for GOING modules

commit 0254127ab977e70798707a7a2b757c9f3c971210 upstream.

During a system boot, it can happen that the kernel receives a burst of
requests to insert the same module but loading it eventually fails
during its init call. For instance, udev can make a request to insert
a frequency module for each individual CPU when another frequency module
is already loaded which causes the init function of the new module to
return an error.

Since commit 6e6de3dee51a ("kernel/module.c: Only return -EEXIST for
modules that have finished loading"), the kernel waits for modules in
MODULE_STATE_GOING state to finish unloading before making another
attempt to load the same module.

This creates unnecessary work in the described scenario and delays the
boot. In the worst case, it can prevent udev from loading drivers for
other devices and might cause timeouts of services waiting on them and
subsequently a failed boot.

This patch attempts a different solution for the problem 6e6de3dee51a
was trying to solve. Rather than waiting for the unloading to complete,
it returns a different error code (-EBUSY) for modules in the GOING
state. This should avoid the error situation that was described in
6e6de3dee51a (user space attempting to load a dependent module because
the -EEXIST error code would suggest to user space that the first module
had been loaded successfully), while avoiding the delay situation too.

This has been tested on linux-next since December 2022 and passes
all kmod selftests except test 0009 with module compression enabled
but it has been confirmed that this issue has existed and has gone
unnoticed since prior to this commit and can also be reproduced without
module compression with a simple usleep(5000000) on tools/modprobe.c [0].
These failures are caused by hitting the kernel mod_concurrent_max and can
happen either due to a self inflicted kernel module auto-loead DoS somehow
or on a system with large CPU count and each CPU count incorrectly triggering
many module auto-loads. Both of those issues need to be fixed in-kernel.

[0] https://lore.kernel.org/all/Y9A4fiobL6IHp%2F%2FP@bombadil.infradead.org/

Fixes: 6e6de3dee51a ("kernel/module.c: Only return -EEXIST for modules that have finished loading")
Co-developed-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Petr Pavlu <petr.pavlu@suse.com>
Cc: stable@vger.kernel.org
Reviewed-by: Petr Mladek <pmladek@suse.com>
[mcgrof: enhance commit log with testing and kmod test result interpretation ]
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

nfsd: don't free files unconditionally in __nfsd_file_cache_purge

[ Upstream commit 4bdbba54e9b1c769da8ded9abd209d765715e1d6 ]

nfsd_file_cache_purge is called when the server is shutting down, in
which case, tearing things down is generally fine, but it also gets
called when the exports cache is flushed.

Instead of walking the cache and freeing everything unconditionally,
handle it the same as when we have a notification of conflicting access.

Fixes: ac3a2585f018 ("nfsd: rework refcounting in filecache")
Reported-by: Ruben Vestergaard <rubenv@drcmr.dk>
Reported-by: Torkil Svensgaard <torkil@drcmr.dk>
Reported-by: Shachar Kagan <skagan@nvidia.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Tested-by: Shachar Kagan <skagan@nvidia.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

kvm/vfio: Fix potential deadlock on vfio group_lock

[ Upstream commit 51cdc8bc120ef6e42f6fb758341f5d91bc955952 ]

Currently it is possible that the final put of a KVM reference comes from
vfio during its device close operation.  This occurs while the vfio group
lock is held; however, if the vfio device is still in the kvm device list,
then the following call chain could result in a deadlock:

VFIO holds group->group_lock/group_rwsem
  -> kvm_put_kvm
   -> kvm_destroy_vm
    -> kvm_destroy_devices
     -> kvm_vfio_destroy
      -> kvm_vfio_file_set_kvm
       -> vfio_file_set_kvm
        -> try to hold group->group_lock/group_rwsem

The key function is the kvm_destroy_devices() which triggers destroy cb
of kvm_device_ops. It calls back to vfio and try to hold group_lock. So
if this path doesn't call back to vfio, this dead lock would be fixed.
Actually, there is a way for it. KVM provides another point to free the
kvm-vfio device which is the point when the device file descriptor is
closed. This can be achieved by providing the release cb instead of the
destroy cb. Also rename kvm_vfio_destroy() to be kvm_vfio_release().

/*
 * Destroy is responsible for freeing dev.
 *
 * Destroy may be called before or after destructors are called
 * on emulated I/O regions, depending on whether a reference is
 * held by a vcpu or other kvm component that gets destroyed
 * after the emulated I/O.
 */
void (*destroy)(struct kvm_device *dev);

/*
 * Release is an alternative method to free the device. It is
 * called when the device file descriptor is closed. Once
 * release is called, the destroy method will not be called
 * anymore as the device is removed from the device list of
 * the VM. kvm->lock is held.
 */
void (*release)(struct kvm_device *dev);

Fixes: 421cfe6596f6 ("vfio: remove VFIO_GROUP_NOTIFY_SET_KVM")
Reported-by: Alex Williamson <alex.williamson@redhat.com>
Suggested-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Link: https://lore.kernel.org/r/20230114000351.115444-1-mjrosato@linux.ibm.com
Link: https://lore.kernel.org/r/20230120150528.471752-1-yi.l.liu@intel.com
[aw: update comment as well, s/destroy/release/]
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: hpsa: Fix allocation size for scsi_host_alloc()

[ Upstream commit bbbd25499100c810ceaf5193c3cfcab9f7402a33 ]

The 'h' is a pointer to struct ctlr_info, so it's just 4 or 8 bytes, while
the structure itself is much bigger.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: edd163687ea5 ("hpsa: add driver for HP Smart Array controllers.")
Link: https://lore.kernel.org/r/20230118031255.GE15213@altlinux.org
Signed-off-by: Alexey V. Vissarionov <gremlin@altlinux.org>
Acked-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

vfio/type1: Respect IOMMU reserved regions in vfio_test_domain_fgsp()

[ Upstream commit 895c0747f726bb50c9b7a805613a61d1b6f9fa06 ]

Since commit cbf7827bc5dc ("iommu/s390: Fix potential s390_domain
aperture shrinking") the s390 IOMMU driver uses reserved regions for the
system provided DMA ranges of PCI devices. Previously it reduced the
size of the IOMMU aperture and checked it on each mapping operation.
On current machines the system denies use of DMA addresses below 2^32 for
all PCI devices.

Usually mapping IOVAs in a reserved regions is harmless until a DMA
actually tries to utilize the mapping. However on s390 there is
a virtual PCI device called ISM which is implemented in firmware and
used for cross LPAR communication. Unlike real PCI devices this device
does not use the hardware IOMMU but inspects IOMMU translation tables
directly on IOTLB flush (s390 RPCIT instruction). If it detects IOVA
mappings outside the allowed ranges it goes into an error state. This
error state then causes the device to be unavailable to the KVM guest.

Analysing this we found that vfio_test_domain_fgsp() maps 2 pages at DMA
address 0 irrespective of the IOMMUs reserved regions. Even if usually
harmless this seems wrong in the general case so instead go through the
freshly updated IOVA list and try to find a range that isn't reserved,
and fits 2 pages, is PAGE_SIZE * 2 aligned. If found use that for
testing for fine grained super pages.

Fixes: af029169b8fd ("vfio/type1: Check reserved region conflict and update iova list")
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20230110164427.4051938-2-schnelle@linux.ibm.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

sched/uclamp: Fix a uninitialized variable warnings

[ Upstream commit e26fd28db82899be71b4b949527373d0a6be1e65 ]

Addresses the following warnings:

> config: riscv-randconfig-m031-20221111
> compiler: riscv64-linux-gcc (GCC) 12.1.0
>
> smatch warnings:
> kernel/sched/fair.c:7263 find_energy_efficient_cpu() error: uninitialized symbol 'util_min'.
> kernel/sched/fair.c:7263 find_energy_efficient_cpu() error: uninitialized symbol 'util_max'.

Fixes: 244226035a1f ("sched/uclamp: Fix fits_capacity() check in feec()")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Qais Yousef (Google) <qyousef@layalina.io>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lore.kernel.org/r/20230112122708.330667-2-qyousef@layalina.io
Signed-off-by: Sasha Levin <sashal@kernel.org>

sched/fair: Check if prev_cpu has highest spare cap in feec()

[ Upstream commit ad841e569f5c88e3332b32a000f251f33ff32187 ]

When evaluating the CPU candidates in the perf domain (pd) containing
the previously used CPU (prev_cpu), find_energy_efficient_cpu()
evaluates the energy of the pd:
- without the task (base_energy)
- with the task placed on prev_cpu (if the task fits)
- with the task placed on the CPU with the highest spare capacity,
  prev_cpu being excluded from this set

If prev_cpu is already the CPU with the highest spare capacity,
max_spare_cap_cpu will be the CPU with the second highest spare
capacity.

On an Arm64 Juno-r2, with a workload of 10 tasks at a 10% duty cycle,
when prev_cpu and max_spare_cap_cpu are both valid candidates,
prev_spare_cap > max_spare_cap at ~82%.
Thus the energy of the pd when placing the task on max_spare_cap_cpu
is computed with no possible positive outcome 82% most of the time.

Do not consider max_spare_cap_cpu as a valid candidate if
prev_spare_cap > max_spare_cap.

Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lore.kernel.org/r/20221006081052.3862167-2-pierre.gondois@arm.com
Stable-dep-of: e26fd28db828 ("sched/uclamp: Fix a uninitialized variable warnings")
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mac80211: Fix iTXQ AMPDU fragmentation handling

commit 592234e941f1addaa598601c9227e3b72d608625 upstream.

mac80211 must not enable aggregation wile transmitting a fragmented
MPDU. Enforce that for mac80211 internal TX queues (iTXQs).

Reported-by: kernel test robot <oliver.sang@intel.com>
Link: https://lore.kernel.org/oe-lkp/202301021738.7cd3e6ae-oliver.sang@intel.com
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Link: https://lore.kernel.org/r/20230106223141.98696-1-alexander@wetzel-home.de
Cc: stable@vger.kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

wifi: mac80211: Proper mark iTXQs for resumption

commit 4444bc2116aecdcde87dce80373540adc8bd478b upstream.

