x86/srso: Disable the mitigation on unaffected configurations

commit e9fbc47b818b964ddff5df5b2d5c0f5f32f4a147 upstream.

Skip the srso cmd line parsing which is not needed on Zen1/2 with SMT
disabled and with the proper microcode applied (latter should be the
case anyway) as those are not affected.

Fixes: 5a15d8348881 ("x86/srso: Tie SBPB bit setting to microcode patch detection")
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230813104517.3346-1-bp@alien8.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/CPU/AMD: Fix the DIV(0) initial fix attempt

commit f58d6fbcb7c848b7f2469be339bc571f2e9d245b upstream.

Initially, it was thought that doing an innocuous division in the #DE
handler would take care to prevent any leaking of old data from the
divider but by the time the fault is raised, the speculation has already
advanced too far and such data could already have been used by younger
operations.

Therefore, do the innocuous division on every exit to userspace so that
userspace doesn't see any potentially old data from integer divisions in
kernel space.

Do the same before VMRUN too, to protect host data from leaking into the
guest too.

Fixes: 77245f1c3c64 ("x86/CPU/AMD: Do not leak quotient data after a division by 0")
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@kernel.org>
Link: https://lore.kernel.org/r/20230811213824.10025-1-bp@alien8.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/retpoline: Don't clobber RFLAGS during srso_safe_ret()

commit ba5ca5e5e6a1d55923e88b4a83da452166f5560e upstream.

Use LEA instead of ADD when adjusting %rsp in srso_safe_ret{,_alias}()
so as to avoid clobbering flags.  Drop one of the INT3 instructions to
account for the LEA consuming one more byte than the ADD.

KVM's emulator makes indirect calls into a jump table of sorts, where
the destination of each call is a small blob of code that performs fast
emulation by executing the target instruction with fixed operands.

E.g. to emulate ADC, fastop() invokes adcb_al_dl():

  adcb_al_dl:
    <+0>:  adc    %dl,%al
    <+2>:  jmp    <__x86_return_thunk>

A major motivation for doing fast emulation is to leverage the CPU to
handle consumption and manipulation of arithmetic flags, i.e. RFLAGS is
both an input and output to the target of the call.  fastop() collects
the RFLAGS result by pushing RFLAGS onto the stack and popping them back
into a variable (held in %rdi in this case):

  asm("push %[flags]; popf; " CALL_NOSPEC " ; pushf; pop %[flags]\n"

  <+71>: mov    0xc0(%r8),%rdx
  <+78>: mov    0x100(%r8),%rcx
  <+85>: push   %rdi
  <+86>: popf
  <+87>: call   *%rsi
  <+89>: nop
  <+90>: nop
  <+91>: nop
  <+92>: pushf
  <+93>: pop    %rdi

and then propagating the arithmetic flags into the vCPU's emulator state:

  ctxt->eflags = (ctxt->eflags & ~EFLAGS_MASK) | (flags & EFLAGS_MASK);

  <+64>:  and    $0xfffffffffffff72a,%r9
  <+94>:  and    $0x8d5,%edi
  <+109>: or     %rdi,%r9
  <+122>: mov    %r9,0x10(%r8)

The failures can be most easily reproduced by running the "emulator"
test in KVM-Unit-Tests.

If you're feeling a bit of deja vu, see commit b63f20a778c8
("x86/retpoline: Don't clobber RFLAGS during CALL_NOSPEC on i386").

In addition, this breaks booting of clang-compiled guest on
a gcc-compiled host where the host contains the %rsp-modifying SRSO
mitigations.

  [ bp: Massage commit message, extend, remove addresses. ]

Fixes: fb3bd914b3ec ("x86/srso: Add a Speculative RAS Overflow mitigation")
Closes: https://lore.kernel.org/all/de474347-122d-54cd-eabf-9dcc95ab9eae@amd.com
Reported-by: Srikanth Aithal <sraithal@amd.com>
Reported-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/20230810013334.GA5354@dev-arch.thelio-3990X/
Link: https://lore.kernel.org/r/20230811155255.250835-1-seanjc@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/static_call: Fix __static_call_fixup()

commit 54097309620ef0dc2d7083783dc521c6a5fef957 upstream.

Christian reported spurious module load crashes after some of Song's
module memory layout patches.

Turns out that if the very last instruction on the very last page of the
module is a 'JMP __x86_return_thunk' then __static_call_fixup() will
trip a fault and die.

And while the module rework made this slightly more likely to happen,
it's always been possible.

Fixes: ee88d363d156 ("x86,static_call: Use alternative RET encoding")
Reported-by: Christian Bricart <christian@bricart.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lkml.kernel.org/r/20230816104419.GA982867@hirez.programming.kicks-ass.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/srso: Explain the untraining sequences a bit more

commit 9dbd23e42ff0b10c9b02c9e649c76e5228241a8e upstream.

The goal is to eventually have a proper documentation about all this.

Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814164447.GFZNpZ/64H4lENIe94@fat_crate.local
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/cpu: Cleanup the untrain mess

commit e7c25c441e9e0fa75b4c83e0b26306b702cfe90d upstream.

Since there can only be one active return_thunk, there only needs be
one (matching) untrain_ret. It fundamentally doesn't make sense to
allow multiple untrain_ret at the same time.

Fold all the 3 different untrain methods into a single (temporary)
helper stub.

