USB: gadget: Fix the memory leak in raw_gadget driver

commit 83e30f2bf86ef7c38fbd476ed81a88522b620628 upstream.

Currently, increasing raw_dev->count happens before invoke the
raw_queue_event(), if the raw_queue_event() return error, invoke
raw_release() will not trigger the dev_free() to be called.

[  268.905865][ T5067] raw-gadget.0 gadget.0: failed to queue event
[  268.912053][ T5067] udc dummy_udc.0: failed to start USB Raw Gadget: -12
[  268.918885][ T5067] raw-gadget.0: probe of gadget.0 failed with error -12
[  268.925956][ T5067] UDC core: USB Raw Gadget: couldn't find an available UDC or it's busy
[  268.934657][ T5067] misc raw-gadget: fail, usb_gadget_register_driver returned -16

BUG: memory leak

[<ffffffff8154bf94>] kmalloc_trace+0x24/0x90 mm/slab_common.c:1076
[<ffffffff8347eb55>] kmalloc include/linux/slab.h:582 [inline]
[<ffffffff8347eb55>] kzalloc include/linux/slab.h:703 [inline]
[<ffffffff8347eb55>] dev_new drivers/usb/gadget/legacy/raw_gadget.c:191 [inline]
[<ffffffff8347eb55>] raw_open+0x45/0x110 drivers/usb/gadget/legacy/raw_gadget.c:385
[<ffffffff827d1d09>] misc_open+0x1a9/0x1f0 drivers/char/misc.c:165

[<ffffffff8154bf94>] kmalloc_trace+0x24/0x90 mm/slab_common.c:1076
[<ffffffff8347cd2f>] kmalloc include/linux/slab.h:582 [inline]
[<ffffffff8347cd2f>] raw_ioctl_init+0xdf/0x410 drivers/usb/gadget/legacy/raw_gadget.c:460
[<ffffffff8347dfe9>] raw_ioctl+0x5f9/0x1120 drivers/usb/gadget/legacy/raw_gadget.c:1250
[<ffffffff81685173>] vfs_ioctl fs/ioctl.c:51 [inline]

[<ffffffff8154bf94>] kmalloc_trace+0x24/0x90 mm/slab_common.c:1076
[<ffffffff833ecc6a>] kmalloc include/linux/slab.h:582 [inline]
[<ffffffff833ecc6a>] kzalloc include/linux/slab.h:703 [inline]
[<ffffffff833ecc6a>] dummy_alloc_request+0x5a/0xe0 drivers/usb/gadget/udc/dummy_hcd.c:665
[<ffffffff833e9132>] usb_ep_alloc_request+0x22/0xd0 drivers/usb/gadget/udc/core.c:196
[<ffffffff8347f13d>] gadget_bind+0x6d/0x370 drivers/usb/gadget/legacy/raw_gadget.c:292

This commit therefore invoke kref_get() under the condition that
raw_queue_event() return success.

Reported-by: syzbot+feb045d335c1fdde5bf7@syzkaller.appspotmail.com
Cc: stable <stable@kernel.org>
Closes: https://syzkaller.appspot.com/bug?extid=feb045d335c1fdde5bf7
Signed-off-by: Zqiang <qiang.zhang1211@gmail.com>
Reviewed-by: Andrey Konovalov <andreyknvl@gmail.com>
Tested-by: Andrey Konovalov <andreyknvl@gmail.com>
Link: https://lore.kernel.org/r/20230714074011.20989-1-qiang.zhang1211@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: gadget: call usb_gadget_check_config() to verify UDC capability

commit f4fc01af5b640bc39bd9403b5fd855345a2ad5f8 upstream.

The legacy gadget driver omitted calling usb_gadget_check_config()
to ensure that the USB device controller (UDC) has adequate resources,
including sufficient endpoint numbers and types, to support the given
configuration.

Previously, usb_add_config() was solely invoked by the legacy gadget
driver. Adds the necessary usb_gadget_check_config() after the bind()
operation to fix the issue.

Fixes: dce49449e04f ("usb: cdns3: allocate TX FIFO size according to composite EP number")
Cc: stable <stable@kernel.org>
Reported-by: Ravi Gunasekaran <r-gunasekaran@ti.com>
Signed-off-by: Frank Li <Frank.Li@nxp.com>
Link: https://lore.kernel.org/r/20230707230015.494999-1-Frank.Li@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Revert "usb: gadget: tegra-xudc: Fix error check in tegra_xudc_powerdomain_init()"

commit a8291be6b5dd465c22af229483dbac543a91e24e upstream.

This reverts commit f08aa7c80dac27ee00fa6827f447597d2fba5465.

The reverted commit was based on static analysis and a misunderstanding
of how PTR_ERR() and NULLs are supposed to work.  When a function
returns both pointer errors and NULL then normally the NULL means
"continue operating without a feature because it was deliberately
turned off".  The NULL should not be treated as a failure.  If a driver
cannot work when that feature is disabled then the KConfig should
enforce that the function cannot return NULL.  We should not need to
test for it.

In this driver, the bug means that probe cannot succeed when CONFIG_PM
is disabled.

Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Fixes: f08aa7c80dac ("usb: gadget: tegra-xudc: Fix error check in tegra_xudc_powerdomain_init()")
Cc: stable <stable@kernel.org>
Link: https://lore.kernel.org/r/ZKQoBa84U/ykEh3C@moroto
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

tracing: Fix warning in trace_buffered_event_disable()

[ Upstream commit dea499781a1150d285c62b26659f62fb00824fce ]

Warning happened in trace_buffered_event_disable() at
  WARN_ON_ONCE(!trace_buffered_event_ref)

  Call Trace:
   ? __warn+0xa5/0x1b0
   ? trace_buffered_event_disable+0x189/0x1b0
   __ftrace_event_enable_disable+0x19e/0x3e0
   free_probe_data+0x3b/0xa0
   unregister_ftrace_function_probe_func+0x6b8/0x800
   event_enable_func+0x2f0/0x3d0
   ftrace_process_regex.isra.0+0x12d/0x1b0
   ftrace_filter_write+0xe6/0x140
   vfs_write+0x1c9/0x6f0
   [...]

The cause of the warning is in __ftrace_event_enable_disable(),
trace_buffered_event_enable() was called once while
trace_buffered_event_disable() was called twice.
Reproduction script show as below, for analysis, see the comments:
 ```
 #!/bin/bash

 cd /sys/kernel/tracing/

 # 1. Register a 'disable_event' command, then:
 #    1) SOFT_DISABLED_BIT was set;
 #    2) trace_buffered_event_enable() was called first time;
 echo 'cmdline_proc_show:disable_event:initcall:initcall_finish' > \
     set_ftrace_filter

 # 2. Enable the event registered, then:
 #    1) SOFT_DISABLED_BIT was cleared;
 #    2) trace_buffered_event_disable() was called first time;
 echo 1 > events/initcall/initcall_finish/enable

 # 3. Try to call into cmdline_proc_show(), then SOFT_DISABLED_BIT was
 #    set again!!!
 cat /proc/cmdline

 # 4. Unregister the 'disable_event' command, then:
 #    1) SOFT_DISABLED_BIT was cleared again;
 #    2) trace_buffered_event_disable() was called second time!!!
 echo '!cmdline_proc_show:disable_event:initcall:initcall_finish' > \
     set_ftrace_filter
 ```

To fix it, IIUC, we can change to call trace_buffered_event_enable() at
fist time soft-mode enabled, and call trace_buffered_event_disable() at
last time soft-mode disabled.

Link: https://lore.kernel.org/linux-trace-kernel/20230726095804.920457-1-zhengyejian1@huawei.com
Cc: <mhiramat@kernel.org>
Fixes: 0fc1b09ff1ff ("tracing: Use temp buffer when filtering events")
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ring-buffer: Fix wrong stat of cpu_buffer->read

[ Upstream commit 2d093282b0d4357373497f65db6a05eb0c28b7c8 ]

When pages are removed in rb_remove_pages(), 'cpu_buffer->read' is set
to 0 in order to make sure any read iterators reset themselves. However,
this will mess 'entries' stating, see following steps:

  # cd /sys/kernel/tracing/
  # 1. Enlarge ring buffer prepare for later reducing:
  # echo 20 > per_cpu/cpu0/buffer_size_kb
  # 2. Write a log into ring buffer of cpu0:
  # taskset -c 0 echo "hello1" > trace_marker
  # 3. Read the log:
  # cat per_cpu/cpu0/trace_pipe
       <...>-332     [000] .....    62.406844: tracing_mark_write: hello1
  # 4. Stop reading and see the stats, now 0 entries, and 1 event readed:
  # cat per_cpu/cpu0/stats
   entries: 0
   [...]
   read events: 1
  # 5. Reduce the ring buffer
  # echo 7 > per_cpu/cpu0/buffer_size_kb
  # 6. Now entries became unexpected 1 because actually no entries!!!
  # cat per_cpu/cpu0/stats
   entries: 1
   [...]
   read events: 0

To fix it, introduce 'page_removed' field to count total removed pages
since last reset, then use it to let read iterators reset themselves
instead of changing the 'read' pointer.

Link: https://lore.kernel.org/linux-trace-kernel/20230724054040.3489499-1-zhengyejian1@huawei.com
Cc: <mhiramat@kernel.org>
Cc: <vnagarnaik@google.com>
Fixes: 83f40318dab0 ("ring-buffer: Make removal of ring buffer pages atomic")
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ata: pata_ns87415: mark ns87560_tf_read static

[ Upstream commit 3fc2febb0f8ffae354820c1772ec008733237cfa ]

The global function triggers a warning because of the missing prototype

drivers/ata/pata_ns87415.c:263:6: warning: no previous prototype for 'ns87560_tf_read' [-Wmissing-prototypes]
  263 | void ns87560_tf_read(struct ata_port *ap, struct ata_taskfile *tf)

There are no other references to this, so just make it static.

Fixes: c4b5b7b6c4423 ("pata_ns87415: Initial cut at 87415/87560 IDE support")
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Reviewed-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ublk: fail to recover device if queue setup is interrupted

[ Upstream commit 0c0cbd4ebc375ceebc75c89df04b74f215fab23a ]

In ublk_ctrl_end_recovery(), if wait_for_completion_interruptible() is
interrupted by signal, queues aren't setup successfully yet, so we
have to fail UBLK_CMD_END_USER_RECOVERY, otherwise kernel oops can be
triggered.

Fixes: c732a852b419 ("ublk_drv: add START_USER_RECOVERY and END_USER_RECOVERY support")
Reported-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20230726144502.566785-3-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ublk: fail to start device if queue setup is interrupted

[ Upstream commit 53e7d08f6d6e214c40db1f51291bb2975c789dc2 ]

In ublk_ctrl_start_dev(), if wait_for_completion_interruptible() is
interrupted by signal, queues aren't setup successfully yet, so we
have to fail UBLK_CMD_START_DEV, otherwise kernel oops can be triggered.

Reported by German when working on qemu-storage-deamon which requires
single thread ublk daemon.

Fixes: 71f28f3136af ("ublk_drv: add io_uring based userspace block driver")
Reported-by: German Maglione <gmaglione@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20230726144502.566785-2-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ublk_drv: move ublk_get_device_from_id into ublk_ctrl_uring_cmd

[ Upstream commit bfbcef036396a73fbf4b3fee385cc670159df5ad ]

It is annoying for each control command handler to get/put ublk
device and deal with failure.

Control command handler is simplified a lot by moving
ublk_get_device_from_id into ublk_ctrl_uring_cmd().

Reviewed-by: ZiyangZhang <ZiyangZhang@linux.alibaba.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20230106041711.914434-4-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stable-dep-of: 53e7d08f6d6e ("ublk: fail to start device if queue setup is interrupted")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm: Disallow submit with fence id 0

[ Upstream commit 1b5d0ddcb34a605835051ae2950d5cfed0373dd8 ]

A fence id of zero is expected to be invalid, and is not removed from
the fence_idr table.  If userspace is requesting to specify the fence
id with the FENCE_SN_IN flag, we need to reject a zero fence id value.

Fixes: 17154addc5c1 ("drm/msm: Add MSM_SUBMIT_FENCE_SN_IN")
Signed-off-by: Rob Clark <robdclark@chromium.org>
Patchwork: https://patchwork.freedesktop.org/patch/549180/
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm: Switch idr_lock to spinlock

[ Upstream commit e4f020c6a05db73eac49b7c3b3650251be374200 ]

Needed to idr_preload() which returns with preemption disabled.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Patchwork: https://patchwork.freedesktop.org/patch/527846/
Link: https://lore.kernel.org/r/20230320144356.803762-11-robdclark@gmail.com
Stable-dep-of: 1b5d0ddcb34a ("drm/msm: Disallow submit with fence id 0")
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/irdma: Report correct WC error

[ Upstream commit ae463563b7a1b7d4a3d0b065b09d37a76b693937 ]

Report the correct WC error if a MW bind is performed
on an already valid/bound window.

Fixes: 44d9e52977a1 ("RDMA/irdma: Implement device initialization definitions")
Signed-off-by: Sindhu Devale <sindhu.devale@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Link: https://lore.kernel.org/r/20230725155439.1057-2-shiraz.saleem@intel.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/irdma: Fix op_type reporting in CQEs

[ Upstream commit 3bfb25fa2b5bb9c29681e6ac861808f4be1331a9 ]

The op_type field CQ poll info structure is incorrectly
filled in with the queue type as opposed to the op_type
received in the CQEs. The wrong opcode could be decoded
and returned to the ULP.

Copy the op_type field received in the CQE in the CQ poll
info structure.

Fixes: 24419777e943 ("RDMA/irdma: Fix RQ completion opcode")
Signed-off-by: Sindhu Devale <sindhu.devale@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Link: https://lore.kernel.org/r/20230725155439.1057-1-shiraz.saleem@intel.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Unlock on error path in dm_handle_mst_sideband_msg_ready_event()

[ Upstream commit 38ac4e8385ffb275b1837986ca6c16f26ea028c5 ]

This error path needs to unlock the "aconnector->handle_mst_msg_ready"
mutex before returning.

Fixes: 4f6d9e38c4d2 ("drm/amd/display: Add polling method to handle MST reply packet")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd: Fix an error handling mistake in psp_sw_init()

[ Upstream commit c01aebeef3ce45f696ffa0a1303cea9b34babb45 ]

If the second call to amdgpu_bo_create_kernel() fails, the memory
allocated from the first call should be cleared.  If the third call
fails, the memory from the second call should be cleared.