When a running wake_tx_queue() call is aborted due to a hw queue stop
the corresponding iTXQ is not always correctly marked for resumption:
wake_tx_push_queue() can stops the queue run without setting
@IEEE80211_TXQ_STOP_NETIF_TX.

Without the @IEEE80211_TXQ_STOP_NETIF_TX flag __ieee80211_wake_txqs()
will not schedule a new queue run and remaining frames in the queue get
stuck till another frame is queued to it.

Fix the issue for all drivers - also the ones with custom wake_tx_queue
callbacks - by moving the logic into ieee80211_tx_dequeue() and drop the
redundant @txqs_stopped.

@IEEE80211_TXQ_STOP_NETIF_TX is also renamed to @IEEE80211_TXQ_DIRTY to
better describe the flag.

Fixes: c850e31f79f0 ("wifi: mac80211: add internal handler for wake_tx_queue")
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Link: https://lore.kernel.org/r/20221230121850.218810-1-alexander@wetzel-home.de
Cc: stable@vger.kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

io_uring/msg_ring: fix remote queue to disabled ring

commit 8579538c89e33ce78be2feb41e07489c8cbf8f31 upstream.

IORING_SETUP_R_DISABLED rings don't have the submitter task set, so
it's not always safe to use ->submitter_task. Disallow posting msg_ring
messaged to disabled rings. Also add task NULL check for loosy sync
around testing for IORING_SETUP_R_DISABLED.

Cc: stable@vger.kernel.org
Fixes: 6d043ee1164ca ("io_uring: do msg_ring in target task via tw")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amdgpu: complete gfxoff allow signal during suspend without delay

commit 4b31b92b143f7d209f3d494c56d4c4673e9fc53d upstream.

change guarantees that gfxoff is allowed before moving further in
s2idle sequence to add more reliablity about gfxoff in amdgpu IP's
suspend flow

Signed-off-by: Harsh Jain <harsh.jain@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: "Limonciello, Mario" <Mario.Limonciello@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/i915: Allow alternate fixed modes always for eDP

[ Upstream commit 55cfeecc2197de68e9cc30f77c711dcbcdf27510 ]

Stop considering VBT's static DRRS support when deciding whether
to use alternate fixed modes or not. It looks like Windows more
or less just uses that to decide whether to automagically switch
refresh rates on AC<->battery changes, or perhaps whether to
even expose a control for that in some UI thing. Either way it
seems happy to always use all EDID modes, and I guess the
DRRS/VRR stuff more or less adjusts how said modes get
actually used.

Let's do the same and just accept all the suitable looking
modes from EDID, whether we have DRRS or VRR.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6323
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6484
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220927180615.25476-3-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/i915: Allow panel fixed modes to have differing sync polarities

[ Upstream commit 2bd0db4b3f0bd529f75b32538fc5a3775e3591c0 ]

Apparently some panels declare multiple modes with random
sync polarities. Seems a bit weird, but looks like Windows/GOP
doesn't care, so let follow suit and accept alternate fixed
modes regardless of their sync polarities.

v2: Don't pollute the DRM_ namespace with a define (Jani)

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6968
Acked-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020093938.27200-1-ville.syrjala@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: hci_sync: cancel cmd_timer if hci_open failed

commit 97dfaf073f5881c624856ef293be307b6166115c upstream.

If a command is already sent, we take care of freeing it, but we
also need to cancel the timeout as well.

Signed-off-by: Archie Pusaka <apusaka@chromium.org>
Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@google.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: efi: Account for the EFI runtime stack in stack unwinder

[ Upstream commit 7ea55715c421d22c1b63f7129cae6a654091b695 ]

The EFI runtime services run from a dedicated stack now, and so the
stack unwinder needs to be informed about this.

Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: efi: Avoid workqueue to check whether EFI runtime is live

[ Upstream commit 8a9a1a18731eb123e35f48176380a18b9782845e ]

Comparing current_work() against efi_rts_work.work is sufficient to
decide whether current is currently running EFI runtime services code at
any level in its call stack.

However, there are other potential users of the EFI runtime stack, such
as the ACPI subsystem, which may invoke efi_call_virt_pointer()
directly, and so any sync exceptions occurring in firmware during those
calls are currently misidentified.

So instead, let's check whether the stashed value of the thread stack
pointer points into current's thread stack. This can only be the case if
current was interrupted while running EFI runtime code. Note that this
implies that we should clear the stashed value after switching back, to
avoid false positives.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: efi: Recover from synchronous exceptions occurring in firmware

[ Upstream commit e8dfdf3162eb549d064b8c10b1564f7e8ee82591 ]

Unlike x86, which has machinery to deal with page faults that occur
during the execution of EFI runtime services, arm64 has nothing like
that, and a synchronous exception raised by firmware code brings down
the whole system.

With more EFI based systems appearing that were not built to run Linux
(such as the Windows-on-ARM laptops based on Qualcomm SOCs), as well as
the introduction of PRM (platform specific firmware routines that are
callable just like EFI runtime services), we are more likely to run into
issues of this sort, and it is much more likely that we can identify and
work around such issues if they don't bring down the system entirely.

Since we already use a EFI runtime services call wrapper in assembler,
we can quite easily add some code that captures the execution state at
the point where the call is made, allowing us to revert to this state
and proceed execution if the call triggered a synchronous exception.

Given that the kernel and the firmware don't share any data structures
that could end up in an indeterminate state, we can happily continue
running, as long as we mark the EFI runtime services as unavailable from
that point on.

Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Stable-dep-of: 8a9a1a18731e ("arm64: efi: Avoid workqueue to check whether EFI runtime is live")
Signed-off-by: Sasha Levin <sashal@kernel.org>

Revert "selftests/bpf: check null propagation only neither reg is PTR_TO_BTF_ID"

This reverts commit 2e5d5c4ae77dedc7eba0e496474ab93f05b25adf.

Signed-off-by: Sasha Levin <sashal@kernel.org>

btrfs: zoned: enable metadata over-commit for non-ZNS setup

[ Upstream commit 85e79ec7b78f863178ca488fd8cb5b3de6347756 ]

The commit 79417d040f4f ("btrfs: zoned: disable metadata overcommit for
zoned") disabled the metadata over-commit to track active zones properly.

However, it also introduced a heavy overhead by allocating new metadata
block groups and/or flushing dirty buffers to release the space
reservations. Specifically, a workload (write only without any sync
operations) worsen its performance from 343.77 MB/sec (v5.19) to 182.89
MB/sec (v6.0).

The performance is still bad on current misc-next which is 187.95 MB/sec.
And, with this patch applied, it improves back to 326.70 MB/sec (+73.82%).

This patch introduces a new fs_info->flag BTRFS_FS_NO_OVERCOMMIT to
indicate it needs to disable the metadata over-commit. The flag is enabled
when a device with max active zones limit is loaded into a file-system.

Fixes: 79417d040f4f ("btrfs: zoned: disable metadata overcommit for zoned")
CC: stable@vger.kernel.org # 6.0+
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

firmware: coreboot: Check size of table entry and use flex-array

[ Upstream commit 3b293487b8752cc42c1cbf8a0447bc6076c075fa ]

The memcpy() of the data following a coreboot_table_entry couldn't
be evaluated by the compiler under CONFIG_FORTIFY_SOURCE. To make it
easier to reason about, add an explicit flexible array member to struct
coreboot_device so the entire entry can be copied at once. Additionally,
validate the sizes before copying. Avoids this run-time false positive
warning:

  memcpy: detected field-spanning write (size 168) of single field "&device->entry" at drivers/firmware/google/coreboot_table.c:103 (size 8)

Reported-by: Paul Menzel <pmenzel@molgen.mpg.de>
Link: https://lore.kernel.org/all/03ae2704-8c30-f9f0-215b-7cdf4ad35a9a@molgen.mpg.de/
Cc: Jack Rosenthal <jrosenth@chromium.org>
Cc: Guenter Roeck <groeck@chromium.org>
Cc: Julius Werner <jwerner@chromium.org>
Cc: Brian Norris <briannorris@chromium.org>
Cc: Stephen Boyd <swboyd@chromium.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Guenter Roeck <groeck@chromium.org>
Link: https://lore.kernel.org/r/20230107031406.gonna.761-kees@kernel.org
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Jack Rosenthal <jrosenth@chromium.org>
Link: https://lore.kernel.org/r/20230112230312.give.446-kees@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

ata: pata_cs5535: Don't build on UML

[ Upstream commit 22eebaa631c40f3dac169ba781e0de471b83bf45 ]

This driver uses MSR functions that aren't implemented under UML.
Avoid building it to prevent tripping up allyesconfig.

e.g.
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: pata_cs5535.c:(.text+0x3a3): undefined reference to `__tracepoint_read_msr'
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: pata_cs5535.c:(.text+0x3d2): undefined reference to `__tracepoint_write_msr'
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: pata_cs5535.c:(.text+0x457): undefined reference to `__tracepoint_write_msr'
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: pata_cs5535.c:(.text+0x481): undefined reference to `do_trace_write_msr'
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: pata_cs5535.c:(.text+0x4d5): undefined reference to `do_trace_write_msr'
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: pata_cs5535.c:(.text+0x4f5): undefined reference to `do_trace_read_msr'
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: pata_cs5535.c:(.text+0x51c): undefined reference to `do_trace_write_msr'

Signed-off-by: Peter Foley <pefoley2@pefoley.com>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

lockref: stop doing cpu_relax in the cmpxchg loop

[ Upstream commit f5fe24ef17b5fbe6db49534163e77499fb10ae8c ]

On the x86-64 architecture even a failing cmpxchg grants exclusive
access to the cacheline, making it preferable to retry the failed op
immediately instead of stalling with the pause instruction.

To illustrate the impact, below are benchmark results obtained by
running various will-it-scale tests on top of the 6.2-rc3 kernel and
Cascade Lake (2 sockets * 24 cores * 2 threads) CPU.

