pstore: Don't use semaphores in always-atomic-context code

commit 8126b1c73108bc691f5643df19071a59a69d0bc6 upstream.

pstore_dump() is *always* invoked in atomic context (nowadays in an RCU
read-side critical section, before that under a spinlock).
It doesn't make sense to try to use semaphores here.

This is mostly a revert of commit ea84b580b955 ("pstore: Convert buf_lock
to semaphore"), except that two parts aren't restored back exactly as they
were:

 - keep the lock initialization in pstore_register
 - in efi_pstore_write(), always set the "block" flag to false
 - omit "is_locked", that was unnecessary since
   commit 959217c84c27 ("pstore: Actually give up during locking failure")
 - fix the bailout message

The actual problem that the buggy commit was trying to address may have
been that the use of preemptible() in efi_pstore_write() was wrong - it
only looks at preempt_count() and the state of IRQs, but __rcu_read_lock()
doesn't touch either of those under CONFIG_PREEMPT_RCU.
(Sidenote: CONFIG_PREEMPT_RCU means that the scheduler can preempt tasks in
RCU read-side critical sections, but you're not allowed to actively
block/reschedule.)

Lockdep probably never caught the problem because it's very rare that you
actually hit the contended case, so lockdep always just sees the
down_trylock(), not the down_interruptible(), and so it can't tell that
there's a problem.

Fixes: ea84b580b955 ("pstore: Convert buf_lock to semaphore")
Cc: stable@vger.kernel.org
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20220314185953.2068993-1-jannh@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

carl9170: fix missing bit-wise or operator for tx_params

commit 02a95374b5eebdbd3b6413fd7ddec151d2ea75a1 upstream.

Currently tx_params is being re-assigned with a new value and the
previous setting IEEE80211_HT_MCS_TX_RX_DIFF is being overwritten.
The assignment operator is incorrect, the original intent was to
bit-wise or the value in. Fix this by replacing the = operator
with |= instead.

Kudos to Christian Lamparter for suggesting the correct fix.

Fixes: fe8ee9ad80b2 ("carl9170: mac80211 glue and command interface")
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Cc: <Stable@vger.kernel.org>
Acked-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20220125004406.344422-1-colin.i.king@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mgag200 fix memmapsl configuration in GCTL6 register

commit 028a73e10705af1ffd51f2537460f616dc58680e upstream.

On some servers with MGA G200_SE_A (rev 42), booting with Legacy BIOS,
the hardware hangs when using kdump and kexec into the kdump kernel.
This happens when the uncompress code tries to write "Decompressing Linux"
to the VGA Console.

It can be reproduced by writing to the VGA console (0xB8000) after
booting to graphic mode, it generates the following error:

kernel:NMI: PCI system error (SERR) for reason a0 on CPU 0.
kernel:Dazed and confused, but trying to continue

The root cause is the configuration of the MGA GCTL6 register

According to the GCTL6 register documentation:

bit 0 is gcgrmode:
    0: Enables alpha mode, and the character generator addressing system is
     activated.
    1: Enables graphics mode, and the character addressing system is not
     used.

bit 1 is chainodd even:
    0: The A0 signal of the memory address bus is used during system memory
     addressing.
    1: Allows A0 to be replaced by either the A16 signal of the system
     address (ifmemmapsl is ‘00’), or by the hpgoddev (MISC<5>, odd/even
     page select) field, described on page 3-294).

bit 3-2 are memmapsl:
    Memory map select bits 1 and 0. VGA.
    These bits select where the video memory is mapped, as shown below:
        00 => A0000h - BFFFFh
        01 => A0000h - AFFFFh
        10 => B0000h - B7FFFh
        11 => B8000h - BFFFFh

bit 7-4 are reserved.

Current code set it to 0x05 => memmapsl to b01 => 0xa0000 (graphic mode)
But on x86, the VGA console is at 0xb8000 (text mode)
In arch/x86/boot/compressed/misc.c debug strings are written to 0xb8000
As the driver doesn't use this mapping at 0xa0000, it is safe to set it to
0xb8000 instead, to avoid kernel hang on G200_SE_A rev42, with kexec/kdump.

Thus changing the value 0x05 to 0x0d

Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Lyude Paul <lyude@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220119102905.1194787-1-jfalempe@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ARM: dts: exynos: add missing HDMI supplies on SMDK5420

commit 453a24ded415f7fce0499c6b0a2c7b28f84911f2 upstream.

Add required VDD supplies to HDMI block on SMDK5420.  Without them, the
HDMI driver won't probe.  Because of lack of schematics, use same
supplies as on Arndale Octa and Odroid XU3 boards (voltage matches).

Cc: <stable@vger.kernel.org> # v3.15+
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com>
Link: https://lore.kernel.org/r/20220208171823.226211-3-krzysztof.kozlowski@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ARM: dts: exynos: add missing HDMI supplies on SMDK5250

commit 60a9914cb2061ba612a3f14f6ad329912b486360 upstream.

Add required VDD supplies to HDMI block on SMDK5250.  Without them, the
HDMI driver won't probe.  Because of lack of schematics, use same
supplies as on Arndale 5250 board (voltage matches).

Cc: <stable@vger.kernel.org> # v3.15+
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com>
Link: https://lore.kernel.org/r/20220208171823.226211-2-krzysztof.kozlowski@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ARM: dts: exynos: fix UART3 pins configuration in Exynos5250

commit 372d7027fed43c8570018e124cf78b89523a1f8e upstream.

The gpa1-4 pin was put twice in UART3 pin configuration of Exynos5250,
instead of proper pin gpa1-5.

Fixes: f8bfe2b050f3 ("ARM: dts: add pin state information in client nodes for Exynos5 platforms")
Cc: <stable@vger.kernel.org>
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com>
Link: https://lore.kernel.org/r/20211230195325.328220-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ARM: dts: at91: sama5d2: Fix PMERRLOC resource size

commit 0fb578a529ac7aca326a9fa475b4a6f58a756fda upstream.

PMERRLOC resource size was set to 0x100, which resulted in HSMC_ERRLOCx
register being truncated to offset x = 21, causing error correction to
fail if more than 22 bit errors and if 24 or 32 bit error correction
was supported.

Fixes: d9c41bf30cf8 ("ARM: dts: at91: Declare EBI/NAND controllers")
Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Cc: <stable@vger.kernel.org> # 4.13.x
Acked-by: Alexander Dahl <ada@thorsis.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20220111132301.906712-1-tudor.ambarus@microchip.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ARM: dts: at91: sama7g5: Remove unused properties in i2c nodes

commit cbb92a7717d2e1c512b7e81c6b22c7298b58a881 upstream.

The "atmel,use-dma-rx", "atmel,use-dma-rx" dt properties are not used by
the i2c-at91 driver, nor they are defined in the bindings file, thus remove
them.

Cc: stable@vger.kernel.org
Fixes: 7540629e2fc7 ("ARM: dts: at91: add sama7g5 SoC DT and sama7g5-ek")
Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Reviewed-by: Eugen Hristev <eugen.hristev@microchip.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20220302161854.32177-1-tudor.ambarus@microchip.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

rfkill: make new event layout opt-in

commit 54f586a9153201c6cff55e1f561990c78bd99aa7 upstream.

Again new complaints surfaced that we had broken the ABI here,
although previously all the userspace tools had agreed that it
was their mistake and fixed it. Yet now there are cases (e.g.
RHEL) that want to run old userspace with newer kernels, and
thus are broken.

Since this is a bit of a whack-a-mole thing, change the whole
extensibility scheme of rfkill to no longer just rely on the
message lengths, but instead require userspace to opt in via a
new ioctl to a given maximum event size that it is willing to
understand.

By default, set that to RFKILL_EVENT_SIZE_V1 (8), so that the
behaviour for userspace not calling the ioctl will look as if
it's just running on an older kernel.

Fixes: 14486c82612a ("rfkill: add a reason to the HW rfkill state")
Cc: stable@vger.kernel.org # 5.11+
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220316212749.16491491b270.Ifcb1950998330a596f29a2a162e00b7546a1d6d0@changeid
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

video: fbdev: atari: Atari 2 bpp (STe) palette bugfix

commit c8be5edbd36ceed2ff3d6b8f8e40643c3f396ea3 upstream.

The code to set the shifter STe palette registers has a long
standing operator precedence bug, manifesting as colors set
on a 2 bits per pixel frame buffer coming up with a distinctive
blue tint.

Add parentheses around the calculation of the per-color palette
data before shifting those into their respective bit field position.

This bug goes back a long way (2.4 days at the very least) so there
won't be a Fixes: tag.

Tested on ARAnyM as well on Falcon030 hardware.

Cc: stable@vger.kernel.org
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Link: https://lore.kernel.org/all/CAMuHMdU3ievhXxKR_xi_v3aumnYW7UNUO6qMdhgfyWTyVSsCkQ@mail.gmail.com
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Tested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

video: fbdev: sm712fb: Fix crash in smtcfb_read()

commit bd771cf5c4254511cc4abb88f3dab3bd58bdf8e8 upstream.

Zheyu Ma reported this crash in the sm712fb driver when reading
three bytes from the framebuffer:

 BUG: unable to handle page fault for address: ffffc90001ffffff
 RIP: 0010:smtcfb_read+0x230/0x3e0
 Call Trace:
  vfs_read+0x198/0xa00
  ? do_sys_openat2+0x27d/0x350
  ? __fget_light+0x54/0x340
  ksys_read+0xce/0x190
  do_syscall_64+0x43/0x90

Fix it by removing the open-coded endianess fixup-code and
by moving the pointer post decrement out the fb_readl() function.

Reported-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: Helge Deller <deller@gmx.de>
Tested-by: Zheyu Ma <zheyuma97@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

fbdev: Hot-unplug firmware fb devices on forced removal

commit 27599aacbaefcbf2af7b06b0029459bbf682000d upstream.

Hot-unplug all firmware-framebuffer devices as part of removing
them via remove_conflicting_framebuffers() et al. Releases all
memory regions to be acquired by native drivers.

Firmware, such as EFI, install a framebuffer while posting the
computer. After removing the firmware-framebuffer device from fbdev,
a native driver takes over the hardware and the firmware framebuffer
becomes invalid.

Firmware-framebuffer drivers, specifically simplefb, don't release
their device from Linux' device hierarchy. It still owns the firmware
framebuffer and blocks the native drivers from loading. This has been
observed in the vmwgfx driver. [1]

Initiating a device removal (i.e., hot unplug) as part of
remove_conflicting_framebuffers() removes the underlying device and
returns the memory range to the system.

[1] https://lore.kernel.org/dri-devel/20220117180359.18114-1-zack@kde.org/

v2:
* rename variable 'dev' to 'device' (Javier)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reported-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
CC: stable@vger.kernel.org # v5.11+
Link: https://patchwork.freedesktop.org/patch/msgid/20220125091222.21457-2-tzimmermann@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/edid: check basic audio support on CEA extension block

commit 5662abf6e21338be6d085d6375d3732ac6147fd2 upstream.

Tag code stored in bit7:5 for CTA block byte[3] is not the same as
CEA extension block definition. Only check CEA block has
basic audio support.

v3: update commit message.

Cc: stable@vger.kernel.org
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Shawn C Lee <shawn.c.lee@intel.com>
Cc: intel-gfx <intel-gfx@lists.freedesktop.org>
Signed-off-by: Cooper Chiou <cooper.chiou@intel.com>
Signed-off-by: Lee Shawn C <shawn.c.lee@intel.com>
Fixes: e28ad544f462 ("drm/edid: parse CEA blocks embedded in DisplayID")
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220324061218.32739-1-shawn.c.lee@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

block: don't merge across cgroup boundaries if blkcg is enabled

commit 6b2b04590b51aa4cf395fcd185ce439cab5961dc upstream.

blk-iocost and iolatency are cgroup aware rq-qos policies but they didn't
disable merges across different cgroups. This obviously can lead to
accounting and control errors but more importantly to priority inversions -
e.g. an IO which belongs to a higher priority cgroup or IO class may end up
getting throttled incorrectly because it gets merged to an IO issued from a
low priority cgroup.

Fix it by adding blk_cgroup_mergeable() which is called from merge paths and
rejects cross-cgroup and cross-issue_as_root merges.

Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: d70675121546 ("block: introduce blk-iolatency io controller")
Cc: stable@vger.kernel.org # v4.19+
Cc: Josef Bacik <jbacik@fb.com>
Link: https://lore.kernel.org/r/Yi/eE/6zFNyWJ+qd@slm.duckdns.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

block: limit request dispatch loop duration

commit 572299f03afd676dd4e20669cdaf5ed0fe1379d4 upstream.

When IO requests are made continuously and the target block device
handles requests faster than request arrival, the request dispatch loop
keeps on repeating to dispatch the arriving requests very long time,
more than a minute. Since the loop runs as a workqueue worker task, the
very long loop duration triggers workqueue watchdog timeout and BUG [1].

To avoid the very long loop duration, break the loop periodically. When
opportunity to dispatch requests still exists, check need_resched(). If
need_resched() returns true, the dispatch loop already consumed its time
slice, then reschedule the dispatch work and break the loop. With heavy
IO load, need_resched() does not return true for 20~30 seconds. To cover
such case, check time spent in the dispatch loop with jiffies. If more
than 1 second is spent, reschedule the dispatch work and break the loop.

[1]

[  609.691437] BUG: workqueue lockup - pool cpus=10 node=1 flags=0x0 nice=-20 stuck for 35s!
[  609.701820] Showing busy workqueues and worker pools:
[  609.707915] workqueue events: flags=0x0
[  609.712615]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  609.712626]     pending: drm_fb_helper_damage_work [drm_kms_helper]
[  609.712687] workqueue events_freezable: flags=0x4
[  609.732943]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  609.732952]     pending: pci_pme_list_scan
[  609.732968] workqueue events_power_efficient: flags=0x80
[  609.751947]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  609.751955]     pending: neigh_managed_work
[  609.752018] workqueue kblockd: flags=0x18
[  609.769480]   pwq 21: cpus=10 node=1 flags=0x0 nice=-20 active=3/256 refcnt=4
[  609.769488]     in-flight: 1020:blk_mq_run_work_fn
[  609.769498]     pending: blk_mq_timeout_work, blk_mq_run_work_fn
[  609.769744] pool 21: cpus=10 node=1 flags=0x0 nice=-20 hung=35s workers=2 idle: 67
[  639.899730] BUG: workqueue lockup - pool cpus=10 node=1 flags=0x0 nice=-20 stuck for 66s!
[  639.909513] Showing busy workqueues and worker pools:
[  639.915404] workqueue events: flags=0x0
[  639.920197]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  639.920215]     pending: drm_fb_helper_damage_work [drm_kms_helper]
[  639.920365] workqueue kblockd: flags=0x18
[  639.939932]   pwq 21: cpus=10 node=1 flags=0x0 nice=-20 active=3/256 refcnt=4
[  639.939942]     in-flight: 1020:blk_mq_run_work_fn
[  639.939955]     pending: blk_mq_timeout_work, blk_mq_run_work_fn
[  639.940212] pool 21: cpus=10 node=1 flags=0x0 nice=-20 hung=66s workers=2 idle: 67

Fixes: 6e6fcbc27e778 ("blk-mq: support batching dispatch in case of io")
Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com>
Cc: stable@vger.kernel.org # v5.10+
Link: https://lore.kernel.org/linux-block/20220310091649.zypaem5lkyfadymg@shindev/
Link: https://lore.kernel.org/r/20220318022641.133484-1-shinichiro.kawasaki@wdc.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mailbox: tegra-hsp: Flush whole channel

commit 60de2d2dc284e0dd1c2c897d08625bde24ef3454 upstream.

The txdone can re-fill the mailbox. Keep polling the mailbox during the
flush until all the messages have been delivered.

This fixes an issue with the Tegra Combined UART (TCU) where output can
get truncated under high traffic load.

Signed-off-by: Pekka Pessi <ppessi@nvidia.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Fixes: 91b1b1c3da8a ("mailbox: tegra-hsp: Add support for shared mailboxes")
Cc: stable@vger.kernel.org
Signed-off-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

landlock: Use square brackets around "landlock-ruleset"

commit aea0b9f2486da8497f35c7114b764bf55e17c7ea upstream.