Fixes: b95b5391684b ("drm/amdgpu/psp: move PSP memory alloc from hw_init to sw_init")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dm raid: protect md_stop() with 'reconfig_mutex'

[ Upstream commit 7d5fff8982a2199d49ec067818af7d84d4f95ca0 ]

__md_stop_writes() and __md_stop() will modify many fields that are
protected by 'reconfig_mutex', and all the callers will grab
'reconfig_mutex' except for md_stop().

Also, update md_stop() to make certain 'reconfig_mutex' is held using
lockdep_assert_held().

Fixes: 9d09e663d550 ("dm: raid456 basic support")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dm raid: clean up four equivalent goto tags in raid_ctr()

[ Upstream commit e74c874eabe2e9173a8fbdad616cd89c70eb8ffd ]

There are four equivalent goto tags in raid_ctr(), clean them up to
use just one.

There is no functional change and this is preparation to fix
raid_ctr()'s unprotected md_stop().

Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
Stable-dep-of: 7d5fff8982a2 ("dm raid: protect md_stop() with 'reconfig_mutex'")
Signed-off-by: Sasha Levin <sashal@kernel.org>

dm raid: fix missing reconfig_mutex unlock in raid_ctr() error paths

[ Upstream commit bae3028799dc4f1109acc4df37c8ff06f2d8f1a0 ]

In the error paths 'bad_stripe_cache' and 'bad_check_reshape',
'reconfig_mutex' is still held after raid_ctr() returns.

Fixes: 9dbd1aa3a81c ("dm raid: add reshaping support to the target")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

xenbus: check xen_domain in xenbus_probe_initcall

[ Upstream commit 0d8f7cc8057890db08c54fe610d8a94af59da082 ]

The same way we already do in xenbus_init.
Fixes the following warning:

[  352.175563] Trying to free already-free IRQ 0
[  352.177355] WARNING: CPU: 1 PID: 88 at kernel/irq/manage.c:1893 free_irq+0xbf/0x350
[...]
[  352.213951] Call Trace:
[  352.214390]  <TASK>
[  352.214717]  ? __warn+0x81/0x170
[  352.215436]  ? free_irq+0xbf/0x350
[  352.215906]  ? report_bug+0x10b/0x200
[  352.216408]  ? prb_read_valid+0x17/0x20
[  352.216926]  ? handle_bug+0x44/0x80
[  352.217409]  ? exc_invalid_op+0x13/0x60
[  352.217932]  ? asm_exc_invalid_op+0x16/0x20
[  352.218497]  ? free_irq+0xbf/0x350
[  352.218979]  ? __pfx_xenbus_probe_thread+0x10/0x10
[  352.219600]  xenbus_probe+0x7a/0x80
[  352.221030]  xenbus_probe_thread+0x76/0xc0

Fixes: 5b3353949e89 ("xen: add support for initializing xenstore later as HVM domain")
Signed-off-by: Stefano Stabellini <stefano.stabellini@amd.com>
Tested-by: Petr Mladek <pmladek@suse.com>
Reviewed-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
Link: https://lore.kernel.org/r/alpine.DEB.2.22.394.2307211609140.3118466@ubuntu-linux-20-04-desktop
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/i915: Fix an error handling path in igt_write_huge()

[ Upstream commit e354f67733115b4453268f61e6e072e9b1ea7a2f ]

All error handling paths go to 'out', except this one. Be consistent and
also branch to 'out' here.

Fixes: c10a652e239e ("drm/i915/selftests: Rework context handling in hugepages selftests")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/7a036b88671312ee9adc01c74ef5b3376f690b76.1689619758.git.christophe.jaillet@wanadoo.fr
(cherry picked from commit 361ecaadb1ce3c5312c7c4c419271326d43899eb)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

smb3: do not set NTLMSSP_VERSION flag for negotiate not auth request

[ Upstream commit 19826558210b9102a7d4681c91784d137d60d71b ]

The NTLMSSP_NEGOTIATE_VERSION flag only needs to be sent during
the NTLMSSP NEGOTIATE (not the AUTH) request, so filter it out for
NTLMSSP AUTH requests. See MS-NLMP 2.2.1.3

This fixes a problem found by the gssntlmssp server.

Link: https://github.com/gssapi/gss-ntlmssp/issues/95
Fixes: 52d005337b2c ("smb3: send NTLMSSP version information")
Acked-by: Roy Shterman <roy.shterman@gmail.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

block: Fix a source code comment in include/uapi/linux/blkzoned.h

[ Upstream commit e0933b526fbfd937c4a8f4e35fcdd49f0e22d411 ]

Fix the symbolic names for zone conditions in the blkzoned.h header
file.

Cc: Hannes Reinecke <hare@suse.de>
Cc: Damien Le Moal <dlemoal@kernel.org>
Fixes: 6a0cb1bc106f ("block: Implement support for zoned block devices")
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Link: https://lore.kernel.org/r/20230706201422.3987341-1-bvanassche@acm.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: fsl_spdif: Silence output on stop

[ Upstream commit 0e4c2b6b0c4a4b4014d9424c27e5e79d185229c5 ]

Clear TX registers on stop to prevent the SPDIF interface from sending
last written word over and over again.

Fixes: a2388a498ad2 ("ASoC: fsl: Add S/PDIF CPU DAI driver")
Signed-off-by: Matus Gajdos <matuszpd@gmail.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Link: https://lore.kernel.org/r/20230719164729.19969-1-matuszpd@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cxl/acpi: Return 'rc' instead of '0' in cxl_parse_cfmws()

[ Upstream commit 91019b5bc7c2c5e6f676cce80ee6d12b2753d018 ]

Driver initialization returned success (return 0) even if the
initialization (cxl_decoder_add() or acpi_table_parse_cedt()) failed.

Return the error instead of swallowing it.

Fixes: f4ce1f766f1e ("cxl/acpi: Convert CFMWS parsing to ACPI sub-table helpers")
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://lore.kernel.org/r/20230714093146.2253438-2-leitao@debian.org
Reviewed-by: Alison Schofield <alison.schofield@intel.com>
Signed-off-by: Vishal Verma <vishal.l.verma@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cxl/acpi: Fix a use-after-free in cxl_parse_cfmws()

[ Upstream commit 4cf67d3cc9994a59cf77bb9c0ccf9007fe916afe ]

KASAN and KFENCE detected an user-after-free in the CXL driver. This
happens in the cxl_decoder_add() fail path. KASAN prints the following
error:

   BUG: KASAN: slab-use-after-free in cxl_parse_cfmws (drivers/cxl/acpi.c:299)

This happens in cxl_parse_cfmws(), where put_device() is called,
releasing cxld, which is accessed later.

Use the local variables in the dev_err() instead of pointing to the
released memory. Since the dev_err() is printing a resource, change the open
coded print format to use the %pr format specifier.

Fixes: e50fe01e1f2a ("cxl/core: Drop ->platform_res attribute for root decoders")
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://lore.kernel.org/r/20230714093146.2253438-1-leitao@debian.org
Reviewed-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Vishal Verma <vishal.l.verma@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm: Fix IS_ERR_OR_NULL() vs NULL check in a5xx_submit_in_rb()

[ Upstream commit 6e8a996563ecbe68e49c49abd4aaeef69f11f2dc ]

The msm_gem_get_vaddr() returns an ERR_PTR() on failure, and a null
is catastrophic here, so we should use IS_ERR_OR_NULL() to check
the return value.

Fixes: 6a8bd08d0465 ("drm/msm: add sudo flag to submit ioctl")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/547712/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/bnxt_re: Prevent handling any completions after qp destroy

[ Upstream commit b5bbc6551297447d3cca55cf907079e206e9cd82 ]

HW may generate completions that indicates QP is destroyed.
Driver should not be scheduling any more completion handlers
for this QP, after the QP is destroyed. Since CQs are active
during the QP destroy, driver may still schedule completion
handlers. This can cause a race where the destroy_cq and poll_cq
running simultaneously.

Snippet of kernel panic while doing bnxt_re driver load unload in loop.
This indicates a poll after the CQ is freed. 

[77786.481636] Call Trace:
[77786.481640]  <TASK>
[77786.481644]  bnxt_re_poll_cq+0x14a/0x620 [bnxt_re]
[77786.481658]  ? kvm_clock_read+0x14/0x30
[77786.481693]  __ib_process_cq+0x57/0x190 [ib_core]
[77786.481728]  ib_cq_poll_work+0x26/0x80 [ib_core]
[77786.481761]  process_one_work+0x1e5/0x3f0
[77786.481768]  worker_thread+0x50/0x3a0
[77786.481785]  ? __pfx_worker_thread+0x10/0x10
[77786.481790]  kthread+0xe2/0x110
[77786.481794]  ? __pfx_kthread+0x10/0x10
[77786.481797]  ret_from_fork+0x2c/0x50

To avoid this, complete all completion handlers before returning the
destroy QP. If free_cq is called soon after destroy_qp,  IB stack
will cancel the CQ work before invoking the destroy_cq verb and
this will prevent any race mentioned.

Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver")
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Link: https://lore.kernel.org/r/1689322969-25402-2-git-send-email-selvin.xavier@broadcom.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/mthca: Fix crash when polling CQ for shared QPs

[ Upstream commit dc52aadbc1849cbe3fcf6bc54d35f6baa396e0a1 ]

Commit 21c2fe94abb2 ("RDMA/mthca: Combine special QP struct with mthca QP")
introduced a new struct mthca_sqp which doesn't contain struct mthca_qp
any longer. Placing a pointer of this new struct into qptable leads
to crashes, because mthca_poll_one() expects a qp pointer. Fix this
by putting the correct pointer into qptable.

Fixes: 21c2fe94abb2 ("RDMA/mthca: Combine special QP struct with mthca QP")
Signed-off-by: Thomas Bogendoerfer <tbogendoerfer@suse.de>
Link: https://lore.kernel.org/r/20230713141658.9426-1-tbogendoerfer@suse.de
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/irdma: Fix data race on CQP request done

[ Upstream commit f0842bb3d38863777e3454da5653d80b5fde6321 ]

KCSAN detects a data race on cqp_request->request_done memory location
which is accessed locklessly in irdma_handle_cqp_op while being
updated in irdma_cqp_ce_handler.

Annotate lockless intent with READ_ONCE/WRITE_ONCE to avoid any
compiler optimizations like load fusing and/or KCSAN warning.

[222808.417128] BUG: KCSAN: data-race in irdma_cqp_ce_handler [irdma] / irdma_wait_event [irdma]

[222808.417532] write to 0xffff8e44107019dc of 1 bytes by task 29658 on cpu 5:
[222808.417610]  irdma_cqp_ce_handler+0x21e/0x270 [irdma]
[222808.417725]  cqp_compl_worker+0x1b/0x20 [irdma]
[222808.417827]  process_one_work+0x4d1/0xa40
[222808.417835]  worker_thread+0x319/0x700
[222808.417842]  kthread+0x180/0x1b0
[222808.417852]  ret_from_fork+0x22/0x30

[222808.417918] read to 0xffff8e44107019dc of 1 bytes by task 29688 on cpu 1:
[222808.417995]  irdma_wait_event+0x1e2/0x2c0 [irdma]
[222808.418099]  irdma_handle_cqp_op+0xae/0x170 [irdma]
[222808.418202]  irdma_cqp_cq_destroy_cmd+0x70/0x90 [irdma]
[222808.418308]  irdma_puda_dele_rsrc+0x46d/0x4d0 [irdma]
[222808.418411]  irdma_rt_deinit_hw+0x179/0x1d0 [irdma]
[222808.418514]  irdma_ib_dealloc_device+0x11/0x40 [irdma]
[222808.418618]  ib_dealloc_device+0x2a/0x120 [ib_core]
[222808.418823]  __ib_unregister_device+0xde/0x100 [ib_core]
[222808.418981]  ib_unregister_device+0x22/0x40 [ib_core]
[222808.419142]  irdma_ib_unregister_device+0x70/0x90 [irdma]
[222808.419248]  i40iw_close+0x6f/0xc0 [irdma]
[222808.419352]  i40e_client_device_unregister+0x14a/0x180 [i40e]
[222808.419450]  i40iw_remove+0x21/0x30 [irdma]
[222808.419554]  auxiliary_bus_remove+0x31/0x50
[222808.419563]  device_remove+0x69/0xb0
[222808.419572]  device_release_driver_internal+0x293/0x360
[222808.419582]  driver_detach+0x7c/0xf0
[222808.419592]  bus_remove_driver+0x8c/0x150
[222808.419600]  driver_unregister+0x45/0x70
[222808.419610]  auxiliary_driver_unregister+0x16/0x30
[222808.419618]  irdma_exit_module+0x18/0x1e [irdma]
[222808.419733]  __do_sys_delete_module.constprop.0+0x1e2/0x310
[222808.419745]  __x64_sys_delete_module+0x1b/0x30
[222808.419755]  do_syscall_64+0x39/0x90
[222808.419763]  entry_SYSCALL_64_after_hwframe+0x63/0xcd

[222808.419829] value changed: 0x01 -> 0x03

Fixes: 915cc7ac0f8e ("RDMA/irdma: Add miscellaneous utility definitions")
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Link: https://lore.kernel.org/r/20230711175253.1289-4-shiraz.saleem@intel.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/irdma: Fix data race on CQP completion stats

[ Upstream commit f2c3037811381f9149243828c7eb9a1631df9f9c ]

CQP completion statistics is read lockesly in irdma_wait_event and
irdma_check_cqp_progress while it can be updated in the completion
thread irdma_sc_ccq_get_cqe_info on another CPU as KCSAN reports.

Make completion statistics an atomic variable to reflect coherent updates
to it. This will also avoid load/store tearing logic bug potentially
possible by compiler optimizations.

