s390/idle: mark arch_cpu_idle() noinstr

[ Upstream commit a9cbc1b471d291c865907542394f1c483b93a811 ]

linux-next commit ("cpuidle: tracing: Warn about !rcu_is_watching()")
adds a new warning which hits on s390's arch_cpu_idle() function:

RCU not on for: arch_cpu_idle+0x0/0x28
WARNING: CPU: 2 PID: 0 at include/linux/trace_recursion.h:162 arch_ftrace_ops_list_func+0x24c/0x258
Modules linked in:
CPU: 2 PID: 0 Comm: swapper/2 Not tainted 6.2.0-rc6-next-20230202 #4
Hardware name: IBM 8561 T01 703 (z/VM 7.3.0)
Krnl PSW : 0404d00180000000 00000000002b55c0 (arch_ftrace_ops_list_func+0x250/0x258)
           R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:3 CC:1 PM:0 RI:0 EA:3
Krnl GPRS: c0000000ffffbfff 0000000080000002 0000000000000026 0000000000000000
           0000037ffffe3a28 0000037ffffe3a20 0000000000000000 0000000000000000
           0000000000000000 0000000000f4acf6 00000000001044f0 0000037ffffe3cb0
           0000000000000000 0000000000000000 00000000002b55bc 0000037ffffe3bb8
Krnl Code: 00000000002b55b0: c02000840051        larl    %r2,0000000001335652
           00000000002b55b6: c0e5fff512d1        brasl   %r14,0000000000157b58
          #00000000002b55bc: af000000            mc      0,0
          >00000000002b55c0: a7f4ffe7            brc     15,00000000002b558e
           00000000002b55c4: 0707                bcr     0,%r7
           00000000002b55c6: 0707                bcr     0,%r7
           00000000002b55c8: eb6ff0480024        stmg    %r6,%r15,72(%r15)
           00000000002b55ce: b90400ef            lgr     %r14,%r15
Call Trace:
 [<00000000002b55c0>] arch_ftrace_ops_list_func+0x250/0x258
([<00000000002b55bc>] arch_ftrace_ops_list_func+0x24c/0x258)
 [<0000000000f5f0fc>] ftrace_common+0x1c/0x20
 [<00000000001044f6>] arch_cpu_idle+0x6/0x28
 [<0000000000f4acf6>] default_idle_call+0x76/0x128
 [<00000000001cc374>] do_idle+0xf4/0x1b0
 [<00000000001cc6ce>] cpu_startup_entry+0x36/0x40
 [<0000000000119d00>] smp_start_secondary+0x140/0x150
 [<0000000000f5d2ae>] restart_int_handler+0x6e/0x90

Mark arch_cpu_idle() noinstr like all other architectures with
CONFIG_ARCH_WANTS_NO_INSTR (should) have it to fix this.

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

uaccess: Add minimum bounds check on kernel buffer size

[ Upstream commit 04ffde1319a715bd0550ded3580d4ea3bc003776 ]

While there is logic about the difference between ksize and usize,
copy_struct_from_user() didn't check the size of the destination buffer
(when it was known) against ksize. Add this check so there is an upper
bounds check on the possible memset() call, otherwise lower bounds
checks made by callers will trigger bounds warnings under -Warray-bounds.
Seen under GCC 13:

In function 'copy_struct_from_user',
    inlined from 'iommufd_fops_ioctl' at
../drivers/iommu/iommufd/main.c:333:8:
../include/linux/fortify-string.h:59:33: warning: '__builtin_memset' offset [57, 4294967294] is out of the bounds [0, 56] of object 'buf' with type 'union ucmd_buffer' [-Warray-bounds=]
   59 | #define __underlying_memset     __builtin_memset
      |                                 ^
../include/linux/fortify-string.h:453:9: note: in expansion of macro '__underlying_memset'
  453 |         __underlying_memset(p, c, __fortify_size); \
      |         ^~~~~~~~~~~~~~~~~~~
../include/linux/fortify-string.h:461:25: note: in expansion of macro '__fortify_memset_chk'
  461 | #define memset(p, c, s) __fortify_memset_chk(p, c, s, \
      |                         ^~~~~~~~~~~~~~~~~~~~
../include/linux/uaccess.h:334:17: note: in expansion of macro 'memset'
  334 |                 memset(dst + size, 0, rest);
      |                 ^~~~~~
../drivers/iommu/iommufd/main.c: In function 'iommufd_fops_ioctl':
../drivers/iommu/iommufd/main.c:311:27: note: 'buf' declared here
  311 |         union ucmd_buffer buf;
      |                           ^~~

Cc: Christian Brauner <brauner@kernel.org>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Dinh Nguyen <dinguyen@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Alexander Potapenko <glider@google.com>
Acked-by: Aleksa Sarai <cyphar@cyphar.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/lkml/20230203193523.never.667-kees@kernel.org/
Signed-off-by: Sasha Levin <sashal@kernel.org>

coda: Avoid partial allocation of sig_inputArgs

[ Upstream commit 48df133578c70185a95a49390d42df1996ddba2a ]

GCC does not like having a partially allocated object, since it cannot
reason about it for bounds checking when it is passed to other code.
Instead, fully allocate sig_inputArgs. (Alternatively, sig_inputArgs
should be defined as a struct coda_in_hdr, if it is actually not using
any other part of the union.) Seen under GCC 13:

../fs/coda/upcall.c: In function 'coda_upcall':
../fs/coda/upcall.c:801:22: warning: array subscript 'union inputArgs[0]' is partly outside array bounds of 'unsigned char[20]' [-Warray-bounds=]
  801 |         sig_inputArgs->ih.opcode = CODA_SIGNAL;
      |                      ^~

Cc: Jan Harkes <jaharkes@cs.cmu.edu>
Cc: coda@cs.cmu.edu
Cc: codalist@coda.cs.cmu.edu
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20230127223921.never.882-kees@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5: fw_tracer: Fix debug print

[ Upstream commit 988c2352273997a242f15c4fc3711773515006a2 ]

The debug message specify tdsn, but takes as an argument the
tmsn. The correct argument is tmsn, hence, fix the print.

Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: video: Fix Lenovo Ideapad Z570 DMI match

[ Upstream commit 2d11eae42d52a131f06061015e49dc0f085c5bfc ]

Multiple Ideapad Z570 variants need acpi_backlight=native to force native
use on these pre Windows 8 machines since acpi_video backlight control
does not work here.

The original DMI quirk matches on a product_name of "102434U" but other
variants may have different product_name-s such as e.g. "1024D9U".

Move to checking product_version instead as is more or less standard for
Lenovo DMI quirks for similar reasons.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: dma: free rx_head in mt76_dma_rx_cleanup

[ Upstream commit 1b88b47e898edef0e56e3a2f4e49f052a136153d ]

Free rx_head skb in mt76_dma_rx_cleanup routine in order to avoid
possible memory leak at module unload.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tools/power/x86/intel-speed-select: Add Emerald Rapid quirk

[ Upstream commit 61f9fdcdcd01f9a996b6db4e7092fcdfe8414ad5 ]

Need memory frequency quirk as Sapphire Rapids in Emerald Rapids.
So add Emerald Rapids CPU model check in is_spr_platform().

Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: Subject, changelog and code edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

gcc-plugins: drop -std=gnu++11 to fix GCC 13 build

[ Upstream commit 5a6b64adc18d9adfb497a529ff004d59b6df151f ]

The latest GCC 13 snapshot (13.0.1 20230129) gives the following:
```
cc1: error: cannot load plugin ./scripts/gcc-plugins/randomize_layout_plugin.so
 :./scripts/gcc-plugins/randomize_layout_plugin.so: undefined symbol: tree_code_type
```

This ends up being because of https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=b0241ce6e37031
upstream in GCC which changes the visibility of some types used by the kernel's
plugin infrastructure like tree_code_type.

After discussion with the GCC folks, we found that the kernel needs to be building
plugins with the same flags used to build GCC - and GCC defaults to gnu++17
right now. The minimum GCC version needed to build the kernel is GCC 5.1
and GCC 5.1 already defaults to gnu++14 anyway, so just drop the flag, as
all GCCs that could be used to build GCC already default to an acceptable
version which was >= the version we forced via flags until now.

Bug: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108634
Signed-off-by: Sam James <sam@gentoo.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20230201230009.2252783-1-sam@gentoo.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

can: isotp: check CAN address family in isotp_bind()

[ Upstream commit c6adf659a8ba85913e16a571d5a9bcd17d3d1234 ]

Add missing check to block non-AF_CAN binds.

Syzbot created some code which matched the right sockaddr struct size
but used AF_XDP (0x2C) instead of AF_CAN (0x1D) in the address family
field:

bind$xdp(r2, &(0x7f0000000540)={0x2c, 0x0, r4, 0x0, r2}, 0x10)
                                ^^^^
This has no funtional impact but the userspace should be notified about
the wrong address family field content.

Link: https://syzkaller.appspot.com/text?tag=CrashLog&x=11ff9d8c480000
Reported-by: syzbot+5aed6c3aaba661f5b917@syzkaller.appspotmail.com
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
Link: https://lore.kernel.org/all/20230104201844.13168-1-socketcan@hartkopp.net
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: nf_tables: NULL pointer dereference in nf_tables_updobj()

[ Upstream commit dac7f50a45216d652887fb92d6cd3b7ca7f006ea ]

static analyzer detect null pointer dereference case for 'type'
function __nft_obj_type_get() can return NULL value which require to handle
if type is NULL pointer return -ENOENT.

This is a theoretical issue, since an existing object has a type, but
better add this failsafe check.

Signed-off-by: Alok Tiwari <alok.a.tiwari@oracle.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/mm,ptdump: avoid Kasan vs Memcpy Real markers swapping

[ Upstream commit 05178996e1a77e2a4664536e6d101a086a905034 ]

---[ Real Memory Copy Area Start ]---
0x001bfffffffff000-0x001c000000000000         4K PTE I
---[ Kasan Shadow Start ]---
---[ Real Memory Copy Area End ]---
0x001c000000000000-0x001c000200000000         8G PMD RW NX
...
---[ Kasan Shadow End ]---

ptdump does a stable sort of markers. Move kasan markers after
memcpy real to avoid swapping.

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

m68k: Check syscall_trace_enter() return code

[ Upstream commit 2ca8a1de4437f21562e57f9ac123914747a8e7a1 ]

Check return code of syscall_trace_enter(), and skip syscall
if -1. Return code will be left at what had been set by
ptrace or seccomp (in regs->d0).

No regression seen in testing with strace on ARAnyM.

Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Link: https://lore.kernel.org/r/20230112035529.13521-2-schmitzmic@gmail.com
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: bcmgenet: Add a check for oversized packets

[ Upstream commit 5c0862c2c962052ed5055220a00ac1cefb92fbcd ]

Occasionnaly we may get oversized packets from the hardware which
exceed the nomimal 2KiB buffer size we allocate SKBs with. Add an early
check which drops the packet to avoid invoking skb_over_panic() and move
on to processing the next packet.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: hisilicon: Wipe entire pool on error

[ Upstream commit aa85923a954e7704bc9d3847dabeb8540aa98d13 ]

To work around a Clang __builtin_object_size bug that shows up under
CONFIG_FORTIFY_SOURCE and UBSAN_BOUNDS, move the per-loop-iteration
mem_block wipe into a single wipe of the entire pool structure after
the loop.

Reported-by: Nathan Chancellor <nathan@kernel.org>
Link: https://github.com/ClangBuiltLinux/linux/issues/1780
Cc: Weili Qian <qianweili@huawei.com>
Cc: Zhou Wang <wangzhou1@hisilicon.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: linux-crypto@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Tested-by: Nathan Chancellor <nathan@kernel.org> # build
Link: https://lore.kernel.org/r/20230106041945.never.831-kees@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

clocksource: Suspend the watchdog temporarily when high read latency detected

[ Upstream commit b7082cdfc464bf9231300605d03eebf943dda307 ]

Bugs have been reported on 8 sockets x86 machines in which the TSC was
wrongly disabled when the system is under heavy workload.

 [ 818.380354] clocksource: timekeeping watchdog on CPU336: hpet wd-wd read-back delay of 1203520ns
 [ 818.436160] clocksource: wd-tsc-wd read-back delay of 181880ns, clock-skew test skipped!
 [ 819.402962] clocksource: timekeeping watchdog on CPU338: hpet wd-wd read-back delay of 324000ns
 [ 819.448036] clocksource: wd-tsc-wd read-back delay of 337240ns, clock-skew test skipped!
 [ 819.880863] clocksource: timekeeping watchdog on CPU339: hpet read-back delay of 150280ns, attempt 3, marking unstable
 [ 819.936243] tsc: Marking TSC unstable due to clocksource watchdog
 [ 820.068173] TSC found unstable after boot, most likely due to broken BIOS. Use 'tsc=unstable'.
 [ 820.092382] sched_clock: Marking unstable (818769414384, 1195404998)
 [ 820.643627] clocksource: Checking clocksource tsc synchronization from CPU 267 to CPUs 0,4,25,70,126,430,557,564.
 [ 821.067990] clocksource: Switched to clocksource hpet

This can be reproduced by running memory intensive 'stream' tests,
or some of the stress-ng subcases such as 'ioport'.

The reason for these issues is the when system is under heavy load, the
read latency of the clocksources can be very high.  Even lightweight TSC
reads can show high latencies, and latencies are much worse for external
clocksources such as HPET or the APIC PM timer.  These latencies can
result in false-positive clocksource-unstable determinations.

These issues were initially reported by a customer running on a production
system, and this problem was reproduced on several generations of Xeon
servers, especially when running the stress-ng test.  These Xeon servers
were not production systems, but they did have the latest steppings
and firmware.

Given that the clocksource watchdog is a continual diagnostic check with
frequency of twice a second, there is no need to rush it when the system
is under heavy load.  Therefore, when high clocksource read latencies
are detected, suspend the watchdog timer for 5 minutes.

Signed-off-by: Feng Tang <feng.tang@intel.com>
Acked-by: Waiman Long <longman@redhat.com>
Cc: John Stultz <jstultz@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Stephen Boyd <sboyd@kernel.org>
Cc: Feng Tang <feng.tang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal: intel: intel_pch: Add support for Wellsburg PCH

[ Upstream commit 40dc1929089fc844ea06d9f8bdb6211ed4517c2e ]

Add the PCI ID for the Wellsburg C610 series chipset PCH.

The driver can read the temperature from the Wellsburg PCH with only
the PCI ID added and no other modifications.

Signed-off-by: Tim Zimmermann <tim@linux4.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf, docs: Fix modulo zero, division by zero, overflow, and underflow

[ Upstream commit 0eb9d19e2201068260e439a5c96dc85f9f3722a2 ]

Fix modulo zero, division by zero, overflow, and underflow. Also clarify how
a negative immediate value is used in unsigned division.

Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230124001218.827-1-dthaler1968@googlemail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: Don't build ACPICA with '-Os'

[ Upstream commit 8f9e0a52810dd83406c768972d022c37e7a18f1f ]

The ACPICA code has been built with '-Os' since the beginning of git
history, though there's no explanatory comment as to why.

This is unfortunate as GCC drops the alignment specificed by
'-falign-functions=N' when '-Os' is used, as reported in GCC bug 88345:

  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88345

This prevents CONFIG_FUNCTION_ALIGNMENT and
CONFIG_DEBUG_FORCE_FUNCTION_ALIGN_64B from having their expected effect
on the ACPICA code. This is doubly unfortunate as in subsequent patches
arm64 will depend upon CONFIG_FUNCTION_ALIGNMENT for its ftrace
implementation.

Drop the '-Os' flag when building the ACPICA code. With this removed,
the code builds cleanly and works correctly in testing so far.

I've tested this by selecting CONFIG_DEBUG_FORCE_FUNCTION_ALIGN_64B=y,
building and booting a kernel using ACPI, and looking for misaligned
text symbols:

* arm64:

  Before, v6.2-rc3:
    # uname -rm
    6.2.0-rc3 aarch64
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | wc -l
    5009

  Before, v6.2-rc3 + fixed __cold:
    # uname -rm
    6.2.0-rc3-00001-g2a2bedf8bfa9 aarch64
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | wc -l
    919

  After:
    # uname -rm
    6.2.0-rc3-00002-g267bddc38572 aarch64
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | wc -l
    323
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | grep acpi | wc -l
    0

* x86_64:

  Before, v6.2-rc3:
    # uname -rm
    6.2.0-rc3 x86_64
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | wc -l
    11537

  Before, v6.2-rc3 + fixed __cold:
    # uname -rm
    6.2.0-rc3-00001-g2a2bedf8bfa9 x86_64
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | wc -l
    2805

  After:
    # uname -rm
    6.2.0-rc3-00002-g267bddc38572 x86_64
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | wc -l
    1357
    # grep ' [Tt] ' /proc/kallsyms | grep -iv '[048c]0 [Tt] ' | grep acpi | wc -l
    0

With the patch applied, the remaining unaligned text labels are a
combination of static call trampolines and labels in assembly, which can
be dealt with in subsequent patches.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Florent Revest <revest@chromium.org>
Cc: Len Brown <lenb@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Moore <robert.moore@intel.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Will Deacon <will@kernel.org>
Cc: linux-acpi@vger.kernel.org
Link: https://lore.kernel.org/r/20230123134603.1064407-4-mark.rutland@arm.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ice: add missing checks for PF vsi type

[ Upstream commit 6a8d013e904ad9a66706fcc926ec9993bed7d190 ]

There were a few places we had missed checking the VSI type to make sure
it was definitely a PF VSI, before calling setup functions intended only
for the PF VSI.

This doesn't fix any explicit bugs but cleans up the code in a few
places and removes one explicit != vsi->type check that can be
superseded by this code (it's a super set)

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ice: restrict PTP HW clock freq adjustments to 100, 000, 000 PPB

[ Upstream commit 8aa4318c3a122b8670bc09af142de3872ca63b88 ]

The PHY provides only 39b timestamp. With current timing
implementation, we discard lower 7b, leaving 32b timestamp.
The driver reconstructs the full 64b timestamp by correlating the
32b timestamp with cached_time for performance. The reconstruction
algorithm does both forward & backward interpolation.

The 32b timeval has overflow duration of 2^32 counts ~= 4.23 second.
Due to interpolation in both direction, its now ~= 2.125 second
IIRC, going with at least half a duration, the cached_time is updated
with periodic thread of 1 second (worst-case) periodicity.