All results in ops/s.  Note there is some variance in re-runs, but the
code is consistently faster when contention is present.

  open3 ("Same file open/close"):
  proc          stock       no-pause
     1         805603         814942       (+%1)
     2        1054980        1054781       (-0%)
     8        1544802        1822858      (+18%)
    24        1191064        2199665      (+84%)
    48         851582        1469860      (+72%)
    96         609481        1427170     (+134%)

  fstat2 ("Same file fstat"):
  proc          stock       no-pause
     1        3013872        3047636       (+1%)
     2        4284687        4400421       (+2%)
     8        3257721        5530156      (+69%)
    24        2239819        5466127     (+144%)
    48        1701072        5256609     (+209%)
    96        1269157        6649326     (+423%)

Additionally, a kernel with a private patch to help access() scalability:
access2 ("Same file access"):

  proc          stock        patched      patched
                                         +nopause
    24        2378041        2005501      5370335  (-15% / +125%)

That is, fixing the problems in access itself *reduces* scalability
after the cacheline ping-pong only happens in lockref with the pause
instruction.

Note that fstat and access benchmarks are not currently integrated into
will-it-scale, but interested parties can find them in pull requests to
said project.

Code at hand has a rather tortured history.  First modification showed
up in commit d472d9d98b46 ("lockref: Relax in cmpxchg loop"), written
with Itanium in mind.  Later it got patched up to use an arch-dependent
macro to stop doing it on s390 where it caused a significant regression.
Said macro had undergone revisions and was ultimately eliminated later,
going back to cpu_relax.

While I intended to only remove cpu_relax for x86-64, I got the
following comment from Linus:

    I would actually prefer just removing it entirely and see if
    somebody else hollers. You have the numbers to prove it hurts on
    real hardware, and I don't think we have any numbers to the
    contrary.

    So I think it's better to trust the numbers and remove it as a
    failure, than say "let's just remove it on x86-64 and leave
    everybody else with the potentially broken code"

Additionally, Will Deacon (maintainer of the arm64 port, one of the
architectures previously benchmarked):

    So, from the arm64 side of the fence, I'm perfectly happy just
    removing the cpu_relax() calls from lockref.

As such, come back full circle in history and whack it altogether.

Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/all/CAGudoHHx0Nqg6DE70zAVA75eV-HXfWyhVMWZ-aSeOofkA_=WdA@mail.gmail.com/
Acked-by: Tony Luck <tony.luck@intel.com> # ia64
Acked-by: Nicholas Piggin <npiggin@gmail.com> # powerpc
Acked-by: Will Deacon <will@kernel.org> # arm64
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: simatic-ipc: add another model

[ Upstream commit d348b1d761e358a4ba03fb34aa7e3dbd278db236 ]

Add IPC PX-39A support.

Signed-off-by: Henning Schild <henning.schild@siemens.com>
Link: https://lore.kernel.org/r/20221222103720.8546-3-henning.schild@siemens.com
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: simatic-ipc: correct name of a model

[ Upstream commit ed058eab22d64c00663563e8e1e112989c65c59f ]

What we called IPC427G should be renamed to BX-39A to be more in line
with the actual product name.

Signed-off-by: Henning Schild <henning.schild@siemens.com>
Link: https://lore.kernel.org/r/20221222103720.8546-2-henning.schild@siemens.com
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: asus-wmi: Ignore fan on E410MA

[ Upstream commit 82cc5c6c624c63f7b57214e325e2ea685d924e89 ]

The ASUS VivoBook has a fan device described in its ACPI tables but does
not actually contain any physical fan.
Use the quirk to inhibit fan handling.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/r/20221221-asus-fan-v1-2-e07f3949725b@weissschuh.net
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: asus-wmi: Add quirk wmi_ignore_fan

[ Upstream commit c874b6de4cdfa2822a07b479887cd5f87fb5d078 ]

Some laptops have a fan device listed in their ACPI tables but do not
actually contain a fan.
Introduce a quirk that can be used to override the fan detection logic.

This was observed with a ASUS VivoBook E410MA running firmware
E410MAB.304.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/r/20221221-asus-fan-v1-1-e07f3949725b@weissschuh.net
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: asus-nb-wmi: Add alternate mapping for KEY_SCREENLOCK

[ Upstream commit db9494895b405bf318dc7e563dee6daa51b3b6ed ]

The 0x33 keycode is emitted by Fn + F6 on a ASUS FX705GE laptop.

Reported-by: Nemcev Aleksey <Nemcev_Aleksey@inbox.ru>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20230112181841.84652-1-hdegoede@redhat.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: asus-nb-wmi: Add alternate mapping for KEY_CAMERA

[ Upstream commit c78a4e191839edc1e8c3e51565cf2e71d40e8883 ]

This keycode is emitted on a Asus VivoBook E410MAB with firmware
E410MAB.304.

The physical key has a strikken-through camera printed on it.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/r/20221216-asus-key-v1-1-45da124119a3@weissschuh.net
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: touchscreen_dmi: Add info for the CSL Panther Tab HD

[ Upstream commit 36c2b9d6710427f802494ba070621cb415198293 ]

Add touchscreen info for the CSL Panther Tab HD.

Signed-off-by: Michael Klein <m.klein@mvz-labor-lb.de>
Link: https://lore.kernel.org/r/20221220121103.uiwn5l7fii2iggct@LLGMVZLB-0037
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

r8152: add vendor/device ID pair for Microsoft Devkit

[ Upstream commit be53771c87f4e322a9835d3faa9cd73a4ecdec5b ]

The Microsoft Devkit 2023 is a an ARM64 based machine featuring a
Realtek 8153 USB3.0-to-GBit Ethernet adapter. As in their other
machines, Microsoft uses a custom USB device ID.

Add the respective ID values to the driver. This makes Ethernet work on
the MS Devkit device. The chip has been visually confirmed to be a
RTL8153.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Link: https://lore.kernel.org/r/20230111133228.190801-1-andre.przywara@arm.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: hisi_sas: Set a port invalid only if there are no devices attached when refreshing port id

[ Upstream commit f58c89700630da6554b24fd3df293a24874c10c1 ]

Currently the driver sets the port invalid if one phy in the port is not
enabled, which may cause issues in expander situation. In directly attached
situation, if phy up doesn't occur in time when refreshing port id, the
port is incorrectly set to invalid which will also cause disk lost.

Therefore set a port invalid only if there are no devices attached to the
port.

Signed-off-by: Yihang Li <liyihang9@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Link: https://lore.kernel.org/r/1672805000-141102-3-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: hisi_sas: Use abort task set to reset SAS disks when discovered

[ Upstream commit 037b48057e8b485a8d72f808122796aeadbbee32 ]

Currently clear task set is used to abort all commands remaining in the
disk when the SAS disk is discovered, and if the disk is discovered by two
initiators, other I_T nexuses are also affected. So use abort task set
instead and take effect only on the specified I_T nexus.

Signed-off-by: Xingui Yang <yangxingui@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Link: https://lore.kernel.org/r/1672805000-141102-2-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

KVM: s390: interrupt: use READ_ONCE() before cmpxchg()

[ Upstream commit 42400d99e9f0728c17240edb9645637ead40f6b9 ]

Use READ_ONCE() before cmpxchg() to prevent that the compiler generates
code that fetches the to be compared old value several times from memory.

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Reviewed-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Link: https://lore.kernel.org/r/20230109145456.2895385-1-hca@linux.ibm.com
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

spi: spidev: remove debug messages that access spidev->spi without locking

[ Upstream commit 6b35b173dbc1711f8d272e3f322d2ad697015919 ]

The two debug messages in spidev_open() dereference spidev->spi without
taking the lock and without checking if it's not null. This can lead to
a crash. Drop the messages as they're not needed - the user-space will
get informed about ENOMEM with the syscall return value.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Link: https://lore.kernel.org/r/20230106100719.196243-2-brgl@bgdev.pl
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cifs: fix potential memory leaks in session setup

[ Upstream commit 2fe58d977ee05da5bb89ef5dc4f5bf2dc15db46f ]

Make sure to free cifs_ses::auth_key.response before allocating it as
we might end up leaking memory in reconnect or mounting.

Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amdkfd: Fix NULL pointer error for GC 11.0.1 on mGPU

[ Upstream commit a6941f89d7c6a6ba49316bbd7da2fb2f719119a7 ]

The point bo->kfd_bo is NULL for queue's write pointer BO
when creating queue on mGPU. To avoid using the pointer
fixes the error.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amdkfd: Add sync after creating vram bo

[ Upstream commit ba029e9991d9be90a28b6a0ceb25e9a6fb348829 ]

There will be data corruption on vram allocated by svm
if the initialization is not complete and application is
writting on the memory. Adding sync to wait for the
initialization completion is to resolve this issue.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: resource: Skip IRQ override on Asus Expertbook B2402CBA

[ Upstream commit 77c7248882385397cd7dffe9e1437f59f32ce2de ]

Like the Asus Expertbook B2502CBA and various Asus Vivobook laptops,
the Asus Expertbook B2402CBA has an ACPI DSDT table that describes IRQ 1
as ActiveLow while the kernel overrides it to Edge_High. This prevents the
keyboard from working. To fix this issue, add this laptop to the
skip_override_table so that the kernel does not override IRQ 1.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=216864
Tested-by: zelenat <zelenat@gmail.com>
Signed-off-by: Tamim Khan <tamim@fusetak.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: fsl-asoc-card: Fix naming of AC'97 CODEC widgets

[ Upstream commit 242fc66ae6e1e2b8519daacc7590a73cd0e8a6e4 ]

The fsl-asoc-card AC'97 support currently tries to route to Playback and
Capture widgets provided by the AC'97 CODEC. This doesn't work since the
generic AC'97 driver registers with an "AC97" at the front of the stream
and hence widget names, update to reflect reality. It's not clear to me
if or how this ever worked.