[77346.170861] BUG: KCSAN: data-race in irdma_handle_cqp_op [irdma] / irdma_sc_ccq_get_cqe_info [irdma]

[77346.171383] write to 0xffff8a3250b108e0 of 8 bytes by task 9544 on cpu 4:
[77346.171483]  irdma_sc_ccq_get_cqe_info+0x27a/0x370 [irdma]
[77346.171658]  irdma_cqp_ce_handler+0x164/0x270 [irdma]
[77346.171835]  cqp_compl_worker+0x1b/0x20 [irdma]
[77346.172009]  process_one_work+0x4d1/0xa40
[77346.172024]  worker_thread+0x319/0x700
[77346.172037]  kthread+0x180/0x1b0
[77346.172054]  ret_from_fork+0x22/0x30

[77346.172136] read to 0xffff8a3250b108e0 of 8 bytes by task 9838 on cpu 2:
[77346.172234]  irdma_handle_cqp_op+0xf4/0x4b0 [irdma]
[77346.172413]  irdma_cqp_aeq_cmd+0x75/0xa0 [irdma]
[77346.172592]  irdma_create_aeq+0x390/0x45a [irdma]
[77346.172769]  irdma_rt_init_hw.cold+0x212/0x85d [irdma]
[77346.172944]  irdma_probe+0x54f/0x620 [irdma]
[77346.173122]  auxiliary_bus_probe+0x66/0xa0
[77346.173137]  really_probe+0x140/0x540
[77346.173154]  __driver_probe_device+0xc7/0x220
[77346.173173]  driver_probe_device+0x5f/0x140
[77346.173190]  __driver_attach+0xf0/0x2c0
[77346.173208]  bus_for_each_dev+0xa8/0xf0
[77346.173225]  driver_attach+0x29/0x30
[77346.173240]  bus_add_driver+0x29c/0x2f0
[77346.173255]  driver_register+0x10f/0x1a0
[77346.173272]  __auxiliary_driver_register+0xbc/0x140
[77346.173287]  irdma_init_module+0x55/0x1000 [irdma]
[77346.173460]  do_one_initcall+0x7d/0x410
[77346.173475]  do_init_module+0x81/0x2c0
[77346.173491]  load_module+0x1232/0x12c0
[77346.173506]  __do_sys_finit_module+0x101/0x180
[77346.173522]  __x64_sys_finit_module+0x3c/0x50
[77346.173538]  do_syscall_64+0x39/0x90
[77346.173553]  entry_SYSCALL_64_after_hwframe+0x63/0xcd

[77346.173634] value changed: 0x0000000000000094 -> 0x0000000000000095

Fixes: 915cc7ac0f8e ("RDMA/irdma: Add miscellaneous utility definitions")
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Link: https://lore.kernel.org/r/20230711175253.1289-3-shiraz.saleem@intel.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/irdma: Add missing read barriers

[ Upstream commit 4984eb51453ff7eddee9e5ce816145be39c0ec5c ]

On code inspection, there are many instances in the driver where
CEQE and AEQE fields written to by HW are read without guaranteeing
that the polarity bit has been read and checked first.

Add a read barrier to avoid reordering of loads on the CEQE/AEQE fields
prior to checking the polarity bit.

Fixes: 3f49d6842569 ("RDMA/irdma: Implement HW Admin Queue OPs")
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Link: https://lore.kernel.org/r/20230711175253.1289-2-shiraz.saleem@intel.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm/adreno: Fix snapshot BINDLESS_DATA size

[ Upstream commit bd846ceee9c478d0397428f02696602ba5eb264a ]

The incorrect size was causing "CP | AHB bus error" when snapshotting
the GPU state on a6xx gen4 (a660 family).

Closes: https://gitlab.freedesktop.org/drm/msm/-/issues/26
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Fixes: 1707add81551 ("drm/msm/a6xx: Add a6xx gpu state")
Patchwork: https://patchwork.freedesktop.org/patch/546763/
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm/dpu: drop enum dpu_core_perf_data_bus_id

[ Upstream commit e8383f5cf1b3573ce140a80bfbfd809278ab16d6 ]

Drop the leftover of bus-client -> interconnect conversion, the enum
dpu_core_perf_data_bus_id.

Fixes: cb88482e2570 ("drm/msm/dpu: clean up references of DPU custom bus scaling")
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/546048/
Link: https://lore.kernel.org/r/20230707193942.3806526-2-dmitry.baryshkov@linaro.org
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/mlx4: Make check for invalid flags stricter

[ Upstream commit d64b1ee12a168030fbb3e0aebf7bce49e9a07589 ]

This code is trying to ensure that only the flags specified in the list
are allowed.  The problem is that ucmd->rx_hash_fields_mask is a u64 and
the flags are an enum which is treated as a u32 in this context.  That
means the test doesn't check whether the highest 32 bits are zero.

Fixes: 4d02ebd9bbbd ("IB/mlx4: Fix RSS hash fields restrictions")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: https://lore.kernel.org/r/233ed975-982d-422a-b498-410f71d8a101@moroto.mountain
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tipc: stop tipc crypto on failure in tipc_node_create

[ Upstream commit de52e17326c3e9a719c9ead4adb03467b8fae0ef ]

If tipc_link_bc_create() fails inside tipc_node_create() for a newly
allocated tipc node then we should stop its tipc crypto and free the
resources allocated with a call to tipc_crypto_start().

As the node ref is initialized to one to that point, just put the ref on
tipc_link_bc_create() error case that would lead to tipc_node_free() be
eventually executed and properly clean the node and its crypto resources.

Found by Linux Verification Center (linuxtesting.org).

Fixes: cb8092d70a6f ("tipc: move bc link creation back to tipc_node_create")
Suggested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Reviewed-by: Xin Long <lucien.xin@gmail.com>
Link: https://lore.kernel.org/r/20230725214628.25246-1-pchelkin@ispras.ru
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tipc: check return value of pskb_trim()

[ Upstream commit e46e06ffc6d667a89b979701288e2264f45e6a7b ]

goto free_skb if an unexpected result is returned by pskb_tirm()
in tipc_crypto_rcv_complete().

Fixes: fc1b6d6de220 ("tipc: introduce TIPC encryption & authentication")
Signed-off-by: Yuanjun Gong <ruc_gongyuanjun@163.com>
Reviewed-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Link: https://lore.kernel.org/r/20230725064810.5820-1-ruc_gongyuanjun@163.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

benet: fix return value check in be_lancer_xmit_workarounds()

[ Upstream commit 5c85f7065718a949902b238a6abd8fc907c5d3e0 ]

in be_lancer_xmit_workarounds(), it should go to label 'tx_drop'
if an unexpected value is returned by pskb_trim().

Fixes: 93040ae5cc8d ("be2net: Fix to trim skb for padded vlan packets to workaround an ASIC Bug")
Signed-off-by: Yuanjun Gong <ruc_gongyuanjun@163.com>
Link: https://lore.kernel.org/r/20230725032726.15002-1-ruc_gongyuanjun@163.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/sched: mqprio: Add length check for TCA_MQPRIO_{MAX/MIN}_RATE64

[ Upstream commit 6c58c8816abb7b93b21fa3b1d0c1726402e5e568 ]

The nla_for_each_nested parsing in function mqprio_parse_nlattr() does
not check the length of the nested attribute. This can lead to an
out-of-attribute read and allow a malformed nlattr (e.g., length 0) to
be viewed as 8 byte integer and passed to priv->max_rate/min_rate.

This patch adds the check based on nla_len() when check the nla_type(),
which ensures that the length of these two attribute must equals
sizeof(u64).

Fixes: 4e8b86c06269 ("mqprio: Introduce new hardware offload mode and shaper in mqprio")
Reviewed-by: Victor Nogueira <victor@mojatatu.com>
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Link: https://lore.kernel.org/r/20230725024227.426561-1-linma@zju.edu.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/sched: mqprio: add extack to mqprio_parse_nlattr()

[ Upstream commit 57f21bf85400abadac0cb2a4db5de1d663f8863f ]

Netlink attribute parsing in mqprio is a minesweeper game, with many
options having the possibility of being passed incorrectly and the user
being none the wiser.

Try to make errors less sour by giving user space some information
regarding what went wrong.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ferenc Fejes <fejes@inf.elte.hu>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 6c58c8816abb ("net/sched: mqprio: Add length check for TCA_MQPRIO_{MAX/MIN}_RATE64")
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/sched: mqprio: refactor nlattr parsing to a separate function

[ Upstream commit feb2cf3dcfb930aec2ca65c66d1365543d5ba943 ]

mqprio_init() is quite large and unwieldy to add more code to.
Split the netlink attribute parsing to a dedicated function.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: 6c58c8816abb ("net/sched: mqprio: Add length check for TCA_MQPRIO_{MAX/MIN}_RATE64")
Signed-off-by: Sasha Levin <sashal@kernel.org>

mm: suppress mm fault logging if fatal signal already pending

[ Upstream commit 5f0bc0b042fc77ff70e14c790abdec960cde4ec1 ]

Commit eda0047296a1 ("mm: make the page fault mmap locking killable")
intentionally made it much easier to trigger the "page fault fails
because a fatal signal is pending" situation, by having the mmap locking
fail early in that case.

We have long aborted page faults in other fatal cases when the actual IO
for a page is interrupted by SIGKILL - which is particularly useful for
the traditional case of NFS hanging due to network issues, but local
filesystems could cause it too if you happened to get the SIGKILL while
waiting for a page to be faulted in (eg lock_folio_maybe_drop_mmap()).

So aborting the page fault wasn't a new condition - but it now triggers
earlier, before we even get to 'handle_mm_fault()'.  And as a result the
error doesn't go through our 'fault_signal_pending()' logic, and doesn't
get filtered away there.

Normally you'd never even notice, because if a fatal signal is pending,
the new SIGSEGV we send ends up being ignored anyway.

But it turns out that there is one very noticeable exception: if you
enable 'show_unhandled_signals', the aborted page fault will be logged
in the kernel messages, and you'll get a scary line looking something
like this in your logs:

  pverados[2183248]: segfault at 55e5a00f9ae0 ip 000055e5a00f9ae0 sp 00007ffc0720bea8 error 14 in perl[55e5a00d4000+195000] likely on CPU 10 (core 4, socket 0)

which is rather misleading.  It's not really a segfault at all, it's
just "the thread was killed before the page fault completed, so we
aborted the page fault".

Fix this by just making it clear that a pending fatal signal means that
any new signal coming in after that is implicitly handled.  This will
avoid the misleading logging, since now the signal isn't 'unhandled' any
more.

Reported-and-tested-by: Fiona Ebner <f.ebner@proxmox.com>
Tested-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Link: https://lore.kernel.org/lkml/8d063a26-43f5-0bb7-3203-c6a04dc159f8@proxmox.com/
Acked-by: Oleg Nesterov <oleg@redhat.com>
Fixes: eda0047296a1 ("mm: make the page fault mmap locking killable")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: nf_tables: disallow rule addition to bound chain via NFTA_RULE_CHAIN_ID

[ Upstream commit 0ebc1064e4874d5987722a2ddbc18f94aa53b211 ]

Bail out with EOPNOTSUPP when adding rule to bound chain via
NFTA_RULE_CHAIN_ID. The following warning splat is shown when
adding a rule to a deleted bound chain:

 WARNING: CPU: 2 PID: 13692 at net/netfilter/nf_tables_api.c:2013 nf_tables_chain_destroy+0x1f7/0x210 [nf_tables]
 CPU: 2 PID: 13692 Comm: chain-bound-rul Not tainted 6.1.39 #1
 RIP: 0010:nf_tables_chain_destroy+0x1f7/0x210 [nf_tables]

Fixes: d0e2c7de92c7 ("netfilter: nf_tables: add NFT_CHAIN_BINDING")
Reported-by: Kevin Rich <kevinrich1337@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: nf_tables: skip immediate deactivate in _PREPARE_ERROR

[ Upstream commit 0a771f7b266b02d262900c75f1e175c7fe76fec2 ]

On error when building the rule, the immediate expression unbinds the
chain, hence objects can be deactivated by the transaction records.

Otherwise, it is possible to trigger the following warning:

 WARNING: CPU: 3 PID: 915 at net/netfilter/nf_tables_api.c:2013 nf_tables_chain_destroy+0x1f7/0x210 [nf_tables]
 CPU: 3 PID: 915 Comm: chain-bind-err- Not tainted 6.1.39 #1
 RIP: 0010:nf_tables_chain_destroy+0x1f7/0x210 [nf_tables]

Fixes: 4bedf9eee016 ("netfilter: nf_tables: fix chain binding transaction logic")
Reported-by: Kevin Rich <kevinrich1337@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: nft_set_rbtree: fix overlap expiration walk

[ Upstream commit f718863aca469a109895cb855e6b81fff4827d71 ]

The lazy gc on insert that should remove timed-out entries fails to release
the other half of the interval, if any.

Can be reproduced with tests/shell/testcases/sets/0044interval_overlap_0
in nftables.git and kmemleak enabled kernel.

Second bug is the use of rbe_prev vs. prev pointer.
If rbe_prev() returns NULL after at least one iteration, rbe_prev points
to element that is not an end interval, hence it should not be removed.

Lastly, check the genmask of the end interval if this is active in the
current generation.

Fixes: c9e6978e2725 ("netfilter: nft_set_rbtree: Switch to node list walk for overlap detection")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

igc: Fix Kernel Panic during ndo_tx_timeout callback

[ Upstream commit d4a7ce642100765119a872d4aba1bf63e3a22c8a ]

The Xeon validation group has been carrying out some loaded tests
with various HW configurations, and they have seen some transmit
queue time out happening during the test. This will cause the
reset adapter function to be called by igc_tx_timeout().
Similar race conditions may arise when the interface is being brought
down and up in igc_reinit_locked(), an interrupt being generated, and
igc_clean_tx_irq() being called to complete the TX.

When the igc_tx_timeout() function is invoked, this patch will turn
off all TX ring HW queues during igc_down() process. TX ring HW queues
will be activated again during the igc_configure_tx_ring() process
when performing the igc_up() procedure later.

This patch also moved existing igc_disable_tx_ring_hw() to avoid using
forward declaration.