But the 1 second periodicity is based on System-timer.
With PPB adjustments, if the 1588 timers increments at say
double the rate, (2s in-place of 1s), the Nyquist rate/half duration
sampling/update of cached_time with 1 second periodic thread will
lead to incorrect interpolations.

Hence we should restrict the PPB adjustments to at least half duration
of cached_time update which translates to 500,000,000 PPB.

Since the periodicity of the cached-time system thread can vary,
it is good to have some buffer time and considering practicality of
PPB adjustments, limiting the max_adj to 100,000,000.

Signed-off-by: Siddaraju DH <siddaraju.dh@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

inet: fix fast path in __inet_hash_connect()

[ Upstream commit 21cbd90a6fab7123905386985e3e4a80236b8714 ]

__inet_hash_connect() has a fast path taken if sk_head(&tb->owners) is
equal to the sk parameter.
sk_head() returns the hlist_entry() with respect to the sk_node field.
However entries in the tb->owners list are inserted with respect to the
sk_bind_node field with sk_add_bind_node().
Thus the check would never pass and the fast path never execute.

This fast path has never been executed or tested as this bug seems
to be present since commit 1da177e4c3f4 ("Linux-2.6.12-rc2"), thus
remove it to reduce code complexity.

Signed-off-by: Pietro Borrello <borrello@diag.uniroma1.it>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230112-inet_hash_connect_bind_head-v3-1-b591fd212b93@diag.uniroma1.it
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt7601u: fix an integer underflow

[ Upstream commit 803f3176c5df3b5582c27ea690f204abb60b19b9 ]

Fix an integer underflow that leads to a null pointer dereference in
'mt7601u_rx_skb_from_seg()'. The variable 'dma_len' in the URB packet
could be manipulated, which could trigger an integer underflow of
'seg_len' in 'mt7601u_rx_process_seg()'. This underflow subsequently
causes the 'bad_frame' checks in 'mt7601u_rx_skb_from_seg()' to be
bypassed, eventually leading to a dereference of the pointer 'p', which
is a null pointer.

Ensure that 'dma_len' is greater than 'min_seg_len'.

Found by a modified version of syzkaller.

KASAN: null-ptr-deref in range [0x0000000000000008-0x000000000000000f]
CPU: 0 PID: 12 Comm: ksoftirqd/0 Tainted: G        W  O      5.14.0+
#139
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
RIP: 0010:skb_add_rx_frag+0x143/0x370
Code: e2 07 83 c2 03 38 ca 7c 08 84 c9 0f 85 86 01 00 00 4c 8d 7d 08 44
89 68 08 48 b8 00 00 00 00 00 fc ff df 4c 89 fa 48 c1 ea 03 <80> 3c 02
00 0f 85 cd 01 00 00 48 8b 45 08 a8 01 0f 85 3d 01 00 00
RSP: 0018:ffffc900000cfc90 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: ffff888115520dc0 RCX: 0000000000000000
RDX: 0000000000000001 RSI: ffff8881118430c0 RDI: ffff8881118430f8
RBP: 0000000000000000 R08: 0000000000000e09 R09: 0000000000000010
R10: ffff888111843017 R11: ffffed1022308602 R12: 0000000000000000
R13: 0000000000000e09 R14: 0000000000000010 R15: 0000000000000008
FS:  0000000000000000(0000) GS:ffff88811a800000(0000)
knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000004035af40 CR3: 00000001157f2000 CR4: 0000000000750ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
 mt7601u_rx_tasklet+0xc73/0x1270
 ? mt7601u_submit_rx_buf.isra.0+0x510/0x510
 ? tasklet_action_common.isra.0+0x79/0x2f0
 tasklet_action_common.isra.0+0x206/0x2f0
 __do_softirq+0x1b5/0x880
 ? tasklet_unlock+0x30/0x30
 run_ksoftirqd+0x26/0x50
 smpboot_thread_fn+0x34f/0x7d0
 ? smpboot_register_percpu_thread+0x370/0x370
 kthread+0x3a1/0x480
 ? set_kthread_struct+0x120/0x120
 ret_from_fork+0x1f/0x30
Modules linked in: 88XXau(O) 88x2bu(O)
---[ end trace 57f34f93b4da0f9b ]---
RIP: 0010:skb_add_rx_frag+0x143/0x370
Code: e2 07 83 c2 03 38 ca 7c 08 84 c9 0f 85 86 01 00 00 4c 8d 7d 08 44
89 68 08 48 b8 00 00 00 00 00 fc ff df 4c 89 fa 48 c1 ea 03 <80> 3c 02
00 0f 85 cd 01 00 00 48 8b 45 08 a8 01 0f 85 3d 01 00 00
RSP: 0018:ffffc900000cfc90 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: ffff888115520dc0 RCX: 0000000000000000
RDX: 0000000000000001 RSI: ffff8881118430c0 RDI: ffff8881118430f8
RBP: 0000000000000000 R08: 0000000000000e09 R09: 0000000000000010
R10: ffff888111843017 R11: ffffed1022308602 R12: 0000000000000000
R13: 0000000000000e09 R14: 0000000000000010 R15: 0000000000000008
FS:  0000000000000000(0000) GS:ffff88811a800000(0000)
knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000004035af40 CR3: 00000001157f2000 CR4: 0000000000750ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
PKRU: 55555554

Signed-off-by: Jisoo Jang <jisoo.jang@yonsei.ac.kr>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20221229092906.2328282-1-jisoo.jang@yonsei.ac.kr
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: brcmfmac: ensure CLM version is null-terminated to prevent stack-out-of-bounds

[ Upstream commit 660145d708be52f946a82e5b633c020f58f996de ]

Fix a stack-out-of-bounds read in brcmfmac that occurs
when 'buf' that is not null-terminated is passed as an argument of
strreplace() in brcmf_c_preinit_dcmds(). This buffer is filled with
a CLM version string by memcpy() in brcmf_fil_iovar_data_get().
Ensure buf is null-terminated.

Found by a modified version of syzkaller.

[   33.004414][ T1896] brcmfmac: brcmf_c_process_clm_blob: no clm_blob available (err=-2), device may have limited channels available
[   33.013486][ T1896] brcmfmac: brcmf_c_preinit_dcmds: Firmware: BCM43236/3 wl0: Nov 30 2011 17:33:42 version 5.90.188.22
[   33.021554][ T1896] ==================================================================
[   33.022379][ T1896] BUG: KASAN: stack-out-of-bounds in strreplace+0xf2/0x110
[   33.023122][ T1896] Read of size 1 at addr ffffc90001d6efc8 by task kworker/0:2/1896
[   33.023852][ T1896]
[   33.024096][ T1896] CPU: 0 PID: 1896 Comm: kworker/0:2 Tainted: G           O      5.14.0+ #132
[   33.024927][ T1896] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
[   33.026065][ T1896] Workqueue: usb_hub_wq hub_event
[   33.026581][ T1896] Call Trace:
[   33.026896][ T1896]  dump_stack_lvl+0x57/0x7d
[   33.027372][ T1896]  print_address_description.constprop.0.cold+0xf/0x334
[   33.028037][ T1896]  ? strreplace+0xf2/0x110
[   33.028403][ T1896]  ? strreplace+0xf2/0x110
[   33.028807][ T1896]  kasan_report.cold+0x83/0xdf
[   33.029283][ T1896]  ? strreplace+0xf2/0x110
[   33.029666][ T1896]  strreplace+0xf2/0x110
[   33.029966][ T1896]  brcmf_c_preinit_dcmds+0xab1/0xc40
[   33.030351][ T1896]  ? brcmf_c_set_joinpref_default+0x100/0x100
[   33.030787][ T1896]  ? rcu_read_lock_sched_held+0xa1/0xd0
[   33.031223][ T1896]  ? rcu_read_lock_bh_held+0xb0/0xb0
[   33.031661][ T1896]  ? lock_acquire+0x19d/0x4e0
[   33.032091][ T1896]  ? find_held_lock+0x2d/0x110
[   33.032605][ T1896]  ? brcmf_usb_deq+0x1a7/0x260
[   33.033087][ T1896]  ? brcmf_usb_rx_fill_all+0x5a/0xf0
[   33.033582][ T1896]  brcmf_attach+0x246/0xd40
[   33.034022][ T1896]  ? wiphy_new_nm+0x1476/0x1d50
[   33.034383][ T1896]  ? kmemdup+0x30/0x40
[   33.034722][ T1896]  brcmf_usb_probe+0x12de/0x1690
[   33.035223][ T1896]  ? brcmf_usbdev_qinit.constprop.0+0x470/0x470
[   33.035833][ T1896]  usb_probe_interface+0x25f/0x710
[   33.036315][ T1896]  really_probe+0x1be/0xa90
[   33.036656][ T1896]  __driver_probe_device+0x2ab/0x460
[   33.037026][ T1896]  ? usb_match_id.part.0+0x88/0xc0
[   33.037383][ T1896]  driver_probe_device+0x49/0x120
[   33.037790][ T1896]  __device_attach_driver+0x18a/0x250
[   33.038300][ T1896]  ? driver_allows_async_probing+0x120/0x120
[   33.038986][ T1896]  bus_for_each_drv+0x123/0x1a0
[   33.039906][ T1896]  ? bus_rescan_devices+0x20/0x20
[   33.041412][ T1896]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   33.041861][ T1896]  ? trace_hardirqs_on+0x1c/0x120
[   33.042330][ T1896]  __device_attach+0x207/0x330
[   33.042664][ T1896]  ? device_bind_driver+0xb0/0xb0
[   33.043026][ T1896]  ? kobject_uevent_env+0x230/0x12c0
[   33.043515][ T1896]  bus_probe_device+0x1a2/0x260
[   33.043914][ T1896]  device_add+0xa61/0x1ce0
[   33.044227][ T1896]  ? __mutex_unlock_slowpath+0xe7/0x660
[   33.044891][ T1896]  ? __fw_devlink_link_to_suppliers+0x550/0x550
[   33.045531][ T1896]  usb_set_configuration+0x984/0x1770
[   33.046051][ T1896]  ? kernfs_create_link+0x175/0x230
[   33.046548][ T1896]  usb_generic_driver_probe+0x69/0x90
[   33.046931][ T1896]  usb_probe_device+0x9c/0x220
[   33.047434][ T1896]  really_probe+0x1be/0xa90
[   33.047760][ T1896]  __driver_probe_device+0x2ab/0x460
[   33.048134][ T1896]  driver_probe_device+0x49/0x120
[   33.048516][ T1896]  __device_attach_driver+0x18a/0x250
[   33.048910][ T1896]  ? driver_allows_async_probing+0x120/0x120
[   33.049437][ T1896]  bus_for_each_drv+0x123/0x1a0
[   33.049814][ T1896]  ? bus_rescan_devices+0x20/0x20
[   33.050164][ T1896]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   33.050579][ T1896]  ? trace_hardirqs_on+0x1c/0x120
[   33.050936][ T1896]  __device_attach+0x207/0x330
[   33.051399][ T1896]  ? device_bind_driver+0xb0/0xb0
[   33.051888][ T1896]  ? kobject_uevent_env+0x230/0x12c0
[   33.052314][ T1896]  bus_probe_device+0x1a2/0x260
[   33.052688][ T1896]  device_add+0xa61/0x1ce0
[   33.053121][ T1896]  ? __fw_devlink_link_to_suppliers+0x550/0x550
[   33.053568][ T1896]  usb_new_device.cold+0x463/0xf66
[   33.053953][ T1896]  ? hub_disconnect+0x400/0x400
[   33.054313][ T1896]  ? rwlock_bug.part.0+0x90/0x90
[   33.054661][ T1896]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   33.055094][ T1896]  hub_event+0x10d5/0x3330
[   33.055530][ T1896]  ? hub_port_debounce+0x280/0x280
[   33.055934][ T1896]  ? __lock_acquire+0x1671/0x5790
[   33.056387][ T1896]  ? wq_calc_node_cpumask+0x170/0x2a0
[   33.056924][ T1896]  ? lock_release+0x640/0x640
[   33.057383][ T1896]  ? rcu_read_lock_sched_held+0xa1/0xd0
[   33.057916][ T1896]  ? rcu_read_lock_bh_held+0xb0/0xb0
[   33.058402][ T1896]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   33.059019][ T1896]  process_one_work+0x873/0x13e0
[   33.059488][ T1896]  ? lock_release+0x640/0x640
[   33.059932][ T1896]  ? pwq_dec_nr_in_flight+0x320/0x320
[   33.060446][ T1896]  ? rwlock_bug.part.0+0x90/0x90
[   33.060898][ T1896]  worker_thread+0x8b/0xd10
[   33.061348][ T1896]  ? __kthread_parkme+0xd9/0x1d0
[   33.061810][ T1896]  ? process_one_work+0x13e0/0x13e0
[   33.062288][ T1896]  kthread+0x379/0x450
[   33.062660][ T1896]  ? _raw_spin_unlock_irq+0x24/0x30
[   33.063148][ T1896]  ? set_kthread_struct+0x100/0x100
[   33.063606][ T1896]  ret_from_fork+0x1f/0x30
[   33.064070][ T1896]
[   33.064313][ T1896]
[   33.064545][ T1896] addr ffffc90001d6efc8 is located in stack of task kworker/0:2/1896 at offset 512 in frame:
[   33.065478][ T1896]  brcmf_c_preinit_dcmds+0x0/0xc40
[   33.065973][ T1896]
[   33.066191][ T1896] this frame has 4 objects:
[   33.066614][ T1896]  [48, 56) 'ptr'
[   33.066618][ T1896]  [80, 148) 'revinfo'
[   33.066957][ T1896]  [192, 210) 'eventmask'
[   33.067338][ T1896]  [256, 512) 'buf'
[   33.067742][ T1896]
[   33.068304][ T1896] Memory state around the buggy address:
[   33.068838][ T1896]  ffffc90001d6ee80: f2 00 00 02 f2 f2 f2 f2 f2 00 00 00 00 00 00 00
[   33.069545][ T1896]  ffffc90001d6ef00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[   33.070626][ T1896] >ffffc90001d6ef80: 00 00 00 00 00 00 00 00 00 f3 f3 f3 f3 f3 f3 f3
[   33.072052][ T1896]                                               ^
[   33.073043][ T1896]  ffffc90001d6f000: f3 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[   33.074230][ T1896]  ffffc90001d6f080: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[   33.074914][ T1896] ==================================================================
[   33.075713][ T1896] Disabling lock debugging due to kernel taint

Reviewed-by: Arend van Spriel<arend.vanspriel@broadcom.com>
Signed-off-by: Jisoo Jang <jisoo.jang@yonsei.ac.kr>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20221230075139.56591-1-jisoo.jang@yonsei.ac.kr
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpftool: Always disable stack protection for BPF objects

[ Upstream commit 878625e1c7a10dfbb1fdaaaae2c4d2a58fbce627 ]

When the clang toolchain has stack protection enabled in order to be
consistent with gcc - which just happens to be the case on Gentoo -
the bpftool build fails:

  [...]
  clang \
-I. \
-I/tmp/portage/dev-util/bpftool-6.0.12/work/linux-6.0/tools/include/uapi/ \
-I/tmp/portage/dev-util/bpftool-6.0.12/work/linux-6.0/tools/bpf/bpftool/bootstrap/libbpf/include \
-g -O2 -Wall -target bpf -c skeleton/pid_iter.bpf.c -o pid_iter.bpf.o
  clang \
-I. \
-I/tmp/portage/dev-util/bpftool-6.0.12/work/linux-6.0/tools/include/uapi/ \
-I/tmp/portage/dev-util/bpftool-6.0.12/work/linux-6.0/tools/bpf/bpftool/bootstrap/libbpf/include \
-g -O2 -Wall -target bpf -c skeleton/profiler.bpf.c -o profiler.bpf.o
  skeleton/profiler.bpf.c:40:14: error: A call to built-in function '__stack_chk_fail' is not supported.
  int BPF_PROG(fentry_XXX)
                ^
  skeleton/profiler.bpf.c:94:14: error: A call to built-in function '__stack_chk_fail' is not supported.
  int BPF_PROG(fexit_XXX)
                ^
  2 errors generated.
  [...]

Since stack-protector makes no sense for the BPF bits just unconditionally
disable it.

Bug: https://bugs.gentoo.org/890638
Signed-off-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/74cd9d2e-6052-312a-241e-2b514a75c92c@applied-asynchrony.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/bugs: Reset speculation control settings on init

[ Upstream commit 0125acda7d76b943ca55811df40ed6ec0ecf670f ]

Currently, x86_spec_ctrl_base is read at boot time and speculative bits
are set if Kconfig items are enabled. For example, IBRS is enabled if
CONFIG_CPU_IBRS_ENTRY is configured, etc. These MSR bits are not cleared
if the mitigations are disabled.

This is a problem when kexec-ing a kernel that has the mitigation
disabled from a kernel that has the mitigation enabled. In this case,
the MSR bits are not cleared during the new kernel boot. As a result,
this might have some performance degradation that is hard to pinpoint.

This problem does not happen if the machine is (hard) rebooted because
the bit will be cleared by default.

  [ bp: Massage. ]

Suggested-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20221128153148.1129350-1-leitao@debian.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

timers: Prevent union confusion from unexpected restart_syscall()

[ Upstream commit 9f76d59173d9d146e96c66886b671c1915a5c5e5 ]

The nanosleep syscalls use the restart_block mechanism, with a quirk:
The `type` and `rmtp`/`compat_rmtp` fields are set up unconditionally on
syscall entry, while the rest of the restart_block is only set up in the
unlikely case that the syscall is actually interrupted by a signal (or
pseudo-signal) that doesn't have a signal handler.

If the restart_block was set up by a previous syscall (futex(...,
FUTEX_WAIT, ...) or poll()) and hasn't been invalidated somehow since then,
this will clobber some of the union fields used by futex_wait_restart() and
do_restart_poll().

If userspace afterwards wrongly calls the restart_syscall syscall,
futex_wait_restart()/do_restart_poll() will read struct fields that have
been clobbered.

This doesn't actually lead to anything particularly interesting because
none of the union fields contain trusted kernel data, and
futex(..., FUTEX_WAIT, ...) and poll() aren't syscalls where it makes much
sense to apply seccomp filters to their arguments.

So the current consequences are just of the "if userspace does bad stuff,
it can damage itself, and that's not a problem" flavor.

But still, it seems like a hazard for future developers, so invalidate the
restart_block when partly setting it up in the nanosleep syscalls.

Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20230105134403.754986-1-jannh@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal: intel: Fix unsigned comparison with less than zero

[ Upstream commit e7fcfe67f9f410736b758969477b17ea285e8e6c ]

The return value from the call to intel_tcc_get_tjmax() is int, which can
be a negative error code. However, the return value is being assigned to
an u32 variable 'tj_max', so making 'tj_max' an int.

Eliminate the following warning:
./drivers/thermal/intel/intel_soc_dts_iosf.c:394:5-11: WARNING: Unsigned expression compared with zero: tj_max < 0

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3637
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Acked-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath11k: debugfs: fix to work with multiple PCI devices

[ Upstream commit 323d91d4684d238f6bc3693fed93caf795378fe0 ]

ath11k fails to load if there are multiple ath11k PCI devices with same name:

 ath11k_pci 0000:01:00.0: Hardware name qcn9074 hw1.0
 debugfs: Directory 'ath11k' with parent '/' already present!
 ath11k_pci 0000:01:00.0: failed to create ath11k debugfs
 ath11k_pci 0000:01:00.0: failed to create soc core: -17
 ath11k_pci 0000:01:00.0: failed to init core: -17
 ath11k_pci: probe of 0000:01:00.0 failed with error -17

Fix this by creating a directory for each ath11k device using schema
<bus>-<devname>, for example "pci-0000:06:00.0". This directory created under
the top-level ath11k directory, for example /sys/kernel/debug/ath11k.

The reference to the toplevel ath11k directory is not stored anymore within ath11k, instead
it's retrieved using debugfs_lookup(). If the directory does not exist it will
be created. After the last directory from the ath11k directory is removed, for
example when doing rmmod ath11k, the empty ath11k directory is left in place,
it's a minor cosmetic issue anyway.

Here's an example hierarchy with one WCN6855:

ath11k
`-- pci-0000:06:00.0
    |-- mac0
    |   |-- dfs_block_radar_events
    |   |-- dfs_simulate_radar
    |   |-- ext_rx_stats
    |   |-- ext_tx_stats
    |   |-- fw_dbglog_config
    |   |-- fw_stats
    |   |   |-- beacon_stats
    |   |   |-- pdev_stats
    |   |   `-- vdev_stats
    |   |-- htt_stats
    |   |-- htt_stats_reset
    |   |-- htt_stats_type
    |   `-- pktlog_filter
    |-- simulate_fw_crash
    `-- soc_dp_stats

I didn't have a test setup where I could connect multiple ath11k devices to the
same the host, so I have only tested this with one device.

Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.9
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.5.0.1-01208-QCAHKSWPL_SILICONZ-1
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.5.0.1-01208-QCAHKSWPL_SILICONZ-1

Tested-by: Robert Marko <robert.marko@sartura.hr>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20221220121231.20120-1-kvalo@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

rcu-tasks: Handle queue-shrink/callback-enqueue race condition

[ Upstream commit a4fcfbee8f6274f9b3f9a71dd5b03e6772ce33f3 ]

The rcu_tasks_need_gpcb() determines whether or not: (1) There are
callbacks needing another grace period, (2) There are callbacks ready
to be invoked, and (3) It would be a good time to shrink back down to a
single-CPU callback list.  This third case is interesting because some
other CPU might be adding new callbacks, which might suddenly make this
a very bad time to be shrinking.

This is currently handled by requiring call_rcu_tasks_generic() to
enqueue callbacks under the protection of rcu_read_lock() and requiring
rcu_tasks_need_gpcb() to wait for an RCU grace period to elapse before
finalizing the transition.  This works well in practice.

Unfortunately, the current code assumes that a grace period whose end is
detected by the poll_state_synchronize_rcu() in the second "if" condition
actually ended before the earlier code counted the callbacks queued on
CPUs other than CPU 0 (local variable "ncbsnz").  Given the current code,
it is possible that a long-delayed call_rcu_tasks_generic() invocation
will queue a callback on a non-zero CPU after these CPUs have had their
callbacks counted and zero has been stored to ncbsnz.  Such a callback
would trigger the WARN_ON_ONCE() in the second "if" statement.

To see this, consider the following sequence of events:

o CPU 0 invokes rcu_tasks_one_gp(), and counts fewer than
rcu_task_collapse_lim callbacks.  It sees at least one
callback queued on some other CPU, thus setting ncbsnz
to a non-zero value.

o CPU 1 invokes call_rcu_tasks_generic() and loads 42 from
->percpu_enqueue_lim.  It therefore decides to enqueue its
callback onto CPU 1's callback list, but is delayed.

o CPU 0 sees the rcu_task_cb_adjust is non-zero and that the number
of callbacks does not exceed rcu_task_collapse_lim.  It therefore
checks percpu_enqueue_lim, and sees that its value is greater
than the value one.  CPU 0 therefore  starts the shift back
to a single callback list.  It sets ->percpu_enqueue_lim to 1,
but CPU 1 has already read the old value of 42.  It also gets
a grace-period state value from get_state_synchronize_rcu().

o CPU 0 sees that ncbsnz is non-zero in its second "if" statement,
so it declines to finalize the shrink operation.

o CPU 0 again invokes rcu_tasks_one_gp(), and counts fewer than
rcu_task_collapse_lim callbacks.  It also sees that there are
no callback queued on any other CPU, and thus sets ncbsnz to zero.

o CPU 1 resumes execution and enqueues its callback onto its own
list.  This invalidates the value of ncbsnz.

o CPU 0 sees the rcu_task_cb_adjust is non-zero and that the number
of callbacks does not exceed rcu_task_collapse_lim.  It therefore
checks percpu_enqueue_lim, but sees that its value is already
unity. It therefore does not get a new grace-period state value.

o CPU 0 sees that rcu_task_cb_adjust is non-zero, ncbsnz is zero,
and that poll_state_synchronize_rcu() says that the grace period
has completed.  it therefore finalizes the shrink operation,
setting ->percpu_dequeue_lim to the value one.

o CPU 0 does a debug check, scanning the other CPUs' callback lists.
It sees that CPU 1's list has a callback, so it (rightly)
triggers the WARN_ON_ONCE().  After all, the new value of
->percpu_dequeue_lim says to not bother looking at CPU 1's
callback list, which means that this callback will never be
invoked.  This can result in hangs and maybe even OOMs.

Based on long experience with rcutorture, this is an extremely
low-probability race condition, but it really can happen, especially in
preemptible kernels or within guest OSes.

This commit therefore checks for completion of the grace period
before counting callbacks.  With this change, in the above failure
scenario CPU 0 would know not to prematurely end the shrink operation
because the grace period would not have completed before the count
operation started.

[ paulmck: Adjust grace-period end rather than adding RCU reader. ]
[ paulmck: Avoid spurious WARN_ON_ONCE() with ->percpu_dequeue_lim check. ]

Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

rcu-tasks: Make rude RCU-Tasks work well with CPU hotplug

[ Upstream commit ea5c8987fef20a8cca07e428aa28bc64649c5104 ]

The synchronize_rcu_tasks_rude() function invokes rcu_tasks_rude_wait_gp()
to wait one rude RCU-tasks grace period.  The rcu_tasks_rude_wait_gp()
function in turn checks if there is only a single online CPU.  If so, it
will immediately return, because a call to synchronize_rcu_tasks_rude()
is by definition a grace period on a single-CPU system.  (We could
have blocked!)

Unfortunately, this check uses num_online_cpus() without synchronization,
which can result in too-short grace periods.  To see this, consider the
following scenario:

        CPU0                                   CPU1 (going offline)
                                          migration/1 task:
                                      cpu_stopper_thread
                                       -> take_cpu_down
                                          -> _cpu_disable
                                           (dec __num_online_cpus)
                                          ->cpuhp_invoke_callback
                                                preempt_disable
                                                access old_data0
           task1
 del old_data0                                  .....
 synchronize_rcu_tasks_rude()
 task1 schedule out
 ....
 task2 schedule in
 rcu_tasks_rude_wait_gp()
     ->__num_online_cpus == 1
       ->return
 ....
 task1 schedule in
 ->free old_data0
                                                preempt_enable

When CPU1 decrements __num_online_cpus, its value becomes 1.  However,
CPU1 has not finished going offline, and will take one last trip through
the scheduler and the idle loop before it actually stops executing
instructions.  Because synchronize_rcu_tasks_rude() is mostly used for
tracing, and because both the scheduler and the idle loop can be traced,
this means that CPU0's prematurely ended grace period might disrupt the
tracing on CPU1.  Given that this disruption might include CPU1 executing
instructions in memory that was just now freed (and maybe reallocated),
this is a matter of some concern.

This commit therefore removes that problematic single-CPU check from the
rcu_tasks_rude_wait_gp() function.  This dispenses with the single-CPU
optimization, but there is no evidence indicating that this optimization
is important.  In addition, synchronize_rcu_tasks_generic() contains a
similar optimization (albeit only for early boot), which also splats.
(As in exactly why are you invoking synchronize_rcu_tasks_rude() so
early in boot, anyway???)

It is OK for the synchronize_rcu_tasks_rude() function's check to be
unsynchronized because the only times that this check can evaluate to
true is when there is only a single CPU running with preemption
disabled.

While in the area, this commit also fixes a minor bug in which a
call to synchronize_rcu_tasks_rude() would instead be attributed to
synchronize_rcu_tasks().

[ paulmck: Add "synchronize_" prefix and "()" suffix. ]

Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

srcu: Delegate work to the boot cpu if using SRCU_SIZE_SMALL

[ Upstream commit 7f24626d6dd844bfc6d1f492d214d29c86d02550 ]

Commit 994f706872e6 ("srcu: Make Tree SRCU able to operate without
snp_node array") assumes that cpu 0 is always online.  However, there
really are situations when some other CPU is the boot CPU, for example,
when booting a kdump kernel with the maxcpus=1 boot parameter.

On PowerPC, the kdump kernel can hang as follows:
...
[    1.740036] systemd[1]: Hostname set to <xyz.com>
[  243.686240] INFO: task systemd:1 blocked for more than 122 seconds.
[  243.686264]       Not tainted 6.1.0-rc1 #1
[  243.686272] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  243.686281] task:systemd         state:D stack:0     pid:1     ppid:0      flags:0x00042000
[  243.686296] Call Trace:
[  243.686301] [c000000016657640] [c000000016657670] 0xc000000016657670 (unreliable)
[  243.686317] [c000000016657830] [c00000001001dec0] __switch_to+0x130/0x220
[  243.686333] [c000000016657890] [c000000010f607b8] __schedule+0x1f8/0x580
[  243.686347] [c000000016657940] [c000000010f60bb4] schedule+0x74/0x140
[  243.686361] [c0000000166579b0] [c000000010f699b8] schedule_timeout+0x168/0x1c0
[  243.686374] [c000000016657a80] [c000000010f61de8] __wait_for_common+0x148/0x360
[  243.686387] [c000000016657b20] [c000000010176bb0] __flush_work.isra.0+0x1c0/0x3d0
[  243.686401] [c000000016657bb0] [c0000000105f2768] fsnotify_wait_marks_destroyed+0x28/0x40
[  243.686415] [c000000016657bd0] [c0000000105f21b8] fsnotify_destroy_group+0x68/0x160
[  243.686428] [c000000016657c40] [c0000000105f6500] inotify_release+0x30/0xa0
[  243.686440] [c000000016657cb0] [c0000000105751a8] __fput+0xc8/0x350
[  243.686452] [c000000016657d00] [c00000001017d524] task_work_run+0xe4/0x170
[  243.686464] [c000000016657d50] [c000000010020e94] do_notify_resume+0x134/0x140
[  243.686478] [c000000016657d80] [c00000001002eb18] interrupt_exit_user_prepare_main+0x198/0x270
[  243.686493] [c000000016657de0] [c00000001002ec60] syscall_exit_prepare+0x70/0x180
[  243.686505] [c000000016657e10] [c00000001000bf7c] system_call_vectored_common+0xfc/0x280
[  243.686520] --- interrupt: 3000 at 0x7fffa47d5ba4
[  243.686528] NIP:  00007fffa47d5ba4 LR: 0000000000000000 CTR: 0000000000000000
[  243.686538] REGS: c000000016657e80 TRAP: 3000   Not tainted  (6.1.0-rc1)
[  243.686548] MSR:  800000000000d033 <SF,EE,PR,ME,IR,DR,RI,LE>  CR: 42044440  XER: 00000000
[  243.686572] IRQMASK: 0
[  243.686572] GPR00: 0000000000000006 00007ffffa606710 00007fffa48e7200 0000000000000000
[  243.686572] GPR04: 0000000000000002 000000000000000a 0000000000000000 0000000000000001
[  243.686572] GPR08: 000001000c172dd0 0000000000000000 0000000000000000 0000000000000000
[  243.686572] GPR12: 0000000000000000 00007fffa4ff4bc0 0000000000000000 0000000000000000
[  243.686572] GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
[  243.686572] GPR20: 0000000132dfdc50 000000000000000e 0000000000189375 0000000000000000
[  243.686572] GPR24: 00007ffffa606ae0 0000000000000005 000001000c185490 000001000c172570
[  243.686572] GPR28: 000001000c172990 000001000c184850 000001000c172e00 00007fffa4fedd98
[  243.686683] NIP [00007fffa47d5ba4] 0x7fffa47d5ba4
[  243.686691] LR [0000000000000000] 0x0
[  243.686698] --- interrupt: 3000
[  243.686708] INFO: task kworker/u16:1:24 blocked for more than 122 seconds.
[  243.686717]       Not tainted 6.1.0-rc1 #1
[  243.686724] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  243.686733] task:kworker/u16:1   state:D stack:0     pid:24    ppid:2      flags:0x00000800
[  243.686747] Workqueue: events_unbound fsnotify_mark_destroy_workfn
[  243.686758] Call Trace:
[  243.686762] [c0000000166736e0] [c00000004fd91000] 0xc00000004fd91000 (unreliable)
[  243.686775] [c0000000166738d0] [c00000001001dec0] __switch_to+0x130/0x220
[  243.686788] [c000000016673930] [c000000010f607b8] __schedule+0x1f8/0x580
[  243.686801] [c0000000166739e0] [c000000010f60bb4] schedule+0x74/0x140
[  243.686814] [c000000016673a50] [c000000010f699b8] schedule_timeout+0x168/0x1c0
[  243.686827] [c000000016673b20] [c000000010f61de8] __wait_for_common+0x148/0x360
[  243.686840] [c000000016673bc0] [c000000010210840] __synchronize_srcu.part.0+0xa0/0xe0
[  243.686855] [c000000016673c30] [c0000000105f2c64] fsnotify_mark_destroy_workfn+0xc4/0x1a0
[  243.686868] [c000000016673ca0] [c000000010174ea8] process_one_work+0x2a8/0x570
[  243.686882] [c000000016673d40] [c000000010175208] worker_thread+0x98/0x5e0
[  243.686895] [c000000016673dc0] [c0000000101828d4] kthread+0x124/0x130
[  243.686908] [c000000016673e10] [c00000001000cd40] ret_from_kernel_thread+0x5c/0x64
[  366.566274] INFO: task systemd:1 blocked for more than 245 seconds.
[  366.566298]       Not tainted 6.1.0-rc1 #1
[  366.566305] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  366.566314] task:systemd         state:D stack:0     pid:1     ppid:0      flags:0x00042000
[  366.566329] Call Trace:
...

The above splat occurs because PowerPC really does use maxcpus=1
instead of nr_cpus=1 in the kernel command line.  Consequently, the
(quite possibly non-zero) kdump CPU is the only online CPU in the kdump
kernel.  SRCU unconditionally queues a sdp->work on cpu 0, for which no
worker thread has been created, so sdp->work will be never executed and
__synchronize_srcu() will never be completed.

This commit therefore replaces CPU ID 0 with get_boot_cpu_id() in key
places in Tree SRCU.  Since the CPU indicated by get_boot_cpu_id()
is guaranteed to be online, this avoids the above splat.

Signed-off-by: Pingfan Liu <kernelfans@gmail.com>
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
To: rcu@vger.kernel.org
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

rcu: Suppress smp_processor_id() complaint in synchronize_rcu_expedited_wait()

[ Upstream commit 2d7f00b2f01301d6e41fd4a28030dab0442265be ]

The normal grace period's RCU CPU stall warnings are invoked from the
scheduling-clock interrupt handler, and can thus invoke smp_processor_id()
with impunity, which allows them to directly invoke dump_cpu_task().
In contrast, the expedited grace period's RCU CPU stall warnings are
invoked from process context, which causes the dump_cpu_task() function's
calls to smp_processor_id() to complain bitterly in debug kernels.

This commit therefore causes synchronize_rcu_expedited_wait() to disable
preemption around its call to dump_cpu_task().

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

rcu: Make RCU_LOCKDEP_WARN() avoid early lockdep checks

[ Upstream commit 0cae5ded535c3a80aed94f119bbd4ee3ae284a65 ]

Currently, RCU_LOCKDEP_WARN() checks the condition before checking
to see if lockdep is still enabled.  This is necessary to avoid the
false-positive splats fixed by commit 3066820034b5dd ("rcu: Reject
RCU_LOCKDEP_WARN() false positives").  However, the current state can
result in false-positive splats during early boot before lockdep is fully
initialized.  This commit therefore checks debug_lockdep_rcu_enabled()
both before and after checking the condition, thus avoiding both sets
of false-positive error reports.

Reported-by: Steven Rostedt <rostedt@goodmis.org>
Reported-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Reported-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: brcmfmac: Fix potential stack-out-of-bounds in brcmf_c_preinit_dcmds()

[ Upstream commit 0a06cadcc2a0044e4a117cc0e61436fc3a0dad69 ]

This patch fixes a stack-out-of-bounds read in brcmfmac that occurs
when 'buf' that is not null-terminated is passed as an argument of
strsep() in brcmf_c_preinit_dcmds(). This buffer is filled with a firmware
version string by memcpy() in brcmf_fil_iovar_data_get().
The patch ensures buf is null-terminated.

Found by a modified version of syzkaller.