Acked-by: Shengjiu Wang <shengjiu.wang@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230106-asoc-udoo-probe-v1-2-a5d7469d4f67@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: fsl_ssi: Rename AC'97 streams to avoid collisions with AC'97 CODEC

[ Upstream commit 8c6a42b5b0ed6f96624f56954e93eeae107440a6 ]

The SSI driver calls the AC'97 playback and transmit streams "AC97 Playback"
and "AC97 Capture" respectively. This is the same name used by the generic
AC'97 CODEC driver in ASoC, creating confusion for the Freescale ASoC card
when it attempts to use these widgets in routing. Add a "CPU" in the name
like the regular DAIs registered by the driver to disambiguate.

Acked-by: Shengjiu Wang <shengjiu.wang@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230106-asoc-udoo-probe-v1-1-a5d7469d4f67@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpufreq: armada-37xx: stop using 0 as NULL pointer

[ Upstream commit 08f0adb193c008de640fde34a2e00a666c01d77c ]

Use NULL for NULL pointer to fix the following sparse warning:
drivers/cpufreq/armada-37xx-cpufreq.c:448:32: sparse: warning: Using plain integer as NULL pointer

Signed-off-by: Miles Chen <miles.chen@mediatek.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tools/nolibc: prevent gcc from making memset() loop over itself

[ Upstream commit 1bfbe1f3e96720daf185f03d101f072d69753f88 ]

When building on ARM in thumb mode with gcc-11.3 at -O2 or -O3,
nolibc-test segfaults during the select() tests. It turns out that at
this level, gcc recognizes an opportunity for using memset() to zero
the fd_set, but it miscompiles it because it also recognizes a memset
pattern as well, and decides to call memset() from the memset() code:

  000122bc <memset>:
     122bc:       b510            push    {r4, lr}
     122be:       0004            movs    r4, r0
     122c0:       2a00            cmp     r2, #0
     122c2:       d003            beq.n   122cc <memset+0x10>
     122c4:       23ff            movs    r3, #255        ; 0xff
     122c6:       4019            ands    r1, r3
     122c8:       f7ff fff8       bl      122bc <memset>
     122cc:       0020            movs    r0, r4
     122ce:       bd10            pop     {r4, pc}

Simply placing an empty asm() statement inside the loop suffices to
avoid this.

Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tools/nolibc: fix missing includes causing build issues at -O0

[ Upstream commit 55abdd1f5e1e07418bf4a46c233a92f83cb5ae97 ]

After the nolibc includes were split to facilitate portability from
standard libcs, programs that include only what they need may miss
some symbols which are needed by libgcc. This is the case for raise()
which is needed by the divide by zero code in some architectures for
example.

Regardless, being able to include only the apparently needed files is
convenient.

Instead of trying to move all exported definitions to a single file,
since this can change over time, this patch takes another approach
consisting in including the nolibc header at the end of all standard
include files. This way their types and functions are already known
at the moment of inclusion, and including any single one of them is
sufficient to bring all the required ones.

Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tools/nolibc: Fix S_ISxxx macros

[ Upstream commit 16f5cea74179b5795af7ce359971f5128d10f80e ]

The mode field has the type encoded as an value in a field, not as a bit
mask. Mask the mode with S_IFMT instead of each type to test. Otherwise,
false positives are possible: eg S_ISDIR will return true for block
devices because S_IFDIR = 0040000 and S_IFBLK = 0060000 since mode is
masked with S_IFDIR instead of S_IFMT. These macros now match the
similar definitions in tools/include/uapi/linux/stat.h.

Signed-off-by: Warner Losh <imp@bsdimp.com>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nolibc: fix fd_set type

[ Upstream commit feaf75658783a919410f8c2039dbc24b6a29603d ]

The kernel uses unsigned long for the fd_set bitmap,
but nolibc use u32. This works fine on little endian
machines, but fails on big endian. Convert to unsigned
long to fix this.

Signed-off-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/x86/intel/uncore: Add Emerald Rapids

[ Upstream commit 5268a2842066c227e6ccd94bac562f1e1000244f ]

From the perspective of the uncore PMU, the new Emerald Rapids is the
same as the Sapphire Rapids. The only difference is the event list,
which will be supported in the perf tool later.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230106160449.3566477-4-kan.liang@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/x86/msr: Add Emerald Rapids

[ Upstream commit 69ced4160969025821f2999ff92163ed26568f1c ]

The same as Sapphire Rapids, the SMI_COUNT MSR is also supported on
Emerald Rapids. Add Emerald Rapids model.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230106160449.3566477-3-kan.liang@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/x86/msr: Add Meteor Lake support

[ Upstream commit 6887a4d3aede084bf08b70fbc9736c69fce05d7f ]

Meteor Lake is Intel's successor to Raptor lake. PPERF and SMI_COUNT MSRs
are also supported.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Link: https://lore.kernel.org/r/20230104201349.1451191-7-kan.liang@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/x86/cstate: Add Meteor Lake support

[ Upstream commit 01f2ea5bcf89dbd7a6530dbce7f2fb4e327e7006 ]

Meteor Lake is Intel's successor to Raptor lake. From the perspective of
Intel cstate residency counters, there is nothing changed compared with
Raptor lake.

Share adl_cstates with Raptor lake.
Update the comments for Meteor Lake.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Link: https://lore.kernel.org/r/20230104201349.1451191-6-kan.liang@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

xen/pvcalls: free active map buffer on pvcalls_front_free_map

[ Upstream commit f57034cedeb6e00256313a2a6ee67f974d709b0b ]

Data buffer for active map is allocated in alloc_active_ring and freed
in free_active_ring function, which is used only for the error
cleanup. pvcalls_front_release is calling pvcalls_front_free_map which
ends foreign access for this buffer, but doesn't free allocated pages.
Call free_active_ring to clean all allocated resources.

Signed-off-by: Oleksii Moisieiev <oleksii_moisieiev@epam.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Link: https://lore.kernel.org/r/6a762ee32dd655cbb09a4aa0e2307e8919761311.1671531297.git.oleksii_moisieiev@epam.com
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390: expicitly align _edata and _end symbols on page boundary

[ Upstream commit 45d619bdaf799196d702a9ae464b07066d6db2f9 ]

Symbols _edata and _end in the linker script are the
only unaligned expicitly on page boundary. Although
_end is aligned implicitly by BSS_SECTION macro that
is still inconsistent and could lead to a bug if a tool
or function would assume that _edata is as aligned as
others.

For example, vmem_map_init() function does not align
symbols _etext, _einittext etc. Should these symbols
be unaligned as well, the size of ranges to update
were short on one page.

Instead of fixing every occurrence of this kind in the
code and external tools just force the alignment on
these two symbols.

Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/debug: add _ASM_S390_ prefix to header guard

[ Upstream commit 0d4d52361b6c29bf771acd4fa461f06d78fb2fac ]

Using DEBUG_H without a prefix is very generic and inconsistent with
other header guards in arch/s390/include/asm. In fact it collides with
the same name in the ath9k wireless driver though that depends on !S390
via disabled wireless support. Let's just use a consistent header guard
name and prevent possible future trouble.

Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm: Add orientation quirk for Lenovo ideapad D330-10IGL

[ Upstream commit 0688773f0710528e1ab302c3d6317e269f2e2e6e ]

Panel is 800x1280 but mounted on a detachable form factor sideways.

Signed-off-by: Patrick Thompson <ptf@google.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20221220205826.178008-1-ptf@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: usb: cdc_ether: add support for Thales Cinterion PLS62-W modem

[ Upstream commit eea8ce81fbb544e3caad1a1c876ba1af467b3d3c ]

This modem has 7 interfaces, 5 of them are serial interfaces and are
driven by cdc_acm, while 2 of them are wwan interfaces and are driven
by cdc_ether:
If 0: Abstract (modem)
If 1: Abstract (modem)
If 2: Abstract (modem)
If 3: Abstract (modem)
If 4: Abstract (modem)
If 5: Ethernet Networking
If 6: Ethernet Networking

Without this change, the 2 network interfaces will be named to usb0
and usb1, our QA think the names are confusing and filed a bug on it.

After applying this change, the name will be wwan0 and wwan1, and
they could work well with modem manager.

Signed-off-by: Hui Wang <hui.wang@canonical.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20230105034249.10433-1-hui.wang@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm/a6xx: Avoid gx gbit halt during rpm suspend

[ Upstream commit f4a75b5933c998e60fd812a7680e0971eb1c7cee ]

As per the downstream driver, gx gbif halt is required only during
recovery sequence. So lets avoid it during regular rpm suspend.

Signed-off-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/515279/
Link: https://lore.kernel.org/r/20221216223253.1.Ice9c47bfeb1fddb8dc377a3491a043a3ee7fca7d@changeid
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: fsl_micfil: Correct the number of steps on SX controls

[ Upstream commit cdfa92eb90f5770b26a79824ef213ebdbbd988b1 ]

The parameter "max" of SOC_SINGLE_SX_TLV() means the number of steps
rather than maximum value. This patch corrects the minimum value to -8
and the number of steps to 15.

Signed-off-by: Chancel Liu <chancel.liu@nxp.com>
Acked-by: Shengjiu Wang <shengjiu.wang@gmail.com>
Link: https://lore.kernel.org/r/20230104025754.3019235-1-chancel.liu@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpufreq: Add SM6375 to cpufreq-dt-platdev blocklist

[ Upstream commit faf28e240dd118d9521c68aeb9388b9b8f02d9d0 ]

The Qualcomm SM6375 platform uses the qcom-cpufreq-hw driver, so add
it to the cpufreq-dt-platdev driver's blocklist.

Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

kcsan: test: don't put the expect array on the stack

[ Upstream commit 5b24ac2dfd3eb3e36f794af3aa7f2828b19035bd ]

Size of the 'expect' array in the __report_matches is 1536 bytes, which
is exactly the default frame size warning limit of the xtensa
architecture.
As a result allmodconfig xtensa kernel builds with the gcc that does not
support the compiler plugins (which otherwise would push the said
warning limit to 2K) fail with the following message:

  kernel/kcsan/kcsan_test.c:257:1: error: the frame size of 1680 bytes
    is larger than 1536 bytes

Fix it by dynamically allocating the 'expect' array.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Marco Elver <elver@google.com>
Tested-by: Marco Elver <elver@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: support machine driver with max98360

[ Upstream commit 810948f45d99c46b60852ef2a5a2777c12d6bb3e ]

Signed-off-by: Mars Chen <chenxiangrui@huaqin.corp-partner.google.com>
Link: https://lore.kernel.org/r/20221228103812.450956-1-chenxiangrui@huaqin.corp-partner.google.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: amd: yc: Add ASUS M5402RA into DMI table

[ Upstream commit a0dd7fcab5cd221fa960f594c586e1f9f16c02c0 ]

ASUS VivoBook 13 OLED (M5402RA) needs this quirk to get the built-in microphone working properly.

Signed-off-by: Aniol Martí <aniol@aniolmarti.cat>
Link: https://lore.kernel.org/r/20221227224932.9771-1-aniol@aniolmarti.cat
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: mediatek: mt8186: Add machine support for max98357a

[ Upstream commit 8a54f666db581bbf07494cca44a0124acbced581 ]

Add support for mt8186 with mt6366 and max98357a.

Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
Link: https://lore.kernel.org/r/20221228115756.28014-1-allen-kh.cheng@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: mediatek: mt8186: support rt5682s_max98360

[ Upstream commit 6e1dbf694d7cd1737ee14866e9e05016ccc9ac40 ]

Add support for using the rt5682s codec together with max98360a on
MT8186-MT6366-RT1019-RT5682S machines.

Signed-off-by: tongjian <tongjian@huaqin.corp-partner.google.com>
Link: https://lore.kernel.org/r/20221228122230.3818533-2-tongjian@huaqin.corp-partner.google.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpufreq: Add Tegra234 to cpufreq-dt-platdev blocklist

[ Upstream commit 01c5bb0cc2a39fbc56ff9a5ef28b79447f0c2351 ]

Tegra234 platform uses the tegra194-cpufreq driver, so add it
to the blocklist in cpufreq-dt-platdev driver to avoid the cpufreq
driver registration from there.

Signed-off-by: Sumit Gupta <sumitg@nvidia.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpufreq: CPPC: Add u64 casts to avoid overflowing

[ Upstream commit f5f94b9c8b805d87ff185caf9779c3a4d07819e3 ]

The fields of the _CPC object are unsigned 32-bits values.
To avoid overflows while using _CPC's values, add 'u64' casts.

Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

spi: cadence: Fix busy cycles calculation

[ Upstream commit e8bb8f19e73a1e855e54788f8673b9b49e46b5cd ]

If xSPI is in x2/x4/x8 mode to calculate busy
cycles, busy bits count must be divided by the number
of lanes.
If opcommand is using 8 busy bits, but SPI is
in x4 mode, there will be only 2 busy cycles.

Signed-off-by: Witold Sadowski <wsadowski@marvell.com>
Reviewed-by: Chandrakala Chavva <cchavva@marvell.com>
Reviewed-by: Sunil Kovvuri Goutham <sgoutham@marvell.com>
Link: https://lore.kernel.org/r/20221219144254.20883-2-wsadowski@marvell.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: amd: yc: Add Razer Blade 14 2022 into DMI table

[ Upstream commit 68506a173dd700c2bd794dcc3489edcdb8ee35c6 ]

Razer Blade 14 (2022) - RZ09-0427 needs the quirk to enable the built in microphone

Signed-off-by: Wim Van Boven <wimvanboven@gmail.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20221216081828.12382-1-wimvanboven@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: SOF: Add FW state to debugfs

[ Upstream commit 9a9134fd56f6ba614ff7b2b3b0bac0bf1d0dc0c9 ]

Allow system health detection mechanisms to check the FW state, this
will allow them to check if the FW is in its "crashed" state going
forward to help automatically diagnose driver state.

Signed-off-by: Curtis Malainey <cujomalainey@chromium.org>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20221220125629.8469-4-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: SOF: pm: Always tear down pipelines before DSP suspend

[ Upstream commit d185e0689abc98ef55fb7a7d75aa0c48a0ed5838 ]

When the DSP is suspended while the firmware is in the crashed state, we
skip tearing down the pipelines. This means that the widget reference
counts will not get to reset to 0 before suspend. This will lead to
errors with resuming audio after system resume. To fix this, invoke the
tear_down_all_pipelines op before skipping to DSP suspend.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Curtis Malainey <cujomalainey@chromium.org>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20221220125629.8469-3-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: SOF: pm: Set target state earlier

[ Upstream commit 6f95eec6fb89e195dbdf30de65553c7fc57d9372 ]

If the DSP crashes before the system suspends, the setting of target state
will be skipped because the firmware state will no longer be
SOF_FW_BOOT_COMPLETE. This leads to the incorrect assumption that the
DSP should suspend to D0I3 instead of suspending to D3. To fix this,
set the target_state before we skip to DSP suspend even when the DSP has
crashed.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Curtis Malainey <cujomalainey@chromium.org>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20221220125629.8469-2-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: iscsi: Fix multiple iSCSI session unbind events sent to userspace

[ Upstream commit a3be19b91ea7121d388084e8c07f5b1b982eb40c ]

It was observed that the kernel would potentially send
ISCSI_KEVENT_UNBIND_SESSION multiple times. Introduce 'target_state' in
iscsi_cls_session() to make sure session will send only one unbind session
event.

This introduces a regression wrt. the issue fixed in commit 13e60d3ba287
("scsi: iscsi: Report unbind session event when the target has been
removed"). If iscsid dies for any reason after sending an unbind session to
kernel, once iscsid is restarted, the kernel's ISCSI_KEVENT_UNBIND_SESSION
event is lost and userspace is then unable to logout. However, the session
is actually in invalid state (its target_id is INVALID) so iscsid should
not sync this session during restart.

Consequently we need to check the session's target state during iscsid
restart.  If session is in unbound state, do not sync this session and
perform session teardown. This is OK because once a session is unbound, we
can not recover it any more (mainly because its target id is INVALID).

Signed-off-by: Wenchao Hao <haowenchao@huawei.com>
Link: https://lore.kernel.org/r/20221126010752.231917-1-haowenchao@huawei.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Reviewed-by: Wu Bo <wubo40@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tcp: fix rate_app_limited to default to 1

[ Upstream commit 300b655db1b5152d6101bcb6801d50899b20c2d6 ]

The initial default value of 0 for tp->rate_app_limited was incorrect,
since a flow is indeed application-limited until it first sends
data. Fixing the default to be 1 is generally correct but also
specifically will help user-space applications avoid using the initial
tcpi_delivery_rate value of 0 that persists until the connection has
some non-zero bandwidth sample.

Fixes: eb8329e0a04d ("tcp: export data delivery rate")
Suggested-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David Morley <morleyd@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Tested-by: David Morley <morleyd@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bnxt: Do not read past the end of test names

[ Upstream commit d3e599c090fc6977331150c5f0a69ab8ce87da21 ]

Test names were being concatenated based on a offset beyond the end of
the first name, which tripped the buffer overflow detection logic:

 detected buffer overflow in strnlen
 [...]
 Call Trace:
 bnxt_ethtool_init.cold+0x18/0x18

Refactor struct hwrm_selftest_qlist_output to use an actual array,
and adjust the concatenation to use snprintf() rather than a series of
strncat() calls.

Reported-by: Niklas Cassel <Niklas.Cassel@wdc.com>
Link: https://lore.kernel.org/lkml/Y8F%2F1w1AZTvLglFX@x1-carbon/
Tested-by: Niklas Cassel <Niklas.Cassel@wdc.com>
Fixes: eb51365846bc ("bnxt_en: Add basic ethtool -t selftest support.")
Cc: Michael Chan <michael.chan@broadcom.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: netdev@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Niklas Cassel <niklas.cassel@wdc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: stmmac: enable all safety features by default

[ Upstream commit fdfc76a116b5e9d3e98e6c96fe83b42d011d21d4 ]

In the original implementation of dwmac5
commit 8bf993a5877e ("net: stmmac: Add support for DWMAC5 and implement Safety Features")
all safety features were enabled by default.

Later it seems some implementations didn't have support for all the
features, so in
commit 5ac712dcdfef ("net: stmmac: enable platform specific safety features")
the safety_feat_cfg structure was added to the callback and defined for
some platforms to selectively enable these safety features.

The problem is that only certain platforms were given that software
support. If the automotive safety package bit is set in the hardware
features register the safety feature callback is called for the platform,
and for platforms that didn't get a safety_feat_cfg defined this results
in the following NULL pointer dereference:

[    7.933303] Call trace:
[    7.935812]  dwmac5_safety_feat_config+0x20/0x170 [stmmac]
[    7.941455]  __stmmac_open+0x16c/0x474 [stmmac]
[    7.946117]  stmmac_open+0x38/0x70 [stmmac]
[    7.950414]  __dev_open+0x100/0x1dc
[    7.954006]  __dev_change_flags+0x18c/0x204
[    7.958297]  dev_change_flags+0x24/0x6c
[    7.962237]  do_setlink+0x2b8/0xfa4
[    7.965827]  __rtnl_newlink+0x4ec/0x840
[    7.969766]  rtnl_newlink+0x50/0x80
[    7.973353]  rtnetlink_rcv_msg+0x12c/0x374
[    7.977557]  netlink_rcv_skb+0x5c/0x130
[    7.981500]  rtnetlink_rcv+0x18/0x2c
[    7.985172]  netlink_unicast+0x2e8/0x340
[    7.989197]  netlink_sendmsg+0x1a8/0x420
[    7.993222]  ____sys_sendmsg+0x218/0x280
[    7.997249]  ___sys_sendmsg+0xac/0x100
[    8.001103]  __sys_sendmsg+0x84/0xe0
[    8.004776]  __arm64_sys_sendmsg+0x24/0x30
[    8.008983]  invoke_syscall+0x48/0x114
[    8.012840]  el0_svc_common.constprop.0+0xcc/0xec
[    8.017665]  do_el0_svc+0x38/0xb0
[    8.021071]  el0_svc+0x2c/0x84
[    8.024212]  el0t_64_sync_handler+0xf4/0x120
[    8.028598]  el0t_64_sync+0x190/0x194

Go back to the original behavior, if the automotive safety package
is found to be supported in hardware enable all the features unless
safety_feat_cfg is passed in saying this particular platform only
supports a subset of the features.