Make the name of the anon inode fd "[landlock-ruleset]" instead of
"landlock-ruleset". This is minor but most anon inode fds already
carry square brackets around their name:

    [eventfd]
    [eventpoll]
    [fanotify]
    [fscontext]
    [io_uring]
    [pidfd]
    [signalfd]
    [timerfd]
    [userfaultfd]

For the sake of consistency lets do the same for the landlock-ruleset anon
inode fd that comes with landlock. We did the same in
1cdc415f1083 ("uapi, fsopen: use square brackets around "fscontext" [ver #2]")
for the new mount api.

Cc: linux-security-module@vger.kernel.org
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/r/20211011133704.1704369-1-brauner@kernel.org
Cc: stable@vger.kernel.org
Signed-off-by: Mickaël Salaün <mic@linux.microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

samples/landlock: Fix path_list memory leak

commit 66b513b7c64a7290c1fbb88e657f7cece992e131 upstream.

Clang static analysis reports this error

sandboxer.c:134:8: warning: Potential leak of memory
  pointed to by 'path_list'
        ret = 0;
              ^
path_list is allocated in parse_path() but never freed.

Signed-off-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/20210428213852.2874324-1-trix@redhat.com
Cc: stable@vger.kernel.org
Signed-off-by: Mickaël Salaün <mic@linux.microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers: hamradio: 6pack: fix UAF bug caused by mod_timer()

commit efe4186e6a1b54bf38b9e05450d43b0da1fd7739 upstream.

When a 6pack device is detaching, the sixpack_close() will act to cleanup
necessary resources. Although del_timer_sync() in sixpack_close()
won't return if there is an active timer, one could use mod_timer() in
sp_xmit_on_air() to wake up timer again by calling userspace syscall such
as ax25_sendmsg(), ax25_connect() and ax25_ioctl().

This unexpected waked handler, sp_xmit_on_air(), realizes nothing about
the undergoing cleanup and may still call pty_write() to use driver layer
resources that have already been released.

One of the possible race conditions is shown below:

      (USE)                      |      (FREE)
ax25_sendmsg()                   |
 ax25_queue_xmit()               |
  ...                            |
  sp_xmit()                      |
   sp_encaps()                   | sixpack_close()
    sp_xmit_on_air()             |  del_timer_sync(&sp->tx_t)
     mod_timer(&sp->tx_t,...)    |  ...
                                 |  unregister_netdev()
                                 |  ...
     (wait a while)              | tty_release()
                                 |  tty_release_struct()
                                 |   release_tty()
    sp_xmit_on_air()             |    tty_kref_put(tty_struct) //FREE
     pty_write(tty_struct) //USE |    ...

The corresponding fail log is shown below:
===============================================================
BUG: KASAN: use-after-free in __run_timers.part.0+0x170/0x470
Write of size 8 at addr ffff88800a652ab8 by task swapper/2/0
...
Call Trace:
  ...
  queue_work_on+0x3f/0x50
  pty_write+0xcd/0xe0pty_write+0xcd/0xe0
  sp_xmit_on_air+0xb2/0x1f0
  call_timer_fn+0x28/0x150
  __run_timers.part.0+0x3c2/0x470
  run_timer_softirq+0x3b/0x80
  __do_softirq+0xf1/0x380
  ...

This patch reorders the del_timer_sync() after the unregister_netdev()
to avoid UAF bugs. Because the unregister_netdev() is well synchronized,
it flushs out any pending queues, waits the refcount of net_device
decreases to zero and removes net_device from kernel. There is not any
running routines after executing unregister_netdev(). Therefore, we could
not arouse timer from userspace again.

Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
Reviewed-by: Lin Ma <linma@zju.edu.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ext4: make mb_optimize_scan performance mount option work with extents

commit 077d0c2c78df6f7260cdd015a991327efa44d8ad upstream.

Currently mb_optimize_scan scan feature which improves filesystem
performance heavily (when FS is fragmented), seems to be not working
with files with extents (ext4 by default has files with extents).

This patch fixes that and makes mb_optimize_scan feature work
for files with extents.

Below are some performance numbers obtained when allocating a 10M and 100M
file with and w/o this patch on a filesytem with no 1M contiguous block.

<perf numbers>
===============
Workload: dd if=/dev/urandom of=test conv=fsync bs=1M count=10/100

Time taken
=====================================================
no.     Size   without-patch     with-patch    Diff(%)
1       10M      0m8.401s         0m5.623s     33.06%
2       100M     1m40.465s        1m14.737s    25.6%

<debug stats>
=============
w/o patch:
  mballoc:
    reqs: 17056
    success: 11407
    groups_scanned: 13643
    cr0_stats:
            hits: 37
            groups_considered: 9472
            useless_loops: 36
            bad_suggestions: 0
    cr1_stats:
            hits: 11418
            groups_considered: 908560
            useless_loops: 1894
            bad_suggestions: 0
    cr2_stats:
            hits: 1873
            groups_considered: 6913
            useless_loops: 21
    cr3_stats:
            hits: 21
            groups_considered: 5040
            useless_loops: 21
    extents_scanned: 417364
            goal_hits: 3707
            2^n_hits: 37
            breaks: 1873
            lost: 0
    buddies_generated: 239/240
    buddies_time_used: 651080
    preallocated: 705
    discarded: 478

with patch:
  mballoc:
    reqs: 12768
    success: 11305
    groups_scanned: 12768
    cr0_stats:
            hits: 1
            groups_considered: 18
            useless_loops: 0
            bad_suggestions: 0
    cr1_stats:
            hits: 5829
            groups_considered: 50626
            useless_loops: 0
            bad_suggestions: 0
    cr2_stats:
            hits: 6938
            groups_considered: 580363
            useless_loops: 0
    cr3_stats:
            hits: 0
            groups_considered: 0
            useless_loops: 0
    extents_scanned: 309059
            goal_hits: 0
            2^n_hits: 1
            breaks: 1463
            lost: 0
    buddies_generated: 239/240
    buddies_time_used: 791392
    preallocated: 673
    discarded: 446

Fixes: 196e402 (ext4: improve cr 0 / cr 1 group scanning)
Cc: stable@kernel.org
Reported-by: Geetika Moolchandani <Geetika.Moolchandani1@ibm.com>
Reported-by: Nageswara R Sastry <rnsastry@linux.ibm.com>
Suggested-by: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Link: https://lore.kernel.org/r/fc9a48f7f8dcfc83891a8b21f6dd8cdf056ed810.1646732698.git.ojaswin@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ext4: fix fs corruption when tring to remove a non-empty directory with IO error

commit 7aab5c84a0f6ec2290e2ba4a6b245178b1bf949a upstream.

We inject IO error when rmdir non empty direcory, then got issue as follows:
step1: mkfs.ext4 -F /dev/sda
step2: mount /dev/sda  test
step3: cd test
step4: mkdir -p 1/2
step5: rmdir 1
[  110.920551] ext4_empty_dir: inject fault
[  110.921926] EXT4-fs warning (device sda): ext4_rmdir:3113: inode #12:
comm rmdir: empty directory '1' has too many links (3)
step6: cd ..
step7: umount test
step8: fsck.ext4 -f /dev/sda
e2fsck 1.42.9 (28-Dec-2013)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Entry '..' in .../??? (13) has deleted/unused inode 12.  Clear<y>? yes
Pass 3: Checking directory connectivity
Unconnected directory inode 13 (...)
Connect to /lost+found<y>? yes
Pass 4: Checking reference counts
Inode 13 ref count is 3, should be 2.  Fix<y>? yes
Pass 5: Checking group summary information

/dev/sda: ***** FILE SYSTEM WAS MODIFIED *****
/dev/sda: 12/131072 files (0.0% non-contiguous), 26157/524288 blocks

ext4_rmdir
if (!ext4_empty_dir(inode))
goto end_rmdir;
ext4_empty_dir
bh = ext4_read_dirblock(inode, 0, DIRENT_HTREE);
if (IS_ERR(bh))
return true;
Now if read directory block failed, 'ext4_empty_dir' will return true, assume
directory is empty. Obviously, it will lead to above issue.
To solve this issue, if read directory block failed 'ext4_empty_dir' just
return false. To avoid making things worse when file system is already
corrupted, 'ext4_empty_dir' also return false.

Signed-off-by: Ye Bin <yebin10@huawei.com>
Cc: stable@kernel.org
Link: https://lore.kernel.org/r/20220228024815.3952506-1-yebin10@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ext4: fix ext4_fc_stats trace point

commit 7af1974af0a9ba8a8ed2e3e947d87dd4d9a78d27 upstream.

ftrace's __print_symbolic() requires that any enum values used in the
symbol to string translation table be wrapped in a TRACE_DEFINE_ENUM
so that the enum value can be decoded from the ftrace ring buffer by
user space tooling.

This patch also fixes few other problems found in this trace point.
e.g. dereferencing structures in TP_printk which should not be done
at any cost.

Also to avoid checkpatch warnings, this patch removes those
whitespaces/tab stops issues.

Cc: stable@kernel.org
Fixes: aa75f4d3daae ("ext4: main fast-commit commit path")
Reported-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Reviewed-by: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
Link: https://lore.kernel.org/r/b4b9691414c35c62e570b723e661c80674169f9a.1647057583.git.riteshh@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coredump: Also dump first pages of non-executable ELF libraries

commit 84158b7f6a0624b81800b4e7c90f7fb7fdecf66c upstream.

When I rewrote the VMA dumping logic for coredumps, I changed it to
recognize ELF library mappings based on the file being executable instead
of the mapping having an ELF header. But turns out, distros ship many ELF
libraries as non-executable, so the heuristic goes wrong...

Restore the old behavior where FILTER(ELF_HEADERS) dumps the first page of
any offset-0 readable mapping that starts with the ELF magic.

This fix is technically layer-breaking a bit, because it checks for
something ELF-specific in fs/coredump.c; but since we probably want to
share this between standard ELF and FDPIC ELF anyway, I guess it's fine?
And this also keeps the change small for backporting.

Cc: stable@vger.kernel.org
Fixes: 429a22e776a2 ("coredump: rework elf/elf_fdpic vma_dump_size() into common helper")
Reported-by: Bill Messmer <wmessmer@microsoft.com>
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20220126025739.2014888-1-jannh@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ACPI: properties: Consistently return -ENOENT if there are no more references

commit babc92da5928f81af951663fc436997352e02d3a upstream.

__acpi_node_get_property_reference() is documented to return -ENOENT if
the caller requests a property reference at an index that does not exist,
not -EINVAL which it actually does.

Fix this by returning -ENOENT consistenly, independently of whether the
property value is a plain reference or a package.

Fixes: c343bc2ce2c6 ("ACPI: properties: Align return codes of __acpi_node_get_property_reference()")
Cc: 4.14+ <stable@vger.kernel.org> # 4.14+
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Revert "ACPI: Pass the same capabilities to the _OSC regardless of the query flag"

commit 2ca8e6285250c07a2e5a22ecbfd59b5a4ef73484 upstream.

Revert commit 159d8c274fd9 ("ACPI: Pass the same capabilities to the
_OSC regardless of the query flag") which caused legitimate usage
scenarios (when the platform firmware does not want the OS to control
certain platform features controlled by the system bus scope _OSC) to
break and was misguided by some misleading language in the _OSC
definition in the ACPI specification (in particular, Section 6.2.11.1.3
"Sequence of _OSC Calls" that contradicts other perts of the _OSC
definition).

Link: https://lore.kernel.org/linux-acpi/CAJZ5v0iStA0JmO0H3z+VgQsVuQONVjKPpw0F5HKfiq=Gb6B5yw@mail.gmail.com
Reported-by: Mario Limonciello <Mario.Limonciello@amd.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Mario Limonciello <mario.limonciello@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ASoC: SOF: Intel: Fix NULL ptr dereference when ENOMEM

commit b7fb0ae09009d076964afe4c1a2bde1ee2bd88a9 upstream.

Do not call snd_dma_free_pages() when snd_dma_alloc_pages() returns
-ENOMEM because it leads to a NULL pointer dereference bug.

The dmesg says:

  [ T1387] sof-audio-pci-intel-tgl 0000:00:1f.3: error: memory alloc failed: -12
  [ T1387] BUG: kernel NULL pointer dereference, address: 0000000000000000
  [ T1387] #PF: supervisor read access in kernel mode
  [ T1387] #PF: error_code(0x0000) - not-present page
  [ T1387] PGD 0 P4D 0
  [ T1387] Oops: 0000 [#1] PREEMPT SMP NOPTI
  [ T1387] CPU: 6 PID: 1387 Comm: alsa-sink-HDA A Tainted: G        W         5.17.0-rc4-superb-owl-00055-g80d47f5de5e3
  [ T1387] Hardware name: HP HP Laptop 14s-dq2xxx/87FD, BIOS F.15 09/15/2021
  [ T1387] RIP: 0010:dma_free_noncontiguous+0x37/0x80
  [ T1387] Code: [... snip ...]
  [ T1387] RSP: 0000:ffffc90002b87770 EFLAGS: 00010246
  [ T1387] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
  [ T1387] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff888101db30d0
  [ T1387] RBP: 00000000fffffff4 R08: 0000000000000000 R09: 0000000000000000
  [ T1387] R10: 0000000000000000 R11: ffffc90002b874d0 R12: 0000000000000001
  [ T1387] R13: 0000000000058000 R14: ffff888105260c68 R15: ffff888105260828
  [ T1387] FS:  00007f42e2ffd640(0000) GS:ffff888466b80000(0000) knlGS:0000000000000000
  [ T1387] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  [ T1387] CR2: 0000000000000000 CR3: 000000014acf0003 CR4: 0000000000770ee0
  [ T1387] PKRU: 55555554
  [ T1387] Call Trace:
  [ T1387]  <TASK>
  [ T1387]  cl_stream_prepare+0x10a/0x120 [snd_sof_intel_hda_common 146addf995b9279ae7f509621078cccbe4f875e1]
  [... snip ...]
  [ T1387]  </TASK>

Cc: Daniel Baluta <daniel.baluta@nxp.com>
Cc: Jaroslav Kysela <perex@perex.cz>
Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Cc: Keyon Jie <yang.jie@linux.intel.com>
Cc: Liam Girdwood <lgirdwood@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Rander Wang <rander.wang@intel.com>
Cc: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Cc: Takashi Iwai <tiwai@suse.com>
Cc: sound-open-firmware@alsa-project.org
Cc: alsa-devel@alsa-project.org
Cc: linux-kernel@vger.kernel.org
Cc: stable@vger.kernel.org # v5.2+
Fixes: d16046ffa6de040bf580a64d5f4d0aa18258a854 ("ASoC: SOF: Intel: Add Intel specific HDA firmware loader")
Link: https://lore.kernel.org/lkml/20220224145124.15985-1-ammarfaizi2@gnuweeb.org/
Link: https://lore.kernel.org/lkml/20220224180850.34592-1-ammarfaizi2@gnuweeb.org/
Link: https://lore.kernel.org/lkml/20220224182818.40301-1-ammarfaizi2@gnuweeb.org/
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Ammar Faizi <ammarfaizi2@gnuweeb.org>
Link: https://lore.kernel.org/r/20220224185836.44907-1-ammarfaizi2@gnuweeb.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: dts: ti: k3-am64: Fix gic-v3 compatible regs

commit de60edf1be3d42d4a1b303b41c7c53b2f865726e upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility,
Cortex-A53 is free to implement the CPU interface as long as it
communicates with the GIC using the stream protocol. This requires
that the SoC integration mark out the PERIPHBASE[1] as reserved area
within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based
on [3]. Without doing this, systems like kvm will not function with
gic-v2 emulation.

[1] https://developer.arm.com/documentation/ddi0500/e/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/87k0e0tirw.wl-maz@kernel.org/
[3] https://developer.arm.com/documentation/ddi0500/e/generic-interrupt-controller-cpu-interface/gic-programmers-model/memory-map

Cc: stable@vger.kernel.org
Fixes: 8abae9389bdb ("arm64: dts: ti: Add support for AM642 SoC")
Reported-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220215201008.15235-5-nm@ti.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: dts: ti: k3-j7200: Fix gic-v3 compatible regs

commit 1a307cc299430dd7139d351a3b8941f493dfa885 upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility,
Cortex-A72 is free to implement the CPU interface as long as it
communicates with the GIC using the stream protocol. This requires
that the SoC integration mark out the PERIPHBASE[1] as reserved area
within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based
on [3]. Without doing this, systems like kvm will not function with
gic-v2 emulation.