[   47.569679][ T1897] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43236b for chip BCM43236/3
[   47.582839][ T1897] brcmfmac: brcmf_c_process_clm_blob: no clm_blob available (err=-2), device may have limited channels available
[   47.601565][ T1897] ==================================================================
[   47.602574][ T1897] BUG: KASAN: stack-out-of-bounds in strsep+0x1b2/0x1f0
[   47.603447][ T1897] Read of size 1 at addr ffffc90001f6f000 by task kworker/0:2/1897
[   47.604336][ T1897]
[   47.604621][ T1897] CPU: 0 PID: 1897 Comm: kworker/0:2 Tainted: G           O      5.14.0+ #131
[   47.605617][ T1897] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
[   47.606907][ T1897] Workqueue: usb_hub_wq hub_event
[   47.607453][ T1897] Call Trace:
[   47.607801][ T1897]  dump_stack_lvl+0x8e/0xd1
[   47.608295][ T1897]  print_address_description.constprop.0.cold+0xf/0x334
[   47.609009][ T1897]  ? strsep+0x1b2/0x1f0
[   47.609434][ T1897]  ? strsep+0x1b2/0x1f0
[   47.609863][ T1897]  kasan_report.cold+0x83/0xdf
[   47.610366][ T1897]  ? strsep+0x1b2/0x1f0
[   47.610882][ T1897]  strsep+0x1b2/0x1f0
[   47.611300][ T1897]  ? brcmf_fil_iovar_data_get+0x3a/0xf0
[   47.611883][ T1897]  brcmf_c_preinit_dcmds+0x995/0xc40
[   47.612434][ T1897]  ? brcmf_c_set_joinpref_default+0x100/0x100
[   47.613078][ T1897]  ? rcu_read_lock_sched_held+0xa1/0xd0
[   47.613662][ T1897]  ? rcu_read_lock_bh_held+0xb0/0xb0
[   47.614208][ T1897]  ? lock_acquire+0x19d/0x4e0
[   47.614704][ T1897]  ? find_held_lock+0x2d/0x110
[   47.615236][ T1897]  ? brcmf_usb_deq+0x1a7/0x260
[   47.615741][ T1897]  ? brcmf_usb_rx_fill_all+0x5a/0xf0
[   47.616288][ T1897]  brcmf_attach+0x246/0xd40
[   47.616758][ T1897]  ? wiphy_new_nm+0x1703/0x1dd0
[   47.617280][ T1897]  ? kmemdup+0x43/0x50
[   47.617720][ T1897]  brcmf_usb_probe+0x12de/0x1690
[   47.618244][ T1897]  ? brcmf_usbdev_qinit.constprop.0+0x470/0x470
[   47.618901][ T1897]  usb_probe_interface+0x2aa/0x760
[   47.619429][ T1897]  ? usb_probe_device+0x250/0x250
[   47.619950][ T1897]  really_probe+0x205/0xb70
[   47.620435][ T1897]  ? driver_allows_async_probing+0x130/0x130
[   47.621048][ T1897]  __driver_probe_device+0x311/0x4b0
[   47.621595][ T1897]  ? driver_allows_async_probing+0x130/0x130
[   47.622209][ T1897]  driver_probe_device+0x4e/0x150
[   47.622739][ T1897]  __device_attach_driver+0x1cc/0x2a0
[   47.623287][ T1897]  bus_for_each_drv+0x156/0x1d0
[   47.623796][ T1897]  ? bus_rescan_devices+0x30/0x30
[   47.624309][ T1897]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   47.624907][ T1897]  ? trace_hardirqs_on+0x46/0x160
[   47.625437][ T1897]  __device_attach+0x23f/0x3a0
[   47.625924][ T1897]  ? device_bind_driver+0xd0/0xd0
[   47.626433][ T1897]  ? kobject_uevent_env+0x287/0x14b0
[   47.627057][ T1897]  bus_probe_device+0x1da/0x290
[   47.627557][ T1897]  device_add+0xb7b/0x1eb0
[   47.628027][ T1897]  ? wait_for_completion+0x290/0x290
[   47.628593][ T1897]  ? __fw_devlink_link_to_suppliers+0x5a0/0x5a0
[   47.629249][ T1897]  usb_set_configuration+0xf59/0x16f0
[   47.629829][ T1897]  usb_generic_driver_probe+0x82/0xa0
[   47.630385][ T1897]  usb_probe_device+0xbb/0x250
[   47.630927][ T1897]  ? usb_suspend+0x590/0x590
[   47.631397][ T1897]  really_probe+0x205/0xb70
[   47.631855][ T1897]  ? driver_allows_async_probing+0x130/0x130
[   47.632469][ T1897]  __driver_probe_device+0x311/0x4b0
[   47.633002][ T1897]  ? usb_generic_driver_match+0x75/0x90
[   47.633573][ T1897]  ? driver_allows_async_probing+0x130/0x130
[   47.634170][ T1897]  driver_probe_device+0x4e/0x150
[   47.634703][ T1897]  __device_attach_driver+0x1cc/0x2a0
[   47.635248][ T1897]  bus_for_each_drv+0x156/0x1d0
[   47.635748][ T1897]  ? bus_rescan_devices+0x30/0x30
[   47.636271][ T1897]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   47.636881][ T1897]  ? trace_hardirqs_on+0x46/0x160
[   47.637396][ T1897]  __device_attach+0x23f/0x3a0
[   47.637904][ T1897]  ? device_bind_driver+0xd0/0xd0
[   47.638426][ T1897]  ? kobject_uevent_env+0x287/0x14b0
[   47.638985][ T1897]  bus_probe_device+0x1da/0x290
[   47.639512][ T1897]  device_add+0xb7b/0x1eb0
[   47.639977][ T1897]  ? __fw_devlink_link_to_suppliers+0x5a0/0x5a0
[   47.640612][ T1897]  ? kfree+0x14a/0x6b0
[   47.641055][ T1897]  ? __usb_get_extra_descriptor+0x116/0x160
[   47.641679][ T1897]  usb_new_device.cold+0x49c/0x1029
[   47.642245][ T1897]  ? hub_disconnect+0x450/0x450
[   47.642756][ T1897]  ? rwlock_bug.part.0+0x90/0x90
[   47.643273][ T1897]  ? _raw_spin_unlock_irq+0x24/0x30
[   47.643822][ T1897]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   47.644445][ T1897]  hub_event+0x1c98/0x3950
[   47.644939][ T1897]  ? hub_port_debounce+0x2e0/0x2e0
[   47.645467][ T1897]  ? check_irq_usage+0x861/0xf20
[   47.645975][ T1897]  ? drain_workqueue+0x280/0x360
[   47.646506][ T1897]  ? lock_release+0x640/0x640
[   47.646994][ T1897]  ? rcu_read_lock_sched_held+0xa1/0xd0
[   47.647572][ T1897]  ? rcu_read_lock_bh_held+0xb0/0xb0
[   47.648111][ T1897]  ? lockdep_hardirqs_on_prepare+0x273/0x3e0
[   47.648735][ T1897]  process_one_work+0x92b/0x1460
[   47.649262][ T1897]  ? pwq_dec_nr_in_flight+0x330/0x330
[   47.649816][ T1897]  ? rwlock_bug.part.0+0x90/0x90
[   47.650336][ T1897]  worker_thread+0x95/0xe00
[   47.650830][ T1897]  ? __kthread_parkme+0x115/0x1e0
[   47.651361][ T1897]  ? process_one_work+0x1460/0x1460
[   47.651904][ T1897]  kthread+0x3a1/0x480
[   47.652329][ T1897]  ? set_kthread_struct+0x120/0x120
[   47.652878][ T1897]  ret_from_fork+0x1f/0x30
[   47.653370][ T1897]
[   47.653608][ T1897]
[   47.653848][ T1897] addr ffffc90001f6f000 is located in stack of task kworker/0:2/1897 at offset 512 in frame:
[   47.654891][ T1897]  brcmf_c_preinit_dcmds+0x0/0xc40
[   47.655442][ T1897]
[   47.655690][ T1897] this frame has 4 objects:
[   47.656151][ T1897]  [48, 56) 'ptr'
[   47.656159][ T1897]  [80, 148) 'revinfo'
[   47.656534][ T1897]  [192, 210) 'eventmask'
[   47.656953][ T1897]  [256, 512) 'buf'
[   47.657410][ T1897]
[   47.658035][ T1897] Memory state around the buggy address:
[   47.658743][ T1897]  ffffc90001f6ef00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[   47.659577][ T1897]  ffffc90001f6ef80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[   47.660394][ T1897] >ffffc90001f6f000: f3 f3 f3 f3 f3 f3 f3 f3 00 00 00 00 00 00 00 00
[   47.661199][ T1897]                    ^
[   47.661625][ T1897]  ffffc90001f6f080: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[   47.662455][ T1897]  ffffc90001f6f100: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 f1 f1
[   47.663318][ T1897] ==================================================================
[   47.664147][ T1897] Disabling lock debugging due to kernel taint

Reported-by: Dokyung Song <dokyungs@yonsei.ac.kr>
Reported-by: Jisoo Jang <jisoo.jang@yonsei.ac.kr>
Reported-by: Minsuk Kang <linuxlovemin@yonsei.ac.kr>
Signed-off-by: Jisoo Jang <jisoo.jang@yonsei.ac.kr>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20221115043458.37562-1-jisoo.jang@yonsei.ac.kr
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath11k: fix monitor mode bringup crash

[ Upstream commit 950b43f8bd8a4d476d2da6d2a083a89bcd3c90d7 ]

When the interface is brought up in monitor mode, it leads
to NULL pointer dereference crash. This crash happens when
the packet type is extracted for a SKB. This extraction
which is present in the received msdu delivery path,is
not needed for the monitor ring packets since they are
all RAW packets. Hence appending the flags with
"RX_FLAG_ONLY_MONITOR" to skip that extraction.

Observed calltrace:

Unable to handle kernel NULL pointer dereference at virtual address
0000000000000064
Mem abort info:
  ESR = 0x0000000096000004
  EC = 0x25: DABT (current EL), IL = 32 bits
  SET = 0, FnV = 0
  EA = 0, S1PTW = 0
  FSC = 0x04: level 0 translation fault
Data abort info:
  ISV = 0, ISS = 0x00000004
  CM = 0, WnR = 0
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000048517000
[0000000000000064] pgd=0000000000000000, p4d=0000000000000000
Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
Modules linked in: ath11k_pci ath11k qmi_helpers
CPU: 2 PID: 1781 Comm: napi/-271 Not tainted
6.1.0-rc5-wt-ath-656295-gef907406320c-dirty #6
Hardware name: Qualcomm Technologies, Inc. IPQ8074/AP-HK10-C2 (DT)
pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : ath11k_hw_qcn9074_rx_desc_get_decap_type+0x34/0x60 [ath11k]
lr : ath11k_hw_qcn9074_rx_desc_get_decap_type+0x5c/0x60 [ath11k]
sp : ffff80000ef5bb10
x29: ffff80000ef5bb10 x28: 0000000000000000 x27: ffff000007baafa0
x26: ffff000014a91ed0 x25: 0000000000000000 x24: 0000000000000000
x23: ffff800002b77378 x22: ffff000014a91ec0 x21: ffff000006c8d600
x20: 0000000000000000 x19: ffff800002b77740 x18: 0000000000000006
x17: 736564203634343a x16: 656e694c20657079 x15: 0000000000000143
x14: 00000000ffffffea x13: ffff80000ef5b8b8 x12: ffff80000ef5b8c8
x11: ffff80000a591d30 x10: ffff80000a579d40 x9 : c0000000ffffefff
x8 : 0000000000000003 x7 : 0000000000017fe8 x6 : ffff80000a579ce8
x5 : 0000000000000000 x4 : 0000000000000000 x3 : 0000000000000000
x2 : 3a35ec12ed7f8900 x1 : 0000000000000000 x0 : 0000000000000052
Call trace:
 ath11k_hw_qcn9074_rx_desc_get_decap_type+0x34/0x60 [ath11k]
 ath11k_dp_rx_deliver_msdu.isra.42+0xa4/0x3d0 [ath11k]
 ath11k_dp_rx_mon_deliver.isra.43+0x2f8/0x458 [ath11k]
 ath11k_dp_rx_process_mon_rings+0x310/0x4c0 [ath11k]
 ath11k_dp_service_srng+0x234/0x338 [ath11k]
 ath11k_pcic_ext_grp_napi_poll+0x30/0xb8 [ath11k]
 __napi_poll+0x5c/0x190
 napi_threaded_poll+0xf0/0x118
 kthread+0xf4/0x110
 ret_from_fork+0x10/0x20

Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1
Reported-by: Florian Schmidt <florian@fls.name>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=216573
Signed-off-by: Nagarajan Maran <quic_nmaran@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20221129142532.23421-1-quic_nmaran@quicinc.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath9k: Fix use-after-free in ath9k_hif_usb_disconnect()

[ Upstream commit f099c5c9e2ba08a379bd354a82e05ef839ae29ac ]

This patch fixes a use-after-free in ath9k that occurs in
ath9k_hif_usb_disconnect() when ath9k_destroy_wmi() is trying to access
'drv_priv' that has already been freed by ieee80211_free_hw(), called by
ath9k_htc_hw_deinit(). The patch moves ath9k_destroy_wmi() before
ieee80211_free_hw(). Note that urbs from the driver should be killed
before freeing 'wmi' with ath9k_destroy_wmi() as their callbacks will
access 'wmi'.

Found by a modified version of syzkaller.

==================================================================
BUG: KASAN: use-after-free in ath9k_destroy_wmi+0x38/0x40
Read of size 8 at addr ffff8881069132a0 by task kworker/0:1/7

CPU: 0 PID: 7 Comm: kworker/0:1 Tainted: G O 5.14.0+ #131
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
Workqueue: usb_hub_wq hub_event
Call Trace:
 dump_stack_lvl+0x8e/0xd1
 print_address_description.constprop.0.cold+0x93/0x334
 ? ath9k_destroy_wmi+0x38/0x40
 ? ath9k_destroy_wmi+0x38/0x40
 kasan_report.cold+0x83/0xdf
 ? ath9k_destroy_wmi+0x38/0x40
 ath9k_destroy_wmi+0x38/0x40
 ath9k_hif_usb_disconnect+0x329/0x3f0
 ? ath9k_hif_usb_suspend+0x120/0x120
 ? usb_disable_interface+0xfc/0x180
 usb_unbind_interface+0x19b/0x7e0
 ? usb_autoresume_device+0x50/0x50
 device_release_driver_internal+0x44d/0x520
 bus_remove_device+0x2e5/0x5a0
 device_del+0x5b2/0xe30
 ? __device_link_del+0x370/0x370
 ? usb_remove_ep_devs+0x43/0x80
 ? remove_intf_ep_devs+0x112/0x1a0
 usb_disable_device+0x1e3/0x5a0
 usb_disconnect+0x267/0x870
 hub_event+0x168d/0x3950
 ? rcu_read_lock_sched_held+0xa1/0xd0
 ? hub_port_debounce+0x2e0/0x2e0
 ? check_irq_usage+0x860/0xf20
 ? drain_workqueue+0x281/0x360
 ? lock_release+0x640/0x640
 ? rcu_read_lock_sched_held+0xa1/0xd0
 ? rcu_read_lock_bh_held+0xb0/0xb0
 ? lockdep_hardirqs_on_prepare+0x273/0x3e0
 process_one_work+0x92b/0x1460
 ? pwq_dec_nr_in_flight+0x330/0x330
 ? rwlock_bug.part.0+0x90/0x90
 worker_thread+0x95/0xe00
 ? __kthread_parkme+0x115/0x1e0
 ? process_one_work+0x1460/0x1460
 kthread+0x3a1/0x480
 ? set_kthread_struct+0x120/0x120
 ret_from_fork+0x1f/0x30

The buggy address belongs to the page:
page:ffffea00041a44c0 refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x106913
flags: 0x200000000000000(node=0|zone=2)
raw: 0200000000000000 0000000000000000 dead000000000122 0000000000000000
raw: 0000000000000000 0000000000000000 00000000ffffffff 0000000000000000
page dumped because: kasan: bad access detected
page_owner tracks the page as freed
page last allocated via order 3, migratetype Unmovable, gfp_mask 0x40dc0(GFP_KERNEL|__GFP_COMP|__GFP_ZERO), pid 7, ts 38347963444, free_ts 41399957635
 prep_new_page+0x1aa/0x240
 get_page_from_freelist+0x159a/0x27c0
 __alloc_pages+0x2da/0x6a0
 alloc_pages+0xec/0x1e0
 kmalloc_order+0x39/0xf0
 kmalloc_order_trace+0x19/0x120
 __kmalloc+0x308/0x390
 wiphy_new_nm+0x6f5/0x1dd0
 ieee80211_alloc_hw_nm+0x36d/0x2230
 ath9k_htc_probe_device+0x9d/0x1e10
 ath9k_htc_hw_init+0x34/0x50
 ath9k_hif_usb_firmware_cb+0x25f/0x4e0
 request_firmware_work_func+0x131/0x240
 process_one_work+0x92b/0x1460
 worker_thread+0x95/0xe00
 kthread+0x3a1/0x480
page last free stack trace:
 free_pcp_prepare+0x3d3/0x7f0
 free_unref_page+0x1e/0x3d0
 device_release+0xa4/0x240
 kobject_put+0x186/0x4c0
 put_device+0x20/0x30
 ath9k_htc_disconnect_device+0x1cf/0x2c0
 ath9k_htc_hw_deinit+0x26/0x30
 ath9k_hif_usb_disconnect+0x2d9/0x3f0
 usb_unbind_interface+0x19b/0x7e0
 device_release_driver_internal+0x44d/0x520
 bus_remove_device+0x2e5/0x5a0
 device_del+0x5b2/0xe30
 usb_disable_device+0x1e3/0x5a0
 usb_disconnect+0x267/0x870
 hub_event+0x168d/0x3950
 process_one_work+0x92b/0x1460

Memory state around the buggy address:
 ffff888106913180: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
 ffff888106913200: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
>ffff888106913280: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
                               ^
 ffff888106913300: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
 ffff888106913380: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
==================================================================

Reported-by: Dokyung Song <dokyungs@yonsei.ac.kr>
Reported-by: Jisoo Jang <jisoo.jang@yonsei.ac.kr>
Reported-by: Minsuk Kang <linuxlovemin@yonsei.ac.kr>
Signed-off-by: Minsuk Kang <linuxlovemin@yonsei.ac.kr>
Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20221205014308.1617597-1-linuxlovemin@yonsei.ac.kr
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/x86/intel/uncore: Add Meteor Lake support

[ Upstream commit c828441f21ddc819a28b5723a72e3c840e9de1c6 ]

The uncore subsystem for Meteor Lake is similar to the previous Alder
Lake. The main difference is that MTL provides PMU support for different
tiles, while ADL only provides PMU support for the whole package. On
ADL, there are CBOX, ARB, and clockbox uncore PMON units. On MTL, they
are split into CBOX/HAC_CBOX, ARB/HAC_ARB, and cncu/sncu which provides
a fixed counter for clockticks. Also, new MSR addresses are introduced
on MTL.