Fixes: 5ac712dcdfef ("net: stmmac: enable platform specific safety features")
Reported-by: Ning Cai <ncai@quicinc.com>
Signed-off-by: Andrew Halaney <ahalaney@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal: core: call put_device() only after device_register() fails

[ Upstream commit 6c54b7bc8a31ce0f7cc7f8deef05067df414f1d8 ]

put_device() shouldn't be called before a prior call to
device_register(). __thermal_cooling_device_register() doesn't follow
that properly and needs fixing. Also
thermal_cooling_device_destroy_sysfs() is getting called unnecessarily
on few error paths.

Fix all this by placing the calls at the right place.

Based on initial work done by Caleb Connolly.

Fixes: 4748f9687caa ("thermal: core: fix some possible name leaks in error paths")
Fixes: c408b3d1d9bb ("thermal: Validate new state in cur_state_store()")
Reported-by: Caleb Connolly <caleb.connolly@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Frank Rowand <frowand.list@gmail.com>
Reviewed-by: Yang Yingliang <yangyingliang@huawei.com>
Tested-by: Caleb Connolly <caleb.connolly@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/core: fix error code in __thermal_cooling_device_register()

[ Upstream commit e49a1e1ee078aee21006192076a8d93335e0daa9 ]

Return an error pointer if ->get_max_state() fails.  The current code
returns NULL which will cause an oops in the callers.

Fixes: c408b3d1d9bb ("thermal: Validate new state in cur_state_store()")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Stable-dep-of: 6c54b7bc8a31 ("thermal: core: call put_device() only after device_register() fails")
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal: Validate new state in cur_state_store()

[ Upstream commit c408b3d1d9bbc7de5fb0304fea424ef2539da616 ]

In cur_state_store(), the new state of the cooling device is received
from user-space and is not validated by the thermal core but the same is
left for the individual drivers to take care of. Apart from duplicating
the code it leaves possibility for introducing bugs where a driver may
not do it right.

Lets make the thermal core check the new state itself and store the max
value in the cooling device structure.

Link: https://lore.kernel.org/all/Y0ltRJRjO7AkawvE@kili/
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Stable-dep-of: 6c54b7bc8a31 ("thermal: core: call put_device() only after device_register() fails")
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: dsa: microchip: ksz9477: port map correction in ALU table entry register

[ Upstream commit 6c977c5c2e4c5d8ad1b604724cc344e38f96fe9b ]

ALU table entry 2 register in KSZ9477 have bit positions reserved for
forwarding port map. This field is referred in ksz9477_fdb_del() for
clearing forward port map and alu table.

But current fdb_del refer ALU table entry 3 register for accessing forward
port map. Update ksz9477_fdb_del() to get forward port map from correct
alu table entry register.

With this bug, issue can be observed while deleting static MAC entries.
Delete any specific MAC entry using "bridge fdb del" command. This should
clear all the specified MAC entries. But it is observed that entries with
self static alone are retained.

Tested on LAN9370 EVB since ksz9477_fdb_del() is used common across
LAN937x and KSZ series.

Fixes: b987e98e50ab ("dsa: add DSA switch driver for Microchip KSZ9477")
Signed-off-by: Rakesh Sankaranarayanan <rakesh.sankaranarayanan@microchip.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Link: https://lore.kernel.org/r/20230118174735.702377-1-rakesh.sankaranarayanan@microchip.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/net: toeplitz: fix race on tpacket_v3 block close

[ Upstream commit 903848249a781d76d59561d51676c95b3a4d7162 ]

Avoid race between process wakeup and tpacket_v3 block timeout.

The test waits for cfg_timeout_msec for packets to arrive. Packets
arrive in tpacket_v3 rings, which pass packets ("frames") to the
process in batches ("blocks"). The sk waits for req3.tp_retire_blk_tov
msec to release a block.

Set the block timeout lower than the process waiting time, else
the process may find that no block has been released by the time it
scans the socket list. Convert to a ring of more than one, smaller,
blocks with shorter timeouts. Blocks must be page aligned, so >= 64KB.

Fixes: 5ebfb4cc3048 ("selftests/net: toeplitz test")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20230118151847.4124260-1-willemdebruijn.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

driver core: Fix test_async_probe_init saves device in wrong array

[ Upstream commit 9be182da0a7526f1b9a3777a336f83baa2e64d23 ]

In test_async_probe_init, second set of asynchronous devices are saved
in sync_dev[sync_id], which should be async_dev[async_id].
This makes these devices not unregistered when exit.

> modprobe test_async_driver_probe && \
> modprobe -r test_async_driver_probe && \
> modprobe test_async_driver_probe
 ...
> sysfs: cannot create duplicate filename '/devices/platform/test_async_driver.4'
> kobject_add_internal failed for test_async_driver.4 with -EEXIST,
  don't try to register things with the same name in the same directory.

Fixes: 57ea974fb871 ("driver core: Rewrite test_async_driver_probe to cover serialization and NUMA affinity")
Signed-off-by: Chen Zhongjin <chenzhongjin@huawei.com>
Link: https://lore.kernel.org/r/20221125063541.241328-1-chenzhongjin@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

w1: fix WARNING after calling w1_process()

[ Upstream commit 36225a7c72e9e3e1ce4001b6ce72849f5c9a2d3b ]

I got the following WARNING message while removing driver(ds2482):

------------[ cut here ]------------
do not call blocking ops when !TASK_RUNNING; state=1 set at [<000000002d50bfb6>] w1_process+0x9e/0x1d0 [wire]
WARNING: CPU: 0 PID: 262 at kernel/sched/core.c:9817 __might_sleep+0x98/0xa0
CPU: 0 PID: 262 Comm: w1_bus_master1 Tainted: G                 N 6.1.0-rc3+ #307
RIP: 0010:__might_sleep+0x98/0xa0
Call Trace:
 exit_signals+0x6c/0x550
 do_exit+0x2b4/0x17e0
 kthread_exit+0x52/0x60
 kthread+0x16d/0x1e0
 ret_from_fork+0x1f/0x30

The state of task is set to TASK_INTERRUPTIBLE in loop in w1_process(),
set it to TASK_RUNNING when it breaks out of the loop to avoid the
warning.

Fixes: 3c52e4e62789 ("W1: w1_process, block or sleep")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221205101558.3599162-1-yangyingliang@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

w1: fix deadloop in __w1_remove_master_device()

[ Upstream commit 25d5648802f12ae486076ceca5d7ddf1fef792b2 ]

I got a deadloop report while doing device(ds2482) add/remove test:

  [  162.241881] w1_master_driver w1_bus_master1: Waiting for w1_bus_master1 to become free: refcnt=1.
  [  163.272251] w1_master_driver w1_bus_master1: Waiting for w1_bus_master1 to become free: refcnt=1.
  [  164.296157] w1_master_driver w1_bus_master1: Waiting for w1_bus_master1 to become free: refcnt=1.
  ...

__w1_remove_master_device() can't return, because the dev->refcnt is not zero.

w1_add_master_device() |
  w1_alloc_dev() |
    atomic_set(&dev->refcnt, 2) |
  kthread_run() |
|__w1_remove_master_device()
|  kthread_stop()
  // KTHREAD_SHOULD_STOP is set, |
  // threadfn(w1_process) won't be |
  // called. |
  kthread() |
|  // refcnt will never be 0, it's deadloop.
|  while (atomic_read(&dev->refcnt)) {...}

After calling w1_add_master_device(), w1_process() is not really
invoked, before w1_process() starting, if kthread_stop() is called
in __w1_remove_master_device(), w1_process() will never be called,
the refcnt can not be decreased, then it causes deadloop in remove
function because of non-zero refcnt.

We need to make sure w1_process() is really started, so move the
set refcnt into w1_process() to fix this problem.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221205080434.3149205-1-yangyingliang@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

device property: fix of node refcount leak in fwnode_graph_get_next_endpoint()

[ Upstream commit 39af728649b05e88a2b40e714feeee6451c3f18e ]

The 'parent' returned by fwnode_graph_get_port_parent()
with refcount incremented when 'prev' is not NULL, it
needs be put when finish using it.

Because the parent is const, introduce a new variable to
store the returned fwnode, then put it before returning
from fwnode_graph_get_next_endpoint().

Fixes: b5b41ab6b0c1 ("device property: Check fwnode->secondary in fwnode_graph_get_next_endpoint()")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-and-tested-by: Daniel Scally <djrscally@gmail.com>
Link: https://lore.kernel.org/r/20221123022542.2999510-1-yangyingliang@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ptdma: pt_core_execute_cmd() should use spinlock

[ Upstream commit 95e5fda3b5f9ed8239b145da3fa01e641cf5d53c ]

The interrupt handler (pt_core_irq_handler()) of the ptdma
driver can be called from interrupt context. The code flow
in this function can lead down to pt_core_execute_cmd() which
will attempt to grab a mutex, which is not appropriate in
interrupt context and ultimately leads to a kernel panic.
The fix here changes this mutex to a spinlock, which has
been verified to resolve the issue.