[1] https://developer.arm.com/documentation/100095/0002/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/87k0e0tirw.wl-maz@kernel.org/
[3] https://developer.arm.com/documentation/100095/0002/way1382452674438

Cc: stable@vger.kernel.org
Fixes: d361ed88455f ("arm64: dts: ti: Add support for J7200 SoC")
Reported-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220215201008.15235-4-nm@ti.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: dts: ti: k3-j721e: Fix gic-v3 compatible regs

commit a06ed27f3bc63ab9e10007dc0118d910908eb045 upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility,
Cortex-A72 is free to implement the CPU interface as long as it
communicates with the GIC using the stream protocol. This requires
that the SoC integration mark out the PERIPHBASE[1] as reserved area
within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based
on [3]. Without doing this, systems like kvm will not function with
gic-v2 emulation.

[1] https://developer.arm.com/documentation/100095/0002/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/87k0e0tirw.wl-maz@kernel.org/
[3] https://developer.arm.com/documentation/100095/0002/way1382452674438

Cc: stable@vger.kernel.org # 5.10+
Fixes: 2d87061e70de ("arm64: dts: ti: Add Support for J721E SoC")
Reported-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220215201008.15235-3-nm@ti.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: dts: ti: k3-am65: Fix gic-v3 compatible regs

commit 8cae268b70f387ff9e697ccd62fb2384079124e7 upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility,
Cortex-A53 is free to implement the CPU interface as long as it
communicates with the GIC using the stream protocol. This requires
that the SoC integration mark out the PERIPHBASE[1] as reserved area
within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based
on [3]. Without doing this, systems like kvm will not function with
gic-v2 emulation.

[1] https://developer.arm.com/documentation/ddi0500/e/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/87k0e0tirw.wl-maz@kernel.org/
[3] https://developer.arm.com/documentation/ddi0500/e/generic-interrupt-controller-cpu-interface/gic-programmers-model/memory-map

Cc: stable@vger.kernel.org # 5.10+
Fixes: ea47eed33a3f ("arm64: dts: ti: Add Support for AM654 SoC")
Reported-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220215201008.15235-2-nm@ti.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: dts: qcom: sm8250: Fix MSI IRQ for PCIe1 and PCIe2

commit 1b7101e8124b450f2d6a35591e9cbb478c143ace upstream.

Fix the MSI IRQ used for PCIe instances 1 and 2.

Cc: stable@vger.kernel.org
Fixes: e53bdfc00977 ("arm64: dts: qcom: sm8250: Add PCIe support")
Reported-by: Jordan Crouse <jordan@cosmicpenguin.net>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220112035556.5108-1-manivannan.sadhasivam@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: Do not defer reserve_crashkernel() for platforms with no DMA memory zones

commit 031495635b4668f94e964e037ca93d0d38bfde58 upstream.

The following patches resulted in deferring crash kernel reservation to
mem_init(), mainly aimed at platforms with DMA memory zones (no IOMMU),
in particular Raspberry Pi 4.

commit 1a8e1cef7603 ("arm64: use both ZONE_DMA and ZONE_DMA32")
commit 8424ecdde7df ("arm64: mm: Set ZONE_DMA size based on devicetree's dma-ranges")
commit 0a30c53573b0 ("arm64: mm: Move reserve_crashkernel() into mem_init()")
commit 2687275a5843 ("arm64: Force NO_BLOCK_MAPPINGS if crashkernel reservation is required")

Above changes introduced boot slowdown due to linear map creation for
all the memory banks with NO_BLOCK_MAPPINGS, see discussion[1].  The proposed
changes restore crash kernel reservation to earlier behavior thus avoids
slow boot, particularly for platforms with IOMMU (no DMA memory zones).

Tested changes to confirm no ~150ms boot slowdown on our SoC with IOMMU
and 8GB memory.  Also tested with ZONE_DMA and/or ZONE_DMA32 configs to confirm
no regression to deferring scheme of crash kernel memory reservation.
In both cases successfully collected kernel crash dump.

[1] https://lore.kernel.org/all/9436d033-579b-55fa-9b00-6f4b661c2dd7@linux.microsoft.com/

Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com>
Cc: stable@vger.kernel.org
Reviewed-by: Pasha Tatashin <pasha.tatashin@soleen.com>
Link: https://lore.kernel.org/r/1646242689-20744-1-git-send-email-vijayb@linux.microsoft.com
[will: Add #ifdef CONFIG_KEXEC_CORE guards to fix 'crashk_res' references in allnoconfig build]
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: signal: nofpsimd: Do not allocate fp/simd context when not available

commit 0a32c88ddb9af30e8a16d41d7b9b824c27d29459 upstream.

Commit 6d502b6ba1b2 ("arm64: signal: nofpsimd: Handle fp/simd context for
signal frames") introduced saving the fp/simd context for signal handling
only when support is available. But setup_sigframe_layout() always
reserves memory for fp/simd context. The additional memory is not touched
because preserve_fpsimd_context() is not called and thus the magic is
invalid.

This may lead to an error when parse_user_sigframe() checks the fp/simd
area and does not find a valid magic number.

Signed-off-by: David Engraf <david.engraf@sysgo.com>
Reviwed-by: Mark Brown <broonie@kernel.org>
Fixes: 6d502b6ba1b267b3 ("arm64: signal: nofpsimd: Handle fp/simd context for signal frames")
Cc: <stable@vger.kernel.org> # 5.6.x
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20220225104008.820289-1-david.engraf@sysgo.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

PCI: fu740: Force 2.5GT/s for initial device probe

commit a382c757ec5ef83137a86125f43a4c43dc2ab50b upstream.

The fu740 PCIe core does not probe any devices on the SiFive Unmatched
board without this fix (or having U-Boot explicitly start the PCIe via
either boot-script or user command). The fix is to start the link at
2.5GT/s speeds and once the link is up then change the maximum speed back
to the default.

The U-Boot driver claims to set the link-speed to 2.5GT/s to get the probe
to work (and U-Boot does print link up at 2.5GT/s) in the following code:
https://source.denx.de/u-boot/u-boot/-/blob/master/drivers/pci/pcie_dw_sifive.c?id=v2022.01#L271

Link: https://lore.kernel.org/r/20220318152430.526320-1-ben.dooks@codethink.co.uk
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Dimitri John Ledkov <dimitri.ledkov@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

powerpc/kvm: Fix kvm_use_magic_page

commit 0c8eb2884a42d992c7726539328b7d3568f22143 upstream.

When switching from __get_user to fault_in_pages_readable, commit
9f9eae5ce717 broke kvm_use_magic_page: like __get_user,
fault_in_pages_readable returns 0 on success.

Fixes: 9f9eae5ce717 ("powerpc/kvm: Prefer fault_in_pages_readable function")
Cc: stable@vger.kernel.org # v4.18+
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drbd: fix potential silent data corruption

commit f4329d1f848ac35757d9cc5487669d19dfc5979c upstream.

Scenario:
---------

bio chain generated by blk_queue_split().
Some split bio fails and propagates its error status to the "parent" bio.
But then the (last part of the) parent bio itself completes without error.

We would clobber the already recorded error status with BLK_STS_OK,
causing silent data corruption.

Reproducer:
-----------

How to trigger this in the real world within seconds:

DRBD on top of degraded parity raid,
small stripe_cache_size, large read_ahead setting.
Drop page cache (sysctl vm.drop_caches=1, fadvise "DONTNEED",
umount and mount again, "reboot").

Cause significant read ahead.

Large read ahead request is split by blk_queue_split().
Parts of the read ahead that are already in the stripe cache,
or find an available stripe cache to use, can be serviced.
Parts of the read ahead that would need "too much work",
would need to wait for a "stripe_head" to become available,
are rejected immediately.

For larger read ahead requests that are split in many pieces, it is very
likely that some "splits" will be serviced, but then the stripe cache is
exhausted/busy, and the remaining ones will be rejected.

Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Christoph Böhmwalder <christoph.boehmwalder@linbit.com>
Cc: <stable@vger.kernel.org> # 4.13.x
Link: https://lore.kernel.org/r/20220330185551.3553196-1-christoph.boehmwalder@linbit.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

tracing: Have trace event string test handle zero length strings

commit eca344a7362e0f34f179298fd8366bcd556eede1 upstream.

If a trace event has in its TP_printk():

 "%*.s", len, len ? __get_str(string) : NULL

It is perfectly valid if len is zero and passing in the NULL.
Unfortunately, the runtime string check at time of reading the trace sees
the NULL and flags it as a bad string and produces a WARN_ON().

Handle this case by passing into the test function if the format has an
asterisk (star) and if so, if the length is zero, then mark it as safe.

Link: https://lore.kernel.org/all/YjsWzuw5FbWPrdqq@bfoster/
Cc: stable@vger.kernel.org
Reported-by: Brian Foster <bfoster@redhat.com>
Tested-by: Brian Foster <bfoster@redhat.com>
Fixes: 9a6944fee68e2 ("tracing: Add a verifier to check string pointers for trace events")
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dm integrity: set journal entry unused when shrinking device

commit cc09e8a9dec4f0e8299e80a7a2a8e6f54164a10b upstream.

Commit f6f72f32c22c ("dm integrity: don't replay journal data past the
end of the device") skips journal replay if the target sector points
beyond the end of the device. Unfortunatelly, it doesn't set the
journal entry unused, which resulted in this BUG being triggered:
BUG_ON(!journal_entry_is_unused(je))

Fix this by calling journal_entry_set_unused() for this case.

Fixes: f6f72f32c22c ("dm integrity: don't replay journal data past the end of the device")
Cc: stable@vger.kernel.org # v5.7+
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Milan Broz <gmazyland@gmail.com>
[snitzer: revised header]
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dm: fix double accounting of flush with data

commit 8d394bc4adf588ca4a0650745167cb83f86c18c9 upstream.

DM handles a flush with data by first issuing an empty flush and then
once it completes the REQ_PREFLUSH flag is removed and the payload is
issued.  The problem fixed by this commit is that both the empty flush
bio and the data payload will account the full extent of the data
payload.

Fix this by factoring out dm_io_acct() and having it wrap all IO
accounting to set the size of  bio with REQ_PREFLUSH to 0, account the
IO, and then restore the original size.

Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dm: interlock pending dm_io and dm_wait_for_bios_completion

commit 9f6dc633761006f974701d4c88da71ab68670749 upstream.

Commit d208b89401e0 ("dm: fix mempool NULL pointer race when
completing IO") didn't go far enough.

When bio_end_io_acct ends the count of in-flight I/Os may reach zero
and the DM device may be suspended. There is a possibility that the
suspend races with dm_stats_account_io.

Fix this by adding percpu "pending_io" counters to track outstanding
dm_io. Move kicking of suspend queue to dm_io_dec_pending(). Also,
rename md_in_flight_bios() to dm_in_flight_bios() and update it to
iterate all pending_io counters.

Fixes: d208b89401e0 ("dm: fix mempool NULL pointer race when completing IO")
Cc: stable@vger.kernel.org
Co-developed-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dm: fix use-after-free in dm_cleanup_zoned_dev()

commit 588b7f5df0cb64f281290c7672470c006abe7160 upstream.

dm_cleanup_zoned_dev() uses queue, so it must be called
before blk_cleanup_disk() starts its killing:

blk_cleanup_disk->blk_cleanup_queue()->kobject_put()->blk_release_queue()->
->...RCU...->blk_free_queue_rcu()->kmem_cache_free()

Otherwise, RCU callback may be executed first and
dm_cleanup_zoned_dev() will touch free'd memory:

 BUG: KASAN: use-after-free in dm_cleanup_zoned_dev+0x33/0xd0
 Read of size 8 at addr ffff88805ac6e430 by task dmsetup/681

 CPU: 4 PID: 681 Comm: dmsetup Not tainted 5.17.0-rc2+ #6
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-2 04/01/2014
 Call Trace:
  <TASK>
  dump_stack_lvl+0x57/0x7d
  print_address_description.constprop.0+0x1f/0x150
  ? dm_cleanup_zoned_dev+0x33/0xd0
  kasan_report.cold+0x7f/0x11b
  ? dm_cleanup_zoned_dev+0x33/0xd0
  dm_cleanup_zoned_dev+0x33/0xd0
  __dm_destroy+0x26a/0x400
  ? dm_blk_ioctl+0x230/0x230
  ? up_write+0xd8/0x270
  dev_remove+0x156/0x1d0
  ctl_ioctl+0x269/0x530
  ? table_clear+0x140/0x140
  ? lock_release+0xb2/0x750
  ? remove_all+0x40/0x40
  ? rcu_read_lock_sched_held+0x12/0x70
  ? lock_downgrade+0x3c0/0x3c0
  ? rcu_read_lock_sched_held+0x12/0x70
  dm_ctl_ioctl+0xa/0x10
  __x64_sys_ioctl+0xb9/0xf0
  do_syscall_64+0x3b/0x90
  entry_SYSCALL_64_after_hwframe+0x44/0xae
 RIP: 0033:0x7fb6dfa95c27

Fixes: bb37d77239af ("dm: introduce zone append emulation")
Cc: stable@vger.kernel.org
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dm stats: fix too short end duration_ns when using precise_timestamps

commit 0cdb90f0f306384ecbc60dfd6dc48cdbc1f2d0d8 upstream.

dm_stats_account_io()'s STAT_PRECISE_TIMESTAMPS support doesn't handle
the fact that with commit b879f915bc48 ("dm: properly fix redundant
bio-based IO accounting") io->start_time _may_ be in the past (meaning
the start_io_acct() was deferred until later).

Add a new dm_stats_recalc_precise_timestamps() helper that will
set/clear a new 'precise_timestamps' flag in the dm_stats struct based
on whether any configured stats enable STAT_PRECISE_TIMESTAMPS.
And update DM core's alloc_io() to use dm_stats_record_start() to set
stats_aux.duration_ns if stats->precise_timestamps is true.

Also, remove unused 'last_sector' and 'last_rw' members from the
dm_stats struct.

Fixes: b879f915bc48 ("dm: properly fix redundant bio-based IO accounting")
Cc: stable@vger.kernel.org
Co-developed-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm/kmemleak: reset tag when compare object pointer

commit bfc8089f00fa526dea983844c880fa8106c33ac4 upstream.

When we use HW-tag based kasan and enable vmalloc support, we hit the
following bug.  It is due to comparison between tagged object and
non-tagged pointer.

We need to reset the kasan tag when we need to compare tagged object and
non-tagged pointer.

  kmemleak: [name:kmemleak&]Scan area larger than object 0xffffffe77076f440
  CPU: 4 PID: 1 Comm: init Tainted: G S      W         5.15.25-android13-0-g5cacf919c2bc #1
  Hardware name: MT6983(ENG) (DT)
  Call trace:
   add_scan_area+0xc4/0x244
   kmemleak_scan_area+0x40/0x9c
   layout_and_allocate+0x1e8/0x288
   load_module+0x2c8/0xf00
   __se_sys_finit_module+0x190/0x1d0
   __arm64_sys_finit_module+0x20/0x30
   invoke_syscall+0x60/0x170
   el0_svc_common+0xc8/0x114
   do_el0_svc+0x28/0xa0
   el0_svc+0x60/0xf8
   el0t_64_sync_handler+0x88/0xec
   el0t_64_sync+0x1b4/0x1b8
  kmemleak: [name:kmemleak&]Object 0xf5ffffe77076b000 (size 32768):
  kmemleak: [name:kmemleak&]  comm "init", pid 1, jiffies 4294894197
  kmemleak: [name:kmemleak&]  min_count = 0
  kmemleak: [name:kmemleak&]  count = 0
  kmemleak: [name:kmemleak&]  flags = 0x1
  kmemleak: [name:kmemleak&]  checksum = 0
  kmemleak: [name:kmemleak&]  backtrace:
       module_alloc+0x9c/0x120
       move_module+0x34/0x19c
       layout_and_allocate+0x1c4/0x288
       load_module+0x2c8/0xf00
       __se_sys_finit_module+0x190/0x1d0
       __arm64_sys_finit_module+0x20/0x30
       invoke_syscall+0x60/0x170
       el0_svc_common+0xc8/0x114
       do_el0_svc+0x28/0xa0
       el0_svc+0x60/0xf8
       el0t_64_sync_handler+0x88/0xec
       el0t_64_sync+0x1b4/0x1b8

Link: https://lkml.kernel.org/r/20220318034051.30687-1-Kuan-Ying.Lee@mediatek.com
Signed-off-by: Kuan-Ying Lee <Kuan-Ying.Lee@mediatek.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: Matthias Brugger <matthias.bgg@gmail.com>
Cc: Chinwen Chang <chinwen.chang@mediatek.com>
Cc: Nicholas Tang <nicholas.tang@mediatek.com>
Cc: Yee Lee <yee.lee@mediatek.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm,hwpoison: unmap poisoned page before invalidation

commit 3149c79f3cb0e2e3bafb7cfadacec090cbd250d3 upstream.