The IMC uncore PMON is the same as Alder Lake. Add new PCIIDs of IMC for
Meteor Lake.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230210190238.1726237-1-kan.liang@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpuidle: lib/bug: Disable rcu_is_watching() during WARN/BUG

[ Upstream commit 5a5d7e9badd2cb8065db171961bd30bd3595e4b6 ]

In order to avoid WARN/BUG from generating nested or even recursive
warnings, force rcu_is_watching() true during
WARN/lockdep_rcu_suspicious().

Notably things like unwinding the stack can trigger rcu_dereference()
warnings, which then triggers more unwinding which then triggers more
warnings etc..

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230126151323.408156109@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpuidle: drivers: firmware: psci: Dont instrument suspend code

[ Upstream commit 393e2ea30aec634b37004d401863428e120d5e1b ]

The PSCI suspend code is currently instrumentable, which is not safe as
instrumentation (e.g. ftrace) may try to make use of RCU during idle
periods when RCU is not watching.

To fix this we need to ensure that psci_suspend_finisher() and anything
it calls are not instrumented. We can do this fairly simply by marking
psci_suspend_finisher() and the psci*_cpu_suspend() functions as
noinstr, and the underlying helper functions as __always_inline.

When CONFIG_DEBUG_VIRTUAL=y, __pa_symbol() can expand to an out-of-line
instrumented function, so we must use __pa_symbol_nodebug() within
psci_suspend_finisher().

The raw SMCCC invocation functions are written in assembly, and are not
subject to compiler instrumentation.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230126151323.349423061@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/fpu: Don't set TIF_NEED_FPU_LOAD for PF_IO_WORKER threads

[ Upstream commit cb3ea4b7671b7cfbac3ee609976b790aebd0bbda ]

We don't set it on PF_KTHREAD threads as they never return to userspace,
and PF_IO_WORKER threads are identical in that regard. As they keep
running in the kernel until they die, skip setting the FPU flag on them.

More of a cosmetic thing that was found while debugging and
issue and pondering why the FPU flag is set on these threads.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/560c844c-f128-555b-40c6-31baff27537f@kernel.dk
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpuidle, intel_idle: Fix CPUIDLE_FLAG_INIT_XSTATE

[ Upstream commit 821ad23d0eaff73ef599ece39ecc77482df20a8c ]

Fix instrumentation bugs objtool found:

  vmlinux.o: warning: objtool: intel_idle_s2idle+0xd5: call to fpu_idle_fpregs() leaves .noinstr.text section
  vmlinux.o: warning: objtool: intel_idle_xstate+0x11: call to fpu_idle_fpregs() leaves .noinstr.text section
  vmlinux.o: warning: objtool: fpu_idle_fpregs+0x9: call to xfeatures_in_use() leaves .noinstr.text section

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195540.494977795@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: zynqmp: Enable hs termination flag for USB dwc3 controller

[ Upstream commit 32405e532d358a2f9d4befae928b9883c8597616 ]

Since we need to support legacy phys with the dwc3 controller,
we enable this quirk on the zynqmp platforms.

Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
Link: https://lore.kernel.org/r/20221023215649.221726-1-m.grzeschik@pengutronix.de
Signed-off-by: Michal Simek <michal.simek@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

btrfs: scrub: improve tree block error reporting

[ Upstream commit 28232909ba43561887508a6ef46d7f33a648f375 ]

[BUG]
When debugging a scrub related metadata error, it turns out that our
metadata error reporting is not ideal.

The only 3 error messages are:

- BTRFS error (device dm-2): bdev /dev/mapper/test-scratch1 errs: wr 0, rd 0, flush 0, corrupt 0, gen 1
  Showing we have metadata generation mismatch errors.

- BTRFS error (device dm-2): unable to fixup (regular) error at logical 7110656 on dev /dev/mapper/test-scratch1
  Showing which tree blocks are corrupted.

- BTRFS warning (device dm-2): checksum/header error at logical 24772608 on dev /dev/mapper/test-scratch2, physical 3801088: metadata node (level 1) in tree 5
  Showing which physical range the corrupted metadata is at.

We have to combine the above 3 to know we have a corrupted metadata with
generation mismatch.

And this is already the better case, if we have other problems, like
fsid mismatch, we can not even know the cause.

[CAUSE]
The problem is caused by the fact that, scrub_checksum_tree_block()
never outputs any error message.

It just return two bits for scrub: sblock->header_error, and
sblock->generation_error.

And later we report error in scrub_print_warning(), but unfortunately we
only have two bits, there is not really much thing we can done to print
any detailed errors.

[FIX]
This patch will do the following to enhance the error reporting of
metadata scrub:

- Add extra warning (ratelimited) for every error we hit
  This can help us to distinguish the different types of errors.
  Some errors can help us to know what's going wrong immediately,
  like bytenr mismatch.

- Re-order the checks
  Currently we check bytenr first, then immediately generation.
  This can lead to false generation mismatch reports, while the fsid
  mismatches.

Here is the new output for the bug I'm debugging (we forgot to
writeback tree blocks for commit roots):

 BTRFS warning (device dm-2): tree block 24117248 mirror 1 has bad fsid, has b77cd862-f150-4c71-90ec-7baf0544d83f want 17df6abf-23cd-445f-b350-5b3e40bfd2fc
 BTRFS warning (device dm-2): tree block 24117248 mirror 0 has bad fsid, has b77cd862-f150-4c71-90ec-7baf0544d83f want 17df6abf-23cd-445f-b350-5b3e40bfd2fc

Now we can immediately know it's some tree blocks didn't even get written
back, other than the original confusing generation mismatch.

Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

trace/blktrace: fix memory leak with using debugfs_lookup()

[ Upstream commit 83e8864fee26f63a7435e941b7c36a20fd6fe93e ]

When calling debugfs_lookup() the result must have dput() called on it,
otherwise the memory will leak over time.  To make things simpler, just
call debugfs_lookup_and_remove() instead which handles all of the logic
at once.

Cc: Jens Axboe <axboe@kernel.dk>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: linux-block@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-trace-kernel@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230202141956.2299521-1-gregkh@linuxfoundation.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

blk-cgroup: synchronize pd_free_fn() from blkg_free_workfn() and blkcg_deactivate_policy()

[ Upstream commit f1c006f1c6850c14040f8337753a63119bba39b9 ]

Currently parent pd can be freed before child pd:

t1: remove cgroup C1
blkcg_destroy_blkgs
 blkg_destroy
  list_del_init(&blkg->q_node)
  // remove blkg from queue list
  percpu_ref_kill(&blkg->refcnt)
   blkg_release
    call_rcu

t2: from t1
__blkg_release
 blkg_free
  schedule_work
t4: deactivate policy
blkcg_deactivate_policy
 pd_free_fn
 // parent of C1 is freed first
t3: from t2
 blkg_free_workfn
  pd_free_fn

If policy(for example, ioc_timer_fn() from iocost) access parent pd from
child pd after pd_offline_fn(), then UAF can be triggered.

Fix the problem by delaying 'list_del_init(&blkg->q_node)' from
blkg_destroy() to blkg_free_workfn(), and using a new disk level mutex to
synchronize blkg_free_workfn() and blkcg_deactivate_policy().

Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20230119110350.2287325-4-yukuai1@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

blk-cgroup: dropping parent refcount after pd_free_fn() is done

[ Upstream commit c7241babf0855d8a6180cd1743ff0ec34de40b4e ]

Some cgroup policies will access parent pd through child pd even
after pd_offline_fn() is done. If pd_free_fn() for parent is called
before child, then UAF can be triggered. Hence it's better to guarantee
the order of pd_free_fn().

Currently refcount of parent blkg is dropped in __blkg_release(), which
is before pd_free_fn() is called in blkg_free_work_fn() while
blkg_free_work_fn() is called asynchronously.

This patch make sure pd_free_fn() called from removing cgroup is ordered
by delaying dropping parent refcount after calling pd_free_fn() for
child.

BTW, pd_free_fn() will also be called from blkcg_deactivate_policy()
from deleting device, and following patches will guarantee the order.

Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20230119110350.2287325-2-yukuai1@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

blk-iocost: fix divide by 0 error in calc_lcoefs()

[ Upstream commit 984af1e66b4126cf145153661cc24c213e2ec231 ]

echo max of u64 to cost.model can cause divide by 0 error.

  # echo 8:0 rbps=18446744073709551615 > /sys/fs/cgroup/io.cost.model

  divide error: 0000 [#1] PREEMPT SMP
  RIP: 0010:calc_lcoefs+0x4c/0xc0
  Call Trace:
   <TASK>
   ioc_refresh_params+0x2b3/0x4f0
   ioc_cost_model_write+0x3cb/0x4c0
   ? _copy_from_iter+0x6d/0x6c0
   ? kernfs_fop_write_iter+0xfc/0x270
   cgroup_file_write+0xa0/0x200
   kernfs_fop_write_iter+0x17d/0x270
   vfs_write+0x414/0x620
   ksys_write+0x73/0x160
   __x64_sys_write+0x1e/0x30
   do_syscall_64+0x35/0x80
   entry_SYSCALL_64_after_hwframe+0x63/0xcd

calc_lcoefs() uses the input value of cost.model in DIV_ROUND_UP_ULL,
overflow would happen if bps plus IOC_PAGE_SIZE is greater than
ULLONG_MAX, it can cause divide by 0 error.

Fix the problem by setting basecost

Signed-off-by: Li Nan <linan122@huawei.com>
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Acked-by: Tejun Heo <tj@kernel.org>
Link: https://lore.kernel.org/r/20230117070806.3857142-5-yukuai1@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

fs: Use CHECK_DATA_CORRUPTION() when kernel bugs are detected

[ Upstream commit 47d586913f2abec4d240bae33417f537fda987ec ]

Currently, filp_close() and generic_shutdown_super() use printk() to log
messages when bugs are detected. This is problematic because infrastructure
like syzkaller has no idea that this message indicates a bug.
In addition, some people explicitly want their kernels to BUG() when kernel
data corruption has been detected (CONFIG_BUG_ON_DATA_CORRUPTION).
And finally, when generic_shutdown_super() detects remaining inodes on a
system without CONFIG_BUG_ON_DATA_CORRUPTION, it would be nice if later
accesses to a busy inode would at least crash somewhat cleanly rather than
walking through freed memory.

To address all three, use CHECK_DATA_CORRUPTION() when kernel bugs are
detected.

Signed-off-by: Jann Horn <jannh@google.com>
Reviewed-by: Christian Brauner (Microsoft) <brauner@kernel.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: exynos: Use Exynos5420 compatible for the MIPI video phy

[ Upstream commit 5d5aa219a790d61cad2c38e1aa32058f16ad2f0b ]

For some reason, the driver adding support for Exynos5420 MIPI phy
back in 2016 wasn't used on Exynos5420, which caused a kernel panic.
Add the proper compatible for it.

Signed-off-by: Markuss Broks <markuss.broks@gmail.com>
Link: https://lore.kernel.org/r/20230121201844.46872-2-markuss.broks@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

exit: Detect and fix irq disabled state in oops

[ Upstream commit 001c28e57187570e4b5aa4492c7a957fb6d65d7b ]

If a task oopses with irqs disabled, this can cause various cascading
problems in the oops path such as sleep-from-invalid warnings, and
potentially worse.

Since commit 0258b5fd7c712 ("coredump: Limit coredumps to a single
thread group"), the unconditional irq enable in coredump_task_exit()
will "fix" the irq state to be enabled early in do_exit(), so currently
this may not be triggerable, but that is coincidental and fragile.

Detect and fix the irqs_disabled() condition in the oops path before
calling do_exit(), similarly to the way in_atomic() is handled.

Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Link: https://lore.kernel.org/lkml/20221004094401.708299-1-npiggin@gmail.com/
Signed-off-by: Sasha Levin <sashal@kernel.org>

context_tracking: Fix noinstr vs KASAN

[ Upstream commit 0e26e1de0032779e43929174339429c16307a299 ]

Low level noinstr context-tracking code is calling out to instrumented
code on KASAN:

  vmlinux.o: warning: objtool: __ct_user_enter+0x72: call to __kasan_check_write() leaves .noinstr.text section
  vmlinux.o: warning: objtool: __ct_user_exit+0x47: call to __kasan_check_write() leaves .noinstr.text section

Use even lower level atomic methods to avoid the instrumentation.

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230112195542.458034262@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

udf: Define EFSCORRUPTED error code

[ Upstream commit 3d2d7e61553dbcc8ba45201d8ae4f383742c8202 ]

Similarly to other filesystems define EFSCORRUPTED error code for
reporting internal filesystem corruption.

Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: qcom: msm8996: Add additional A2NoC clocks

[ Upstream commit 67fb53745e0b38275fa0b422b6a3c6c1c028c9a2 ]

On eMMC devices, the UFS clocks aren't started in the bootloader (or well,
at least it should not be, as that would just leak power..), which results
in platform reboots when trying to access the unclocked UFS hardware,
which unfortunately happens on each and every boot, as interconnect calls
sync_state and goes over each and every path.

Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> #db820c
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221210200353.418391-6-konrad.dybcio@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: OMAP2+: omap4-common: Fix refcount leak bug

[ Upstream commit 7c32919a378782c95c72bc028b5c30dfe8c11f82 ]

In omap4_sram_init(), of_find_compatible_node() will return a node
pointer with refcount incremented. We should use of_node_put() when
it is not used anymore.

Signed-off-by: Liang He <windhl@126.com>
Message-Id: <20220628112939.160737-1-windhl@126.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

rpmsg: glink: Release driver_override

[ Upstream commit fb80ef67e8ff6a00d3faad4cb348dafdb8eccfd8 ]

Upon termination of the rpmsg_device, driver_override needs to be freed
to avoid leaking the potentially assigned string.

Fixes: 42cd402b8fd4 ("rpmsg: Fix kfree() of static memory on setting driver_override")
Fixes: 39e47767ec9b ("rpmsg: Add driver_override device attribute for rpmsg_device")
Reviewed-by: Chris Lew <quic_clew@quicinc.com>
Signed-off-by: Bjorn Andersson <quic_bjorande@quicinc.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230109223931.1706429-1-quic_bjorande@quicinc.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

rpmsg: glink: Avoid infinite loop on intent for missing channel

[ Upstream commit 3e74ec2f39362bffbd42854acbb67c7f4cb808f9 ]

In the event that an intent advertisement arrives on an unknown channel
the fifo is not advanced, resulting in the same message being handled
over and over.

Fixes: dacbb35e930f ("rpmsg: glink: Receive and store the remote intent buffers")
Signed-off-by: Bjorn Andersson <quic_bjorande@quicinc.com>
Reviewed-by: Chris Lew <quic_clew@quicinc.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230214234231.2069751-1-quic_bjorande@quicinc.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: saa7134: Use video_unregister_device for radio_dev

[ Upstream commit bc7635c6435c77a0c168e2cc6535740adfaff4e4 ]

The radio device doesn't use vb2, thus calling vb2_video_unregister_device()
which results in the following warning being printed on module unload.

WARNING: CPU: 1 PID: 215963 at drivers/media/common/videobuf2/videobuf2-v4l2.c:1236 vb2_video_unregister_device+0xc6/0xe0 [videobuf2_v4l2]

Fixes: 11788d9b7e91 ("media: media/pci: use vb2_video_unregister_device()")
Signed-off-by: Tasos Sahanidis <tasos@tasossah.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: usb: siano: Fix use after free bugs caused by do_submit_urb

[ Upstream commit ebad8e731c1c06adf04621d6fd327b860c0861b5 ]

There are UAF bugs caused by do_submit_urb(). One of the KASan reports
is shown below:

[   36.403605] BUG: KASAN: use-after-free in worker_thread+0x4a2/0x890
[   36.406105] Read of size 8 at addr ffff8880059600e8 by task kworker/0:2/49
[   36.408316]
[   36.408867] CPU: 0 PID: 49 Comm: kworker/0:2 Not tainted 6.2.0-rc3-15798-g5a41237ad1d4-dir8
[   36.411696] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g15584
[   36.416157] Workqueue:  0x0 (events)
[   36.417654] Call Trace:
[   36.418546]  <TASK>
[   36.419320]  dump_stack_lvl+0x96/0xd0
[   36.420522]  print_address_description+0x75/0x350
[   36.421992]  print_report+0x11b/0x250
[   36.423174]  ? _raw_spin_lock_irqsave+0x87/0xd0
[   36.424806]  ? __virt_addr_valid+0xcf/0x170
[   36.426069]  ? worker_thread+0x4a2/0x890
[   36.427355]  kasan_report+0x131/0x160
[   36.428556]  ? worker_thread+0x4a2/0x890
[   36.430053]  worker_thread+0x4a2/0x890
[   36.431297]  ? worker_clr_flags+0x90/0x90
[   36.432479]  kthread+0x166/0x190
[   36.433493]  ? kthread_blkcg+0x50/0x50
[   36.434669]  ret_from_fork+0x22/0x30
[   36.435923]  </TASK>
[   36.436684]
[   36.437215] Allocated by task 24:
[   36.438289]  kasan_set_track+0x50/0x80
[   36.439436]  __kasan_kmalloc+0x89/0xa0
[   36.440566]  smsusb_probe+0x374/0xc90
[   36.441920]  usb_probe_interface+0x2d1/0x4c0
[   36.443253]  really_probe+0x1d5/0x580
[   36.444539]  __driver_probe_device+0xe3/0x130
[   36.446085]  driver_probe_device+0x49/0x220
[   36.447423]  __device_attach_driver+0x19e/0x1b0
[   36.448931]  bus_for_each_drv+0xcb/0x110
[   36.450217]  __device_attach+0x132/0x1f0
[   36.451470]  bus_probe_device+0x59/0xf0
[   36.452563]  device_add+0x4ec/0x7b0
[   36.453830]  usb_set_configuration+0xc63/0xe10
[   36.455230]  usb_generic_driver_probe+0x3b/0x80
[   36.456166] printk: console [ttyGS0] disabled
[   36.456569]  usb_probe_device+0x90/0x110
[   36.459523]  really_probe+0x1d5/0x580
[   36.461027]  __driver_probe_device+0xe3/0x130
[   36.462465]  driver_probe_device+0x49/0x220
[   36.463847]  __device_attach_driver+0x19e/0x1b0
[   36.465229]  bus_for_each_drv+0xcb/0x110
[   36.466466]  __device_attach+0x132/0x1f0
[   36.467799]  bus_probe_device+0x59/0xf0
[   36.469010]  device_add+0x4ec/0x7b0
[   36.470125]  usb_new_device+0x863/0xa00
[   36.471374]  hub_event+0x18c7/0x2220
[   36.472746]  process_one_work+0x34c/0x5b0
[   36.474041]  worker_thread+0x4b7/0x890
[   36.475216]  kthread+0x166/0x190
[   36.476267]  ret_from_fork+0x22/0x30
[   36.477447]
[   36.478160] Freed by task 24:
[   36.479239]  kasan_set_track+0x50/0x80
[   36.480512]  kasan_save_free_info+0x2b/0x40
[   36.481808]  ____kasan_slab_free+0x122/0x1a0
[   36.483173]  __kmem_cache_free+0xc4/0x200
[   36.484563]  smsusb_term_device+0xcd/0xf0
[   36.485896]  smsusb_probe+0xc85/0xc90
[   36.486976]  usb_probe_interface+0x2d1/0x4c0
[   36.488303]  really_probe+0x1d5/0x580
[   36.489498]  __driver_probe_device+0xe3/0x130
[   36.491140]  driver_probe_device+0x49/0x220
[   36.492475]  __device_attach_driver+0x19e/0x1b0
[   36.493988]  bus_for_each_drv+0xcb/0x110
[   36.495171]  __device_attach+0x132/0x1f0
[   36.496617]  bus_probe_device+0x59/0xf0
[   36.497875]  device_add+0x4ec/0x7b0
[   36.498972]  usb_set_configuration+0xc63/0xe10
[   36.500264]  usb_generic_driver_probe+0x3b/0x80
[   36.501740]  usb_probe_device+0x90/0x110
[   36.503084]  really_probe+0x1d5/0x580
[   36.504241]  __driver_probe_device+0xe3/0x130
[   36.505548]  driver_probe_device+0x49/0x220
[   36.506766]  __device_attach_driver+0x19e/0x1b0
[   36.508368]  bus_for_each_drv+0xcb/0x110
[   36.509646]  __device_attach+0x132/0x1f0
[   36.510911]  bus_probe_device+0x59/0xf0
[   36.512103]  device_add+0x4ec/0x7b0
[   36.513215]  usb_new_device+0x863/0xa00
[   36.514736]  hub_event+0x18c7/0x2220
[   36.516130]  process_one_work+0x34c/0x5b0
[   36.517396]  worker_thread+0x4b7/0x890
[   36.518591]  kthread+0x166/0x190
[   36.519599]  ret_from_fork+0x22/0x30
[   36.520851]
[   36.521405] Last potentially related work creation:
[   36.523143]  kasan_save_stack+0x3f/0x60
[   36.524275]  kasan_record_aux_stack_noalloc+0x9d/0xb0
[   36.525831]  insert_work+0x25/0x130
[   36.527039]  __queue_work+0x4d4/0x620
[   36.528236]  queue_work_on+0x72/0xb0
[   36.529344]  __usb_hcd_giveback_urb+0x13f/0x1b0
[   36.530819]  dummy_timer+0x350/0x1a40
[   36.532149]  call_timer_fn+0x2c/0x190
[   36.533567]  expire_timers+0x69/0x1f0
[   36.534736]  __run_timers+0x289/0x2d0
[   36.535841]  run_timer_softirq+0x2d/0x60
[   36.537110]  __do_softirq+0x116/0x380
[   36.538377]
[   36.538950] Second to last potentially related work creation:
[   36.540855]  kasan_save_stack+0x3f/0x60
[   36.542084]  kasan_record_aux_stack_noalloc+0x9d/0xb0
[   36.543592]  insert_work+0x25/0x130
[   36.544891]  __queue_work+0x4d4/0x620
[   36.546168]  queue_work_on+0x72/0xb0
[   36.547328]  __usb_hcd_giveback_urb+0x13f/0x1b0
[   36.548805]  dummy_timer+0x350/0x1a40
[   36.550116]  call_timer_fn+0x2c/0x190
[   36.551570]  expire_timers+0x69/0x1f0
[   36.552762]  __run_timers+0x289/0x2d0
[   36.553916]  run_timer_softirq+0x2d/0x60
[   36.555118]  __do_softirq+0x116/0x380
[   36.556239]
[   36.556807] The buggy address belongs to the object at ffff888005960000
[   36.556807]  which belongs to the cache kmalloc-4k of size 4096
[   36.560652] The buggy address is located 232 bytes inside of
[   36.560652]  4096-byte region [ffff888005960000, ffff888005961000)
[   36.564791]
[   36.565355] The buggy address belongs to the physical page:
[   36.567212] page:000000004f0a0731 refcount:1 mapcount:0 mapping:0000000000000000 index:0x00
[   36.570534] head:000000004f0a0731 order:3 compound_mapcount:0 subpages_mapcount:0 compound0
[   36.573717] flags: 0x100000000010200(slab|head|node=0|zone=1)
[   36.575481] raw: 0100000000010200 ffff888001042140 dead000000000122 0000000000000000
[   36.577842] raw: 0000000000000000 0000000000040004 00000001ffffffff 0000000000000000
[   36.580175] page dumped because: kasan: bad access detected
[   36.581994]
[   36.582548] Memory state around the buggy address:
[   36.583983]  ffff88800595ff80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   36.586240]  ffff888005960000: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[   36.588884] >ffff888005960080: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[   36.591071]                                                           ^
[   36.593295]  ffff888005960100: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[   36.595705]  ffff888005960180: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[   36.598026] ==================================================================
[   36.600224] Disabling lock debugging due to kernel taint
[   36.602681] general protection fault, probably for non-canonical address 0x43600a000000060I
[   36.607129] CPU: 0 PID: 49 Comm: kworker/0:2 Tainted: G    B              6.2.0-rc3-15798-8
[   36.611115] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g15584
[   36.615026] Workqueue: events do_submit_urb
[   36.616290] RIP: 0010:_raw_spin_lock_irqsave+0x8a/0xd0
[   36.618107] Code: 24 00 00 00 00 48 89 df be 04 00 00 00 e8 9e b5 c6 fe 48 89 ef be 04 00 5
[   36.623522] RSP: 0018:ffff888004b6fcf0 EFLAGS: 00010046
[   36.625072] RAX: 0000000000000000 RBX: 043600a000000060 RCX: ffffffff9fc0e0d7
[   36.627206] RDX: 0000000000000000 RSI: dffffc0000000000 RDI: ffff888004b6fcf0
[   36.629813] RBP: ffff888004b6fcf0 R08: dffffc0000000000 R09: ffffed100096df9f
[   36.631974] R10: dfffe9100096dfa0 R11: 1ffff1100096df9e R12: ffff888005960020
[   36.634285] R13: ffff8880059600f0 R14: 0000000000000246 R15: 0000000000000001
[   36.636438] FS:  0000000000000000(0000) GS:ffff88806d600000(0000) knlGS:0000000000000000
[   36.639092] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   36.640951] CR2: 00007f07476819a3 CR3: 0000000004a34000 CR4: 00000000000006f0
[   36.643411] Call Trace:
[   36.644215]  <TASK>
[   36.644902]  smscore_getbuffer+0x3e/0x1e0
[   36.646147]  do_submit_urb+0x4f/0x190
[   36.647449]  process_one_work+0x34c/0x5b0
[   36.648777]  worker_thread+0x4b7/0x890
[   36.649984]  ? worker_clr_flags+0x90/0x90
[   36.651166]  kthread+0x166/0x190
[   36.652151]  ? kthread_blkcg+0x50/0x50
[   36.653547]  ret_from_fork+0x22/0x30
[   36.655051]  </TASK>
[   36.655733] Modules linked in:
[   36.656787] ---[ end trace 0000000000000000 ]---
[   36.658328] RIP: 0010:_raw_spin_lock_irqsave+0x8a/0xd0
[   36.660045] Code: 24 00 00 00 00 48 89 df be 04 00 00 00 e8 9e b5 c6 fe 48 89 ef be 04 00 5
[   36.665730] RSP: 0018:ffff888004b6fcf0 EFLAGS: 00010046
[   36.667448] RAX: 0000000000000000 RBX: 043600a000000060 RCX: ffffffff9fc0e0d7
[   36.669675] RDX: 0000000000000000 RSI: dffffc0000000000 RDI: ffff888004b6fcf0
[   36.672645] RBP: ffff888004b6fcf0 R08: dffffc0000000000 R09: ffffed100096df9f
[   36.674921] R10: dfffe9100096dfa0 R11: 1ffff1100096df9e R12: ffff888005960020
[   36.677034] R13: ffff8880059600f0 R14: 0000000000000246 R15: 0000000000000001
[   36.679184] FS:  0000000000000000(0000) GS:ffff88806d600000(0000) knlGS:0000000000000000
[   36.681655] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   36.683383] CR2: 00007f07476819a3 CR3: 0000000004a34000 CR4: 00000000000006f0
[   36.685733] Kernel panic - not syncing: Fatal exception
[   36.688585] Kernel Offset: 0x1d400000 from 0xffffffff81000000 (relocation range: 0xfffffff)
[   36.692199] ---[ end Kernel panic - not syncing: Fatal exception ]---

When the siano device is plugged in, it may call the following functions
to initialize the device.

smsusb_probe()-->smsusb_init_device()-->smscore_start_device().

When smscore_start_device() gets failed, the function smsusb_term_device()
will be called and smsusb_device_t will be deallocated. Although we use
usb_kill_urb() in smsusb_stop_streaming() to cancel transfer requests
and wait for them to finish, the worker threads that are scheduled by
smsusb_onresponse() may be still running. As a result, the UAF bugs
could happen.

We add cancel_work_sync() in smsusb_stop_streaming() in order that the
worker threads could finish before the smsusb_device_t is deallocated.

Fixes: dd47fbd40e6e ("[media] smsusb: don't sleep while atomic")
Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: i2c: ov7670: 0 instead of -EINVAL was returned

[ Upstream commit 6a4c664539e6de9b32b65ddcf767ec1bcc1d7f8a ]

If the media bus is unsupported, then return -EINVAL. Instead it
returned 'ret' which happened to be 0.

This fixes a smatch warning:

ov7670.c:1843 ov7670_parse_dt() warn: missing error code? 'ret'

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Fixes: 01b8444828fc ("media: v4l2: i2c: ov7670: Implement OF mbus configuration")
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: atomisp: Only set default_run_mode on first open of a stream/asd

[ Upstream commit 60ec70a71a9f9975a5d2dd4a7d97c20da0e41976 ]

Calling v4l2_ctrl_s_ctrl(asd->run_mode, pipe->default_run_mode) when
the stream is already active (through another /dev/video# node) causes
the stream to stop.

Move the call to set the default run-mode so that it is only done
on the first open of one of the 4 /dev/video# nodes of one of
the 2 streams (atomisp-sub-devices / asd-s).

Fixes: 2c45e343c581 ("media: atomisp: set per-device's default mode")
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: rc: Fix use-after-free bugs caused by ene_tx_irqsim()

[ Upstream commit 29b0589a865b6f66d141d79b2dd1373e4e50fe17 ]

When the ene device is detaching, function ene_remove() will
be called. But there is no function to cancel tx_sim_timer
in ene_remove(), the timer handler ene_tx_irqsim() could race
with ene_remove(). As a result, the UAF bugs could happen,
the process is shown below.

    (cleanup routine)          |        (timer routine)
                               | mod_timer(&dev->tx_sim_timer, ..)
ene_remove()                   | (wait a time)
                               | ene_tx_irqsim()
                               |   dev->hw_lock //USE
                               |   ene_tx_sample(dev) //USE

Fix by adding del_timer_sync(&dev->tx_sim_timer) in ene_remove(),
The tx_sim_timer could stop before ene device is deallocated.

What's more, The rc_unregister_device() and del_timer_sync()
should be called first in ene_remove() and the deallocated
functions such as free_irq(), release_region() and so on
should be called behind them. Because the rc_unregister_device()
is well synchronized. Otherwise, race conditions may happen. The
situations that may lead to race conditions are shown below.

Firstly, the rx receiver is disabled with ene_rx_disable()
before rc_unregister_device() in ene_remove(), which means it
can be enabled again if a process opens /dev/lirc0 between
ene_rx_disable() and rc_unregister_device().

Secondly, the irqaction descriptor is freed by free_irq()
before the rc device is unregistered, which means irqaction
descriptor may be accessed again after it is deallocated.

Thirdly, the timer can call ene_tx_sample() that can write
to the io ports, which means the io ports could be accessed
again after they are deallocated by release_region().

Therefore, the rc_unregister_device() and del_timer_sync()
should be called first in ene_remove().

Suggested by: Sean Young <sean@mess.org>

Fixes: 9ea53b74df9c ("V4L/DVB: STAGING: remove lirc_ene0100 driver")
Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: drivers/media/v4l2-core/v4l2-h264 : add detection of null pointers

[ Upstream commit be3ae7cf4326e95bb1d5413b63baabc26f4a1324 ]

When the pointer variable is judged to be null, null is returned
directly.

[hverkuil: fix two checkpatch warnings]

Signed-off-by: Dong Chuanjian <chuanjian@nfschina.com>
Acked-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Fixes: d3f756ad629b ("media: v4l2: Trace calculated p/b0/b1 initial reflist")
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: amphion: correct the unspecified color space

[ Upstream commit 809060c8a357e020010dd8f797a5efd3c5432b13 ]

in the E.2.1 of Rec. ITU-T H.264 (06/2019),
0 of colour primaries is reserved, and 2 is unspecified.
driver can map V4L2_COLORSPACE_LAST to 0,
and map V4L2_COLORSPACE_DEFAULT to 2.

v4l2_xfer_func and v4l2_ycbcr_encoding are similar case.

Fixes: 3cd084519c6f ("media: amphion: add vpu v4l2 m2m support")
Signed-off-by: Ming Qian <ming.qian@nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: imx-jpeg: Apply clk_bulk api instead of operating specific clk

[ Upstream commit 61fe43dc9f454bc3caa99dbdd8f5fa3ba813981a ]

using the api of clk_bulk can simplify the code.
and the clock of the jpeg codec may be changed,
the clk_bulk api can be compatible with the future change.

Fixes: 4c2e5156d9fa ("media: imx-jpeg: Add pm-runtime support for imx-jpeg")
Signed-off-by: Ming Qian <ming.qian@nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: hantro: Fix JPEG encoder ENUM_FRMSIZE on RK3399

[ Upstream commit 29bd426764dee14a09e37700406f4a5920825fcc ]

Since 79c987de8b354, enumerating framesize on format set with "MODE_NONE"
(any raw formats) is reporting an invalid frmsize.

  Size: Stepwise 0x0 - 0x0 with step 0/0

Before this change, the driver would return EINVAL, which is also invalid
but worked in GStreamer. The original intent was not to implement it, hence
the -ENOTTY return in this change. While drivers should implement
ENUM_FRMSIZE for all formats and queues, this change is limited in scope to
fix the regression.

This fixes taking picture in Gnome Cheese software, or any software using
GSteamer to encode JPEG with hardware acceleration.

Fixes: 79c987de8b35 ("media: hantro: Use post processor scaling capacities")
Reported-by: Robert Mader <robert.mader@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Reviewed-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Tested-by: Robert Mader <robert.mader@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: v4l2-jpeg: ignore the unknown APP14 marker

[ Upstream commit 251c0ea6efd3c3ea0f8a55fdd96c749a98639bd3 ]

The legal identifier of APP14 is "Adobe\0",
but sometimes it may be
"This is an unknown APP marker . Compliant decoders must ignore it."
In this case, just ignore it.
It won't affect the decode result.

Fixes: b8035f7988a8 ("media: Add parsing for APP14 data segment in jpeg helpers")
Signed-off-by: Ming Qian <ming.qian@nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: v4l2-jpeg: correct the skip count in jpeg_parse_app14_data

[ Upstream commit 41959c4f973b837a12061b84d3a436fc64c73a30 ]

The curr pointer has advanced 14 bytes in jpeg_parse_app14_data.
1. jpeg_get_word_be(stream), it goes forward 2 bytes.
2. jpeg_skip(stream, 11), it goes forward 11 bytes.
3. jpeg_get_byte(stream), it goes forward 1 bytes.

so the remain bytes of this segment should be (lp - 2 - 11 - 1),
but not (lp - 2 - 11).

if driver skip 1 extra bytes, the following parsing may go wrong.

Fixes: b8035f7988a8 ("media: Add parsing for APP14 data segment in jpeg helpers")
Signed-off-by: Ming Qian <ming.qian@nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: platform: mtk-mdp3: fix Kconfig dependencies

[ Upstream commit e3f7feb6d89311f369dd4ad903ea62e45328cdbe ]

The new mdp3 driver uses 'select' to force-enable a couple of drivers
it depends on. This is error-prone and likely to cause dependency
loops as well as warnings like:

WARNING: unmet direct dependencies detected for VIDEO_MEDIATEK_VPU
  Depends on [n]: MEDIA_SUPPORT [=m] && MEDIA_PLATFORM_SUPPORT [=y] && MEDIA_PLATFORM_DRIVERS [=y] && V4L_MEM2MEM_DRIVERS [=n] && VIDEO_DEV [=m] && (ARCH_MEDIATEK [=y] || COMPILE_TEST [=y])
  Selected by [m]:
  - VIDEO_MEDIATEK_MDP3 [=m] && MEDIA_SUPPORT [=m] && MEDIA_PLATFORM_SUPPORT [=y] && MEDIA_PLATFORM_DRIVERS [=y] && (MTK_IOMMU [=m] || COMPILE_TEST [=y]) && VIDEO_DEV [=m] && (ARCH_MEDIATEK [=y] || COMPILE_TEST [=y]) && HAS_DMA [=y] && REMOTEPROC [=y]

This specific warning was already addressed in a previous patch,
but there are similar unnecessary 'select' statements, so turn those
into 'depends on'. This also means the dependency on ARCH_MEDIATEK
is redundant and can be dropped.