Kernel trace:
[ 7678.747813] ------------[ cut here ]------------
[ 7678.757914] NETDEV WATCHDOG: enp1s0 (igc): transmit queue 2 timed out
[ 7678.770117] WARNING: CPU: 0 PID: 13 at net/sched/sch_generic.c:525 dev_watchdog+0x1ae/0x1f0
[ 7678.784459] Modules linked in: xt_conntrack nft_chain_nat xt_MASQUERADE xt_addrtype nft_compat
nf_tables nfnetlink br_netfilter bridge stp llc overlay dm_mod emrcha(PO) emriio(PO) rktpm(PO)
cegbuf_mod(PO) patch_update(PO) se(PO) sgx_tgts(PO) mktme(PO) keylocker(PO) svtdx(PO) svfs_pci_hotplug(PO)
vtd_mod(PO) davemem(PO) svmabort(PO) svindexio(PO) usbx2(PO) ehci_sched(PO) svheartbeat(PO) ioapic(PO)
sv8259(PO) svintr(PO) lt(PO) pcierootport(PO) enginefw_mod(PO) ata(PO) smbus(PO) spiflash_cdf(PO) arden(PO)
dsa_iax(PO) oobmsm_punit(PO) cpm(PO) svkdb(PO) ebg_pch(PO) pch(PO) sviotargets(PO) svbdf(PO) svmem(PO)
svbios(PO) dram(PO) svtsc(PO) targets(PO) superio(PO) svkernel(PO) cswitch(PO) mcf(PO) pentiumIII_mod(PO)
fs_svfs(PO) mdevdefdb(PO) svfs_os_services(O) ixgbe mdio mdio_devres libphy emeraldrapids_svdefs(PO)
regsupport(O) libnvdimm nls_cp437 snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_hda_intel
snd_intel_dspcfg snd_hda_codec snd_hwdep x86_pkg_temp_thermal snd_hda_core snd_pcm snd_timer isst_if_mbox_pci
[ 7678.784496]  input_leds isst_if_mmio sg snd isst_if_common soundcore wmi button sad9(O) drm fuse backlight
configfs efivarfs ip_tables x_tables vmd sdhci led_class rtl8150 r8152 hid_generic pegasus mmc_block usbhid
mmc_core hid megaraid_sas ixgb igb i2c_algo_bit ice i40e hpsa scsi_transport_sas e1000e e1000 e100 ax88179_178a
usbnet xhci_pci sd_mod xhci_hcd t10_pi crc32c_intel crc64_rocksoft igc crc64 crc_t10dif usbcore
crct10dif_generic ptp crct10dif_common usb_common pps_core
[ 7679.200403] RIP: 0010:dev_watchdog+0x1ae/0x1f0
[ 7679.210201] Code: 28 e9 53 ff ff ff 4c 89 e7 c6 05 06 42 b9 00 01 e8 17 d1 fb ff 44 89 e9 4c
89 e6 48 c7 c7 40 ad fb 81 48 89 c2 e8 52 62 82 ff <0f> 0b e9 72 ff ff ff 65 8b 05 80 7d 7c 7e
89 c0 48 0f a3 05 0a c1
[ 7679.245438] RSP: 0018:ffa00000001f7d90 EFLAGS: 00010282
[ 7679.256021] RAX: 0000000000000000 RBX: ff11000109938440 RCX: 0000000000000000
[ 7679.268710] RDX: ff11000361e26cd8 RSI: ff11000361e1b880 RDI: ff11000361e1b880
[ 7679.281314] RBP: ffa00000001f7da8 R08: ff1100035f8fffe8 R09: 0000000000027ffb
[ 7679.293840] R10: 0000000000001f0a R11: ff1100035f840000 R12: ff11000109938000
[ 7679.306276] R13: 0000000000000002 R14: dead000000000122 R15: ffa00000001f7e18
[ 7679.318648] FS:  0000000000000000(0000) GS:ff11000361e00000(0000) knlGS:0000000000000000
[ 7679.332064] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 7679.342757] CR2: 00007ffff7fca168 CR3: 000000013b08a006 CR4: 0000000000471ef8
[ 7679.354984] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 7679.367207] DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
[ 7679.379370] PKRU: 55555554
[ 7679.386446] Call Trace:
[ 7679.393152]  <TASK>
[ 7679.399363]  ? __pfx_dev_watchdog+0x10/0x10
[ 7679.407870]  call_timer_fn+0x31/0x110
[ 7679.415698]  expire_timers+0xb2/0x120
[ 7679.423403]  run_timer_softirq+0x179/0x1e0
[ 7679.431532]  ? __schedule+0x2b1/0x820
[ 7679.439078]  __do_softirq+0xd1/0x295
[ 7679.446426]  ? __pfx_smpboot_thread_fn+0x10/0x10
[ 7679.454867]  run_ksoftirqd+0x22/0x30
[ 7679.462058]  smpboot_thread_fn+0xb7/0x160
[ 7679.469670]  kthread+0xcd/0xf0
[ 7679.476097]  ? __pfx_kthread+0x10/0x10
[ 7679.483211]  ret_from_fork+0x29/0x50
[ 7679.490047]  </TASK>
[ 7679.495204] ---[ end trace 0000000000000000 ]---
[ 7679.503179] igc 0000:01:00.0 enp1s0: Register Dump
[ 7679.511230] igc 0000:01:00.0 enp1s0: Register Name   Value
[ 7679.519892] igc 0000:01:00.0 enp1s0: CTRL            181c0641
[ 7679.528782] igc 0000:01:00.0 enp1s0: STATUS          40280683
[ 7679.537551] igc 0000:01:00.0 enp1s0: CTRL_EXT        10000040
[ 7679.546284] igc 0000:01:00.0 enp1s0: MDIC            180a3800
[ 7679.554942] igc 0000:01:00.0 enp1s0: ICR             00000081
[ 7679.563503] igc 0000:01:00.0 enp1s0: RCTL            04408022
[ 7679.571963] igc 0000:01:00.0 enp1s0: RDLEN[0-3]      00001000 00001000 00001000 00001000
[ 7679.583075] igc 0000:01:00.0 enp1s0: RDH[0-3]        00000068 000000b6 0000000f 00000031
[ 7679.594162] igc 0000:01:00.0 enp1s0: RDT[0-3]        00000066 000000b2 0000000e 00000030
[ 7679.605174] igc 0000:01:00.0 enp1s0: RXDCTL[0-3]     02040808 02040808 02040808 02040808
[ 7679.616196] igc 0000:01:00.0 enp1s0: RDBAL[0-3]      1bb7c000 1bb7f000 1bb82000 0ef33000
[ 7679.627242] igc 0000:01:00.0 enp1s0: RDBAH[0-3]      00000001 00000001 00000001 00000001
[ 7679.638256] igc 0000:01:00.0 enp1s0: TCTL            a503f0fa
[ 7679.646607] igc 0000:01:00.0 enp1s0: TDBAL[0-3]      2ba4a000 1bb6f000 1bb74000 1bb79000
[ 7679.657609] igc 0000:01:00.0 enp1s0: TDBAH[0-3]      00000001 00000001 00000001 00000001
[ 7679.668551] igc 0000:01:00.0 enp1s0: TDLEN[0-3]      00001000 00001000 00001000 00001000
[ 7679.679470] igc 0000:01:00.0 enp1s0: TDH[0-3]        000000a7 0000002d 000000bf 000000d9
[ 7679.690406] igc 0000:01:00.0 enp1s0: TDT[0-3]        000000a7 0000002d 000000bf 000000d9
[ 7679.701264] igc 0000:01:00.0 enp1s0: TXDCTL[0-3]     02100108 02100108 02100108 02100108
[ 7679.712123] igc 0000:01:00.0 enp1s0: Reset adapter
[ 7683.085967] igc 0000:01:00.0 enp1s0: NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX
[ 8086.945561] ------------[ cut here ]------------
Entering kdb (current=0xffffffff8220b200, pid 0) on processor 0
Oops: (null) due to oops @ 0xffffffff81573888
RIP: 0010:dql_completed+0x148/0x160
Code: c9 00 48 89 57 58 e9 46 ff ff ff 45 85 e4 41 0f 95 c4 41 39 db 0f 95
c1 41 84 cc 74 05 45 85 ed 78 0a 44 89 c1 e9 27 ff ff ff <0f> 0b 01 f6 44 89
c1 29 f1 0f 48 ca eb 8c cc cc cc cc cc cc cc cc
RSP: 0018:ffa0000000003e00 EFLAGS: 00010287
RAX: 000000000000006c RBX: ffa0000003eb0f78 RCX: ff11000109938000
RDX: 0000000000000003 RSI: 0000000000000160 RDI: ff110001002e9480
RBP: ffa0000000003ed8 R08: ff110001002e93c0 R09: ffa0000000003d28
R10: 0000000000007cc0 R11: 0000000000007c54 R12: 00000000ffffffd9
R13: ff1100037039cb00 R14: 00000000ffffffd9 R15: ff1100037039c048
FS:  0000000000000000(0000) GS:ff11000361e00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007ffff7fca168 CR3: 000000013b08a003 CR4: 0000000000471ef8
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
 <IRQ>
 ? igc_poll+0x1a9/0x14d0 [igc]
 __napi_poll+0x2e/0x1b0
 net_rx_action+0x126/0x250
 __do_softirq+0xd1/0x295
 irq_exit_rcu+0xc5/0xf0
 common_interrupt+0x86/0xa0
 </IRQ>
 <TASK>
 asm_common_interrupt+0x27/0x40
RIP: 0010:cpuidle_enter_state+0xd3/0x3e0
Code: 73 f1 ff ff 49 89 c6 8b 05 e2 ca a7 00 85 c0 0f 8f b3 02 00 00 31 ff e8 1b
de 75 ff 80 7d d7 00 0f 85 cd 01 00 00 fb 45 85 ff <0f> 88 fd 00 00 00 49 63 cf
4c 2b 75 c8 48 8d 04 49 48 89 ca 48 8d
RSP: 0018:ffffffff82203df0 EFLAGS: 00000202
RAX: ff11000361e2a200 RBX: 0000000000000002 RCX: 000000000000001f
RDX: 0000000000000000 RSI: 000000003cf3cf3d RDI: 0000000000000000
RBP: ffffffff82203e28 R08: 0000075ae38471c8 R09: 0000000000000018
R10: 000000000000031a R11: ffffffff8238dca0 R12: ffd1ffffff200000
R13: ffffffff8238dca0 R14: 0000075ae38471c8 R15: 0000000000000002
 cpuidle_enter+0x2e/0x50
 call_cpuidle+0x23/0x40
 do_idle+0x1be/0x220
 cpu_startup_entry+0x20/0x30
 rest_init+0xb5/0xc0
 arch_call_rest_init+0xe/0x30
 start_kernel+0x448/0x760
 x86_64_start_kernel+0x109/0x150
 secondary_startup_64_no_verify+0xe0/0xeb
 </TASK>
more>
[0]kdb>