In some cases it appears the invalidation of a hwpoisoned page fails
because the page is still mapped in another process.  This can cause a
program to be continuously restarted and die when it page faults on the
page that was not invalidated.  Avoid that problem by unmapping the
hwpoisoned page when we find it.

Another issue is that sometimes we end up oopsing in finish_fault, if
the code tries to do something with the now-NULL vmf->page.  I did not
hit this error when submitting the previous patch because there are
several opportunities for alloc_set_pte to bail out before accessing
vmf->page, and that apparently happened on those systems, and most of
the time on other systems, too.

However, across several million systems that error does occur a handful
of times a day.  It can be avoided by returning VM_FAULT_NOPAGE which
will cause do_read_fault to return before calling finish_fault.

Link: https://lkml.kernel.org/r/20220325161428.5068d97e@imladris.surriel.com
Fixes: e53ac7374e64 ("mm: invalidate hwpoison page cache page in fault path")
Signed-off-by: Rik van Riel <riel@surriel.com>
Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Tested-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Revert "mm: madvise: skip unmapped vma holes passed to process_madvise"

commit e6b0a7b357659c332231621e4315658d062c23ee upstream.

This reverts commit 08095d6310a7 ("mm: madvise: skip unmapped vma holes
passed to process_madvise") as process_madvise() fails to return the
exact processed bytes in other cases too.

As an example: if process_madvise() hits mlocked pages after processing
some initial bytes passed in [start, end), it just returns EINVAL
although some bytes are processed.  Thus making an exception only for
ENOMEM is partially fixing the problem of returning the proper advised
bytes.

Thus revert this patch and return proper bytes advised.

Link: https://lkml.kernel.org/r/e73da1304a88b6a8a11907045117cccf4c2b8374.1648046642.git.quic_charante@quicinc.com
Fixes: 08095d6310a7ce ("mm: madvise: skip unmapped vma holes passed to process_madvise")
Signed-off-by: Charan Teja Kalla <quic_charante@quicinc.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: Nadav Amit <nadav.amit@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm: madvise: return correct bytes advised with process_madvise

commit 5bd009c7c9a9e888077c07535dc0c70aeab242c3 upstream.

Patch series "mm: madvise: return correct bytes processed with
process_madvise", v2.  With the process_madvise(), always choose to return
non zero processed bytes over an error.  This can help the user to know on
which VMA, passed in the 'struct iovec' vector list, is failed to advise
thus can take the decission of retrying/skipping on that VMA.

This patch (of 2):

The process_madvise() system call returns error even after processing some
VMA's passed in the 'struct iovec' vector list which leaves the user
confused to know where to restart the advise next.  It is also against
this syscall man page[1] documentation where it mentions that "return
value may be less than the total number of requested bytes, if an error
occurred after some iovec elements were already processed.".

Consider a user passed 10 VMA's in the 'struct iovec' vector list of which
9 are processed but one.  Then it just returns the error caused on that
failed VMA despite the first 9 VMA's processed, leaving the user confused
about on which VMA it is failed.  Returning the number of bytes processed
here can help the user to know which VMA it is failed on and thus can
retry/skip the advise on that VMA.

[1]https://man7.org/linux/man-pages/man2/process_madvise.2.html.

Link: https://lkml.kernel.org/r/cover.1647008754.git.quic_charante@quicinc.com
Link: https://lkml.kernel.org/r/125b61a0edcee5c2db8658aed9d06a43a19ccafc.1647008754.git.quic_charante@quicinc.com
Fixes: ecb8ac8b1f14("mm/madvise: introduce process_madvise() syscall: an external memory hinting API")
Signed-off-by: Charan Teja Kalla <quic_charante@quicinc.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Nadav Amit <nadav.amit@gmail.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm: madvise: skip unmapped vma holes passed to process_madvise

commit 08095d6310a7ce43256b4251577bc66a25c6e1a6 upstream.

The process_madvise() system call is expected to skip holes in vma passed
through 'struct iovec' vector list.  But do_madvise, which
process_madvise() calls for each vma, returns ENOMEM in case of unmapped
holes, despite the VMA is processed.

Thus process_madvise() should treat ENOMEM as expected and consider the
VMA passed to as processed and continue processing other vma's in the
vector list.  Returning -ENOMEM to user, despite the VMA is processed,
will be unable to figure out where to start the next madvise.

Link: https://lkml.kernel.org/r/4f091776142f2ebf7b94018146de72318474e686.1647008754.git.quic_charante@quicinc.com
Fixes: ecb8ac8b1f14("mm/madvise: introduce process_madvise() syscall: an external memory hinting API")
Signed-off-by: Charan Teja Kalla <quic_charante@quicinc.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Nadav Amit <nadav.amit@gmail.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/simpledrm: Add "panel orientation" property on non-upright mounted LCD panels

commit 94fa115f7b28a3f02611499175e134f0a823b686 upstream.

Some devices use e.g. a portrait panel in a standard laptop casing made
for landscape panels. efifb calls drm_get_panel_orientation_quirk() and
sets fb_info.fbcon_rotate_hint to make fbcon rotate the console so that
it shows up-right instead of on its side.

When switching to simpledrm the fbcon renders on its side. Call the
drm_connector_set_panel_orientation_with_quirk() helper to add
a "panel orientation" property on devices listed in the quirk table,
to make the fbcon (and aware userspace apps) rotate the image to
display properly.

Cc: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220221220045.11958-1-hdegoede@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ocfs2: fix crash when mount with quota enabled

commit de19433423c7bedabbd4f9a25f7dbc62c5e78921 upstream.

There is a reported crash when mounting ocfs2 with quota enabled.

  RIP: 0010:ocfs2_qinfo_lock_res_init+0x44/0x50 [ocfs2]
  Call Trace:
    ocfs2_local_read_info+0xb9/0x6f0 [ocfs2]
    dquot_load_quota_sb+0x216/0x470
    dquot_load_quota_inode+0x85/0x100
    ocfs2_enable_quotas+0xa0/0x1c0 [ocfs2]
    ocfs2_fill_super.cold+0xc8/0x1bf [ocfs2]
    mount_bdev+0x185/0x1b0
    legacy_get_tree+0x27/0x40
    vfs_get_tree+0x25/0xb0
    path_mount+0x465/0xac0
    __x64_sys_mount+0x103/0x140

It is caused by when initializing dqi_gqlock, the corresponding dqi_type
and dqi_sb are not properly initialized.

This issue is introduced by commit 6c85c2c72819, which wants to avoid
accessing uninitialized variables in error cases.  So make global quota
info properly initialized.

Link: https://lkml.kernel.org/r/20220323023644.40084-1-joseph.qi@linux.alibaba.com
Link: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1007141
Fixes: 6c85c2c72819 ("ocfs2: quota_local: fix possible uninitialized-variable access in ocfs2_local_read_info()")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Reported-by: Dayvison <sathlerds@gmail.com>
Tested-by: Valentin Vidic <vvidic@valentin-vidic.from.hr>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

rtc: pl031: fix rtc features null pointer dereference

commit ea6af39f3da50c86367a71eb3cc674ade3ed244c upstream.

When there is no interrupt line, rtc alarm feature is disabled.

The clearing of the alarm feature bit was being done prior to allocations
of ldata->rtc device, resulting in a null pointer dereference.

Clear RTC_FEATURE_ALARM after the rtc device is allocated.

Fixes: d9b0dd54a194 ("rtc: pl031: use RTC_FEATURE_ALARM")
Cc: stable@vger.kernel.org
Signed-off-by: Ali Pouladi <quic_apouladi@quicinc.com>
Signed-off-by: Elliot Berman <quic_eberman@quicinc.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/20220225161924.274141-1-quic_eberman@quicinc.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

rtc: mc146818-lib: fix locking in mc146818_set_time

commit 811f5559270f25c34c338d6eaa2ece2544c3d3bd upstream.

In mc146818_set_time(), CMOS_READ(RTC_CONTROL) was performed without the
rtc_lock taken, which is required for CMOS accesses. Fix this.

Nothing in kernel modifies RTC_DM_BINARY, so a separate critical section
is allowed here.

Fixes: dcf257e92622 ("rtc: mc146818: Reduce spinlock section in mc146818_set_time()")
Signed-off-by: Mateusz Jończyk <mat.jonczyk@o2.pl>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Alexandre Belloni <alexandre.belloni@bootlin.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/20220220090403.153928-1-mat.jonczyk@o2.pl
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ALSA: hda/realtek: Fix audio regression on Mi Notebook Pro 2020

commit f30741cded62f87bb4b1cc58bc627f076abcaba8 upstream.

Commit 5aec98913095 ("ALSA: hda/realtek - ALC236 headset MIC recording
issue") is to solve recording issue met on AL236, by matching codec
variant ALC269_TYPE_ALC257 and ALC269_TYPE_ALC256.

This match can be too broad and Mi Notebook Pro 2020 is broken by the
patch.

Instead, use codec ID to be narrow down the scope, in order to make
ALC256 unaffected.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=215484
Fixes: 5aec98913095 ("ALSA: hda/realtek - ALC236 headset MIC recording issue")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20220330061335.1015533-1-kai.heng.feng@canonical.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ALSA: pcm: Fix potential AB/BA lock with buffer_mutex and mmap_lock

commit bc55cfd5718c7c23e5524582e9fa70b4d10f2433 upstream.

syzbot caught a potential deadlock between the PCM
runtime->buffer_mutex and the mm->mmap_lock.  It was brought by the
recent fix to cover the racy read/write and other ioctls, and in that
commit, I overlooked a (hopefully only) corner case that may take the
revert lock, namely, the OSS mmap.  The OSS mmap operation
exceptionally allows to re-configure the parameters inside the OSS
mmap syscall, where mm->mmap_mutex is already held.  Meanwhile, the
copy_from/to_user calls at read/write operations also take the
mm->mmap_lock internally, hence it may lead to a AB/BA deadlock.

A similar problem was already seen in the past and we fixed it with a
refcount (in commit b248371628aa).  The former fix covered only the
call paths with OSS read/write and OSS ioctls, while we need to cover
the concurrent access via both ALSA and OSS APIs now.

This patch addresses the problem above by replacing the buffer_mutex
lock in the read/write operations with a refcount similar as we've
used for OSS.  The new field, runtime->buffer_accessing, keeps the
number of concurrent read/write operations.  Unlike the former
buffer_mutex protection, this protects only around the
copy_from/to_user() calls; the other codes are basically protected by
the PCM stream lock.  The refcount can be a negative, meaning blocked
by the ioctls.  If a negative value is seen, the read/write aborts
with -EBUSY.  In the ioctl side, OTOH, they check this refcount, too,
and set to a negative value for blocking unless it's already being
accessed.

Reported-by: syzbot+6e5c88838328e99c7e1c@syzkaller.appspotmail.com
Fixes: dca947d4d26d ("ALSA: pcm: Fix races among concurrent read/write and buffer changes")
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/000000000000381a0d05db622a81@google.com
Link: https://lore.kernel.org/r/20220330120903.4738-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ALSA: hda: Avoid unsol event during RPM suspending

commit 6ddc2f749621d5d45ca03edc9f0616bcda136d29 upstream.

There is a corner case with unsol event handling during codec runtime
suspending state. When the codec runtime suspend call initiated, the
codec->in_pm atomic variable would be 0, currently the codec runtime
suspend function calls snd_hdac_enter_pm() which will just increments
the codec->in_pm atomic variable. Consider unsol event happened just
after this step and before snd_hdac_leave_pm() in the codec runtime
suspend function. The snd_hdac_power_up_pm() in the unsol event
flow in hdmi_present_sense_via_verbs() function would just increment
the codec->in_pm atomic variable without calling pm_runtime_get_sync
function.

As codec runtime suspend flow is already in progress and in parallel
unsol event is also accessing the codec verbs, as soon as codec
suspend flow completes and clocks are  switched off before completing
the unsol event handling as both functions doesn't wait for each other.
This will result in below errors

[  589.428020] tegra-hda 3510000.hda: azx_get_response timeout, switching
to polling mode: last cmd=0x505f2f57
[  589.428344] tegra-hda 3510000.hda: spurious response 0x80000074:0x5,
last cmd=0x505f2f57
[  589.428547] tegra-hda 3510000.hda: spurious response 0x80000065:0x5,
last cmd=0x505f2f57

To avoid this, the unsol event flow should not perform any codec verb
related operations during RPM_SUSPENDING state.

Signed-off-by: Mohan Kumar <mkumard@nvidia.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20220329155940.26331-1-mkumard@nvidia.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ALSA: cs4236: fix an incorrect NULL check on list iterator

commit 0112f822f8a6d8039c94e0bc9b264d7ffc5d4704 upstream.

The bug is here:
err = snd_card_cs423x_pnp(dev, card->private_data, pdev, cdev);

The list iterator value 'cdev' will *always* be set and non-NULL
by list_for_each_entry(), so it is incorrect to assume that the
iterator value will be NULL if the list is empty or no element
is found.

To fix the bug, use a new variable 'iter' as the list iterator,
while use the original variable 'cdev' as a dedicated pointer
to point to the found element. And snd_card_cs423x_pnp() itself
has NULL check for cdev.

Cc: stable@vger.kernel.org
Fixes: c2b73d1458014 ("ALSA: cs4236: cs4232 and cs4236 driver merge to solve PnP BIOS detection")
Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com>
Link: https://lore.kernel.org/r/20220327060822.4735-1-xiam0nd.tong@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

cifs: fix NULL ptr dereference in smb2_ioctl_query_info()

commit d6f5e358452479fa8a773b5c6ccc9e4ec5a20880 upstream.

When calling smb2_ioctl_query_info() with invalid
smb_query_info::flags, a NULL ptr dereference is triggered when trying
to kfree() uninitialised rqst[n].rq_iov array.

This also fixes leaked paths that are created in SMB2_open_init()
which required SMB2_open_free() to properly free them.