Fixes: fa5d823b16a9 ("dmaengine: ptdma: Initial driver for the AMD PTDMA")
Signed-off-by: Eric Pilmore <epilmore@gigaio.com>
Link: https://lore.kernel.org/r/20230119033907.35071-1-epilmore@gigaio.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: dwc3: fix extcon dependency

[ Upstream commit 7d80dbd708c18c683dd34f79b600a05307707ce8 ]

The dwc3 core support now links against the extcon subsystem,
so it cannot be built-in when extcon is a loadable module:

arm-linux-gnueabi-ld: drivers/usb/dwc3/core.o: in function `dwc3_get_extcon':
core.c:(.text+0x572): undefined reference to `extcon_get_edev_by_phandle'
arm-linux-gnueabi-ld: core.c:(.text+0x596): undefined reference to `extcon_get_extcon_dev'
arm-linux-gnueabi-ld: core.c:(.text+0x5ea): undefined reference to `extcon_find_edev_by_node'

There was already a Kconfig dependency in the dual-role support,
but this is now needed for the entire dwc3 driver.

It is still possible to build dwc3 without extcon, but this
prevents it from being set to built-in when extcon is a loadable
module.

Fixes: d182c2e1bc92 ("usb: dwc3: Don't switch OTG -> peripheral if extcon is present")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
Link: https://lore.kernel.org/r/20230118090147.2126563-1-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tcp: avoid the lookup process failing to get sk in ehash table

[ Upstream commit 3f4ca5fafc08881d7a57daa20449d171f2887043 ]

While one cpu is working on looking up the right socket from ehash
table, another cpu is done deleting the request socket and is about
to add (or is adding) the big socket from the table. It means that
we could miss both of them, even though it has little chance.

Let me draw a call trace map of the server side.
   CPU 0                           CPU 1
   -----                           -----
tcp_v4_rcv()                  syn_recv_sock()
                            inet_ehash_insert()
                            -> sk_nulls_del_node_init_rcu(osk)
__inet_lookup_established()
                            -> __sk_nulls_add_node_rcu(sk, list)

Notice that the CPU 0 is receiving the data after the final ack
during 3-way shakehands and CPU 1 is still handling the final ack.

Why could this be a real problem?
This case is happening only when the final ack and the first data
receiving by different CPUs. Then the server receiving data with
ACK flag tries to search one proper established socket from ehash
table, but apparently it fails as my map shows above. After that,
the server fetches a listener socket and then sends a RST because
it finds a ACK flag in the skb (data), which obeys RST definition
in RFC 793.

Besides, Eric pointed out there's one more race condition where it
handles tw socket hashdance. Only by adding to the tail of the list
before deleting the old one can we avoid the race if the reader has
already begun the bucket traversal and it would possibly miss the head.

Many thanks to Eric for great help from beginning to end.

Fixes: 5e0724d027f0 ("tcp/dccp: fix hashdance race for passive sessions")
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/lkml/20230112065336.41034-1-kerneljasonxing@gmail.com/
Link: https://lore.kernel.org/r/20230118015941.1313-1-kerneljasonxing@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvme-pci: fix timeout request state check

[ Upstream commit 1c5842085851f786eba24a39ecd02650ad892064 ]

Polling the completion can progress the request state to IDLE, either
inline with the completion, or through softirq. Either way, the state
may not be COMPLETED, so don't check for that. We only care if the state
isn't IN_FLIGHT.

This is fixing an issue where the driver aborts an IO that we just
completed. Seeing the "aborting" message instead of "polled" is very
misleading as to where the timeout problem resides.

Fixes: bf392a5dc02a9b ("nvme-pci: Remove tag from process cq")
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: sched: gred: prevent races when adding offloads to stats

[ Upstream commit 339346d49ae0859fe19b860998867861d37f1a76 ]

Naresh reports seeing a warning that gred is calling
u64_stats_update_begin() with preemption enabled.
Arnd points out it's coming from _bstats_update().

We should be holding the qdisc lock when writing
to stats, they are also updated from the datapath.

Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>
Link: https://lore.kernel.org/all/CA+G9fYsTr9_r893+62u6UGD3dVaCE-kN9C-Apmb2m=hxjc1Cqg@mail.gmail.com/
Fixes: e49efd5288bd ("net: sched: gred: support reporting stats from offloads")
Link: https://lore.kernel.org/r/20230113044137.1383067-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: fix issues with driver unload

[ Upstream commit e433adc60f7f847e734c56246b09291532f29b6d ]

Currently, we run into a number of WARN()s when attempting to unload the
amdgpu driver (e.g. using "modprobe -r amdgpu"). These all stem from
calling drm_encoder_cleanup() too early. So, to fix this we can stop
calling drm_encoder_cleanup() from amdgpu_dm_fini() and instead have it
be called from amdgpu_dm_encoder_destroy(). Also, we don't need to free
in amdgpu_dm_encoder_destroy() since mst_encoders[] isn't explicitly
allocated by the slab allocator.

Fixes: f74367e492ba ("drm/amdgpu/display: create fake mst encoders ahead of time (v4)")
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

phy: phy-can-transceiver: Skip warning if no "max-bitrate"

[ Upstream commit bc30c15f275484f9b9fe27c2fa0895f3022d9943 ]

According to the DT bindings, the "max-bitrate" property is optional.
However, when it is not present, a warning is printed.
Fix this by adding a missing check for -EINVAL.

Fixes: a4a86d273ff1b6f7 ("phy: phy-can-transceiver: Add support for generic CAN transceiver driver")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/88e158f97dd52ebaa7126cd9631f34764b9c0795.1674037334.git.geert+renesas@glider.be
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: tegra: Fix memory leak in terminate_all()

[ Upstream commit a7a7ee6f5a019ad72852c001abbce50d35e992f2 ]

Terminate vdesc when terminating an ongoing transfer.
This will ensure that the vdesc is present in the desc_terminated list
The descriptor will be freed later in desc_free_list().

This fixes the memory leaks which can happen when terminating an
ongoing transfer.

Fixes: ee17028009d4 ("dmaengine: tegra: Add tegra gpcdma driver")
Signed-off-by: Akhil R <akhilrajeev@nvidia.com>
Link: https://lore.kernel.org/r/20230118115801.15210-1-akhilrajeev@nvidia.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: xilinx_dma: call of_node_put() when breaking out of for_each_child_of_node()

[ Upstream commit 596b53ccc36a546ab28e8897315c5b4d1d5a0200 ]

Since for_each_child_of_node() will increase the refcount of node, we need
to call of_node_put() manually when breaking out of the iteration.

Fixes: 9cd4360de609 ("dma: Add Xilinx AXI Video Direct Memory Access Engine driver support")
Signed-off-by: Liu Shixin <liushixin2@huawei.com>
Acked-by: Peter Korsgaard <peter@korsgaard.com>
Link: https://lore.kernel.org/r/20221122021612.1908866-1-liushixin2@huawei.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cifs: fix potential deadlock in cache_refresh_path()

[ Upstream commit 9fb0db40513e27537fde63287aea920b60557a69 ]

Avoid getting DFS referral from an exclusive lock in
cache_refresh_path() because the tcon IPC used for getting the
referral could be disconnected and thus causing a deadlock as shown
below:

task A                       task B
======                       ======
cifs_demultiplex_thread()    dfs_cache_find()
 cifs_handle_standard()       cache_refresh_path()
  reconnect_dfs_server()       down_write()
   dfs_cache_noreq_find()       get_dfs_referral()
    down_read() <- deadlock      smb2_get_dfs_refer()
                                  SMB2_ioctl()
   cifs_send_recv()
    compound_send_recv()
     wait_for_response()

where task A cannot wake up task B because it is blocked on
down_read() due to the exclusive lock held in cache_refresh_path() and
therefore not being able to make progress.

Fixes: c9f711039905 ("cifs: keep referral server sessions alive")
Reviewed-by: Aurélien Aptel <aurelien.aptel@gmail.com>
Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/i915/selftests: Unwind hugepages to drop wakeref on error

[ Upstream commit 93eea624526fc7d070cdae463408665824075f54 ]

Make sure that upon error after we have acquired the wakeref we do
release it again.

v2: add another missing "goto out_wf"(Andi).

Fixes: 027c38b4121e ("drm/i915/selftests: Grab the runtime pm in shrink_thp")
Cc: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Chris Wilson <chris.p.wilson@linux.intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Nirmoy Das <nirmoy.das@intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230117123234.26487-1-nirmoy.das@intel.com
(cherry picked from commit 14ec40a88210151296fff3e981c1a7196ad9bf55)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

HID: betop: check shape of output reports

[ Upstream commit 3782c0d6edf658b71354a64d60aa7a296188fc90 ]

betopff_init() only checks the total sum of the report counts for each
report field to be at least 4, but hid_betopff_play() expects 4 report
fields.
A device advertising an output report with one field and 4 report counts
would pass the check but crash the kernel with a NULL pointer dereference
in hid_betopff_play().

Fixes: 52cd7785f3cd ("HID: betop: add drivers/hid/hid-betopff.c")
Signed-off-by: Pietro Borrello <borrello@diag.uniroma1.it>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Sasha Levin <sashal@kernel.org>

l2tp: prevent lockdep issue in l2tp_tunnel_register()

[ Upstream commit b9fb10d131b8c84af9bb14e2078d5c63600c7dea ]

lockdep complains with the following lock/unlock sequence:

     lock_sock(sk);
     write_lock_bh(&sk->sk_callback_lock);
[1]  release_sock(sk);
[2]  write_unlock_bh(&sk->sk_callback_lock);

We need to swap [1] and [2] to fix this issue.