[0]kdb>
[0]kdb> go
Catastrophic error detected
kdb_continue_catastrophic=0, type go a second time if you really want to
continue
[0]kdb> go
Catastrophic error detected
kdb_continue_catastrophic=0, attempting to continue
[ 8086.955689] refcount_t: underflow; use-after-free.
[ 8086.955697] WARNING: CPU: 0 PID: 0 at lib/refcount.c:28 refcount_warn_saturate+0xc2/0x110
[ 8086.955706] Modules linked in: xt_conntrack nft_chain_nat xt_MASQUERADE xt_addrtype nft_compat
nf_tables nfnetlink br_netfilter bridge stp llc overlay dm_mod emrcha(PO) emriio(PO) rktpm(PO)
cegbuf_mod(PO) patch_update(PO) se(PO) sgx_tgts(PO) mktme(PO) keylocker(PO) svtdx(PO)
svfs_pci_hotplug(PO) vtd_mod(PO) davemem(PO) svmabort(PO) svindexio(PO) usbx2(PO) ehci_sched(PO)
svheartbeat(PO) ioapic(PO) sv8259(PO) svintr(PO) lt(PO) pcierootport(PO) enginefw_mod(PO) ata(PO)
smbus(PO) spiflash_cdf(PO) arden(PO) dsa_iax(PO) oobmsm_punit(PO) cpm(PO) svkdb(PO) ebg_pch(PO)
pch(PO) sviotargets(PO) svbdf(PO) svmem(PO) svbios(PO) dram(PO) svtsc(PO) targets(PO) superio(PO)
svkernel(PO) cswitch(PO) mcf(PO) pentiumIII_mod(PO) fs_svfs(PO) mdevdefdb(PO) svfs_os_services(O)
ixgbe mdio mdio_devres libphy emeraldrapids_svdefs(PO) regsupport(O) libnvdimm nls_cp437
snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_hda_intel snd_intel_dspcfg
snd_hda_codec snd_hwdep x86_pkg_temp_thermal snd_hda_core snd_pcm snd_timer isst_if_mbox_pci
[ 8086.955751]  input_leds isst_if_mmio sg snd isst_if_common soundcore wmi button sad9(O) drm
fuse backlight configfs efivarfs ip_tables x_tables vmd sdhci led_class rtl8150 r8152 hid_generic
pegasus mmc_block usbhid mmc_core hid megaraid_sas ixgb igb i2c_algo_bit ice i40e hpsa
scsi_transport_sas e1000e e1000 e100 ax88179_178a usbnet xhci_pci sd_mod xhci_hcd t10_pi
crc32c_intel crc64_rocksoft igc crc64 crc_t10dif usbcore crct10dif_generic ptp crct10dif_common
usb_common pps_core
[ 8086.955784] RIP: 0010:refcount_warn_saturate+0xc2/0x110
[ 8086.955788] Code: 01 e8 82 e7 b4 ff 0f 0b 5d c3 cc cc cc cc 80 3d 68 c6 eb 00 00 75 81
48 c7 c7 a0 87 f6 81 c6 05 58 c6 eb 00 01 e8 5e e7 b4 ff <0f> 0b 5d c3 cc cc cc cc 80 3d
42 c6 eb 00 00 0f 85 59 ff ff ff 48
[ 8086.955790] RSP: 0018:ffa0000000003da0 EFLAGS: 00010286
[ 8086.955793] RAX: 0000000000000000 RBX: ff1100011da40ee0 RCX: ff11000361e1b888
[ 8086.955794] RDX: 00000000ffffffd8 RSI: 0000000000000027 RDI: ff11000361e1b880
[ 8086.955795] RBP: ffa0000000003da0 R08: 80000000ffff9f45 R09: ffa0000000003d28
[ 8086.955796] R10: ff1100035f840000 R11: 0000000000000028 R12: ff11000319ff8000
[ 8086.955797] R13: ff1100011bb79d60 R14: 00000000ffffffd6 R15: ff1100037039cb00
[ 8086.955798] FS:  0000000000000000(0000) GS:ff11000361e00000(0000) knlGS:0000000000000000
[ 8086.955800] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 8086.955801] CR2: 00007ffff7fca168 CR3: 000000013b08a003 CR4: 0000000000471ef8
[ 8086.955803] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 8086.955803] DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
[ 8086.955804] PKRU: 55555554
[ 8086.955805] Call Trace:
[ 8086.955806]  <IRQ>
[ 8086.955808]  tcp_wfree+0x112/0x130
[ 8086.955814]  skb_release_head_state+0x24/0xa0
[ 8086.955818]  napi_consume_skb+0x9c/0x160
[ 8086.955821]  igc_poll+0x5d8/0x14d0 [igc]
[ 8086.955835]  __napi_poll+0x2e/0x1b0
[ 8086.955839]  net_rx_action+0x126/0x250
[ 8086.955843]  __do_softirq+0xd1/0x295
[ 8086.955846]  irq_exit_rcu+0xc5/0xf0
[ 8086.955851]  common_interrupt+0x86/0xa0
[ 8086.955857]  </IRQ>
[ 8086.955857]  <TASK>
[ 8086.955858]  asm_common_interrupt+0x27/0x40
[ 8086.955862] RIP: 0010:cpuidle_enter_state+0xd3/0x3e0
[ 8086.955866] Code: 73 f1 ff ff 49 89 c6 8b 05 e2 ca a7 00 85 c0 0f 8f b3 02 00 00 31 ff e8
1b de 75 ff 80 7d d7 00 0f 85 cd 01 00 00 fb 45 85 ff <0f> 88 fd 00 00 00 49 63 cf 4c 2b 75
c8 48 8d 04 49 48 89 ca 48 8d
[ 8086.955867] RSP: 0018:ffffffff82203df0 EFLAGS: 00000202
[ 8086.955869] RAX: ff11000361e2a200 RBX: 0000000000000002 RCX: 000000000000001f
[ 8086.955870] RDX: 0000000000000000 RSI: 000000003cf3cf3d RDI: 0000000000000000
[ 8086.955871] RBP: ffffffff82203e28 R08: 0000075ae38471c8 R09: 0000000000000018
[ 8086.955872] R10: 000000000000031a R11: ffffffff8238dca0 R12: ffd1ffffff200000
[ 8086.955873] R13: ffffffff8238dca0 R14: 0000075ae38471c8 R15: 0000000000000002
[ 8086.955875]  cpuidle_enter+0x2e/0x50
[ 8086.955880]  call_cpuidle+0x23/0x40
[ 8086.955884]  do_idle+0x1be/0x220
[ 8086.955887]  cpu_startup_entry+0x20/0x30
[ 8086.955889]  rest_init+0xb5/0xc0
[ 8086.955892]  arch_call_rest_init+0xe/0x30
[ 8086.955895]  start_kernel+0x448/0x760
[ 8086.955898]  x86_64_start_kernel+0x109/0x150
[ 8086.955900]  secondary_startup_64_no_verify+0xe0/0xeb
[ 8086.955904]  </TASK>
[ 8086.955904] ---[ end trace 0000000000000000 ]---
[ 8086.955912] ------------[ cut here ]------------
[ 8086.955913] kernel BUG at lib/dynamic_queue_limits.c:27!
[ 8086.955918] invalid opcode: 0000 [#1] SMP
[ 8086.955922] RIP: 0010:dql_completed+0x148/0x160
[ 8086.955925] Code: c9 00 48 89 57 58 e9 46 ff ff ff 45 85 e4 41 0f 95 c4 41 39 db
0f 95 c1 41 84 cc 74 05 45 85 ed 78 0a 44 89 c1 e9 27 ff ff ff <0f> 0b 01 f6 44 89
c1 29 f1 0f 48 ca eb 8c cc cc cc cc cc cc cc cc
[ 8086.955927] RSP: 0018:ffa0000000003e00 EFLAGS: 00010287
[ 8086.955928] RAX: 000000000000006c RBX: ffa0000003eb0f78 RCX: ff11000109938000
[ 8086.955929] RDX: 0000000000000003 RSI: 0000000000000160 RDI: ff110001002e9480
[ 8086.955930] RBP: ffa0000000003ed8 R08: ff110001002e93c0 R09: ffa0000000003d28
[ 8086.955931] R10: 0000000000007cc0 R11: 0000000000007c54 R12: 00000000ffffffd9
[ 8086.955932] R13: ff1100037039cb00 R14: 00000000ffffffd9 R15: ff1100037039c048
[ 8086.955933] FS:  0000000000000000(0000) GS:ff11000361e00000(0000) knlGS:0000000000000000
[ 8086.955934] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 8086.955935] CR2: 00007ffff7fca168 CR3: 000000013b08a003 CR4: 0000000000471ef8
[ 8086.955936] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 8086.955937] DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
[ 8086.955938] PKRU: 55555554
[ 8086.955939] Call Trace:
[ 8086.955939]  <IRQ>
[ 8086.955940]  ? igc_poll+0x1a9/0x14d0 [igc]
[ 8086.955949]  __napi_poll+0x2e/0x1b0
[ 8086.955952]  net_rx_action+0x126/0x250
[ 8086.955956]  __do_softirq+0xd1/0x295
[ 8086.955958]  irq_exit_rcu+0xc5/0xf0
[ 8086.955961]  common_interrupt+0x86/0xa0
[ 8086.955964]  </IRQ>
[ 8086.955965]  <TASK>
[ 8086.955965]  asm_common_interrupt+0x27/0x40
[ 8086.955968] RIP: 0010:cpuidle_enter_state+0xd3/0x3e0
[ 8086.955971] Code: 73 f1 ff ff 49 89 c6 8b 05 e2 ca a7 00 85 c0 0f 8f b3 02 00 00
31 ff e8 1b de 75 ff 80 7d d7 00 0f 85 cd 01 00 00 fb 45 85 ff <0f> 88 fd 00 00 00
49 63 cf 4c 2b 75 c8 48 8d 04 49 48 89 ca 48 8d
[ 8086.955972] RSP: 0018:ffffffff82203df0 EFLAGS: 00000202
[ 8086.955973] RAX: ff11000361e2a200 RBX: 0000000000000002 RCX: 000000000000001f
[ 8086.955974] RDX: 0000000000000000 RSI: 000000003cf3cf3d RDI: 0000000000000000
[ 8086.955974] RBP: ffffffff82203e28 R08: 0000075ae38471c8 R09: 0000000000000018
[ 8086.955975] R10: 000000000000031a R11: ffffffff8238dca0 R12: ffd1ffffff200000
[ 8086.955976] R13: ffffffff8238dca0 R14: 0000075ae38471c8 R15: 0000000000000002
[ 8086.955978]  cpuidle_enter+0x2e/0x50
[ 8086.955981]  call_cpuidle+0x23/0x40
[ 8086.955984]  do_idle+0x1be/0x220
[ 8086.955985]  cpu_startup_entry+0x20/0x30
[ 8086.955987]  rest_init+0xb5/0xc0
[ 8086.955990]  arch_call_rest_init+0xe/0x30
[ 8086.955992]  start_kernel+0x448/0x760
[ 8086.955994]  x86_64_start_kernel+0x109/0x150
[ 8086.955996]  secondary_startup_64_no_verify+0xe0/0xeb
[ 8086.955998]  </TASK>
[ 8086.955999] Modules linked in: xt_conntrack nft_chain_nat xt_MASQUERADE xt_addrtype
nft_compat nf_tables nfnetlink br_netfilter bridge stp llc overlay dm_mod emrcha(PO) emriio(PO)
rktpm(PO) cegbuf_mod(PO) patch_update(PO) se(PO) sgx_tgts(PO) mktme(PO) keylocker(PO) svtdx(PO)
svfs_pci_hotplug(PO) vtd_mod(PO) davemem(PO) svmabort(PO) svindexio(PO) usbx2(PO) ehci_sched(PO)
svheartbeat(PO) ioapic(PO) sv8259(PO) svintr(PO) lt(PO) pcierootport(PO) enginefw_mod(PO) ata(PO)
smbus(PO) spiflash_cdf(PO) arden(PO) dsa_iax(PO) oobmsm_punit(PO) cpm(PO) svkdb(PO) ebg_pch(PO)
pch(PO) sviotargets(PO) svbdf(PO) svmem(PO) svbios(PO) dram(PO) svtsc(PO) targets(PO) superio(PO)
svkernel(PO) cswitch(PO) mcf(PO) pentiumIII_mod(PO) fs_svfs(PO) mdevdefdb(PO) svfs_os_services(O)
ixgbe mdio mdio_devres libphy emeraldrapids_svdefs(PO) regsupport(O) libnvdimm nls_cp437
snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_hda_intel snd_intel_dspcfg
snd_hda_codec snd_hwdep x86_pkg_temp_thermal snd_hda_core snd_pcm snd_timer isst_if_mbox_pci
[ 8086.956029]  input_leds isst_if_mmio sg snd isst_if_common soundcore wmi button sad9(O) drm
fuse backlight configfs efivarfs ip_tables x_tables vmd sdhci led_class rtl8150 r8152 hid_generic
pegasus mmc_block usbhid mmc_core hid megaraid_sas ixgb igb i2c_algo_bit ice i40e hpsa
scsi_transport_sas e1000e e1000 e100 ax88179_178a usbnet xhci_pci sd_mod xhci_hcd t10_pi
crc32c_intel crc64_rocksoft igc crc64 crc_t10dif usbcore crct10dif_generic ptp crct10dif_common
usb_common pps_core
[16762.543675] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.593 msecs
[16762.543678] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.595 msecs
[16762.543673] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.495 msecs
[16762.543679] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.599 msecs
[16762.543678] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.598 msecs
[16762.543690] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.605 msecs
[16762.543684] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.599 msecs
[16762.543693] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 8675587.613 msecs
[16762.543784] ---[ end trace 0000000000000000 ]---
[16762.849099] RIP: 0010:dql_completed+0x148/0x160
PANIC: Fatal exception in interrupt

Fixes: 9b275176270e ("igc: Add ndo_tx_timeout support")
Tested-by: Alejandra Victoria Alcaraz <alejandra.victoria.alcaraz@intel.com>
Signed-off-by: Muhammad Husaini Zulkifli <muhammad.husaini.zulkifli@intel.com>
Acked-by: Sasha Neftin <sasha.neftin@intel.com>
Tested-by: Naama Meir <naamax.meir@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/traps: Fix load_unaligned_zeropad() handling for shared TDX memory

[ Upstream commit 9f9116406120638b4d8db3831ffbc430dd2e1e95 ]

Commit c4e34dd99f2e ("x86: simplify load_unaligned_zeropad()
implementation") changes how exceptions around load_unaligned_zeropad()
handled.  The kernel now uses the fault_address in fixup_exception() to
verify the address calculations for the load_unaligned_zeropad().

It works fine for #PF, but breaks on #VE since no fault address is
passed down to fixup_exception().

Propagating ve_info.gla down to fixup_exception() resolves the issue.

See commit 1e7769653b06 ("x86/tdx: Handle load_unaligned_zeropad()
page-cross to a shared page") for more context.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-by: Michael Kelley <mikelley@microsoft.com>
Fixes: c4e34dd99f2e ("x86: simplify load_unaligned_zeropad() implementation")
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86: msi-laptop: Fix rfkill out-of-sync on MSI Wind U100

[ Upstream commit ad084a6d99bc182bf109c190c808e2ea073ec57b ]

Only the HW rfkill state is toggled on laptops with quirks->ec_read_only
(so far only MSI Wind U90/U100). There are, however, a few issues with
the implementation:

1. The initial HW state is always unblocked, regardless of the actual
   state on boot, because msi_init_rfkill only sets the SW state,
   regardless of ec_read_only.

2. The initial SW state corresponds to the actual state on boot, but it
   can't be changed afterwards, because set_device_state returns
   -EOPNOTSUPP. It confuses the userspace, making Wi-Fi and/or Bluetooth
   unusable if it was blocked on boot, and breaking the airplane mode if
   the rfkill was unblocked on boot.

Address the above issues by properly initializing the HW state on
ec_read_only laptops and by allowing the userspace to toggle the SW
state. Don't set the SW state ourselves and let the userspace fully
control it. Toggling the SW state is a no-op, however, it allows the
userspace to properly toggle the airplane mode. The actual SW radio
disablement is handled by the corresponding rtl818x_pci and btusb
drivers that have their own rfkills.

Tested on MSI Wind U100 Plus, BIOS ver 1.0G, EC ver 130.

Fixes: 0816392b97d4 ("msi-laptop: merge quirk tables to one")
Fixes: 0de6575ad0a8 ("msi-laptop: Add MSI Wind U90/U100 support")
Signed-off-by: Maxim Mikityanskiy <maxtram95@gmail.com>
Link: https://lore.kernel.org/r/20230721145423.161057-1-maxtram95@gmail.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>

net: stmmac: Apply redundant write work around on 4.xx too

[ Upstream commit 284779dbf4e98753458708783af8c35630674a21 ]

commit a3a57bf07de23fe1ff779e0fdf710aa581c3ff73 ("net: stmmac: work
around sporadic tx issue on link-up") worked around a problem with TX
sometimes not working after a link-up by avoiding a redundant write to
MAC_CTRL_REG (aka GMAC_CONFIG), since the IP appeared to have problems
with handling multiple writes to that register in some cases.

That commit however only added the work around to dwmac_lib.c (apart
from the common code in stmmac_main.c), but my systems with version
4.21a of the IP exhibit the same problem, so add the work around to
dwmac4_lib.c too.

Fixes: a3a57bf07de2 ("net: stmmac: work around sporadic tx issue on link-up")
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230721-stmmac-tx-workaround-v1-1-9411cbd5ee07@axis.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

octeontx2-af: Fix hash extraction enable configuration

[ Upstream commit 4e62c99d71e56817c934caa2a709a775c8cee078 ]

As of today, hash extraction support is enabled for all the silicons.
Because of which we are facing initialization issues when the silicon
does not support hash extraction. During creation of the hardware
parsing table for IPv6 address, we need to consider if hash extraction
is enabled then extract only 32 bit, otherwise 128 bit needs to be
extracted. This patch fixes the issue and configures the hardware parser
based on the availability of the feature.

Fixes: a95ab93550d3 ("octeontx2-af: Use hashed field in MCAM key")
Signed-off-by: Suman Ghosh <sumang@marvell.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230721061222.2632521-1-sumang@marvell.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

octeontx2-af: Removed unnecessary debug messages.

[ Upstream commit 609aa68d60965f70485655def733d533f99b341b ]

NPC exact match feature is supported only on one silicon
variant, removed debug messages which print that this
feature is not available on all other silicon variants.

Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: Ratheesh Kannoth <rkannoth@marvell.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20230201040301.1034843-1-rkannoth@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 4e62c99d71e5 ("octeontx2-af: Fix hash extraction enable configuration")
Signed-off-by: Sasha Levin <sashal@kernel.org>

team: reset team's flags when down link is P2P device

[ Upstream commit fa532bee17d15acf8bba4bc8e2062b7a093ba801 ]

When adding a point to point downlink to team device, we neglected to reset
the team's flags, which were still using flags like BROADCAST and
MULTICAST. Consequently, this would initiate ARP/DAD for P2P downlink
interfaces, such as when adding a GRE device to team device. Fix this by
remove multicast/broadcast flags and add p2p and noarp flags.

After removing the none ethernet interface and adding an ethernet interface
to team, we need to reset team interface flags. Unlike bonding interface,
team do not need restore IFF_MASTER, IFF_SLAVE flags.