Here is a small C reproducer that triggers it

#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/ioctl.h>

#define die(s) perror(s), exit(1)
#define QUERY_INFO 0xc018cf07

int main(int argc, char *argv[])
{
int fd;

if (argc < 2)
exit(1);
fd = open(argv[1], O_RDONLY);
if (fd == -1)
die("open");
if (ioctl(fd, QUERY_INFO, (uint32_t[]) { 0, 0, 0, 4, 0, 0}) == -1)
die("ioctl");
close(fd);
return 0;
}

mount.cifs //srv/share /mnt -o ...
gcc repro.c && ./a.out /mnt/f0

[ 1832.124468] CIFS: VFS: \\w22-dc.zelda.test\test Invalid passthru query flags: 0x4
[ 1832.125043] general protection fault, probably for non-canonical address 0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN NOPTI
[ 1832.125764] KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
[ 1832.126241] CPU: 3 PID: 1133 Comm: a.out Not tainted 5.17.0-rc8 #2
[ 1832.126630] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.15.0-0-g2dd4b9b-rebuilt.opensuse.org 04/01/2014
[ 1832.127322] RIP: 0010:smb2_ioctl_query_info+0x7a3/0xe30 [cifs]
[ 1832.127749] Code: 00 00 00 fc ff df 48 c1 ea 03 80 3c 02 00 0f 85 6c 05 00 00 48 b8 00 00 00 00 00 fc ff df 4d 8b 74 24 28 4c 89 f2 48 c1 ea 03 <80> 3c 02 00 0f 85 cb 04 00 00 49 8b 3e e8 bb fc fa ff 48 89 da 48
[ 1832.128911] RSP: 0018:ffffc90000957b08 EFLAGS: 00010256
[ 1832.129243] RAX: dffffc0000000000 RBX: ffff888117e9b850 RCX: ffffffffa020580d
[ 1832.129691] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffffffffa043a2c0
[ 1832.130137] RBP: ffff888117e9b878 R08: 0000000000000001 R09: 0000000000000003
[ 1832.130585] R10: fffffbfff4087458 R11: 0000000000000001 R12: ffff888117e9b800
[ 1832.131037] R13: 00000000ffffffea R14: 0000000000000000 R15: ffff888117e9b8a8
[ 1832.131485] FS:  00007fcee9900740(0000) GS:ffff888151a00000(0000) knlGS:0000000000000000
[ 1832.131993] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1832.132354] CR2: 00007fcee9a1ef5e CR3: 0000000114cd2000 CR4: 0000000000350ee0
[ 1832.132801] Call Trace:
[ 1832.132962]  <TASK>
[ 1832.133104]  ? smb2_query_reparse_tag+0x890/0x890 [cifs]
[ 1832.133489]  ? cifs_mapchar+0x460/0x460 [cifs]
[ 1832.133822]  ? rcu_read_lock_sched_held+0x3f/0x70
[ 1832.134125]  ? cifs_strndup_to_utf16+0x15b/0x250 [cifs]
[ 1832.134502]  ? lock_downgrade+0x6f0/0x6f0
[ 1832.134760]  ? cifs_convert_path_to_utf16+0x198/0x220 [cifs]
[ 1832.135170]  ? smb2_check_message+0x1080/0x1080 [cifs]
[ 1832.135545]  cifs_ioctl+0x1577/0x3320 [cifs]
[ 1832.135864]  ? lock_downgrade+0x6f0/0x6f0
[ 1832.136125]  ? cifs_readdir+0x2e60/0x2e60 [cifs]
[ 1832.136468]  ? rcu_read_lock_sched_held+0x3f/0x70
[ 1832.136769]  ? __rseq_handle_notify_resume+0x80b/0xbe0
[ 1832.137096]  ? __up_read+0x192/0x710
[ 1832.137327]  ? __ia32_sys_rseq+0xf0/0xf0
[ 1832.137578]  ? __x64_sys_openat+0x11f/0x1d0
[ 1832.137850]  __x64_sys_ioctl+0x127/0x190
[ 1832.138103]  do_syscall_64+0x3b/0x90
[ 1832.138378]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[ 1832.138702] RIP: 0033:0x7fcee9a253df
[ 1832.138937] Code: 00 48 89 44 24 18 31 c0 48 8d 44 24 60 c7 04 24 10 00 00 00 48 89 44 24 08 48 8d 44 24 20 48 89 44 24 10 b8 10 00 00 00 0f 05 <41> 89 c0 3d 00 f0 ff ff 77 1f 48 8b 44 24 18 64 48 2b 04 25 28 00
[ 1832.140107] RSP: 002b:00007ffeba94a8a0 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[ 1832.140606] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fcee9a253df
[ 1832.141058] RDX: 00007ffeba94a910 RSI: 00000000c018cf07 RDI: 0000000000000003
[ 1832.141503] RBP: 00007ffeba94a930 R08: 00007fcee9b24db0 R09: 00007fcee9b45c4e
[ 1832.141948] R10: 00007fcee9918d40 R11: 0000000000000246 R12: 00007ffeba94aa48
[ 1832.142396] R13: 0000000000401176 R14: 0000000000403df8 R15: 00007fcee9b78000
[ 1832.142851]  </TASK>
[ 1832.142994] Modules linked in: cifs cifs_arc4 cifs_md4 bpf_preload [last unloaded: cifs]

Cc: stable@vger.kernel.org
Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

cifs: prevent bad output lengths in smb2_ioctl_query_info()

commit b92e358757b91c2827af112cae9af513f26a3f34 upstream.

When calling smb2_ioctl_query_info() with
smb_query_info::flags=PASSTHRU_FSCTL and
smb_query_info::output_buffer_length=0, the following would return
0x10

buffer = memdup_user(arg + sizeof(struct smb_query_info),
     qi.output_buffer_length);
if (IS_ERR(buffer)) {
kfree(vars);
return PTR_ERR(buffer);
}

rather than a valid pointer thus making IS_ERR() check fail.  This
would then cause a NULL ptr deference in @buffer when accessing it
later in smb2_ioctl_query_ioctl().  While at it, prevent having a
@buffer smaller than 8 bytes to correctly handle SMB2_SET_INFO
FileEndOfFileInformation requests when
smb_query_info::flags=PASSTHRU_SET_INFO.

Here is a small C reproducer which triggers a NULL ptr in @buffer when
passing an invalid smb_query_info::flags

#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/ioctl.h>

#define die(s) perror(s), exit(1)
#define QUERY_INFO 0xc018cf07

int main(int argc, char *argv[])
{
int fd;

if (argc < 2)
exit(1);
fd = open(argv[1], O_RDONLY);
if (fd == -1)
die("open");
if (ioctl(fd, QUERY_INFO, (uint32_t[]) { 0, 0, 0, 4, 0, 0}) == -1)
die("ioctl");
close(fd);
return 0;
}

mount.cifs //srv/share /mnt -o ...
gcc repro.c && ./a.out /mnt/f0

[  114.138620] general protection fault, probably for non-canonical address 0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN NOPTI
[  114.139310] KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
[  114.139775] CPU: 2 PID: 995 Comm: a.out Not tainted 5.17.0-rc8 #1
[  114.140148] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.15.0-0-g2dd4b9b-rebuilt.opensuse.org 04/01/2014
[  114.140818] RIP: 0010:smb2_ioctl_query_info+0x206/0x410 [cifs]
[  114.141221] Code: 00 00 00 00 fc ff df 48 c1 ea 03 80 3c 02 00 0f 85 c8 01 00 00 48 b8 00 00 00 00 00 fc ff df 4c 8b 7b 28 4c 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 9c 01 00 00 49 8b 3f e8 58 02 fb ff 48 8b 14 24
[  114.142348] RSP: 0018:ffffc90000b47b00 EFLAGS: 00010256
[  114.142692] RAX: dffffc0000000000 RBX: ffff888115503200 RCX: ffffffffa020580d
[  114.143119] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffffffffa043a380
[  114.143544] RBP: ffff888115503278 R08: 0000000000000001 R09: 0000000000000003
[  114.143983] R10: fffffbfff4087470 R11: 0000000000000001 R12: ffff888115503288
[  114.144424] R13: 00000000ffffffea R14: ffff888115503228 R15: 0000000000000000
[  114.144852] FS:  00007f7aeabdf740(0000) GS:ffff888151600000(0000) knlGS:0000000000000000
[  114.145338] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  114.145692] CR2: 00007f7aeacfdf5e CR3: 000000012000e000 CR4: 0000000000350ee0
[  114.146131] Call Trace:
[  114.146291]  <TASK>
[  114.146432]  ? smb2_query_reparse_tag+0x890/0x890 [cifs]
[  114.146800]  ? cifs_mapchar+0x460/0x460 [cifs]
[  114.147121]  ? rcu_read_lock_sched_held+0x3f/0x70
[  114.147412]  ? cifs_strndup_to_utf16+0x15b/0x250 [cifs]
[  114.147775]  ? dentry_path_raw+0xa6/0xf0
[  114.148024]  ? cifs_convert_path_to_utf16+0x198/0x220 [cifs]
[  114.148413]  ? smb2_check_message+0x1080/0x1080 [cifs]
[  114.148766]  ? rcu_read_lock_sched_held+0x3f/0x70
[  114.149065]  cifs_ioctl+0x1577/0x3320 [cifs]
[  114.149371]  ? lock_downgrade+0x6f0/0x6f0
[  114.149631]  ? cifs_readdir+0x2e60/0x2e60 [cifs]
[  114.149956]  ? rcu_read_lock_sched_held+0x3f/0x70
[  114.150250]  ? __rseq_handle_notify_resume+0x80b/0xbe0
[  114.150562]  ? __up_read+0x192/0x710
[  114.150791]  ? __ia32_sys_rseq+0xf0/0xf0
[  114.151025]  ? __x64_sys_openat+0x11f/0x1d0
[  114.151296]  __x64_sys_ioctl+0x127/0x190
[  114.151549]  do_syscall_64+0x3b/0x90
[  114.151768]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[  114.152079] RIP: 0033:0x7f7aead043df
[  114.152306] Code: 00 48 89 44 24 18 31 c0 48 8d 44 24 60 c7 04 24 10 00 00 00 48 89 44 24 08 48 8d 44 24 20 48 89 44 24 10 b8 10 00 00 00 0f 05 <41> 89 c0 3d 00 f0 ff ff 77 1f 48 8b 44 24 18 64 48 2b 04 25 28 00
[  114.153431] RSP: 002b:00007ffc2e0c1f80 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[  114.153890] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f7aead043df
[  114.154315] RDX: 00007ffc2e0c1ff0 RSI: 00000000c018cf07 RDI: 0000000000000003
[  114.154747] RBP: 00007ffc2e0c2010 R08: 00007f7aeae03db0 R09: 00007f7aeae24c4e
[  114.155192] R10: 00007f7aeabf7d40 R11: 0000000000000246 R12: 00007ffc2e0c2128
[  114.155642] R13: 0000000000401176 R14: 0000000000403df8 R15: 00007f7aeae57000
[  114.156071]  </TASK>
[  114.156218] Modules linked in: cifs cifs_arc4 cifs_md4 bpf_preload
[  114.156608] ---[ end trace 0000000000000000 ]---
[  114.156898] RIP: 0010:smb2_ioctl_query_info+0x206/0x410 [cifs]
[  114.157792] Code: 00 00 00 00 fc ff df 48 c1 ea 03 80 3c 02 00 0f 85 c8 01 00 00 48 b8 00 00 00 00 00 fc ff df 4c 8b 7b 28 4c 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 9c 01 00 00 49 8b 3f e8 58 02 fb ff 48 8b 14 24
[  114.159293] RSP: 0018:ffffc90000b47b00 EFLAGS: 00010256
[  114.159641] RAX: dffffc0000000000 RBX: ffff888115503200 RCX: ffffffffa020580d
[  114.160093] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffffffffa043a380
[  114.160699] RBP: ffff888115503278 R08: 0000000000000001 R09: 0000000000000003
[  114.161196] R10: fffffbfff4087470 R11: 0000000000000001 R12: ffff888115503288
[  114.155642] R13: 0000000000401176 R14: 0000000000403df8 R15: 00007f7aeae57000
[  114.156071]  </TASK>
[  114.156218] Modules linked in: cifs cifs_arc4 cifs_md4 bpf_preload
[  114.156608] ---[ end trace 0000000000000000 ]---
[  114.156898] RIP: 0010:smb2_ioctl_query_info+0x206/0x410 [cifs]
[  114.157792] Code: 00 00 00 00 fc ff df 48 c1 ea 03 80 3c 02 00 0f 85 c8 01 00 00 48 b8 00 00 00 00 00 fc ff df 4c 8b 7b 28 4c 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 9c 01 00 00 49 8b 3f e8 58 02 fb ff 48 8b 14 24
[  114.159293] RSP: 0018:ffffc90000b47b00 EFLAGS: 00010256
[  114.159641] RAX: dffffc0000000000 RBX: ffff888115503200 RCX: ffffffffa020580d
[  114.160093] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffffffffa043a380
[  114.160699] RBP: ffff888115503278 R08: 0000000000000001 R09: 0000000000000003
[  114.161196] R10: fffffbfff4087470 R11: 0000000000000001 R12: ffff888115503288
[  114.161823] R13: 00000000ffffffea R14: ffff888115503228 R15: 0000000000000000
[  114.162274] FS:  00007f7aeabdf740(0000) GS:ffff888151600000(0000) knlGS:0000000000000000
[  114.162853] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  114.163218] CR2: 00007f7aeacfdf5e CR3: 000000012000e000 CR4: 0000000000350ee0
[  114.163691] Kernel panic - not syncing: Fatal exception
[  114.164087] Kernel Offset: disabled
[  114.164316] ---[ end Kernel panic - not syncing: Fatal exception ]---

Cc: stable@vger.kernel.org
Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Revert "Input: clear BTN_RIGHT/MIDDLE on buttonpads"

commit 8b188fba75195745026e11d408e4a7e94e01d701 upstream.

This reverts commit 37ef4c19b4c659926ce65a7ac709ceaefb211c40.

The touchpad present in the Dell Precision 7550 and 7750 laptops
reports a HID_DG_BUTTONTYPE of type MT_BUTTONTYPE_CLICKPAD. However,
the device is not a clickpad, it is a touchpad with physical buttons.

In order to fix this issue, a quirk for the device was introduced in
libinput [1] [2] to disable the INPUT_PROP_BUTTONPAD property:

[Precision 7x50 Touchpad]
MatchBus=i2c
MatchUdevType=touchpad
MatchDMIModalias=dmi:*svnDellInc.:pnPrecision7?50*
AttrInputPropDisable=INPUT_PROP_BUTTONPAD

However, because of the change introduced in 37ef4c19b4 ("Input: clear
BTN_RIGHT/MIDDLE on buttonpads") the BTN_RIGHT key bit is not mapped
anymore breaking the device right click button and making impossible to
workaround it in user space.

In order to avoid breakage on other present or future devices, revert
the patch causing the issue.

Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20220321184404.20025-1-jose.exposito89@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

riscv: Increase stack size under KASAN

commit b81d591386c3a50b96dddcf663628ea0df0bf2b3 upstream.

KASAN requires more stack space because of compiler instrumentation.
Increase stack size as other arches do.

Signed-off-by: Dmitry Vyukov <dvyukov@google.com>
Reported-by: syzbot+0600986d88e2d4d7ebb8@syzkaller.appspotmail.com
Fixes: 8ad8b72721d0 ("riscv: Add KASAN support")
Cc: stable@vger.kernel.org
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

riscv: Fix fill_callchain return value

commit 2b2b574ac587ec5bd7716a356492a85ab8b0ce9f upstream.

perf_callchain_store return 0 on success, -1 otherwise,
fix fill_callchain to return correct bool value.

Fixes: dbeb90b0c1eb ("riscv: Add perf callchain support")
Signed-off-by: Nikita Shubin <n.shubin@yadro.com>
Cc: stable@vger.kernel.org
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

riscv: dts: canaan: Fix SPI3 bus width

commit 6846d656106add3aeefcd6eda0dc885787deaa6e upstream.

According to the K210 Standalone SDK Programming guide:
https://canaan-creative.com/wp-content/uploads/2020/03/kendryte_standalone_programming_guide_20190311144158_en.pdf

Section 15.4.3.3:
SPI0 and SPI1 supports: standard, dual, quad and octal transfers.
SPI3 supports: standard, dual and quad transfers (octal is not supported).

In order to support quad transfers (Quad SPI), SPI3 must have four IO wires
connected to the SPI flash.

Update the device tree to specify the correct bus width.

Tested on maix bit, maix dock and maixduino, which all have the same
SPI flash (gd25lq128d) connected to SPI3. maix go is untested, but it
would not make sense for this k210 board to be designed differently.

Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Fixes: 8f5b0e79f3e5 ("riscv: Add SiPeed MAIXDUINO board device tree")
Fixes: 8194f08bda18 ("riscv: Add SiPeed MAIX GO board device tree")
Fixes: a40f920964c4 ("riscv: Add SiPeed MAIX DOCK board device tree")
Fixes: 97c279bcf813 ("riscv: Add SiPeed MAIX BiT board device tree")
Cc: stable@vger.kernel.org
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

qed: validate and restrict untrusted VFs vlan promisc mode

commit cbcc44db2cf7b836896733acc0e5ea966136ed22 upstream.

Today when VFs are put in promiscuous mode, they can request PF
to configure device for them to receive all VLANs traffic regardless
of what vlan is configured by the PF (via ip link) and PF allows this
config request regardless of whether VF is trusted or not.