Fixes: 61890ccaefaf ("media: platform: mtk-mdp3: add MediaTek MDP3 driver")
Fixes: 9195a860ef0a ("media: platform: mtk-mdp3: remove unused VIDEO_MEDIATEK_VPU config")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: platform: mtk-mdp3: remove unused VIDEO_MEDIATEK_VPU config

[ Upstream commit 9195a860ef0a384d2ca2065cc61a0cc80d620de5 ]

Since REMOTEPROC completely replaces the VIDEO_MEDIATEK_VPU in MDP3,
unused config should be removed to avoid compilation warnings
reported on i386 or x86_64.

Warning messages:
    WARNING: unmet direct dependencies detected for VIDEO_MEDIATEK_VPU
          Depends on [n]: MEDIA_SUPPORT [=y] && MEDIA_PLATFORM_SUPPORT [=y]
        && MEDIA_PLATFORM_DRIVERS [=y] && V4L_MEM2MEM_DRIVERS [=n] &&
        VIDEO_DEV [=y] && (ARCH_MEDIATEK || COMPILE_TEST [=y])
          Selected by [y]:
          - VIDEO_MEDIATEK_MDP3 [=y] && MEDIA_SUPPORT [=y] &&
        MEDIA_PLATFORM_SUPPORT [=y] && MEDIA_PLATFORM_DRIVERS [=y] &&
        (MTK_IOMMU [=n] || COMPILE_TEST [=y]) && VIDEO_DEV [=y] &&
        (ARCH_MEDIATEK || COMPILE_TEST [=y]) && HAS_DMA [=y] && REMOTEPROC
        [=y]

Fixes: 61890ccaefaf ("media: platform: mtk-mdp3: add MediaTek MDP3 driver")
Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Stable-dep-of: e3f7feb6d893 ("media: platform: mtk-mdp3: fix Kconfig dependencies")
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: camss: csiphy-3ph: avoid undefined behavior

[ Upstream commit 05fb9ace34b8645cb76f7e3a21b5c7b754329cae ]

Marking a case of the switch statement as unreachable means the
compiler treats it as undefined behavior, which is then caught by
an objtool warning:

drivers/media/platform/qcom/camss/camss-csiphy-3ph-1-0.o: warning: objtool: csiphy_lanes_enable() falls through to next function csiphy_lanes_disable()

Instead of simply continuing execution at a random place of the
driver, print a warning and return from to the caller, which
makes it possible to understand what happens and avoids the
warning.

Fixes: 53655d2a0ff2 ("media: camss: csiphy-3ph: add support for SM8250 CSI DPHY")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: platform: mtk-mdp3: Fix return value check in mdp_probe()

[ Upstream commit 1963689bed4d500236938d90c91cdd5e63c1eb28 ]

In case of error, the function mtk_mutex_get()
returns ERR_PTR() and never returns NULL. The NULL test in the
return value check should be replaced with IS_ERR().
And also fix the err_free_mutex case.

Fixes: 61890ccaefaf ("media: platform: mtk-mdp3: add MediaTek MDP3 driver")
Signed-off-by: Qiheng Lin <linqiheng@huawei.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: i2c: imx219: Fix binning for RAW8 capture

[ Upstream commit ef86447e775fb1f2ced00d4c7fff2c0a1c63f165 ]

2x2 binning works fine for RAW10 capture, but for RAW8 1232p mode it
leads to corrupted frames [1][2].

Using the special 2x2 analog binning mode fixes the issue, but causes
artefacts for RAW10 1232p capture. So here we choose the binning mode
depending upon the frame format selected.

As both binning modes work fine for 480p RAW8 and RAW10 capture, it can
share the same code path as 1232p for selecting binning mode.

[1] https://forums.raspberrypi.com/viewtopic.php?t=332103
[2] https://github.com/raspberrypi/libcamera-apps/issues/281

Fixes: 22da1d56e982 ("media: i2c: imx219: Add support for RAW8 bit bayer format")
Signed-off-by: Jai Luthra <j-luthra@ti.com>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: i2c: imx219: Split common registers from mode tables

[ Upstream commit 8508455961d5a9e8907bcfd8dcd58f19d9b6ce47 ]

There are four modes, and each mode has a table of registers.
Some of the registers are common to all modes, so create new
tables for these common registers to reduce duplicate code.

Signed-off-by: Adam Ford <aford173@gmail.com>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Stable-dep-of: ef86447e775f ("media: i2c: imx219: Fix binning for RAW8 capture")
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: i2c: ov772x: Fix memleak in ov772x_probe()

[ Upstream commit 7485edb2b6ca5960205c0a49bedfd09bba30e521 ]

A memory leak was reported when testing ov772x with bpf mock device:

AssertionError: unreferenced object 0xffff888109afa7a8 (size 8):
  comm "python3", pid 279, jiffies 4294805921 (age 20.681s)
  hex dump (first 8 bytes):
    80 22 88 15 81 88 ff ff                          ."......
  backtrace:
    [<000000009990b438>] __kmalloc_node+0x44/0x1b0
    [<000000009e32f7d7>] kvmalloc_node+0x34/0x180
    [<00000000faf48134>] v4l2_ctrl_handler_init_class+0x11d/0x180 [videodev]
    [<00000000da376937>] ov772x_probe+0x1c3/0x68c [ov772x]
    [<000000003f0d225e>] i2c_device_probe+0x28d/0x680
    [<00000000e0b6db89>] really_probe+0x17c/0x3f0
    [<000000001b19fcee>] __driver_probe_device+0xe3/0x170
    [<0000000048370519>] driver_probe_device+0x49/0x120
    [<000000005ead07a0>] __device_attach_driver+0xf7/0x150
    [<0000000043f452b8>] bus_for_each_drv+0x114/0x180
    [<00000000358e5596>] __device_attach+0x1e5/0x2d0
    [<0000000043f83c5d>] bus_probe_device+0x126/0x140
    [<00000000ee0f3046>] device_add+0x810/0x1130
    [<00000000e0278184>] i2c_new_client_device+0x359/0x4f0
    [<0000000070baf34f>] of_i2c_register_device+0xf1/0x110
    [<00000000a9f2159d>] of_i2c_notify+0x100/0x160
unreferenced object 0xffff888119825c00 (size 256):
  comm "python3", pid 279, jiffies 4294805921 (age 20.681s)
  hex dump (first 32 bytes):
    00 b4 a5 17 81 88 ff ff 00 5e 82 19 81 88 ff ff  .........^......
    10 5c 82 19 81 88 ff ff 10 5c 82 19 81 88 ff ff  .\.......\......
  backtrace:
    [<000000009990b438>] __kmalloc_node+0x44/0x1b0
    [<000000009e32f7d7>] kvmalloc_node+0x34/0x180
    [<0000000073d88e0b>] v4l2_ctrl_new.cold+0x19b/0x86f [videodev]
    [<00000000b1f576fb>] v4l2_ctrl_new_std+0x16f/0x210 [videodev]
    [<00000000caf7ac99>] ov772x_probe+0x1fa/0x68c [ov772x]
    [<000000003f0d225e>] i2c_device_probe+0x28d/0x680
    [<00000000e0b6db89>] really_probe+0x17c/0x3f0
    [<000000001b19fcee>] __driver_probe_device+0xe3/0x170
    [<0000000048370519>] driver_probe_device+0x49/0x120
    [<000000005ead07a0>] __device_attach_driver+0xf7/0x150
    [<0000000043f452b8>] bus_for_each_drv+0x114/0x180
    [<00000000358e5596>] __device_attach+0x1e5/0x2d0
    [<0000000043f83c5d>] bus_probe_device+0x126/0x140
    [<00000000ee0f3046>] device_add+0x810/0x1130
    [<00000000e0278184>] i2c_new_client_device+0x359/0x4f0
    [<0000000070baf34f>] of_i2c_register_device+0xf1/0x110

The reason is that if priv->hdl.error is set, ov772x_probe() jumps to the
error_mutex_destroy without doing v4l2_ctrl_handler_free(), and all
resources allocated in v4l2_ctrl_handler_init() and v4l2_ctrl_new_std()
are leaked.

Fixes: 1112babde214 ("media: i2c: Copy ov772x soc_camera sensor driver")
Signed-off-by: Yuan Can <yuancan@huawei.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: mc: Get media_device directly from pad

[ Upstream commit a967a3a788028f541e4db54beabcebc3648997db ]

Various functions access the media_device from a pad by going through
the entity the pad belongs to. Remove the level of indirection and get
the media_device from the pad directly.

Fixes: 9e3576a1ae2b ("media: mc: convert pipeline funcs to take media_pad")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: ov5640: Handle delays when no reset_gpio set

[ Upstream commit d7ff69139908842adf824be4f50c7e9ac5886c04 ]

Some module manufacturers [1][2] don't expose the RESETB and PWDN pins
of the sensor directly through the 15-pin FFC connector. Instead wiring
~PWDN gpio to the sensor pins with appropriate delays.

In such cases, reset_gpio will not be available to the driver, but it
will still be toggled when the sensor is powered on, and thus we should
still honor the wait time of >= 5ms + 1ms + 20ms (see figure 2-3 in [3])
before attempting any i/o operations over SCCB.

Also, rename the function to ov5640_powerup_sequence to better match the
datasheet (section 2.7).

[1] https://digilent.com/reference/_media/reference/add-ons/pcam-5c/pcam_5c_sch.pdf
[2] https://www.alinx.com/public/upload/file/AN5641_User_Manual.pdf
[3] https://cdn.sparkfun.com/datasheets/Sensors/LightImaging/OV5640_datasheet.pdf

Fixes: 19a81c1426c1 ("[media] add Omnivision OV5640 sensor driver")
Reported-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Jai Luthra <j-luthra@ti.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: ov5640: Fix soft reset sequence and timings

[ Upstream commit decea0a98b7ac04536c7d659f74783e8d67a06c0 ]

Move the register-based reset out of the init_setting[] and into the
powerup_sequence function. The sensor is power cycled and reset using
the gpio pins so the soft reset is not always necessary.

This also ensures that soft reset honors the timing sequence
from the datasheet [1].

[1] https://cdn.sparkfun.com/datasheets/Sensors/LightImaging/OV5640_datasheet.pdf

Fixes: 19a81c1426c1 ("[media] add Omnivision OV5640 sensor driver")
Reported-by: Nishanth Menon <nm@ti.com>
Suggested-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jai Luthra <j-luthra@ti.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonaboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: ov5675: Fix memleak in ov5675_init_controls()

[ Upstream commit dd74ed6c213003533e3abf4c204374ef01d86978 ]

There is a kmemleak when testing the media/i2c/ov5675.c with bpf mock
device:

AssertionError: unreferenced object 0xffff888107362160 (size 16):
  comm "python3", pid 277, jiffies 4294832798 (age 20.722s)
  hex dump (first 16 bytes):
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<00000000abe7d67c>] __kmalloc_node+0x44/0x1b0
    [<000000008a725aac>] kvmalloc_node+0x34/0x180
    [<000000009a53cd11>] v4l2_ctrl_handler_init_class+0x11d/0x180
[videodev]
    [<0000000055b46db0>] ov5675_probe+0x38b/0x897 [ov5675]
    [<00000000153d886c>] i2c_device_probe+0x28d/0x680
    [<000000004afb7e8f>] really_probe+0x17c/0x3f0
    [<00000000ff2f18e4>] __driver_probe_device+0xe3/0x170
    [<000000000a001029>] driver_probe_device+0x49/0x120
    [<00000000e39743c7>] __device_attach_driver+0xf7/0x150
    [<00000000d32fd070>] bus_for_each_drv+0x114/0x180
    [<000000009083ac41>] __device_attach+0x1e5/0x2d0
    [<0000000015b4a830>] bus_probe_device+0x126/0x140
    [<000000007813deaf>] device_add+0x810/0x1130
    [<000000007becb867>] i2c_new_client_device+0x386/0x540
    [<000000007f9cf4b4>] of_i2c_register_device+0xf1/0x110
    [<00000000ebfdd032>] of_i2c_notify+0xfc/0x1f0

ov5675_init_controls() won't clean all the allocated resources in fail
path, which may causes the memleaks. Add v4l2_ctrl_handler_free() to
prevent memleak.

Fixes: bf27502b1f3b ("media: ov5675: Add support for OV5675 sensor")
Signed-off-by: Shang XiaoJing <shangxiaojing@huawei.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: ov2740: Fix memleak in ov2740_init_controls()

[ Upstream commit 2d899592ed7829d0d5140853bac4d58742a6b8af ]

There is a kmemleak when testing the media/i2c/ov2740.c with bpf mock
device:

unreferenced object 0xffff8881090e19e0 (size 16):
  comm "51-i2c-ov2740", pid 278, jiffies 4294781584 (age 23.613s)
  hex dump (first 16 bytes):
    00 f3 7c 0b 81 88 ff ff 80 75 6a 09 81 88 ff ff  ..|......uj.....
  backtrace:
    [<000000004e9fad8f>] __kmalloc_node+0x44/0x1b0
    [<0000000039c802f4>] kvmalloc_node+0x34/0x180
    [<000000009b8b5c63>] v4l2_ctrl_handler_init_class+0x11d/0x180
[videodev]
    [<0000000038644056>] ov2740_probe+0x37d/0x84f [ov2740]
    [<0000000092489f59>] i2c_device_probe+0x28d/0x680
    [<000000001038babe>] really_probe+0x17c/0x3f0
    [<0000000098c7af1c>] __driver_probe_device+0xe3/0x170
    [<00000000e1b3dc24>] device_driver_attach+0x34/0x80
    [<000000005a04a34d>] bind_store+0x10b/0x1a0
    [<00000000ce25d4f2>] drv_attr_store+0x49/0x70
    [<000000007d9f4e9a>] sysfs_kf_write+0x8c/0xb0
    [<00000000be6cff0f>] kernfs_fop_write_iter+0x216/0x2e0
    [<0000000031ddb40a>] vfs_write+0x658/0x810
    [<0000000041beecdd>] ksys_write+0xd6/0x1b0
    [<0000000023755840>] do_syscall_64+0x38/0x90
    [<00000000b2cc2da2>] entry_SYSCALL_64_after_hwframe+0x63/0xcd

ov2740_init_controls() won't clean all the allocated resources in fail
path, which may causes the memleaks. Add v4l2_ctrl_handler_free() to
prevent memleak.

Fixes: 866edc895171 ("media: i2c: Add ov2740 image sensor driver")
Signed-off-by: Shang XiaoJing <shangxiaojing@huawei.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: max9286: Fix memleak in max9286_v4l2_register()

[ Upstream commit 8636c5fc7658c7c6299fb8b352d24ea4b9ba99e2 ]

There is a kmemleak when testing the media/i2c/max9286.c with bpf mock
device:

kmemleak: 5 new suspected memory leaks (see /sys/kernel/debug/kmemleak)

unreferenced object 0xffff88810defc400 (size 256):
  comm "python3", pid 278, jiffies 4294737563 (age 31.978s)
  hex dump (first 32 bytes):
    28 06 a7 0a 81 88 ff ff 00 fe 22 12 81 88 ff ff  (.........".....
    10 c4 ef 0d 81 88 ff ff 10 c4 ef 0d 81 88 ff ff  ................
  backtrace:
    [<00000000191de6a7>] __kmalloc_node+0x44/0x1b0
    [<000000002f4912b7>] kvmalloc_node+0x34/0x180
    [<0000000057dc4cae>] v4l2_ctrl_new+0x325/0x10f0 [videodev]
    [<0000000026030272>] v4l2_ctrl_new_std+0x16f/0x210 [videodev]
    [<00000000f0d9ea2f>] max9286_probe+0x76e/0xbff [max9286]
    [<00000000ea8f6455>] i2c_device_probe+0x28d/0x680
    [<0000000087529af3>] really_probe+0x17c/0x3f0
    [<00000000b08be526>] __driver_probe_device+0xe3/0x170
    [<000000004382edea>] driver_probe_device+0x49/0x120
    [<000000007bde528a>] __device_attach_driver+0xf7/0x150
    [<000000009f9c6ab4>] bus_for_each_drv+0x114/0x180
    [<00000000c8aaf588>] __device_attach+0x1e5/0x2d0
    [<0000000041cc06b9>] bus_probe_device+0x126/0x140
    [<000000002309860d>] device_add+0x810/0x1130
    [<000000002827bf98>] i2c_new_client_device+0x359/0x4f0
    [<00000000593bdc85>] of_i2c_register_device+0xf1/0x110

max9286_v4l2_register() calls v4l2_ctrl_new_std(), but won't free the
created v412_ctrl when fwnode_graph_get_endpoint_by_id() failed, which
causes the memleak. Call v4l2_ctrl_handler_free() to free the v412_ctrl.

Fixes: 66d8c9d2422d ("media: i2c: Add MAX9286 driver")
Signed-off-by: Shang XiaoJing <shangxiaojing@huawei.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

builddeb: clean generated package content

[ Upstream commit c9f9cf2560e40b62015c6c4a04be60f55ce5240e ]

For each binary Debian package, a directory with the package name is
created in the debian directory. Correct the generated file matches in the
package's clean target, which were renamed without adjusting the target.

Fixes: 1694e94e4f46 ("builddeb: match temporary directory name to the package name")
Signed-off-by: Bastian Germann <bage@linutronix.de>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/vdso: Drop '-shared' from KBUILD_CFLAGS_64

[ Upstream commit fd8589dce8107e2ce62e92f76089654462dd67b4 ]

When clang's -Qunused-arguments is dropped from KBUILD_CPPFLAGS, it
points out that there is a linking phase flag added to CFLAGS, which
will only be used for compiling

  clang-16: error: argument unused during compilation: '-shared' [-Werror,-Wunused-command-line-argument]

'-shared' is already present in ldflags-y so it can just be dropped.

Fixes: 2b2a25845d53 ("s390/vdso: Use $(LD) instead of $(CC) to link vDSO")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Acked-by: Heiko Carstens <hca@linux.ibm.com>
Reviewed-by: Sven Schnelle <svens@linux.ibm.com>
Tested-by: Linux Kernel Functional Testing <lkft@linaro.org>
Tested-by: Anders Roxell <anders.roxell@linaro.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc: Remove linker flag from KBUILD_AFLAGS

[ Upstream commit 31f48f16264bc70962fb3e7ec62da64d0a2ba04a ]

When clang's -Qunused-arguments is dropped from KBUILD_CPPFLAGS, it
points out that KBUILD_AFLAGS contains a linker flag, which will be
unused:

  clang: error: -Wl,-a32: 'linker' input unused [-Werror,-Wunused-command-line-argument]

This was likely supposed to be '-Wa,-a$(BITS)'. However, this change is
unnecessary, as all supported versions of clang and gcc will pass '-a64'
or '-a32' to GNU as based on the value of '-m'; the behavior of the
latest stable release of the oldest supported major version of each
compiler is shown below and each compiler's latest release exhibits the
same behavior (GCC 12.2.0 and Clang 15.0.6).