Reported-by: Liang Li <liali@redhat.com>
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=2221438
Fixes: 1d76efe1577b ("team: add support for non-ethernet devices")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bonding: reset bond's flags when down link is P2P device

[ Upstream commit da19a2b967cf1e2c426f50d28550d1915214a81d ]

When adding a point to point downlink to the bond, we neglected to reset
the bond's flags, which were still using flags like BROADCAST and
MULTICAST. Consequently, this would initiate ARP/DAD for P2P downlink
interfaces, such as when adding a GRE device to the bonding.

To address this issue, let's reset the bond's flags for P2P interfaces.

Before fix:
7: gre0@NONE: <POINTOPOINT,NOARP,SLAVE,UP,LOWER_UP> mtu 1500 qdisc noqueue master bond0 state UNKNOWN group default qlen 1000
    link/gre6 2006:70:10::1 peer 2006:70:10::2 permaddr 167f:18:f188::
8: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/gre6 2006:70:10::1 brd 2006:70:10::2
    inet6 fe80::200:ff:fe00:0/64 scope link
       valid_lft forever preferred_lft forever

After fix:
7: gre0@NONE: <POINTOPOINT,NOARP,SLAVE,UP,LOWER_UP> mtu 1500 qdisc noqueue master bond2 state UNKNOWN group default qlen 1000
    link/gre6 2006:70:10::1 peer 2006:70:10::2 permaddr c29e:557a:e9d9::
8: bond0: <POINTOPOINT,NOARP,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/gre6 2006:70:10::1 peer 2006:70:10::2
    inet6 fe80::1/64 scope link
       valid_lft forever preferred_lft forever

Reported-by: Liang Li <liali@redhat.com>
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=2221438
Fixes: 872254dd6b1f ("net/bonding: Enable bonding to enslave non ARPHRD_ETHER")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ice: Fix memory management in ice_ethtool_fdir.c

[ Upstream commit a3336056504d780590ac6d6ac94fbba829994594 ]

Fix ethtool FDIR logic to not use memory after its release.
In the ice_ethtool_fdir.c file there are 2 spots where code can
refer to pointers which may be missing.

In the ice_cfg_fdir_xtrct_seq() function seg may be freed but
even then may be still used by memcpy(&tun_seg[1], seg, sizeof(*seg)).

In the ice_add_fdir_ethtool() function struct ice_fdir_fltr *input
may first fail to be added via ice_fdir_update_list_entry() but then
may be deleted by ice_fdir_update_list_entry.

Terminate in both cases when the returned value of the previous
operation is other than 0, free memory and don't use it anymore.

Reported-by: Michal Schmidt <mschmidt@redhat.com>
Link: https://bugzilla.redhat.com/show_bug.cgi?id=2208423
Fixes: cac2a27cd9ab ("ice: Support IPv4 Flow Director filters")
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Jedrzej Jagielski <jedrzej.jagielski@intel.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Link: https://lore.kernel.org/r/20230721155854.1292805-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tcp: Reduce chance of collisions in inet6_hashfn().

[ Upstream commit d11b0df7ddf1831f3e170972f43186dad520bfcc ]

For both IPv4 and IPv6 incoming TCP connections are tracked in a hash
table with a hash over the source & destination addresses and ports.
However, the IPv6 hash is insufficient and can lead to a high rate of
collisions.

The IPv6 hash used an XOR to fit everything into the 96 bits for the
fast jenkins hash, meaning it is possible for an external entity to
ensure the hash collides, thus falling back to a linear search in the
bucket, which is slow.

We take the approach of hash the full length of IPv6 address in
__ipv6_addr_jhash() so that all users can benefit from a more secure
version.

While this may look like it adds overhead, the reality of modern CPUs
means that this is unmeasurable in real world scenarios.

In simulating with llvm-mca, the increase in cycles for the hashing
code was ~16 cycles on Skylake (from a base of ~155), and an extra ~9
on Nehalem (base of ~173).

In commit dd6d2910c5e0 ("netfilter: conntrack: switch to siphash")
netfilter switched from a jenkins hash to a siphash, but even the faster
hsiphash is a more significant overhead (~20-30%) in some preliminary
testing.  So, in this patch, we keep to the more conservative approach to
ensure we don't add much overhead per SYN.

In testing, this results in a consistently even spread across the
connection buckets.  In both testing and real-world scenarios, we have
not found any measurable performance impact.

Fixes: 08dcdbf6a7b9 ("ipv6: use a stronger hash for tcp")
Signed-off-by: Stewart Smith <trawets@amazon.com>
Signed-off-by: Samuel Mendoza-Jonas <samjonas@amazon.com>
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230721222410.17914-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ipv6 addrconf: fix bug where deleting a mngtmpaddr can create a new temporary address

[ Upstream commit 69172f0bcb6a09110c5d2a6d792627f5095a9018 ]

currently on 6.4 net/main:

  # ip link add dummy1 type dummy
  # echo 1 > /proc/sys/net/ipv6/conf/dummy1/use_tempaddr
  # ip link set dummy1 up
  # ip -6 addr add 2000::1/64 mngtmpaddr dev dummy1
  # ip -6 addr show dev dummy1

  11: dummy1: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000
      inet6 2000::44f3:581c:8ca:3983/64 scope global temporary dynamic
         valid_lft 604800sec preferred_lft 86172sec
      inet6 2000::1/64 scope global mngtmpaddr
         valid_lft forever preferred_lft forever
      inet6 fe80::e8a8:a6ff:fed5:56d4/64 scope link
         valid_lft forever preferred_lft forever

  # ip -6 addr del 2000::44f3:581c:8ca:3983/64 dev dummy1

  (can wait a few seconds if you want to, the above delete isn't [directly] the problem)

  # ip -6 addr show dev dummy1

  11: dummy1: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000
      inet6 2000::1/64 scope global mngtmpaddr
         valid_lft forever preferred_lft forever
      inet6 fe80::e8a8:a6ff:fed5:56d4/64 scope link
         valid_lft forever preferred_lft forever

  # ip -6 addr del 2000::1/64 mngtmpaddr dev dummy1
  # ip -6 addr show dev dummy1

  11: dummy1: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000
      inet6 2000::81c9:56b7:f51a:b98f/64 scope global temporary dynamic
         valid_lft 604797sec preferred_lft 86169sec
      inet6 fe80::e8a8:a6ff:fed5:56d4/64 scope link
         valid_lft forever preferred_lft forever

This patch prevents this new 'global temporary dynamic' address from being
created by the deletion of the related (same subnet prefix) 'mngtmpaddr'
(which is triggered by there already being no temporary addresses).

Cc: Jiri Pirko <jiri@resnulli.us>
Fixes: 53bd67491537 ("ipv6 addrconf: introduce IFA_F_MANAGETEMPADDR to tell kernel to manage temporary addresses")
Reported-by: Xiao Ma <xiaom@google.com>
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20230720160022.1887942-1-maze@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ethernet: atheros: fix return value check in atl1e_tso_csum()

[ Upstream commit 69a184f7a372aac588babfb0bd681aaed9779f5b ]

in atl1e_tso_csum, it should check the return value of pskb_trim(),
and return an error code if an unexpected value is returned
by pskb_trim().

Fixes: a6a5325239c2 ("atl1e: Atheros L1E Gigabit Ethernet driver")
Signed-off-by: Yuanjun Gong <ruc_gongyuanjun@163.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230720144219.39285-1-ruc_gongyuanjun@163.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

atheros: fix return value check in atl1_tso()

[ Upstream commit ed96824b71ed67664390890441b229423a25317f ]

in atl1_tso(), it should check the return value of pskb_trim(),
and return an error code if an unexpected value is returned
by pskb_trim().

Fixes: 401c0aabec4b ("atl1: simplify tx packet descriptor")
Signed-off-by: Yuanjun Gong <ruc_gongyuanjun@163.com>
Link: https://lore.kernel.org/r/20230722142511.12448-1-ruc_gongyuanjun@163.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

phy: hisilicon: Fix an out of bounds check in hisi_inno_phy_probe()

[ Upstream commit 13c088cf3657d70893d75cf116be937f1509cc0f ]

The size of array 'priv->ports[]' is INNO_PHY_PORT_NUM.

In the for loop, 'i' is used as the index for array 'priv->ports[]'
with a check (i > INNO_PHY_PORT_NUM) which indicates that
INNO_PHY_PORT_NUM is allowed value for 'i' in the same loop.

This > comparison needs to be changed to >=, otherwise it potentially leads
to an out of bounds write on the next iteration through the loop

Fixes: ba8b0ee81fbb ("phy: add inno-usb2-phy driver for hi3798cv200 SoC")
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
Link: https://lore.kernel.org/r/20230721090558.3588613-1-harshit.m.mogalapalli@oracle.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

vxlan: fix GRO with VXLAN-GPE

[ Upstream commit b0b672c4d0957e5897685667fc848132b8bd2d71 ]

In VXLAN-GPE, there may not be an Ethernet header following the VXLAN
header. But in GRO, the vxlan driver calls eth_gro_receive
unconditionally, which means the following header is incorrectly parsed
as Ethernet.

Introduce GPE specific GRO handling.

For better performance, do not check for GPE during GRO but rather
install a different set of functions at setup time.

Fixes: e1e5314de08ba ("vxlan: implement GPE")
Reported-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

vxlan: generalize vxlan_parse_gpe_hdr and remove unused args

[ Upstream commit 17a0a64448b568442a101de09575f81ffdc45d15 ]

The vxlan_parse_gpe_hdr function extracts the next protocol value from
the GPE header and marks GPE bits as parsed.

In order to be used in the next patch, split the function into protocol
extraction and bit marking. The bit marking is meaningful only in
vxlan_rcv; move it directly there.

Rename the function to vxlan_parse_gpe_proto to reflect what it now
does. Remove unused arguments skb and vxflags. Move the function earlier
in the file to allow it to be called from more places in the next patch.

Signed-off-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: b0b672c4d095 ("vxlan: fix GRO with VXLAN-GPE")
Signed-off-by: Sasha Levin <sashal@kernel.org>

vxlan: calculate correct header length for GPE

[ Upstream commit 94d166c5318c6edd1e079df8552233443e909c33 ]

VXLAN-GPE does not add an extra inner Ethernet header. Take that into
account when calculating header length.

This causes problems in skb_tunnel_check_pmtu, where incorrect PMTU is
cached.

In the collect_md mode (which is the only mode that VXLAN-GPE
supports), there's no magic auto-setting of the tunnel interface MTU.
It can't be, since the destination and thus the underlying interface
may be different for each packet.

So, the administrator is responsible for setting the correct tunnel
interface MTU. Apparently, the administrators are capable enough to
calculate that the maximum MTU for VXLAN-GPE is (their_lower_MTU - 36).
They set the tunnel interface MTU to 1464. If you run a TCP stream over
such interface, it's then segmented according to the MTU 1464, i.e.
producing 1514 bytes frames. Which is okay, this still fits the lower
MTU.

However, skb_tunnel_check_pmtu (called from vxlan_xmit_one) uses 50 as
the header size and thus incorrectly calculates the frame size to be
1528. This leads to ICMP too big message being generated (locally),
PMTU of 1450 to be cached and the TCP stream to be resegmented.

The fix is to use the correct actual header size, especially for
skb_tunnel_check_pmtu calculation.

Fixes: e1e5314de08ba ("vxlan: implement GPE")
Signed-off-by: Jiri Benc <jbenc@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: hns3: fix wrong bw weight of disabled tc issue

[ Upstream commit 882481b1c55fc44861d7e2d54b4e0936b1b39f2c ]

In dwrr mode, the default bandwidth weight of disabled tc is set to 0.
If the bandwidth weight is 0, the mode will change to sp.
Therefore, disabled tc default bandwidth weight need changed to 1,
and 0 is returned when query the bandwidth weight of disabled tc.
In addition, driver need stop configure bandwidth weight if tc is disabled.

Fixes: 848440544b41 ("net: hns3: Add support of TX Scheduler & Shaper to HNS3 driver")
Signed-off-by: Jie Wang <wangjie125@huawei.com>
Signed-off-by: Jijie Shao <shaojijie@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: hns3: fix wrong tc bandwidth weight data issue

[ Upstream commit 116d9f732eef634abbd871f2c6f613a5b4677742 ]

Currently, the weight saved by the driver is used as the query result,
which may be different from the actual weight in the register.
Therefore, the register value read from the firmware is used
as the query result

Fixes: 0e32038dc856 ("net: hns3: refactor dump tc of debugfs")
Signed-off-by: Jijie Shao <shaojijie@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: hns3: fix the imp capability bit cannot exceed 32 bits issue

[ Upstream commit b27d0232e8897f7c896dc8ad80c9907dd57fd3f3 ]

Current only the first 32 bits of the capability flag bit are considered.
When the matching capability flag bit is greater than 31 bits,
it will get an error bit.This patch use bitmap to solve this issue.
It can handle each capability bit whitout bit width limit.

Fixes: da77aef9cc58 ("net: hns3: create common cmdq resource allocate/free/query APIs")
Signed-off-by: Hao Lan <lanhao@huawei.com>
Signed-off-by: Jijie Shao <shaojijie@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: phy: marvell10g: fix 88x3310 power up

[ Upstream commit c7b75bea853daeb64fc831dbf39a6bbabcc402ac ]

Clear MV_V2_PORT_CTRL_PWRDOWN bit to set power up for 88x3310 PHY,
it sometimes does not take effect immediately. And a read of this
register causes the bit not to clear. This will cause mv3310_reset()
to time out, which will fail the config initialization. So add a delay
before the next access.

Fixes: c9cc1c815d36 ("net: phy: marvell10g: place in powersave mode at probe")
Signed-off-by: Jiawen Wu <jiawenwu@trustnetic.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iavf: check for removal state before IAVF_FLAG_PF_COMMS_FAILED

[ Upstream commit 91896c8acce23d33ed078cffd46a9534b1f82be5 ]

In iavf_adminq_task(), if the function can't acquire the
adapter->crit_lock, it checks if the driver is removing. If so, it simply
exits without re-enabling the interrupt. This is done to ensure that the
task stops processing as soon as possible once the driver is being removed.

However, if the IAVF_FLAG_PF_COMMS_FAILED is set, the function checks this
before attempting to acquire the lock. In this case, the function exits
early and re-enables the interrupt. This will happen even if the driver is
already removing.

Avoid this, by moving the check to after the adapter->crit_lock is
acquired. This way, if the driver is removing, we will not re-enable the
interrupt.