From security POV, when VLAN is configured for VF through PF (via ip link),
honour such config requests from VF only when they are configured to be
trusted, otherwise restrict such VFs vlan promisc mode config.

Cc: stable@vger.kernel.org
Fixes: f990c82c385b ("qed*: Add support for ndo_set_vf_trust")
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

qed: display VF trust config

commit 4e6e6bec7440b9b76f312f28b1f4e944eebb3abc upstream.

Driver does support SR-IOV VFs trust configuration but
it does not display it when queried via ip link utility.

Cc: stable@vger.kernel.org
Fixes: f990c82c385b ("qed*: Add support for ndo_set_vf_trust")
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

scsi: libsas: Fix sas_ata_qc_issue() handling of NCQ NON DATA commands

commit 8454563e4c2aafbfb81a383ab423ea8b9b430a25 upstream.

To detect for the DMA_NONE (no data transfer) DMA direction,
sas_ata_qc_issue() tests if the command protocol is ATA_PROT_NODATA.  This
test does not include the ATA_CMD_NCQ_NON_DATA command as this command
protocol is defined as ATA_PROT_NCQ_NODATA (equal to ATA_PROT_FLAG_NCQ) and
not as ATA_PROT_NODATA.

To include both NCQ and non-NCQ commands when testing for the DMA_NONE DMA
direction, use "!ata_is_data()".

Link: https://lore.kernel.org/r/20220220031810.738362-2-damien.lemoal@opensource.wdc.com
Fixes: 176ddd89171d ("scsi: libsas: Reset num_scatter if libata marks qc as NODATA")
Cc: stable@vger.kernel.org
Reviewed-by: John Garry <john.garry@huawei.com>
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

scsi: scsi_transport_fc: Fix FPIN Link Integrity statistics counters

commit 07e0984b96ec1ba8c6de1c092b986b00ea0c114c upstream.

In the original FPIN commit, stats were incremented by the event_count.
Event_count is the minimum # of events that must occur before an FPIN is
sent. Thus, its not the actual number of events, and could be significantly
off (too low) as it doesn't reflect anything not reported.  Rather than
attempt to count events, have the statistic count how many FPINS cross the
threshold and were reported.

Link: https://lore.kernel.org/r/20220301175536.60250-1-jsmart2021@gmail.com
Fixes: 3dcfe0de5a97 ("scsi: fc: Parse FPIN packets and update statistics")
Cc: <stable@vger.kernel.org> # v5.11+
Cc: Shyam Sundar <ssundar@marvell.com>
Cc: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

scsi: ufs: Fix runtime PM messages never-ending cycle

commit 71bb9ab6e3511b7bb98678a19eb8cf1ccbf3ca2f upstream.

Kernel messages produced during runtime PM can cause a never-ending cycle
because user space utilities (e.g. journald or rsyslog) write the messages
back to storage, causing runtime resume, more messages, and so on.

Messages that tell of things that are expected to happen, are arguably
unnecessary, so suppress them.

UFS driver messages are changes to from dev_err() to dev_dbg() which means
they will not display unless activated by dynamic debug of building with
-DDEBUG.

sdev->silence_suspend is set to skip messages from sd_suspend_common()
"Synchronizing SCSI cache", "Stopping disk" and scsi_report_sense()
"Power-on or device reset occurred" message (Note, that message appears
when the LUN is accessed after runtime PM, not during runtime PM)

 Example messages from Ubuntu 21.10:

 $ dmesg | tail
 [ 1620.380071] ufshcd 0000:00:12.5: ufshcd_print_pwr_info:[RX, TX]: gear=[1, 1], lane[1, 1], pwr[SLOWAUTO_MODE, SLOWAUTO_MODE], rate = 0
 [ 1620.408825] ufshcd 0000:00:12.5: ufshcd_print_pwr_info:[RX, TX]: gear=[4, 4], lane[2, 2], pwr[FAST MODE, FAST MODE], rate = 2
 [ 1620.409020] ufshcd 0000:00:12.5: ufshcd_find_max_sup_active_icc_level: Regulator capability was not set, actvIccLevel=0
 [ 1620.409524] sd 0:0:0:0: Power-on or device reset occurred
 [ 1622.938794] sd 0:0:0:0: [sda] Synchronizing SCSI cache
 [ 1622.939184] ufs_device_wlun 0:0:0:49488: Power-on or device reset occurred
 [ 1625.183175] ufshcd 0000:00:12.5: ufshcd_print_pwr_info:[RX, TX]: gear=[1, 1], lane[1, 1], pwr[SLOWAUTO_MODE, SLOWAUTO_MODE], rate = 0
 [ 1625.208041] ufshcd 0000:00:12.5: ufshcd_print_pwr_info:[RX, TX]: gear=[4, 4], lane[2, 2], pwr[FAST MODE, FAST MODE], rate = 2
 [ 1625.208311] ufshcd 0000:00:12.5: ufshcd_find_max_sup_active_icc_level: Regulator capability was not set, actvIccLevel=0
 [ 1625.209035] sd 0:0:0:0: Power-on or device reset occurred

Note for stable: depends on patch "scsi: core: sd: Add silence_suspend flag
to suppress some PM messages".

Link: https://lore.kernel.org/r/20220228113652.970857-3-adrian.hunter@intel.com
Cc: stable@vger.kernel.org
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

scsi: core: sd: Add silence_suspend flag to suppress some PM messages

commit af4edb1d50c6d1044cb34bc43621411b7ba2cffe upstream.

Kernel messages produced during runtime PM can cause a never-ending cycle
because user space utilities (e.g. journald or rsyslog) write the messages
back to storage, causing runtime resume, more messages, and so on.

Messages that tell of things that are expected to happen are arguably
unnecessary, so add a flag to suppress them. This flag is used by the UFS
driver.

Link: https://lore.kernel.org/r/20220228113652.970857-2-adrian.hunter@intel.com
Cc: stable@vger.kernel.org
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mempolicy: mbind_range() set_policy() after vma_merge()

commit 4e0906008cdb56381638aa17d9c32734eae6d37a upstream.

v2.6.34 commit 9d8cebd4bcd7 ("mm: fix mbind vma merge problem") introduced
vma_merge() to mbind_range(); but unlike madvise, mlock and mprotect, it
put a "continue" to next vma where its precedents go to update flags on
current vma before advancing: that left vma with the wrong setting in the
infamous vma_merge() case 8.

v3.10 commit 1444f92c8498 ("mm: merging memory blocks resets mempolicy")
tried to fix that in vma_adjust(), without fully understanding the issue.

v3.11 commit 3964acd0dbec ("mm: mempolicy: fix mbind_range() &&
vma_adjust() interaction") reverted that, and went about the fix in the
right way, but chose to optimize out an unnecessary mpol_dup() with a
prior mpol_equal() test.  But on tmpfs, that also pessimized out the vital
call to its ->set_policy(), leaving the new mbind unenforced.

The user visible effect was that the pages got allocated on the local
node (happened to be 0), after the mbind() caller had specifically
asked for them to be allocated on node 1.  There was not any page
migration involved in the case reported: the pages simply got allocated
on the wrong node.

Just delete that optimization now (though it could be made conditional on
vma not having a set_policy).  Also remove the "next" variable: it turned
out to be blameless, but also pointless.

Link: https://lkml.kernel.org/r/319e4db9-64ae-4bca-92f0-ade85d342ff@google.com
Fixes: 3964acd0dbec ("mm: mempolicy: fix mbind_range() && vma_adjust() interaction")
Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm: invalidate hwpoison page cache page in fault path

commit e53ac7374e64dede04d745ff0e70ff5048378d1f upstream.

Sometimes the page offlining code can leave behind a hwpoisoned clean
page cache page.  This can lead to programs being killed over and over
and over again as they fault in the hwpoisoned page, get killed, and
then get re-spawned by whatever wanted to run them.

This is particularly embarrassing when the page was offlined due to
having too many corrected memory errors.  Now we are killing tasks due
to them trying to access memory that probably isn't even corrupted.

This problem can be avoided by invalidating the page from the page fault
handler, which already has a branch for dealing with these kinds of
pages.  With this patch we simply pretend the page fault was successful
if the page was invalidated, return to userspace, incur another page
fault, read in the file from disk (to a new memory page), and then
everything works again.

Link: https://lkml.kernel.org/r/20220212213740.423efcea@imladris.surriel.com
Signed-off-by: Rik van Riel <riel@surriel.com>
Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Acked-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm/pages_alloc.c: don't create ZONE_MOVABLE beyond the end of a node

commit ddbc84f3f595cf1fc8234a191193b5d20ad43938 upstream.

ZONE_MOVABLE uses the remaining memory in each node.  Its starting pfn
is also aligned to MAX_ORDER_NR_PAGES.  It is possible for the remaining
memory in a node to be less than MAX_ORDER_NR_PAGES, meaning there is
not enough room for ZONE_MOVABLE on that node.

Unfortunately this condition is not checked for.  This leads to
zone_movable_pfn[] getting set to a pfn greater than the last pfn in a
node.

calculate_node_totalpages() then sets zone->present_pages to be greater
than zone->spanned_pages which is invalid, as spanned_pages represents
the maximum number of pages in a zone assuming no holes.

Subsequently it is possible free_area_init_core() will observe a zone of
size zero with present pages.  In this case it will skip setting up the
zone, including the initialisation of free_lists[].

However populated_zone() checks zone->present_pages to see if a zone has
memory available.  This is used by iterators such as
walk_zones_in_node().  pagetypeinfo_showfree() uses this to walk the
free_list of each zone in each node, which are assumed to be initialised
due to the zone not being empty.

As free_area_init_core() never initialised the free_lists[] this results
in the following kernel crash when trying to read /proc/pagetypeinfo:

  BUG: kernel NULL pointer dereference, address: 0000000000000000
  #PF: supervisor read access in kernel mode
  #PF: error_code(0x0000) - not-present page
  PGD 0 P4D 0
  Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC NOPTI
  CPU: 0 PID: 456 Comm: cat Not tainted 5.16.0 #461
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.14.0-2 04/01/2014
  RIP: 0010:pagetypeinfo_show+0x163/0x460
  Code: 9e 82 e8 80 57 0e 00 49 8b 06 b9 01 00 00 00 4c 39 f0 75 16 e9 65 02 00 00 48 83 c1 01 48 81 f9 a0 86 01 00 0f 84 48 02 00 00 <48> 8b 00 4c 39 f0 75 e7 48 c7 c2 80 a2 e2 82 48 c7 c6 79 ef e3 82
  RSP: 0018:ffffc90001c4bd10 EFLAGS: 00010003
  RAX: 0000000000000000 RBX: ffff88801105f638 RCX: 0000000000000001
  RDX: 0000000000000001 RSI: 000000000000068b RDI: ffff8880163dc68b
  RBP: ffffc90001c4bd90 R08: 0000000000000001 R09: ffff8880163dc67e
  R10: 656c6261766f6d6e R11: 6c6261766f6d6e55 R12: ffff88807ffb4a00
  R13: ffff88807ffb49f8 R14: ffff88807ffb4580 R15: ffff88807ffb3000
  FS:  00007f9c83eff5c0(0000) GS:ffff88807dc00000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 0000000000000000 CR3: 0000000013c8e000 CR4: 0000000000350ef0
  Call Trace:
   seq_read_iter+0x128/0x460
   proc_reg_read_iter+0x51/0x80
   new_sync_read+0x113/0x1a0
   vfs_read+0x136/0x1d0
   ksys_read+0x70/0xf0
   __x64_sys_read+0x1a/0x20
   do_syscall_64+0x3b/0xc0
   entry_SYSCALL_64_after_hwframe+0x44/0xae

Fix this by checking that the aligned zone_movable_pfn[] does not exceed
the end of the node, and if it does skip creating a movable zone on this
node.

Link: https://lkml.kernel.org/r/20220215025831.2113067-1-apopple@nvidia.com
Fixes: 2a1e274acf0b ("Create the ZONE_MOVABLE zone")
Signed-off-by: Alistair Popple <apopple@nvidia.com>
Acked-by: David Hildenbrand <david@redhat.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Zi Yan <ziy@nvidia.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm: fs: fix lru_cache_disabled race in bh_lru

commit c0226eb8bde854e016a594a16f5c0d98aca426fa upstream.

Check lru_cache_disabled under bh_lru_lock.  Otherwise, it could introduce
race below and it fails to migrate pages containing buffer_head.

   CPU 0 CPU 1

bh_lru_install
                                       lru_cache_disable
  lru_cache_disabled = false
                                       atomic_inc(&lru_disable_count);
       invalidate_bh_lrus_cpu of CPU 0
       bh_lru_lock
       __invalidate_bh_lrus
       bh_lru_unlock
  bh_lru_lock
  install the bh
  bh_lru_unlock

WHen this race happens a CMA allocation fails, which is critical for
the workload which depends on CMA.

Link: https://lkml.kernel.org/r/20220308180709.2017638-1-minchan@kernel.org
Fixes: 8cc621d2f45d ("mm: fs: invalidate BH LRU during page migration")
Signed-off-by: Minchan Kim <minchan@kernel.org>
Cc: Chris Goldsworthy <cgoldswo@codeaurora.org>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: John Dias <joaodias@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

jffs2: fix memory leak in jffs2_scan_medium

commit 9cdd3128874f5fe759e2c4e1360ab7fb96a8d1df upstream.

If an error is returned in jffs2_scan_eraseblock() and some memory
has been added to the jffs2_summary *s, we can observe the following
kmemleak report:

--------------------------------------------
unreferenced object 0xffff88812b889c40 (size 64):
  comm "mount", pid 692, jiffies 4294838325 (age 34.288s)
  hex dump (first 32 bytes):
    40 48 b5 14 81 88 ff ff 01 e0 31 00 00 00 50 00  @H........1...P.
    00 00 01 00 00 00 01 00 00 00 02 00 00 00 09 08  ................
  backtrace:
    [<ffffffffae93a3a3>] __kmalloc+0x613/0x910
    [<ffffffffaf423b9c>] jffs2_sum_add_dirent_mem+0x5c/0xa0
    [<ffffffffb0f3afa8>] jffs2_scan_medium.cold+0x36e5/0x4794
    [<ffffffffb0f3dbe1>] jffs2_do_mount_fs.cold+0xa7/0x2267
    [<ffffffffaf40acf3>] jffs2_do_fill_super+0x383/0xc30
    [<ffffffffaf40c00a>] jffs2_fill_super+0x2ea/0x4c0
    [<ffffffffb0315d64>] mtd_get_sb+0x254/0x400
    [<ffffffffb0315f5f>] mtd_get_sb_by_nr+0x4f/0xd0
    [<ffffffffb0316478>] get_tree_mtd+0x498/0x840
    [<ffffffffaf40bd15>] jffs2_get_tree+0x25/0x30
    [<ffffffffae9f358d>] vfs_get_tree+0x8d/0x2e0
    [<ffffffffaea7a98f>] path_mount+0x50f/0x1e50
    [<ffffffffaea7c3d7>] do_mount+0x107/0x130
    [<ffffffffaea7c5c5>] __se_sys_mount+0x1c5/0x2f0
    [<ffffffffaea7c917>] __x64_sys_mount+0xc7/0x160
    [<ffffffffb10142f5>] do_syscall_64+0x45/0x70
unreferenced object 0xffff888114b54840 (size 32):
  comm "mount", pid 692, jiffies 4294838325 (age 34.288s)
  hex dump (first 32 bytes):
    c0 75 b5 14 81 88 ff ff 02 e0 02 00 00 00 02 00  .u..............
    00 00 84 00 00 00 44 00 00 00 6b 6b 6b 6b 6b a5  ......D...kkkkk.
  backtrace:
    [<ffffffffae93be24>] kmem_cache_alloc_trace+0x584/0x880
    [<ffffffffaf423b04>] jffs2_sum_add_inode_mem+0x54/0x90
    [<ffffffffb0f3bd44>] jffs2_scan_medium.cold+0x4481/0x4794
    [...]
unreferenced object 0xffff888114b57280 (size 32):
  comm "mount", pid 692, jiffies 4294838393 (age 34.357s)
  hex dump (first 32 bytes):
    10 d5 6c 11 81 88 ff ff 08 e0 05 00 00 00 01 00  ..l.............
    00 00 38 02 00 00 28 00 00 00 6b 6b 6b 6b 6b a5  ..8...(...kkkkk.
  backtrace:
    [<ffffffffae93be24>] kmem_cache_alloc_trace+0x584/0x880
    [<ffffffffaf423c34>] jffs2_sum_add_xattr_mem+0x54/0x90
    [<ffffffffb0f3a24f>] jffs2_scan_medium.cold+0x298c/0x4794
    [...]
unreferenced object 0xffff8881116cd510 (size 16):
  comm "mount", pid 692, jiffies 4294838395 (age 34.355s)
  hex dump (first 16 bytes):
    00 00 00 00 00 00 00 00 09 e0 60 02 00 00 6b a5  ..........`...k.
  backtrace:
    [<ffffffffae93be24>] kmem_cache_alloc_trace+0x584/0x880
    [<ffffffffaf423cc4>] jffs2_sum_add_xref_mem+0x54/0x90
    [<ffffffffb0f3b2e3>] jffs2_scan_medium.cold+0x3a20/0x4794
    [...]
--------------------------------------------

Therefore, we should call jffs2_sum_reset_collected(s) on exit to
release the memory added in s. In addition, a new tag "out_buf" is
added to prevent the NULL pointer reference caused by s being NULL.
(thanks to Zhang Yi for this analysis)

Fixes: e631ddba5887 ("[JFFS2] Add erase block summary support (mount time improvement)")
Cc: stable@vger.kernel.org
Co-developed-with: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

jffs2: fix memory leak in jffs2_do_mount_fs

commit d051cef784de4d54835f6b6836d98a8f6935772c upstream.