Fixes: 0b2c59720e65 ("l2tp: close all race conditions in l2tp_tunnel_register()")
Reported-by: syzbot+bbd35b345c7cab0d9a08@syzkaller.appspotmail.com
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/netdev/20230114030137.672706-1-xiyou.wangcong@gmail.com/T/#m1164ff20628671b0f326a24cb106ab3239c70ce3
Cc: Cong Wang <cong.wang@bytedance.com>
Cc: Guillaume Nault <gnault@redhat.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

virtio-net: correctly enable callback during start_xmit

[ Upstream commit d71ebe8114b4bf622804b810f5e274069060a174 ]

Commit a7766ef18b33("virtio_net: disable cb aggressively") enables
virtqueue callback via the following statement:

        do {
if (use_napi)
virtqueue_disable_cb(sq->vq);

free_old_xmit_skbs(sq, false);

} while (use_napi && kick &&
               unlikely(!virtqueue_enable_cb_delayed(sq->vq)));

When NAPI is used and kick is false, the callback won't be enabled
here. And when the virtqueue is about to be full, the tx will be
disabled, but we still don't enable tx interrupt which will cause a TX
hang. This could be observed when using pktgen with burst enabled.

TO be consistent with the logic that tries to disable cb only for
NAPI, fixing this by trying to enable delayed callback only when NAPI
is enabled when the queue is about to be full.

Fixes: a7766ef18b33 ("virtio_net: disable cb aggressively")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Tested-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: macb: fix PTP TX timestamp failure due to packet padding

[ Upstream commit 7b90f5a665acd46efbbfa677a3a3a18d01ad6487 ]

PTP TX timestamp handling was observed to be broken with this driver
when using the raw Layer 2 PTP encapsulation. ptp4l was not receiving
the expected TX timestamp after transmitting a packet, causing it to
enter a failure state.

The problem appears to be due to the way that the driver pads packets
which are smaller than the Ethernet minimum of 60 bytes. If headroom
space was available in the SKB, this caused the driver to move the data
back to utilize it. However, this appears to cause other data references
in the SKB to become inconsistent. In particular, this caused the
ptp_one_step_sync function to later (in the TX completion path) falsely
detect the packet as a one-step SYNC packet, even when it was not, which
caused the TX timestamp to not be processed when it should be.

Using the headroom for this purpose seems like an unnecessary complexity
as this is not a hot path in the driver, and in most cases it appears
that there is sufficient tailroom to not require using the headroom
anyway. Remove this usage of headroom to prevent this inconsistency from
occurring and causing other problems.

Fixes: 653e92a9175e ("net: macb: add support for padding and fcs computation")
Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Claudiu Beznea <claudiu.beznea@microchip.com> # on SAMA7G5
Reviewed-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: Fix double increment of client_count in dma_chan_get()

[ Upstream commit f3dc1b3b4750851a94212dba249703dd0e50bb20 ]

The first time dma_chan_get() is called for a channel the channel
client_count is incorrectly incremented twice for public channels,
first in balance_ref_count(), and again prior to returning. This
results in an incorrect client count which will lead to the
channel resources not being freed when they should be. A simple
 test of repeated module load and unload of async_tx on a Dell
 Power Edge R7425 also shows this resulting in a kref underflow
 warning.

[  124.329662] async_tx: api initialized (async)
[  129.000627] async_tx: api initialized (async)
[  130.047839] ------------[ cut here ]------------
[  130.052472] refcount_t: underflow; use-after-free.
[  130.057279] WARNING: CPU: 3 PID: 19364 at lib/refcount.c:28
refcount_warn_saturate+0xba/0x110
[  130.065811] Modules linked in: async_tx(-) rfkill intel_rapl_msr
intel_rapl_common amd64_edac edac_mce_amd ipmi_ssif kvm_amd dcdbas kvm
mgag200 drm_shmem_helper acpi_ipmi irqbypass drm_kms_helper ipmi_si
syscopyarea sysfillrect rapl pcspkr ipmi_devintf sysimgblt fb_sys_fops
k10temp i2c_piix4 ipmi_msghandler acpi_power_meter acpi_cpufreq vfat
fat drm fuse xfs libcrc32c sd_mod t10_pi sg ahci crct10dif_pclmul
libahci crc32_pclmul crc32c_intel ghash_clmulni_intel igb megaraid_sas
i40e libata i2c_algo_bit ccp sp5100_tco dca dm_mirror dm_region_hash
dm_log dm_mod [last unloaded: async_tx]
[  130.117361] CPU: 3 PID: 19364 Comm: modprobe Kdump: loaded Not
tainted 5.14.0-185.el9.x86_64 #1
[  130.126091] Hardware name: Dell Inc. PowerEdge R7425/02MJ3T, BIOS
1.18.0 01/17/2022
[  130.133806] RIP: 0010:refcount_warn_saturate+0xba/0x110
[  130.139041] Code: 01 01 e8 6d bd 55 00 0f 0b e9 72 9d 8a 00 80 3d
26 18 9c 01 00 75 85 48 c7 c7 f8 a3 03 9d c6 05 16 18 9c 01 01 e8 4a
bd 55 00 <0f> 0b e9 4f 9d 8a 00 80 3d 01 18 9c 01 00 0f 85 5e ff ff ff
48 c7
[  130.157807] RSP: 0018:ffffbf98898afe68 EFLAGS: 00010286
[  130.163036] RAX: 0000000000000000 RBX: ffff9da06028e598 RCX: 0000000000000000
[  130.170172] RDX: ffff9daf9de26480 RSI: ffff9daf9de198a0 RDI: ffff9daf9de198a0
[  130.177316] RBP: ffff9da7cddf3970 R08: 0000000000000000 R09: 00000000ffff7fff
[  130.184459] R10: ffffbf98898afd00 R11: ffffffff9d9e8c28 R12: ffff9da7cddf1970
[  130.191596] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
[  130.198739] FS:  00007f646435c740(0000) GS:ffff9daf9de00000(0000)
knlGS:0000000000000000
[  130.206832] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  130.212586] CR2: 00007f6463b214f0 CR3: 00000008ab98c000 CR4: 00000000003506e0
[  130.219729] Call Trace:
[  130.222192]  <TASK>
[  130.224305]  dma_chan_put+0x10d/0x110
[  130.227988]  dmaengine_put+0x7a/0xa0
[  130.231575]  __do_sys_delete_module.constprop.0+0x178/0x280
[  130.237157]  ? syscall_trace_enter.constprop.0+0x145/0x1d0
[  130.242652]  do_syscall_64+0x5c/0x90
[  130.246240]  ? exc_page_fault+0x62/0x150
[  130.250178]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
[  130.255243] RIP: 0033:0x7f6463a3f5ab
[  130.258830] Code: 73 01 c3 48 8b 0d 75 a8 1b 00 f7 d8 64 89 01 48
83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa b8 b0 00 00
00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 45 a8 1b 00 f7 d8 64 89
01 48
[  130.277591] RSP: 002b:00007fff22f972c8 EFLAGS: 00000206 ORIG_RAX:
00000000000000b0
[  130.285164] RAX: ffffffffffffffda RBX: 000055b6786edd40 RCX: 00007f6463a3f5ab
[  130.292303] RDX: 0000000000000000 RSI: 0000000000000800 RDI: 000055b6786edda8
[  130.299443] RBP: 000055b6786edd40 R08: 0000000000000000 R09: 0000000000000000
[  130.306584] R10: 00007f6463b9eac0 R11: 0000000000000206 R12: 000055b6786edda8
[  130.313731] R13: 0000000000000000 R14: 000055b6786edda8 R15: 00007fff22f995f8
[  130.320875]  </TASK>
[  130.323081] ---[ end trace eff7156d56b5cf25 ]---

cat /sys/class/dma/dma0chan*/in_use would get the wrong result.
2
2
2

Fixes: d2f4f99db3e9 ("dmaengine: Rework dma_chan_get")
Signed-off-by: Koba Ko <koba.ko@canonical.com>
Reviewed-by: Jie Hai <haijie1@huawei.com>
Test-by: Jie Hai <haijie1@huawei.com>
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Tested-by: Joel Savitz <jsavitz@redhat.com>
Link: https://lore.kernel.org/r/20221201030050.978595-1-koba.ko@canonical.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/panfrost: fix GENERIC_ATOMIC64 dependency

[ Upstream commit 6437a549ae178a3f5a5c03e983f291ebcdc2bbc7 ]

On ARMv5 and earlier, a randconfig build can still run into

WARNING: unmet direct dependencies detected for IOMMU_IO_PGTABLE_LPAE
  Depends on [n]: IOMMU_SUPPORT [=y] && (ARM [=y] || ARM64 || COMPILE_TEST [=y]) && !GENERIC_ATOMIC64 [=y]
  Selected by [y]:
  - DRM_PANFROST [=y] && HAS_IOMEM [=y] && DRM [=y] && (ARM [=y] || ARM64 || COMPILE_TEST [=y] && !GENERIC_ATOMIC64 [=y]) && MMU [=y]

Rework the dependencies to always require a working cmpxchg64.

Fixes: db594ba3fcf9 ("drm/panfrost: depend on !GENERIC_ATOMIC64 when using COMPILE_TEST")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Steven Price <steven.price@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230117164456.1591901-1-arnd@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: mlx5: eliminate anonymous module_init & module_exit

[ Upstream commit 2c1e1b949024989e20907b84e11a731a50778416 ]

Eliminate anonymous module_init() and module_exit(), which can lead to
confusion or ambiguity when reading System.map, crashes/oops/bugs,
or an initcall_debug log.

Give each of these init and exit functions unique driver-specific
names to eliminate the anonymous names.

Example 1: (System.map)
 ffffffff832fc78c t init
 ffffffff832fc79e t init
 ffffffff832fc8f8 t init

Example 2: (initcall_debug log)
 calling  init+0x0/0x12 @ 1
 initcall init+0x0/0x12 returned 0 after 15 usecs
 calling  init+0x0/0x60 @ 1
 initcall init+0x0/0x60 returned 0 after 2 usecs
 calling  init+0x0/0x9a @ 1
 initcall init+0x0/0x9a returned 0 after 74 usecs

Fixes: e126ba97dba9 ("mlx5: Add driver for Mellanox Connect-IB adapters")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Eli Cohen <eli@mellanox.com>
Cc: Saeed Mahameed <saeedm@nvidia.com>
Cc: Leon Romanovsky <leon@kernel.org>
Cc: linux-rdma@vger.kernel.org
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>