Fixes: fc2e6b3b132a ("iavf: Rework mutexes for better synchronisation")
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iavf: fix potential deadlock on allocation failure

[ Upstream commit a2f054c10bef0b54600ec9cb776508443e941343 ]

In iavf_adminq_task(), if kzalloc() fails to allocate the event.msg_buf,
the function will exit without releasing the adapter->crit_lock.

This is unlikely, but if it happens, the next access to that mutex will
deadlock.

Fix this by moving the unlock to the end of the function, and adding a new
label to allow jumping to the unlock portion of the function exit flow.

Fixes: fc2e6b3b132a ("iavf: Rework mutexes for better synchronisation")
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

i40e: Fix an NULL vs IS_ERR() bug for debugfs_create_dir()

[ Upstream commit 043b1f185fb0f3939b7427f634787706f45411c4 ]

The debugfs_create_dir() function returns error pointers.
It never returns NULL. Most incorrect error checks were fixed,
but the one in i40e_dbg_init() was forgotten.

Fix the remaining error check.

Fixes: 02e9c290814c ("i40e: debugfs interface")
Signed-off-by: Wang Ming <machel@vivo.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: amphion: Fix firmware path to match linux-firmware

[ Upstream commit dcff0b56f661b6b42e828012b464d22cc2068c38 ]

The path did not match the one it was submitted into linux-firmware
which prevented generic distribution from having working CODEC.

Fixes: 9f599f351e86 ("media: amphion: add vpu core driver")
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: staging: atomisp: select V4L2_FWNODE

[ Upstream commit bf4c985707d3168ebb7d87d15830de66949d979c ]

Select V4L2_FWNODE as the driver depends on it.

Reported-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Fixes: aa31f6514047 ("media: atomisp: allow building the driver again")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

soundwire: qcom: update status correctly with mask

[ Upstream commit f84d41b2a083b990cbdf70f3b24b6b108b9678ad ]

SoundWire device status can be incorrectly updated without
proper mask, fix this by adding a mask before updating the status.

Fixes: c7d49c76d1d5 ("soundwire: qcom: add support to new interrupts")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20230525133812.30841-2-srinivas.kandagatla@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

phy: qcom-snps-femto-v2: properly enable ref clock

[ Upstream commit 8a0eb8f9b9a002291a3934acfd913660b905249e ]

The driver is not enabling the ref clock, which thus gets disabled by
the clk_disable_unused() initcall. This leads to the dwc3 controller
failing to initialize if probed after clk_disable_unused() is called,
for instance when the driver is built as a module.

To fix this, switch to the clk_bulk API to handle both cfg_ahb and ref
clocks at the proper places.

Note that the cfg_ahb clock is currently not used by any device tree
instantiation of the PHY. Work needs to be done separately to fix this.

Link: https://lore.kernel.org/linux-arm-msm/ZEqvy+khHeTkC2hf@fedora/
Fixes: 51e8114f80d0 ("phy: qcom-snps: Add SNPS USB PHY driver for QCOM based SOCs")
Signed-off-by: Adrien Thierry <athierry@redhat.com>
Link: https://lore.kernel.org/r/20230629144542.14906-3-athierry@redhat.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

phy: qcom-snps-femto-v2: keep cfg_ahb_clk enabled during runtime suspend

[ Upstream commit 45d89a344eb46db9dce851c28e14f5e3c635c251 ]

In the dwc3 core, both system and runtime suspend end up calling
dwc3_suspend_common(). From there, what happens for the PHYs depends on
the USB mode and whether the controller is entering system or runtime
suspend.

HOST mode:
  (1) system suspend on a non-wakeup-capable controller

  The [1] if branch is taken. dwc3_core_exit() is called, which ends up
  calling phy_power_off() and phy_exit(). Those two functions decrease the
  PM runtime count at some point, so they will trigger the PHY runtime
  sleep (assuming the count is right).

  (2) runtime suspend / system suspend on a wakeup-capable controller

  The [1] branch is not taken. dwc3_suspend_common() calls
  phy_pm_runtime_put_sync(). Assuming the ref count is right, the PHY
  runtime suspend op is called.

DEVICE mode:
  dwc3_core_exit() is called on both runtime and system sleep
  unless the controller is already runtime suspended.

OTG mode:
  (1) system suspend : dwc3_core_exit() is called

  (2) runtime suspend : do nothing

In host mode, the code seems to make a distinction between 1) runtime
sleep / system sleep for wakeup-capable controller, and 2) system sleep
for non-wakeup-capable controller, where phy_power_off() and phy_exit()
are only called for the latter. This suggests the PHY is not supposed to
be in a fully powered-off state for runtime sleep and system sleep for
wakeup-capable controller.

Moreover, downstream, cfg_ahb_clk only gets disabled for system suspend.
The clocks are disabled by phy->set_suspend() [2] which is only called
in the system sleep path through dwc3_core_exit() [3].

With that in mind, don't disable the clocks during the femto PHY runtime
suspend callback. The clocks will only be disabled during system suspend
for non-wakeup-capable controllers, through dwc3_core_exit().

[1] https://elixir.bootlin.com/linux/v6.4/source/drivers/usb/dwc3/core.c#L1988
[2] https://git.codelinaro.org/clo/la/kernel/msm-5.4/-/blob/LV.AU.1.2.1.r2-05300-gen3meta.0/drivers/usb/phy/phy-msm-snps-hs.c#L524
[3] https://git.codelinaro.org/clo/la/kernel/msm-5.4/-/blob/LV.AU.1.2.1.r2-05300-gen3meta.0/drivers/usb/dwc3/core.c#L1915

Signed-off-by: Adrien Thierry <athierry@redhat.com>
Link: https://lore.kernel.org/r/20230629144542.14906-2-athierry@redhat.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Stable-dep-of: 8a0eb8f9b9a0 ("phy: qcom-snps-femto-v2: properly enable ref clock")
Signed-off-by: Sasha Levin <sashal@kernel.org>

phy: qcom-snps: correct struct qcom_snps_hsphy kerneldoc

[ Upstream commit 2a881183dc5ab2474ef602e48fe7af34db460d95 ]

Update kerneldoc of struct qcom_snps_hsphy to fix:

  drivers/phy/qualcomm/phy-qcom-snps-femto-v2.c:135: warning: Function parameter or member 'update_seq_cfg' not described in 'qcom_snps_hsphy'

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Link: https://lore.kernel.org/r/20230507144818.193039-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Stable-dep-of: 8a0eb8f9b9a0 ("phy: qcom-snps-femto-v2: properly enable ref clock")
Signed-off-by: Sasha Levin <sashal@kernel.org>

phy: phy-mtk-dp: Fix an error code in probe()

[ Upstream commit 5782017cc4d0c8f3425d55b893675bb8a20c33e9 ]

Negative -EINVAL was intended instead of positive EINVAL.

Fixes: 6a23afad443a ("phy: phy-mtk-dp: Add driver for DP phy")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Link: https://lore.kernel.org/r/3c699e00-2883-40d9-92c3-0da1dc38fdd4@moroto.mountain
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Prevent vtotal from being set to 0

[ Upstream commit 2a9482e55968ed7368afaa9c2133404069117320 ]

[Why]
In dcn314 DML the destination pipe vtotal was being set
to the crtc adjustment vtotal_min value even in cases
where that value is 0.

[How]
Only set vtotal to the crtc adjustment vtotal_min value
in cases where the value is non-zero.

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Daniel Miess <daniel.miess@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Fix possible underflow for displays with large vblank

[ Upstream commit 1a4bcdbea4319efeb26cc4b05be859a7867e02dc ]

[Why]
Underflow observed when using a display with a large vblank region
and low refresh rate

[How]
Simplify calculation of vblank_nom

Increase value for VBlankNomDefaultUS to 800us

Reviewed-by: Jun Lei <jun.lei@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Daniel Miess <daniel.miess@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stable-dep-of: 2a9482e55968 ("drm/amd/display: Prevent vtotal from being set to 0")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: update extended blank for dcn314 onwards

[ Upstream commit 469a62938a45ef382c9cb7b9fec6c6c1fcd781c0 ]

[Why]
Flickering and underflow was observed when testing extended
blank on dcn314.

[What]
Vstartup is contrainted by vblank_nom, so adjusting it to include
non-adjusted vtotal in its calculation during freesync video
means that Vstartup is not changed when vtotal changes.
This fixed the flickering + underflow.

dc_extended_blank_supported function was removed
because extended blank is only relevant to when
zstate is supported. The increased vtotal during
freesync can be passed to dml regardless of whether
extended blank is supported or not, so this function is
not needed.

Updates were made recently in dml to the calculation of
min_dst_y_next_start. Dml input for dcn314 will now
always use the newer calculation for min_dst_y_next_start.
Dml input for older dcn versions remains untouched.

The variable optimized_min_dst_y_next_start
is replaced everywhere with min_dst_y_next_start,
and the updated dml allows min_dst_y_next_start to
increase to an optimized value during freesync video,
then return to default when freesync is disengaged.

Also removed registry key for controlling
extended blank feature.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Gabe Teeger <gabe.teeger@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stable-dep-of: 2a9482e55968 ("drm/amd/display: Prevent vtotal from being set to 0")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Add FAMS validation before trying to use it

[ Upstream commit e3416e872f84086667df21daf166506fab97358d ]

To ensure that FAMS can be used, DC must check if there is VRR support.
This commit adds the required configuration to ensure FAMS can be executed in the target system.

Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stable-dep-of: 2a9482e55968 ("drm/amd/display: Prevent vtotal from being set to 0")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: fix dc/core/dc.c kernel-doc

[ Upstream commit db4107e92a817502ad19fdd30250f87dcb6f6331 ]

Fix all kernel-doc warnings in dc/core/dc.c:

dc.c:385: warning: missing initial short description on line:
 *  dc_stream_adjust_vmin_vmax:
dc.c:392: warning: contents before sections
dc.c:399: warning: No description found for return value of 'dc_stream_adjust_vmin_vmax'
dc.c:434: warning: Excess function parameter 'adjust' description in 'dc_stream_get_last_used_drr_vtotal'
dc.c:434: warning: No description found for return value of 'dc_stream_get_last_used_drr_vtotal'
dc.c:574: warning: No description found for return value of 'dc_stream_configure_crc'
dc.c:1746: warning: No description found for return value of 'dc_commit_state_no_check'
dc.c:4991: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
 * dc_extended_blank_supported 0 Decide whether extended blank is supported
dc.c:4991: warning: missing initial short description on line:
 * dc_extended_blank_supported 0 Decide whether extended blank is supported
dc.c:4723: warning: Function parameter or member 'dc' not described in 'dc_enable_dmub_outbox'
dc.c:4926: warning: Function parameter or member 'dc' not described in 'dc_process_dmub_dpia_hpd_int_enable'
dc.c:4926: warning: Function parameter or member 'hpd_int_enable' not described in 'dc_process_dmub_dpia_hpd_int_enable'
12 warnings

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com>
Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Hamza Mahfooz <hamza.mahfooz@amd.com>
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Leo Li <sunpeng.li@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: "Pan, Xinhui" <Xinhui.Pan@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stable-dep-of: 2a9482e55968 ("drm/amd/display: Prevent vtotal from being set to 0")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Rework comments on dc file

[ Upstream commit e366f36958f60c431a7430c8c421c9db0ec6738d ]

The file dc.c has multiple comments that do not follow the kernel-doc or
are made in a distracting way. This commit alleviates part of this issue
by reorganizing some comments inside the dc file.

Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Reviewed-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stable-dep-of: 2a9482e55968 ("drm/amd/display: Prevent vtotal from being set to 0")
Signed-off-by: Sasha Levin <sashal@kernel.org>

maple_tree: fix 32 bit mas_next testing

[ Upstream commit 7a93c71a6714ca1a9c03d70432dac104b0cfb815 ]

The test setup of mas_next is dependent on node entry size to create a 2
level tree, but the tests did not account for this in the expected value
when shifting beyond the scope of the tree.

Fix this by setting up the test to succeed depending on the node entries
which is dependent on the 32/64 bit setup.

Link: https://lkml.kernel.org/r/20230712173916.168805-1-Liam.Howlett@oracle.com
Fixes: 120b116208a0 ("maple_tree: reorganize testing to restore module testing")
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
  Closes: https://lore.kernel.org/linux-mm/CAMuHMdV4T53fOw7VPoBgPR7fP6RYqf=CBhD_y_vOg53zZX_DnA@mail.gmail.com/
Tested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

maple_tree: add __init and __exit to test module

[ Upstream commit eaf9790d3bc6e157a2134c01c7d707a5a712fab1 ]

The test functions are not needed after the module is removed, so mark
them as such.  Add __exit to the module removal function.  Some other
variables have been marked as const static as well.

Link: https://lkml.kernel.org/r/20230518145544.1722059-20-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Suggested-by: Andrew Morton <akpm@linux-foundation.org>
Cc: David Binderman <dcb314@hotmail.com>
Cc: Peng Zhang <zhangpeng.00@bytedance.com>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Vernon Yang <vernon2gm@gmail.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stable-dep-of: 7a93c71a6714 ("maple_tree: fix 32 bit mas_next testing")
Signed-off-by: Sasha Levin <sashal@kernel.org>

test_maple_tree: test modifications while iterating

[ Upstream commit 5159d64b335401fa83f18c27e2267f1eafc41bd3 ]

Add a testcase to ensure the iterator detects bad states on modifications
and does what the user expects

Link: https://lkml.kernel.org/r/20230120162650.984577-5-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stable-dep-of: 7a93c71a6714 ("maple_tree: fix 32 bit mas_next testing")
Signed-off-by: Sasha Levin <sashal@kernel.org>

tracing/probes: Fix to record 0-length data_loc in fetch_store_string*() if fails

[ Upstream commit 797311bce5c2ac90b8d65e357603cfd410d36ebb ]

Fix to record 0-length data to data_loc in fetch_store_string*() if it fails
to get the string data.
Currently those expect that the data_loc is updated by store_trace_args() if
it returns the error code. However, that does not work correctly if the
argument is an array of strings. In that case, store_trace_args() only clears
the first entry of the array (which may have no error) and leaves other
entries. So it should be cleared by fetch_store_string*() itself.
Also, 'dyndata' and 'maxlen' in store_trace_args() should be updated
only if it is used (ret > 0 and argument is a dynamic data.)

Link: https://lore.kernel.org/all/168908496683.123124.4761206188794205601.stgit@devnote2/
Fixes: 40b53b771806 ("tracing: probeevent: Add array type support")
Cc: stable@vger.kernel.org
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Revert "tracing: Add "(fault)" name injection to kernel probes"

[ Upstream commit 4ed8f337dee32df71435689c19d22e4ee846e15a ]

This reverts commit 2e9906f84fc7c99388bb7123ade167250d50f1c0.