If jffs2_build_filesystem() in jffs2_do_mount_fs() returns an error,
we can observe the following kmemleak report:

--------------------------------------------
unreferenced object 0xffff88811b25a640 (size 64):
  comm "mount", pid 691, jiffies 4294957728 (age 71.952s)
  hex dump (first 32 bytes):
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<ffffffffa493be24>] kmem_cache_alloc_trace+0x584/0x880
    [<ffffffffa5423a06>] jffs2_sum_init+0x86/0x130
    [<ffffffffa5400e58>] jffs2_do_mount_fs+0x798/0xac0
    [<ffffffffa540acf3>] jffs2_do_fill_super+0x383/0xc30
    [<ffffffffa540c00a>] jffs2_fill_super+0x2ea/0x4c0
    [...]
unreferenced object 0xffff88812c760000 (size 65536):
  comm "mount", pid 691, jiffies 4294957728 (age 71.952s)
  hex dump (first 32 bytes):
    bb bb bb bb bb bb bb bb bb bb bb bb bb bb bb bb  ................
    bb bb bb bb bb bb bb bb bb bb bb bb bb bb bb bb  ................
  backtrace:
    [<ffffffffa493a449>] __kmalloc+0x6b9/0x910
    [<ffffffffa5423a57>] jffs2_sum_init+0xd7/0x130
    [<ffffffffa5400e58>] jffs2_do_mount_fs+0x798/0xac0
    [<ffffffffa540acf3>] jffs2_do_fill_super+0x383/0xc30
    [<ffffffffa540c00a>] jffs2_fill_super+0x2ea/0x4c0
    [...]
--------------------------------------------

This is because the resources allocated in jffs2_sum_init() are not
released. Call jffs2_sum_exit() to release these resources to solve
the problem.

Fixes: e631ddba5887 ("[JFFS2] Add erase block summary support (mount time improvement)")
Cc: stable@vger.kernel.org
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

jffs2: fix use-after-free in jffs2_clear_xattr_subsystem

commit 4c7c44ee1650677fbe89d86edbad9497b7679b5c upstream.

When we mount a jffs2 image, assume that the first few blocks of
the image are normal and contain at least one xattr-related inode,
but the next block is abnormal. As a result, an error is returned
in jffs2_scan_eraseblock(). jffs2_clear_xattr_subsystem() is then
called in jffs2_build_filesystem() and then again in
jffs2_do_fill_super().

Finally we can observe the following report:
 ==================================================================
 BUG: KASAN: use-after-free in jffs2_clear_xattr_subsystem+0x95/0x6ac
 Read of size 8 at addr ffff8881243384e0 by task mount/719

 Call Trace:
  dump_stack+0x115/0x16b
  jffs2_clear_xattr_subsystem+0x95/0x6ac
  jffs2_do_fill_super+0x84f/0xc30
  jffs2_fill_super+0x2ea/0x4c0
  mtd_get_sb+0x254/0x400
  mtd_get_sb_by_nr+0x4f/0xd0
  get_tree_mtd+0x498/0x840
  jffs2_get_tree+0x25/0x30
  vfs_get_tree+0x8d/0x2e0
  path_mount+0x50f/0x1e50
  do_mount+0x107/0x130
  __se_sys_mount+0x1c5/0x2f0
  __x64_sys_mount+0xc7/0x160
  do_syscall_64+0x45/0x70
  entry_SYSCALL_64_after_hwframe+0x44/0xa9

 Allocated by task 719:
  kasan_save_stack+0x23/0x60
  __kasan_kmalloc.constprop.0+0x10b/0x120
  kasan_slab_alloc+0x12/0x20
  kmem_cache_alloc+0x1c0/0x870
  jffs2_alloc_xattr_ref+0x2f/0xa0
  jffs2_scan_medium.cold+0x3713/0x4794
  jffs2_do_mount_fs.cold+0xa7/0x2253
  jffs2_do_fill_super+0x383/0xc30
  jffs2_fill_super+0x2ea/0x4c0
 [...]

 Freed by task 719:
  kmem_cache_free+0xcc/0x7b0
  jffs2_free_xattr_ref+0x78/0x98
  jffs2_clear_xattr_subsystem+0xa1/0x6ac
  jffs2_do_mount_fs.cold+0x5e6/0x2253
  jffs2_do_fill_super+0x383/0xc30
  jffs2_fill_super+0x2ea/0x4c0
 [...]

 The buggy address belongs to the object at ffff8881243384b8
  which belongs to the cache jffs2_xattr_ref of size 48
 The buggy address is located 40 bytes inside of
  48-byte region [ffff8881243384b8, ffff8881243384e8)
 [...]
 ==================================================================

The triggering of the BUG is shown in the following stack:
-----------------------------------------------------------
jffs2_fill_super
  jffs2_do_fill_super
    jffs2_do_mount_fs
      jffs2_build_filesystem
        jffs2_scan_medium
          jffs2_scan_eraseblock        <--- ERROR
        jffs2_clear_xattr_subsystem    <--- free
    jffs2_clear_xattr_subsystem        <--- free again
-----------------------------------------------------------

An error is returned in jffs2_do_mount_fs(). If the error is returned
by jffs2_sum_init(), the jffs2_clear_xattr_subsystem() does not need to
be executed. If the error is returned by jffs2_build_filesystem(), the
jffs2_clear_xattr_subsystem() also does not need to be executed again.
So move jffs2_clear_xattr_subsystem() from 'out_inohash' to 'out_root'
to fix this UAF problem.

Fixes: aa98d7cf59b5 ("[JFFS2][XATTR] XATTR support on JFFS2 (version. 5)")
Cc: stable@vger.kernel.org
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

can: usb_8dev: usb_8dev_start_xmit(): fix double dev_kfree_skb() in error path

commit 3d3925ff6433f98992685a9679613a2cc97f3ce2 upstream.

There is no need to call dev_kfree_skb() when usb_submit_urb() fails
because can_put_echo_skb() deletes original skb and
can_free_echo_skb() deletes the cloned skb.

Fixes: 0024d8ad1639 ("can: usb_8dev: Add support for USB2CAN interface from 8 devices")
Link: https://lore.kernel.org/all/20220311080614.45229-1-hbh25y@gmail.com
Cc: stable@vger.kernel.org
Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

can: m_can: m_can_tx_handler(): fix use after free of skb

commit 2e8e79c416aae1de224c0f1860f2e3350fa171f8 upstream.

can_put_echo_skb() will clone skb then free the skb. Move the
can_put_echo_skb() for the m_can version 3.0.x directly before the
start of the xmit in hardware, similar to the 3.1.x branch.

Fixes: 80646733f11c ("can: m_can: update to support CAN FD features")
Link: https://lore.kernel.org/all/20220317081305.739554-1-mkl@pengutronix.de
Cc: stable@vger.kernel.org
Reported-by: Hangyu Hua <hbh25y@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

can: ems_usb: ems_usb_start_xmit(): fix double dev_kfree_skb() in error path

commit c70222752228a62135cee3409dccefd494a24646 upstream.

There is no need to call dev_kfree_skb() when usb_submit_urb() fails
beacause can_put_echo_skb() deletes the original skb and
can_free_echo_skb() deletes the cloned skb.

Link: https://lore.kernel.org/all/20220228083639.38183-1-hbh25y@gmail.com
Fixes: 702171adeed3 ("ems_usb: Added support for EMS CPC-USB/ARM7 CAN/USB interface")
Cc: stable@vger.kernel.org
Cc: Sebastian Haas <haas@ems-wuensche.com>
Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mtd: rawnand: protect access to rawnand devices while in suspend

commit 8cba323437a49a45756d661f500b324fc2d486fe upstream.

Prevent rawnand access while in a suspended state.

Commit 013e6292aaf5 ("mtd: rawnand: Simplify the locking") allows the
rawnand layer to return errors rather than waiting in a blocking wait.

Tested on a iMX6ULL.

Fixes: 013e6292aaf5 ("mtd: rawnand: Simplify the locking")
Signed-off-by: Sean Nyekjaer <sean@geanix.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Cc: stable@vger.kernel.org
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20220208085213.1838273-1-sean@geanix.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

spi: mxic: Fix the transmit path

commit 5fd6739e0df7e320bcac103dfb95fe75941fea17 upstream.

By working with external hardware ECC engines, we figured out that
Under certain circumstances, it is needed for the SPI controller to
check INT_TX_EMPTY and INT_RX_NOT_EMPTY in both receive and transmit
path (not only in the receive path). The delay penalty being
negligible, move this code in the common path.

Fixes: b942d80b0a39 ("spi: Add MXIC controller driver")
Cc: stable@vger.kernel.org
Suggested-by: Mason Yang <masonccyang@mxic.com.tw>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Zhengxun Li <zhengxunli@mxic.com.tw>
Reviewed-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/linux-mtd/20220127091808.1043392-10-miquel.raynal@bootlin.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

net: bnxt_ptp: fix compilation error

commit dcf500065fabe27676dfe7b4ba521a4f1e0fc8ac upstream.

The Broadcom bnxt_ptp driver does not compile with GCC 11.2.2 when
CONFIG_WERROR is enabled. The following error is generated:

drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c: In function ‘bnxt_ptp_enable’:
drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c:400:43: error: array
subscript 255 is above array bounds of ‘struct pps_pin[4]’
[-Werror=array-bounds]
  400 |  ptp->pps_info.pins[pin_id].event = BNXT_PPS_EVENT_EXTERNAL;
      |  ~~~~~~~~~~~~~~~~~~^~~~~~~~
In file included from drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c:20:
drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.h:75:24: note: while
referencing ‘pins’
   75 |         struct pps_pin pins[BNXT_MAX_TSIO_PINS];
      |                        ^~~~
cc1: all warnings being treated as errors

This is due to the function ptp_find_pin() returning a pin ID of -1 when
a valid pin is not found and this error never being checked.
Change the TSIO_PIN_VALID() function to also check that a pin ID is not
negative and use this macro in bnxt_ptp_enable() to check the result of
the calls to ptp_find_pin() to return an error early for invalid pins.
This fixes the compilation error.

Cc: <stable@vger.kernel.org>
Fixes: 9e518f25802c ("bnxt_en: 1PPS functions to configure TSIO pins")
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20220328062708.207079-1-damien.lemoal@opensource.wdc.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

pinctrl: samsung: drop pin banks references on error paths

commit 50ebd19e3585b9792e994cfa8cbee8947fe06371 upstream.

The driver iterates over its devicetree children with
for_each_child_of_node() and stores for later found node pointer.  This
has to be put in error paths to avoid leak during re-probing.

Fixes: ab663789d697 ("pinctrl: samsung: Match pin banks with their device nodes")
Cc: <stable@vger.kernel.org>
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Chanho Park <chanho61.park@samsung.com>
Link: https://lore.kernel.org/r/20220111201426.326777-2-krzysztof.kozlowski@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

pinctrl: ingenic: Fix regmap on X series SoCs

commit 9279c00fa40250e5cb23a8423dce7dbc6516a0ea upstream.

The X series Ingenic SoCs have a shadow GPIO group which is at a higher
offset than the other groups, and is used for all GPIO configuration.
The regmap did not take this offset into account and set max_register
too low, so the regmap API blocked writes to the shadow group, which
made the pinctrl driver unable to configure any pins.

Fix this by adding regmap access tables to the chip info. The way that
max_register was computed was also off by one, since max_register is an
inclusive bound, not an exclusive bound; this has been fixed.

Cc: stable@vger.kernel.org
Signed-off-by: Aidan MacDonald <aidanmacdonald.0x0@gmail.com>
Fixes: 6626a76ef857 ("pinctrl: ingenic: Add .max_register in  regmap_config")
Reviewed-by: Paul Cercueil <paul@crapouillou.net>
Link: https://lore.kernel.org/r/20220317000740.1045204-1-aidanmacdonald.0x0@gmail.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm/mlock: fix two bugs in user_shm_lock()

commit e97824ff663ce3509fe040431c713182c2f058b1 upstream.

user_shm_lock forgets to set allowed to 0 when get_ucounts fails. So the
later user_shm_unlock might do the extra dec_rlimit_ucounts. Also in the
RLIM_INFINITY case, user_shm_lock will success regardless of the value of
memlock where memblock == LONG_MAX && !capable(CAP_IPC_LOCK) should fail.
Fix all of these by changing the code to leave lock_limit at ULONG_MAX aka
RLIM_INFINITY, leave "allowed" initialized to 0 and remove the special case
of RLIM_INFINITY as nothing can be greater than ULONG_MAX.

Credit goes to Eric W. Biederman for proposing simplifying the code and
thus catching the later bug.

Fixes: d7c9e99aee48 ("Reimplement RLIMIT_MEMLOCK on top of ucounts")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: stable@vger.kernel.org
v1: https://lkml.kernel.org/r/20220310132417.41189-1-linmiaohe@huawei.com
v2: https://lkml.kernel.org/r/20220314064039.62972-1-linmiaohe@huawei.com
Link: https://lkml.kernel.org/r/20220322080918.59861-1-linmiaohe@huawei.com
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

remoteproc: Fix count check in rproc_coredump_write()

commit f89672cc3681952f2d06314981a6b45f8b0045d1 upstream.

Check count for 0, to avoid a potential underflow. Make the check the
same as the one in rproc_recovery_write().

Fixes: 3afdc59e4390 ("remoteproc: Add coredump debugfs entry")
Signed-off-by: Alistair Delva <adelva@google.com>
Cc: Rishabh Bhatnagar <rishabhb@codeaurora.org>
Cc: stable@vger.kernel.org
Cc: Ohad Ben-Cohen <ohad@wizery.com>
Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Cc: Sibi Sankar <sibis@codeaurora.org>
Cc: linux-remoteproc@vger.kernel.org
Cc: kernel-team@android.com
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220119232139.1125908-1-adelva@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

f2fs: fix to do sanity check on .cp_pack_total_block_count

commit 5b5b4f85b01604389f7a0f11ef180a725bf0e2d4 upstream.