  $ powerpc64-linux-gcc --version | head -1
  powerpc64-linux-gcc (GCC) 5.5.0

  $ powerpc64-linux-gcc -m64 -### -x assembler-with-cpp -c -o /dev/null /dev/null &| grep 'as '
  .../as -a64 -mppc64 -many -mbig -o /dev/null /tmp/cctwuBzZ.s

  $ powerpc64-linux-gcc -m32 -### -x assembler-with-cpp -c -o /dev/null /dev/null &| grep 'as '
  .../as -a32 -mppc -many -mbig -o /dev/null /tmp/ccaZP4mF.sg

  $ clang --version | head -1
  Ubuntu clang version 11.1.0-++20211011094159+1fdec59bffc1-1~exp1~20211011214622.5

  $ clang --target=powerpc64-linux-gnu -fno-integrated-as -m64 -### \
    -x assembler-with-cpp -c -o /dev/null /dev/null &| grep gnu-as
   "/usr/bin/powerpc64-linux-gnu-as" "-a64" "-mppc64" "-many" "-o" "/dev/null" "/tmp/null-80267c.s"

  $ clang --target=powerpc64-linux-gnu -fno-integrated-as -m64 -### \
    -x assembler-with-cpp -c -o /dev/null /dev/null &| grep gnu-as
   "/usr/bin/powerpc64-linux-gnu-as" "-a32" "-mppc" "-many" "-o" "/dev/null" "/tmp/null-ab8f8d.s"

Remove this flag altogether to avoid future issues.

Fixes: 1421dc6d4829 ("powerpc/kbuild: Use flags variables rather than overriding LD/CC/AS")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Linux Kernel Functional Testing <lkft@linaro.org>
Tested-by: Anders Roxell <anders.roxell@linaro.org>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: imx: imx7-media-csi: fix missing clk_disable_unprepare() in imx7_csi_init()

[ Upstream commit cea606d9e996a77eed57fc60709e0728341450e3 ]

Add missing clk_disable_unprepare(), if imx7_csi_dma_setup() fails
in imx7_csi_init().

Fixes: ff43ca911978 ("media: imx: imx7-media-csi: Move CSI configuration before source start")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: platform: ti: Add missing check for devm_regulator_get

[ Upstream commit da8e05f84a11c3cc3b0ba0a3c62d20e358002d99 ]

Add check for the return value of devm_regulator_get since it may return
error pointer.

Fixes: 448de7e7850b ("[media] omap3isp: OMAP3 ISP core")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: ti: cal: fix possible memory leak in cal_ctx_create()

[ Upstream commit 7acd650a0484d92985a0d6d867d980c6dd019885 ]

The memory of ctx is allocated in cal_ctx_create(), but it will
not be freed when cal_ctx_v4l2_init() fails, so add kfree() when
cal_ctx_v4l2_init() fails to fix it.

Fixes: d68a94e98a89 ("media: ti-vpe: cal: Split video device initialization and registration")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

remoteproc: qcom_q6v5_mss: Use a carveout to authenticate modem headers

[ Upstream commit 57f72170a2b2a362c35bb9407fc844eac5afdec1 ]

Any access to the dynamically allocated metadata region by the application
processor after assigning it to the remote Q6 will result in a XPU
violation. Fix this by replacing the dynamically allocated memory region
with a no-map carveout and unmap the modem metadata memory region before
passing control to the remote Q6.

Reported-and-tested-by: Amit Pundir <amit.pundir@linaro.org>
Fixes: 6c5a9dc2481b ("remoteproc: qcom: Make secure world call for mem ownership switch")
Signed-off-by: Sibi Sankar <quic_sibis@quicinc.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230117085840.32356-7-quic_sibis@quicinc.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

Revert "remoteproc: qcom_q6v5_mss: map/unmap metadata region before/after use"

[ Upstream commit a899d542b687c9b04ccbd9eefabc829ba5fef791 ]

This reverts commit fc156629b23a21181e473e60341e3a78af25a1d4.

This commit manages to do three API violations at once:

 - dereference the return value of dma_alloc_attrs with the
   DMA_ATTR_NO_KERNEL_MAPPING mapping, which is clearly forbidden and
   will do the wrong thing on various dma mapping implementations.  The
   fact that dma-direct uses a struct page as a cookie is an undocumented
   implementation detail
 - include dma-map-ops.h and use pgprot_dmacoherent despite a clear
   comment documenting that this is not acceptable
 - use of the VM_DMA_COHERENT for something that is not the dma-mapping
   code
 - use of VM_FLUSH_RESET_PERMS for vmap, while it is only supported for
   vmalloc

Acked-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sibi Sankar <quic_sibis@quicinc.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230117085840.32356-6-quic_sibis@quicinc.com
Stable-dep-of: 57f72170a2b2 ("remoteproc: qcom_q6v5_mss: Use a carveout to authenticate modem headers")
Signed-off-by: Sasha Levin <sashal@kernel.org>

IB/hfi1: Fix sdma.h tx->num_descs off-by-one errors

[ Upstream commit fd8958efe8779d3db19c9124fce593ce681ac709 ]

Fix three sources of error involving struct sdma_txreq.num_descs.

When _extend_sdma_tx_descs() extends the descriptor array, it uses the
value of tx->num_descs to determine how many existing entries from the
tx's original, internal descriptor array to copy to the newly allocated
one.  As this value was incremented before the call, the copy loop will
access one entry past the internal descriptor array, copying its contents
into the corresponding slot in the new array.

If the call to _extend_sdma_tx_descs() fails, _pad_smda_tx_descs() then
invokes __sdma_tx_clean() which uses the value of tx->num_desc to drive a
loop that unmaps all descriptor entries in use.  As this value was
incremented before the call, the unmap loop will invoke sdma_unmap_desc()
on a descriptor entry whose contents consist of whatever random data was
copied into it during (1), leading to cascading further calls into the
kernel and driver using arbitrary data.

_sdma_close_tx() was using tx->num_descs instead of tx->num_descs - 1.

Fix all of the above by:
- Only increment .num_descs after .descp is extended.
- Use .num_descs - 1 instead of .num_descs for last .descp entry.

Fixes: f4d26d81ad7f ("staging/rdma/hfi1: Add coalescing support for SDMA TX descriptors")
Link: https://lore.kernel.org/r/167656658879.2223096.10026561343022570690.stgit@awfm-02.cornelisnetworks.com
Signed-off-by: Brendan Cunningham <bcunningham@cornelisnetworks.com>
Signed-off-by: Patrick Kelsey <pat.kelsey@cornelisnetworks.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

IB/hfi1: Fix math bugs in hfi1_can_pin_pages()

[ Upstream commit a0d198f79a8d033bd46605b779859193649f1f99 ]

Fix arithmetic and logic errors in hfi1_can_pin_pages() that  would allow
hfi1 to attempt pinning pages in cases where it should not because of
resource limits or lack of required capability.

Fixes: 2c97ce4f3c29 ("IB/hfi1: Add pin query function")
Link: https://lore.kernel.org/r/167656658362.2223096.10954762619837718026.stgit@awfm-02.cornelisnetworks.com
Signed-off-by: Brendan Cunningham <bcunningham@cornelisnetworks.com>
Signed-off-by: Patrick Kelsey <pat.kelsey@cornelisnetworks.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/rxe: Fix missing memory barriers in rxe_queue.h

[ Upstream commit a77a52385e9a761f896a88a4162e69fb7ccafe3f ]

An earlier patch which introduced smp_load_acquire/smp_store_release
into rxe_queue.h incorrectly assumed that surrounding spin-locks in
rxe_verbs.c around queue updates for kernel ulps was sufficient to
protect the passing of data through the queues between the ulp and
the rxe tasklets. But this was incorrect. The typical sequence was

ulp rxe requester tasklet
------------------------ ---------------------
spin_lock_irqsave() wqe = queue_head(queue)
if (!queue_full(q)) { if (!wqe)
spin_unlock_irqrestore return;
return -ENOMEM
} <process wqe>
wqe = queue_producer_addr(q)
<fill in wqe> queue_advance_consumer(queue)
queue_advance_producer(q)
spin_unlock_irqrestore()

queue_head() calls queue_empty() which calls smp_load_acquire()
For user space apps queue_advance_producer() calls smp_store_release()
so that there is a memory barrier between the producer and the
consumer but for kernel ulps queue_advance_produce() just incremented
the producer index because the lock function is a release function.
But to work the barrier has to come between filling in the wqe and
updating the producer index. This patch adds the missing barriers.
It also changes the enum names for the ulp queue types to
QUEUE_TYPE_FROM/TO_ULP instead of QUEUE_TYPE_TO/FROM_DRIVER
which is very ambiguous. This bug is suspected as the cause of very
rare lockups in a very high scale storage application. It is a bug
in any case and should be corrected.

Fixes: 0a67c46d2e99 ("RDMA/rxe: Protect user space index loads/stores")
Link: https://lore.kernel.org/r/20230214071053.5395-1-rpearsonhpe@gmail.com
Signed-off-by: Bob Pearson <rpearsonhpe@gmail.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/rxe: cleanup some error handling in rxe_verbs.c

[ Upstream commit 692373d186205dfb1b56f35f22702412d94d9420 ]

Instead of 'goto and return', just return directly to
simplify the error handling, and avoid some unnecessary
return value check.

Link: https://lore.kernel.org/r/20221028075053.3990467-1-xuhaoyue1@hisilicon.com
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Haoyue Xu <xuhaoyue1@hisilicon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Stable-dep-of: a77a52385e9a ("RDMA/rxe: Fix missing memory barriers in rxe_queue.h")
Signed-off-by: Sasha Levin <sashal@kernel.org>

iommu/vt-d: Allow to use flush-queue when first level is default

[ Upstream commit 257ec290741924f8df678927d0dfecb1deebb9c5 ]

Commit 29b32839725f ("iommu/vt-d: Do not use flush-queue when caching-mode
is on") forced default domains to be strict mode as long as IOMMU
caching-mode is flagged. The reason for doing this is that when vIOMMU
uses VT-d caching mode to synchronize shadowing page tables, the strict
mode shows better performance.

However, this optimization is orthogonal to the first-level page table
because the Intel VT-d architecture does not define the caching mode of
the first-level page table. Refer to VT-d spec, section 6.1, "When the
CM field is reported as Set, any software updates to remapping
structures other than first-stage mapping (including updates to not-
present entries or present entries whose programming resulted in
translation faults) requires explicit invalidation of the caches."
Exclude the first-level page table from this optimization.

Generally using first-stage translation in vIOMMU implies nested
translation enabled in the physical IOMMU. In this case the first-stage
page table is wholly captured by the guest. The vIOMMU only needs to
transfer the cache invalidations on vIOMMU to the physical IOMMU.
Forcing the default domain to strict mode will cause more frequent
cache invalidations, resulting in performance degradation. In a real
performance benchmark test measured by iperf receive, the performance
result on Sapphire Rapids 100Gb NIC shows:
w/ this fix ~51 Gbits/s, w/o this fix ~39.3 Gbits/s.

Theoretically a first-stage IOMMU page table can still be shadowed
in absence of the caching mode, e.g. with host write-protecting guest
IOMMU page table to synchronize changed PTEs with the physical
IOMMU page table. In this case the shadowing overhead is decoupled
from emulating IOTLB invalidation then the overhead of the latter part
is solely decided by the frequency of IOTLB invalidations. Hence
allowing guest default dma domain to be lazy can also benefit the
overall performance by reducing the total VM-exit numbers.

Fixes: 29b32839725f ("iommu/vt-d: Do not use flush-queue when caching-mode is on")
Reported-by: Sanjay Kumar <sanjay.k.kumar@intel.com>
Suggested-by: Sanjay Kumar <sanjay.k.kumar@intel.com>
Signed-off-by: Tina Zhang <tina.zhang@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20230214025618.2292889-1-tina.zhang@intel.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iommu/vt-d: Fix error handling in sva enable/disable paths

[ Upstream commit 60b1daa3b168fbc648ae2ad28a84759223e49e18 ]

Roll back all previous actions in error paths of intel_iommu_enable_sva()
and intel_iommu_disable_sva().

Fixes: d5b9e4bfe0d8 ("iommu/vt-d: Report prq to io-pgfault framework")
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20230208051559.700109-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: ptdma: check for null desc before calling pt_cmd_callback

[ Upstream commit 928469986171a6f763b34b039427f5667ba3fd50 ]

Resolves a panic that can occur on AMD systems, typically during host
shutdown, after the PTDMA driver had been exercised. The issue was
the pt_issue_pending() function is mistakenly assuming that there will
be at least one descriptor in the Submitted queue when the function
is called. However, it is possible that both the Submitted and Issued
queues could be empty, which could result in pt_cmd_callback() being
mistakenly called with a NULL pointer.
Ref: Bugzilla Bug 216856.

Fixes: 6fa7e0e836e2 ("dmaengine: ptdma: fix concurrency issue with multiple dma transfer")
Signed-off-by: Eric Pilmore <epilmore@gigaio.com>
Link: https://lore.kernel.org/r/20230210075142.58253-1-epilmore@gigaio.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: dw-axi-dmac: Do not dereference NULL structure

[ Upstream commit be4d46edeee4b2459d2f53f37ada88bbfb634b6c ]

If "vdesc" is NULL, it cannot be used with vd_to_axi_desc(). Leave
"bytes" unchanged at 0. Seen under GCC 13 with -Warray-bounds:

../drivers/dma/dw-axi-dmac/dw-axi-dmac-platform.c: In function 'dma_chan_tx_status':
../drivers/dma/dw-axi-dmac/dw-axi-dmac-platform.c:329:46: warning: array subscript 0 is outside array bounds of 'struct
virt_dma_desc[46116860184273879]' [-Warray-bounds=]
  329 |                 bytes = vd_to_axi_desc(vdesc)->length;
      |                                              ^~

Fixes: 8e55444da65c ("dmaengine: dw-axi-dmac: Support burst residue granularity")
Cc: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Cc: Vinod Koul <vkoul@kernel.org>
Cc: dmaengine@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230127223623.never.507-kees@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: sf-pdma: pdma_desc memory leak fix

[ Upstream commit b02e07015a5ac7bbc029da931ae17914b8ae0339 ]

Commit b2cc5c465c2c ("dmaengine: sf-pdma: Add multithread support for a
DMA channel") changed sf_pdma_prep_dma_memcpy() to unconditionally
allocate a new sf_pdma_desc each time it is called.

The driver previously recycled descs, by checking the in_use flag, only
allocating additional descs if the existing one was in use. This logic
was removed in commit b2cc5c465c2c ("dmaengine: sf-pdma: Add multithread
support for a DMA channel"), but sf_pdma_free_desc() was not changed to
handle the new behaviour.

As a result, each time sf_pdma_prep_dma_memcpy() is called, the previous
descriptor is leaked, over time leading to memory starvation:

  unreferenced object 0xffffffe008447300 (size 192):
  comm "irq/39-mchp_dsc", pid 343, jiffies 4294906910 (age 981.200s)
  hex dump (first 32 bytes):
    00 00 00 ff 00 00 00 00 b8 c1 00 00 00 00 00 00  ................
    00 00 70 08 10 00 00 00 00 00 00 c0 00 00 00 00  ..p.............
  backtrace:
    [<00000000064a04f4>] kmemleak_alloc+0x1e/0x28
    [<00000000018927a7>] kmem_cache_alloc+0x11e/0x178
    [<000000002aea8d16>] sf_pdma_prep_dma_memcpy+0x40/0x112

Add the missing kfree() to sf_pdma_free_desc(), and remove the redundant
in_use flag.

Fixes: b2cc5c465c2c ("dmaengine: sf-pdma: Add multithread support for a DMA channel")
Signed-off-by: Shravan Chippa <shravan.chippa@microchip.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://lore.kernel.org/r/20230120100623.3530634-1-shravan.chippa@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iommu/amd: Do not identity map v2 capable device when snp is enabled

[ Upstream commit 18792e99ea2fea27c72eb1ecca1879e5e6be304d ]

Flow:
  - Booted system with SNP enabled, memory encryption off and
    IOMMU DMA translation mode
  - AMD driver detects v2 capable device and amd_iommu_def_domain_type()
    returns identity mode
  - amd_iommu_domain_alloc() returns NULL an SNP is enabled
  - System will fail to register device

On SNP enabled system, passthrough mode is not supported. IOMMU default
domain is set to translation mode. We need to return zero from
amd_iommu_def_domain_type() so that it allocates translation domain.

Fixes: fb2accadaa94 ("iommu/amd: Introduce function to check and enable SNP")
CC: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Vasant Hegde <vasant.hegde@amd.com>
Link: https://lore.kernel.org/r/20230207091752.7656-1-vasant.hegde@amd.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iommu: Fix error unwind in iommu_group_alloc()

[ Upstream commit 4daa861174d56023c2068ddb03de0752f07fa199 ]

If either iommu_group_grate_file() fails then the
iommu_group is leaked.

Destroy it on these error paths.

Found by kselftest/iommu/iommufd_fail_nth

Fixes: bc7d12b91bd3 ("iommu: Implement reserved_regions iommu-group sysfs file")
Fixes: c52c72d3dee8 ("iommu: Add sysfs attribyte for domain type")
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/0-v1-8f616bee028d+8b-iommu_group_alloc_leak_jgg@nvidia.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iw_cxgb4: Fix potential NULL dereference in c4iw_fill_res_cm_id_entry()

[ Upstream commit 4ca446b127c568b59cb8d9748b6f70499624bb18 ]

This condition needs to match the previous "if (epcp->state == LISTEN) {"
exactly to avoid a NULL dereference of either "listen_ep" or "ep". The
problem is that "epcp" has been re-assigned so just testing
"if (epcp->state == LISTEN) {" a second time is not sufficient.

Fixes: 116aeb887371 ("iw_cxgb4: provide detailed provider-specific CM_ID information")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Link: https://lore.kernel.org/r/Y+usKuWIKr4dimZh@kili
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>