It was turned out that commit 2e9906f84fc7 ("tracing: Add "(fault)"
name injection to kernel probes") did not work correctly and probe
events still show just '(fault)' (instead of '"(fault)"'). Also,
current '(fault)' is more explicit that it faulted.

This also moves FAULT_STRING macro to trace.h so that synthetic
event can keep using it, and uses it in trace_probe.c too.

Link: https://lore.kernel.org/all/168908495772.123124.1250788051922100079.stgit@devnote2/
Link: https://lore.kernel.org/all/20230706230642.3793a593@rorschach.local.home/
Cc: stable@vger.kernel.org
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Tom Zanussi <zanussi@kernel.org>
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Stable-dep-of: 797311bce5c2 ("tracing/probes: Fix to record 0-length data_loc in fetch_store_string*() if fails")
Signed-off-by: Sasha Levin <sashal@kernel.org>

tracing: Allow synthetic events to pass around stacktraces

[ Upstream commit 00cf3d672a9dd409418647e9f98784c339c3ff63 ]

Allow a stacktrace from one event to be displayed by the end event of a
synthetic event. This is very useful when looking for the longest latency
of a sleep or something blocked on I/O.

 # cd /sys/kernel/tracing/
 # echo 's:block_lat pid_t pid; u64 delta; unsigned long[] stack;' > dynamic_events
 # echo 'hist:keys=next_pid:ts=common_timestamp.usecs,st=stacktrace  if prev_state == 1||prev_state == 2' > events/sched/sched_switch/trigger
 # echo 'hist:keys=prev_pid:delta=common_timestamp.usecs-$ts,s=$st:onmax($delta).trace(block_lat,prev_pid,$delta,$s)' >> events/sched/sched_switch/trigger

The above creates a "block_lat" synthetic event that take the stacktrace of
when a task schedules out in either the interruptible or uninterruptible
states, and on a new per process max $delta (the time it was scheduled
out), will print the process id and the stacktrace.

  # echo 1 > events/synthetic/block_lat/enable
  # cat trace
 #           TASK-PID     CPU#  |||||  TIMESTAMP  FUNCTION
 #              | |         |   |||||     |         |
    kworker/u16:0-767     [006] d..4.   560.645045: block_lat: pid=767 delta=66 stack=STACK:
 => __schedule
 => schedule
 => pipe_read
 => vfs_read
 => ksys_read
 => do_syscall_64
 => 0x966000aa

           <idle>-0       [003] d..4.   561.132117: block_lat: pid=0 delta=413787 stack=STACK:
 => __schedule
 => schedule
 => schedule_hrtimeout_range_clock
 => do_sys_poll
 => __x64_sys_poll
 => do_syscall_64
 => 0x966000aa

            <...>-153     [006] d..4.   562.068407: block_lat: pid=153 delta=54 stack=STACK:
 => __schedule
 => schedule
 => io_schedule
 => rq_qos_wait
 => wbt_wait
 => __rq_qos_throttle
 => blk_mq_submit_bio
 => submit_bio_noacct_nocheck
 => ext4_bio_write_page
 => mpage_submit_page
 => mpage_process_page_bufs
 => mpage_prepare_extent_to_map
 => ext4_do_writepages
 => ext4_writepages
 => do_writepages
 => __writeback_single_inode

Link: https://lkml.kernel.org/r/20230117152236.010941267@goodmis.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Tom Zanussi <zanussi@kernel.org>
Cc: Ross Zwisler <zwisler@google.com>
Cc: Ching-lin Yu <chinglinyu@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Stable-dep-of: 797311bce5c2 ("tracing/probes: Fix to record 0-length data_loc in fetch_store_string*() if fails")
Signed-off-by: Sasha Levin <sashal@kernel.org>

tracing/probes: Fix to avoid double count of the string length on the array

[ Upstream commit 66bcf65d6cf0ca6540e2341e88ee7ef02dbdda08 ]

If an array is specified with the ustring or symstr, the length of the
strings are accumlated on both of 'ret' and 'total', which means the
length is double counted.
Just set the length to the 'ret' value for avoiding double counting.

Link: https://lore.kernel.org/all/168908492917.123124.15076463491122036025.stgit@devnote2/
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/all/8819b154-2ba1-43c3-98a2-cbde20892023@moroto.mountain/
Fixes: 88903c464321 ("tracing/probe: Add ustring type for user-space string")
Cc: stable@vger.kernel.org
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tracing/probes: Add symstr type for dynamic events

[ Upstream commit b26a124cbfa80f42bfc4e63e1d5643ca98159d66 ]

Add 'symstr' type for storing the kernel symbol as a string data
instead of the symbol address. This allows us to filter the
events by wildcard symbol name.

e.g.
  # echo 'e:wqfunc workqueue.workqueue_execute_start symname=$function:symstr' >> dynamic_events
  # cat events/eprobes/wqfunc/format
  name: wqfunc
  ID: 2110
  format:
   field:unsigned short common_type; offset:0; size:2; signed:0;
   field:unsigned char common_flags; offset:2; size:1; signed:0;
   field:unsigned char common_preempt_count; offset:3; size:1; signed:0;
   field:int common_pid; offset:4; size:4; signed:1;

   field:__data_loc char[] symname; offset:8; size:4; signed:1;

  print fmt: " symname=\"%s\"", __get_str(symname)

Note that there is already 'symbol' type which just change the
print format (so it still stores the symbol address in the tracing
ring buffer.) On the other hand, 'symstr' type stores the actual
"symbol+offset/size" data as a string.

Link: https://lore.kernel.org/all/166679930847.1528100.4124308529180235965.stgit@devnote3/
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Stable-dep-of: 66bcf65d6cf0 ("tracing/probes: Fix to avoid double count of the string length on the array")
Signed-off-by: Sasha Levin <sashal@kernel.org>

mptcp: do not rely on implicit state check in mptcp_listen()

[ Upstream commit 0226436acf2495cde4b93e7400e5a87305c26054 ]

Since the blamed commit, closing the first subflow resets the first
subflow socket state to SS_UNCONNECTED.

The current mptcp listen implementation relies only on such
state to prevent touching not-fully-disconnected sockets.

Incoming mptcp fastclose (or paired endpoint removal) unconditionally
closes the first subflow.

All the above allows an incoming fastclose followed by a listen() call
to successfully race with a blocking recvmsg(), potentially causing the
latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window().

Address the issue explicitly checking the msk socket state in
mptcp_listen(). An alternative solution would be moving the first
subflow socket state update into mptcp_disconnect(), but in the long
term the first subflow socket should be removed: better avoid relaying
on it for internal consistency check.

Fixes: b29fcfb54cd7 ("mptcp: full disconnect implementation")
Cc: stable@vger.kernel.org
Reported-by: Christoph Paasch <cpaasch@apple.com>
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/414
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mptcp: introduce 'sk' to replace 'sock->sk' in mptcp_listen()

[ Upstream commit cfdcfeed6449d702825d249cb85346ecf56236fc ]

'sock->sk' is used frequently in mptcp_listen(). Therefore, we can
introduce the 'sk' and replace 'sock->sk' with it.

Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: 0226436acf24 ("mptcp: do not rely on implicit state check in mptcp_listen()")
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: errata: Mitigate Ampere1 erratum AC03_CPU_38 at stage-2

[ Upstream commit 6df696cd9bc1ceed0e92e36908f88bbd16d18255 ]

AmpereOne has an erratum in its implementation of FEAT_HAFDBS that
required disabling the feature on the design. This was done by reporting
the feature as not implemented in the ID register, although the
corresponding control bits were not actually RES0. This does not align
well with the requirements of the architecture, which mandates these
bits be RES0 if HAFDBS isn't implemented.

The kernel's use of stage-1 is unaffected, as the HA and HD bits are
only set if HAFDBS is detected in the ID register. KVM, on the other
hand, relies on the RES0 behavior at stage-2 to use the same value for
VTCR_EL2 on any cpu in the system. Mitigate the non-RES0 behavior by
leaving VTCR_EL2.HA clear on affected systems.

Cc: stable@vger.kernel.org
Cc: D Scott Phillips <scott@os.amperecomputing.com>
Cc: Darren Hart <darren@os.amperecomputing.com>
Acked-by: D Scott Phillips <scott@os.amperecomputing.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20230609220104.1836988-2-oliver.upton@linux.dev
Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
Signed-off-by: Sasha Levin <sashal@kernel.org>

KVM: arm64: Condition HW AF updates on config option

[ Upstream commit 1dfc3e905089a0bcada268fb5691a605655e0319 ]

As it currently stands, KVM makes use of FEAT_HAFDBS unconditionally.
Use of the feature in the rest of the kernel is guarded by an associated
Kconfig option.

Align KVM with the rest of the kernel and only enable VTCR_HA when
ARM64_HW_AFDBM is enabled. This can be helpful for testing changes to
the stage-2 access fault path on Armv8.1+ implementations.

Link: https://lore.kernel.org/r/20221202185156.696189-7-oliver.upton@linux.dev
Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
Stable-dep-of: 6df696cd9bc1 ("arm64: errata: Mitigate Ampere1 erratum AC03_CPU_38 at stage-2")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/ttm: never consider pinned BOs for eviction&swap

[ Upstream commit a2848d08742c8e8494675892c02c0d22acbe3cf8 ]

There is a small window where we have already incremented the pin count
but not yet moved the bo from the lru to the pinned list.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reported-by: Pelloux-Prayer, Pierre-Eric <Pierre-eric.Pelloux-prayer@amd.com>
Tested-by: Pelloux-Prayer, Pierre-Eric <Pierre-eric.Pelloux-prayer@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Link: https://patchwork.freedesktop.org/patch/msgid/20230707120826.3701-1-christian.koenig@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

tty: fix hang on tty device with no_room set

[ Upstream commit 4903fde8047a28299d1fc79c1a0dcc255e928f12 ]

It is possible to hang pty devices in this case, the reader was
blocking at epoll on master side, the writer was sleeping at
wait_woken inside n_tty_write on slave side, and the write buffer
on tty_port was full, we found that the reader and writer would
never be woken again and blocked forever.

The problem was caused by a race between reader and kworker:
n_tty_read(reader):  n_tty_receive_buf_common(kworker):
copy_from_read_buf()|
                    |room = N_TTY_BUF_SIZE - (ldata->read_head - tail)
                    |room <= 0
n_tty_kick_worker() |
                    |ldata->no_room = true

After writing to slave device, writer wakes up kworker to flush
data on tty_port to reader, and the kworker finds that reader
has no room to store data so room <= 0 is met. At this moment,
reader consumes all the data on reader buffer and calls
n_tty_kick_worker to check ldata->no_room which is false and
reader quits reading. Then kworker sets ldata->no_room=true
and quits too.

If write buffer is not full, writer will wake kworker to flush data
again after following writes, but if write buffer is full and writer
goes to sleep, kworker will never be woken again and tty device is
blocked.

This problem can be solved with a check for read buffer size inside
n_tty_receive_buf_common, if read buffer is empty and ldata->no_room
is true, a call to n_tty_kick_worker is necessary to keep flushing
data to reader.

Cc: <stable@vger.kernel.org>
Fixes: 42458f41d08f ("n_tty: Ensure reader restarts worker for next reader")
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Hui Li <caelli@tencent.com>
Message-ID: <1680749090-14106-1-git-send-email-caelli@tencent.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

n_tty: Rename tail to old_tail in n_tty_read()

[ Upstream commit 947d66b68f3c4e7cf8f3f3500807b9d2a0de28ce ]

The local tail variable in n_tty_read() is used for one purpose, it
keeps the old tail. Thus, rename it appropriately to improve code
readability.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
Link: https://lore.kernel.org/r/22b37499-ff9a-7fc1-f6e0-58411328d122@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Stable-dep-of: 4903fde8047a ("tty: fix hang on tty device with no_room set")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/ttm: Don't leak a resource on eviction error

[ Upstream commit e8188c461ee015ba0b9ab2fc82dbd5ebca5a5532 ]

On eviction errors other than -EMULTIHOP we were leaking a resource.
Fix.

v2:
- Avoid yet another goto (Andi Shyti)

Fixes: 403797925768 ("drm/ttm: Fix multihop assert on eviction.")
Cc: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Christian Koenig <christian.koenig@amd.com>
Cc: Huang Rui <ray.huang@amd.com>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v5.15+
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Nirmoy Das <nirmoy.das@intel.com> #v1
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230626091450.14757-4-thomas.hellstrom@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/ttm: Don't print error message if eviction was interrupted

[ Upstream commit 8ab3b0663e279ab550bc2c0b5d602960e8b94e02 ]

Avoid printing an error message if eviction was interrupted by,
for example, the user pressing CTRL-C. That may happen if eviction
is waiting for something, like for example a free batch-buffer.

Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230307144621.10748-6-thomas.hellstrom@linux.intel.com
Stable-dep-of: e8188c461ee0 ("drm/ttm: Don't leak a resource on eviction error")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Set minimum requirement for using PSR-SU on Phoenix

[ Upstream commit cd2e31a9ab93d13c412a36c6e26811e0f830985b ]

The same parade TCON issue can potentially happen on Phoenix, and the same
PSR resilience changes have been ported into the DMUB firmware.

Don't allow running PSR-SU unless on the newer firmware.

Cc: stable@vger.kernel.org
Cc: Sean Wang <sean.ns.wang@amd.com>
Cc: Marc Rossi <Marc.Rossi@amd.com>
Cc: Hamza Mahfooz <Hamza.Mahfooz@amd.com>
Cc: Tsung-hua (Ryan) Lin <Tsung-hua.Lin@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Set minimum requirement for using PSR-SU on Rembrandt

[ Upstream commit c35b6ea8f2ecfa9d775530b70d4e727869099a9c ]

A number of parade TCONs are causing system hangs when utilized with
older DMUB firmware and PSR-SU. Some changes have been introduced into
DMUB firmware to add resilience against these failures.

Don't allow running PSR-SU unless on the newer firmware.

Cc: stable@vger.kernel.org
Cc: Sean Wang <sean.ns.wang@amd.com>
Cc: Marc Rossi <Marc.Rossi@amd.com>
Cc: Hamza Mahfooz <Hamza.Mahfooz@amd.com>
Cc: Tsung-hua (Ryan) Lin <Tsung-hua.Lin@amd.com>
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2443
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stable-dep-of: cd2e31a9ab93 ("drm/amd/display: Set minimum requirement for using PSR-SU on Phoenix")
Signed-off-by: Sasha Levin <sashal@kernel.org>