As bughunter reported in bugzilla:

https://bugzilla.kernel.org/show_bug.cgi?id=215709

f2fs may hang when mounting a fuzzed image, the dmesg shows as below:

__filemap_get_folio+0x3a9/0x590
pagecache_get_page+0x18/0x60
__get_meta_page+0x95/0x460 [f2fs]
get_checkpoint_version+0x2a/0x1e0 [f2fs]
validate_checkpoint+0x8e/0x2a0 [f2fs]
f2fs_get_valid_checkpoint+0xd0/0x620 [f2fs]
f2fs_fill_super+0xc01/0x1d40 [f2fs]
mount_bdev+0x18a/0x1c0
f2fs_mount+0x15/0x20 [f2fs]
legacy_get_tree+0x28/0x50
vfs_get_tree+0x27/0xc0
path_mount+0x480/0xaa0
do_mount+0x7c/0xa0
__x64_sys_mount+0x8b/0xe0
do_syscall_64+0x38/0xc0
entry_SYSCALL_64_after_hwframe+0x44/0xae

The root cause is cp_pack_total_block_count field in checkpoint was fuzzed
to one, as calcuated, two cp pack block locates in the same block address,
so then read latter cp pack block, it will block on the page lock due to
the lock has already held when reading previous cp pack block, fix it by
adding sanity check for cp_pack_total_block_count.

Cc: stable@vger.kernel.org
Signed-off-by: Chao Yu <chao.yu@oppo.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

f2fs: quota: fix loop condition at f2fs_quota_sync()

commit 680af5b824a52faa819167628665804a14f0e0df upstream.

cnt should be passed to sb_has_quota_active() instead of type to check
active quota properly.

Moreover, when the type is -1, the compiler with enough inline knowledge
can discard sb_has_quota_active() check altogether, causing a NULL pointer
dereference at the following inode_lock(dqopt->files[cnt]):

[    2.796010] Unable to handle kernel NULL pointer dereference at virtual address 00000000000000a0
[    2.796024] Mem abort info:
[    2.796025]   ESR = 0x96000005
[    2.796028]   EC = 0x25: DABT (current EL), IL = 32 bits
[    2.796029]   SET = 0, FnV = 0
[    2.796031]   EA = 0, S1PTW = 0
[    2.796032] Data abort info:
[    2.796034]   ISV = 0, ISS = 0x00000005
[    2.796035]   CM = 0, WnR = 0
[    2.796046] user pgtable: 4k pages, 39-bit VAs, pgdp=00000003370d1000
[    2.796048] [00000000000000a0] pgd=0000000000000000, pud=0000000000000000
[    2.796051] Internal error: Oops: 96000005 [#1] PREEMPT SMP
[    2.796056] CPU: 7 PID: 640 Comm: f2fs_ckpt-259:7 Tainted: G S                5.4.179-arter97-r8-64666-g2f16e087f9d8 #1
[    2.796057] Hardware name: Qualcomm Technologies, Inc. Lahaina MTP lemonadep (DT)
[    2.796059] pstate: 80c00005 (Nzcv daif +PAN +UAO)
[    2.796065] pc : down_write+0x28/0x70
[    2.796070] lr : f2fs_quota_sync+0x100/0x294
[    2.796071] sp : ffffffa3f48ffc30
[    2.796073] x29: ffffffa3f48ffc30 x28: 0000000000000000
[    2.796075] x27: ffffffa3f6d718b8 x26: ffffffa415fe9d80
[    2.796077] x25: ffffffa3f7290048 x24: 0000000000000001
[    2.796078] x23: 0000000000000000 x22: ffffffa3f7290000
[    2.796080] x21: ffffffa3f72904a0 x20: ffffffa3f7290110
[    2.796081] x19: ffffffa3f77a9800 x18: ffffffc020aae038
[    2.796083] x17: ffffffa40e38e040 x16: ffffffa40e38e6d0
[    2.796085] x15: ffffffa40e38e6cc x14: ffffffa40e38e6d0
[    2.796086] x13: 00000000000004f6 x12: 00162c44ff493000
[    2.796088] x11: 0000000000000400 x10: ffffffa40e38c948
[    2.796090] x9 : 0000000000000000 x8 : 00000000000000a0
[    2.796091] x7 : 0000000000000000 x6 : 0000d1060f00002a
[    2.796093] x5 : ffffffa3f48ff718 x4 : 000000000000000d
[    2.796094] x3 : 00000000060c0000 x2 : 0000000000000001
[    2.796096] x1 : 0000000000000000 x0 : 00000000000000a0
[    2.796098] Call trace:
[    2.796100]  down_write+0x28/0x70
[    2.796102]  f2fs_quota_sync+0x100/0x294
[    2.796104]  block_operations+0x120/0x204
[    2.796106]  f2fs_write_checkpoint+0x11c/0x520
[    2.796107]  __checkpoint_and_complete_reqs+0x7c/0xd34
[    2.796109]  issue_checkpoint_thread+0x6c/0xb8
[    2.796112]  kthread+0x138/0x414
[    2.796114]  ret_from_fork+0x10/0x18
[    2.796117] Code: aa0803e0 aa1f03e1 52800022 aa0103e9 (c8e97d02)
[    2.796120] ---[ end trace 96e942e8eb6a0b53 ]---
[    2.800116] Kernel panic - not syncing: Fatal exception
[    2.800120] SMP: stopping secondary CPUs

Fixes: 9de71ede81e6 ("f2fs: quota: fix potential deadlock")
Cc: <stable@vger.kernel.org> # v5.15+
Signed-off-by: Juhyung Park <qkrwngud825@gmail.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

f2fs: fix to unlock page correctly in error path of is_alive()

commit 6d18762ed5cd549fde74fd0e05d4d87bac5a3beb upstream.

As Pavel Machek reported in below link [1]:

After commit 77900c45ee5c ("f2fs: fix to do sanity check in is_alive()"),
node page should be unlock via calling f2fs_put_page() in the error path
of is_alive(), otherwise, f2fs may hang when it tries to lock the node
page, fix it.

[1] https://lore.kernel.org/stable/20220124203637.GA19321@duo.ucw.cz/

Fixes: 77900c45ee5c ("f2fs: fix to do sanity check in is_alive()")
Cc: <stable@vger.kernel.org>
Reported-by: Pavel Machek <pavel@denx.de>
Signed-off-by: Pavel Machek <pavel@denx.de>
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

NFSD: prevent integer overflow on 32 bit systems

commit 23a9dbbe0faf124fc4c139615633b9d12a3a89ef upstream.

On a 32 bit system, the "len * sizeof(*p)" operation can have an
integer overflow.

Cc: stable@vger.kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

NFSD: prevent underflow in nfssvc_decode_writeargs()

commit 184416d4b98509fb4c3d8fc3d6dc1437896cc159 upstream.

Smatch complains:

fs/nfsd/nfsxdr.c:341 nfssvc_decode_writeargs()
warn: no lower bound on 'args->len'

Change the type to unsigned to prevent this issue.

Cc: stable@vger.kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

NFS: NFSv2/v3 clients should never be setting NFS_CAP_XATTR

commit b622ffe1d9ecbac71f0cddb52ff0831efdf8fb83 upstream.

Ensure that we always initialise the 'xattr_support' field in struct
nfs_fsinfo, so that nfs_server_set_fsinfo() doesn't declare our NFSv2/v3
client to be capable of supporting the NFSv4.2 xattr protocol by setting
the NFS_CAP_XATTR capability.

This configuration can cause nfs_do_access() to set access mode bits
that are unsupported by the NFSv3 ACCESS call, which may confuse
spec-compliant servers.

Reported-by: Olga Kornievskaia <kolga@netapp.com>
Fixes: b78ef845c35d ("NFSv4.2: query the server for extended attribute support")
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

SUNRPC: avoid race between mod_timer() and del_timer_sync()

commit 3848e96edf4788f772d83990022fa7023a233d83 upstream.

xprt_destory() claims XPRT_LOCKED and then calls del_timer_sync().
Both xprt_unlock_connect() and xprt_release() call
 ->release_xprt()
which drops XPRT_LOCKED and *then* xprt_schedule_autodisconnect()
which calls mod_timer().

This may result in mod_timer() being called *after* del_timer_sync().
When this happens, the timer may fire long after the xprt has been freed,
and run_timer_softirq() will probably crash.

The pairing of ->release_xprt() and xprt_schedule_autodisconnect() is
always called under ->transport_lock.  So if we take ->transport_lock to
call del_timer_sync(), we can be sure that mod_timer() will run first
(if it runs at all).

Cc: stable@vger.kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

HID: intel-ish-hid: Use dma_alloc_coherent for firmware update

commit f97ec5d75e9261a5da78dc28a8955b7cc0c4468b upstream.

Allocating memory with kmalloc and GPF_DMA32 is not allowed, the
allocator will ignore the attribute.

Instead, use dma_alloc_coherent() API as we allocate a small amount of
memory to transfer firmware fragment to the ISH.

On Arcada chromebook, after the patch the warning:
"Unexpected gfp: 0x4 (GFP_DMA32). Fixing up to gfp: 0xcc0 (GFP_KERNEL).  Fix your code!"
is gone. The ISH firmware is loaded properly and we can interact with
the ISH:
> ectool  --name cros_ish version
...
Build info:    arcada_ish_v2.0.3661+3c1a1c1ae0 2022-02-08 05:37:47 @localhost
Tool version:  v2.0.12300-900b03ec7f 2022-02-08 10:01:48 @localhost

Fixes: commit 91b228107da3 ("HID: intel-ish-hid: ISH firmware loader client driver")
Signed-off-by: Gwendal Grignou <gwendal@chromium.org>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

firmware: sysfb: fix platform-device leak in error path

commit 202c08914ba50dd324e42d5ad99535a89f242560 upstream.

Make sure to free the platform device also in the unlikely event that
registration fails.

Fixes: 0589e8889dce ("drivers/firmware: Add missing platform_device_put() in sysfb_create_simplefb")
Fixes: 8633ef82f101 ("drivers/firmware: consolidate EFI framebuffer setup for all arches")
Cc: stable@vger.kernel.org # 5.14
Cc: Miaoqian Lin <linmq006@gmail.com>
Cc: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20220303180519.3117-1-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

firmware: stratix10-svc: add missing callback parameter on RSU

commit b850b7a8b369322adf699ef48ceff4d902525c8c upstream.

Fix a bug whereby, the return response of parameter a1 from an
SMC call is not properly set to the callback data during an
INTEL_SIP_SMC_RSU_ERROR command.

Link: https://lore.kernel.org/lkml/20220216081513.28319-1-tien.sung.ang@intel.com
Fixes: 6b50d882d38d ("firmware: add remote status update client support")
Cc: stable@vger.kernel.org
Signed-off-by: Ang Tien Sung <tien.sung.ang@intel.com>
Signed-off-by: Dinh Nguyen <dinguyen@kernel.org>
Link: https://lore.kernel.org/r/20220223144146.399263-1-dinguyen@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Documentation: update stable tree link

commit 555d44932c67e617d89bc13c81c7efac5b51fcfa upstream.

The link to stable tree is redirected to
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git. Update
accordingly.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Sasha Levin <sashal@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: stable@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Link: https://lore.kernel.org/r/20220314113329.485372-6-bagasdotme@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Documentation: add link to stable release candidate tree

commit 587d39b260c4d090166314d64be70b1f6a26b0b5 upstream.

There is also stable release candidate tree. Mention it, however with a
warning that the tree is for testing purposes.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Sasha Levin <sashal@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: stable@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Link: https://lore.kernel.org/r/20220314113329.485372-5-bagasdotme@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

KEYS: asymmetric: properly validate hash_algo and encoding

commit 590bfb57b2328951d5833979e7ca1d5fde2e609a upstream.

It is insecure to allow arbitrary hash algorithms and signature
encodings to be used with arbitrary signature algorithms.  Notably,
ECDSA, ECRDSA, and SM2 all sign/verify raw hash values and don't
disambiguate between different hash algorithms like RSA PKCS#1 v1.5
padding does.  Therefore, they need to be restricted to certain sets of
hash algorithms (ideally just one, but in practice small sets are used).
Additionally, the encoding is an integral part of modern signature
algorithms, and is not supposed to vary.

Therefore, tighten the checks of hash_algo and encoding done by
software_key_determine_akcipher().

Also rearrange the parameters to software_key_determine_akcipher() to
put the public_key first, as this is the most important parameter and it
often determines everything else.

Fixes: 299f561a6693 ("x509: Add support for parsing x509 certs with ECDSA keys")
Fixes: 215525639631 ("X.509: support OSCCA SM2-with-SM3 certificate verification")
Fixes: 0d7a78643f69 ("crypto: ecrdsa - add EC-RDSA (GOST 34.10) algorithm")
Cc: stable@vger.kernel.org
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Vitaly Chikunov <vt@altlinux.org>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

KEYS: asymmetric: enforce that sig algo matches key algo

commit 2abc9c246e0548e52985b10440c9ea3e9f65f793 upstream.

Most callers of public_key_verify_signature(), including most indirect
callers via verify_signature() as well as pkcs7_verify_sig_chain(),
don't check that public_key_signature::pkey_algo matches
public_key::pkey_algo.  These should always match.  However, a malicious
signature could intentionally declare an unintended algorithm.  It is
essential that such signatures be rejected outright, or that the
algorithm of the *key* be used -- not the algorithm of the signature as
that would allow attackers to choose the algorithm used.

Currently, public_key_verify_signature() correctly uses the key's
algorithm when deciding which akcipher to allocate.  That's good.
However, it uses the signature's algorithm when deciding whether to do
the first step of SM2, which is incorrect.  Also, v4.19 and older
kernels used the signature's algorithm for the entire process.

Prevent such errors by making public_key_verify_signature() enforce that
the signature's algorithm (if given) matches the key's algorithm.

Also remove two checks of this done by callers, which are now redundant.

Cc: stable@vger.kernel.org
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Vitaly Chikunov <vt@altlinux.org>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

KEYS: fix length validation in keyctl_pkey_params_get_2()

commit c51abd96837f600d8fd940b6ab8e2da578575504 upstream.

In many cases, keyctl_pkey_params_get_2() is validating the user buffer
lengths against the wrong algorithm properties.  Fix it to check against
the correct properties.

Probably this wasn't noticed before because for all asymmetric keys of
the "public_key" subtype, max_data_size == max_sig_size == max_enc_size
== max_dec_size.  However, this isn't necessarily true for the
"asym_tpm" subtype (it should be, but it's not strictly validated).  Of
course, future key types could have different values as well.

Fixes: 00d60fd3b932 ("KEYS: Provide keyctls to drive the new key type ops for asymmetric keys [ver #2]")
Cc: <stable@vger.kernel.org> # v4.20+
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

cifs: we do not need a spinlock around the tree access during umount

commit 9a14b65d590105d393b63f5320e1594edda7c672 upstream.

Remove the spinlock around the tree traversal as we are calling possibly
sleeping functions.
We do not need a spinlock here as there will be no modifications to this
tree at this point.

This prevents warnings like this to occur in dmesg:
[  653.774996] BUG: sleeping function called from invalid context at kernel/loc\
king/mutex.c:280
[  653.775088] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1827, nam\
e: umount
[  653.775152] preempt_count: 1, expected: 0
[  653.775191] CPU: 0 PID: 1827 Comm: umount Tainted: G        W  OE     5.17.0\
-rc7-00006-g4eb628dd74df #135
[  653.775195] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.14.0-\
1.fc33 04/01/2014
[  653.775197] Call Trace:
[  653.775199]  <TASK>
[  653.775202]  dump_stack_lvl+0x34/0x44
[  653.775209]  __might_resched.cold+0x13f/0x172
[  653.775213]  mutex_lock+0x75/0xf0
[  653.775217]  ? __mutex_lock_slowpath+0x10/0x10
[  653.775220]  ? _raw_write_lock_irq+0xd0/0xd0
[  653.775224]  ? dput+0x6b/0x360
[  653.775228]  cifs_kill_sb+0xff/0x1d0 [cifs]
[  653.775285]  deactivate_locked_super+0x85/0x130
[  653.775289]  cleanup_mnt+0x32c/0x4d0
[  653.775292]  ? path_umount+0x228/0x380
[  653.775296]  task_work_run+0xd8/0x180
[  653.775301]  exit_to_user_mode_loop+0x152/0x160
[  653.775306]  exit_to_user_mode_prepare+0x89/0xd0
[  653.775315]  syscall_exit_to_user_mode+0x12/0x30
[  653.775322]  do_syscall_64+0x48/0x90
[  653.775326]  entry_SYSCALL_64_after_hwframe+0x44/0xae

Fixes: 187af6e98b44e5d8f25e1d41a92db138eb54416f ("cifs: fix handlecache and multiuser")
Reported-by: kernel test robot <oliver.sang@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>