Fixes: fb3bd914b3ec ("x86/srso: Add a Speculative RAS Overflow mitigation")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814121149.042774962@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/cpu: Rename srso_(.*)_alias to srso_alias_\1

commit 42be649dd1f2eee6b1fb185f1a231b9494cf095f upstream.

For a more consistent namespace.

  [ bp: Fixup names in the doc too. ]

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814121148.976236447@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/cpu: Rename original retbleed methods

commit d025b7bac07a6e90b6b98b487f88854ad9247c39 upstream.

Rename the original retbleed return thunk and untrain_ret to
retbleed_return_thunk() and retbleed_untrain_ret().

No functional changes.

Suggested-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814121148.909378169@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/cpu: Clean up SRSO return thunk mess

commit d43490d0ab824023e11d0b57d0aeec17a6e0ca13 upstream.

Use the existing configurable return thunk. There is absolute no
justification for having created this __x86_return_thunk alternative.

To clarify, the whole thing looks like:

Zen3/4 does:

  srso_alias_untrain_ret:
  nop2
  lfence
  jmp srso_alias_return_thunk
  int3

  srso_alias_safe_ret: // aliasses srso_alias_untrain_ret just so
  add $8, %rsp
  ret
  int3

  srso_alias_return_thunk:
  call srso_alias_safe_ret
  ud2

While Zen1/2 does:

  srso_untrain_ret:
  movabs $foo, %rax
  lfence
  call srso_safe_ret           (jmp srso_return_thunk ?)
  int3

  srso_safe_ret: // embedded in movabs instruction
  add $8,%rsp
          ret
          int3

  srso_return_thunk:
  call srso_safe_ret
  ud2

While retbleed does:

  zen_untrain_ret:
  test $0xcc, %bl
  lfence
  jmp zen_return_thunk
          int3

  zen_return_thunk: // embedded in the test instruction
  ret
          int3

Where Zen1/2 flush the BTB entry using the instruction decoder trick
(test,movabs) Zen3/4 use BTB aliasing. SRSO adds a return sequence
(srso_safe_ret()) which forces the function return instruction to
speculate into a trap (UD2).  This RET will then mispredict and
execution will continue at the return site read from the top of the
stack.

Pick one of three options at boot (evey function can only ever return
once).

  [ bp: Fixup commit message uarch details and add them in a comment in
    the code too. Add a comment about the srso_select_mitigation()
    dependency on retbleed_select_mitigation(). Add moar ifdeffery for
    32-bit builds. Add a dummy srso_untrain_ret_alias() definition for
    32-bit alternatives needing the symbol. ]

Fixes: fb3bd914b3ec ("x86/srso: Add a Speculative RAS Overflow mitigation")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814121148.842775684@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/alternative: Make custom return thunk unconditional

commit 095b8303f3835c68ac4a8b6d754ca1c3b6230711 upstream.

There is infrastructure to rewrite return thunks to point to any
random thunk one desires, unwrap that from CALL_THUNKS, which up to
now was the sole user of that.

  [ bp: Make the thunks visible on 32-bit and add ifdeffery for the
    32-bit builds. ]

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814121148.775293785@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/cpu: Fix up srso_safe_ret() and __x86_return_thunk()

commit af023ef335f13c8b579298fc432daeef609a9e60 upstream.

  vmlinux.o: warning: objtool: srso_untrain_ret() falls through to next function __x86_return_skl()
  vmlinux.o: warning: objtool: __x86_return_thunk() falls through to next function __x86_return_skl()

This is because these functions (can) end with CALL, which objtool
does not consider a terminating instruction. Therefore, replace the
INT3 instruction (which is a non-fatal trap) with UD2 (which is a
fatal-trap).

This indicates execution will not continue past this point.

Fixes: fb3bd914b3ec ("x86/srso: Add a Speculative RAS Overflow mitigation")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814121148.637802730@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/cpu: Fix __x86_return_thunk symbol type

commit 77f67119004296a9b2503b377d610e08b08afc2a upstream.

Commit

  fb3bd914b3ec ("x86/srso: Add a Speculative RAS Overflow mitigation")

reimplemented __x86_return_thunk with a mix of SYM_FUNC_START and
SYM_CODE_END, this is not a sane combination.

Since nothing should ever actually 'CALL' this, make it consistently
CODE.

Fixes: fb3bd914b3ec ("x86/srso: Add a Speculative RAS Overflow mitigation")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230814121148.571027074@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Linux 6.1.47

Link: https://lore.kernel.org/r/20230821194122.695845670@linuxfoundation.org
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Tested-by: Takeshi Ogasawara <takeshi.ogasawara@futuring-girl.com>
Tested-by: Bagas Sanjaya <bagasdotme@gmail.com>
Tested-by: SeongJae Park <sj@kernel.org>
Tested-by: Shuah Khan <skhan@linuxfoundation.org>
Tested-by: Florian Fainelli <florian.fainelli@broadcom.com>
Tested-by: Ron Economos <re@w6rz.net>
Tested-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mmc: f-sdh30: fix order of function calls in sdhci_f_sdh30_remove

commit 58abdd80b93b09023ca03007b608685c41e3a289 upstream.

The order of function calls in sdhci_f_sdh30_remove is wrong,
let's call sdhci_pltfm_unregister first.

Cc: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Fixes: 5def5c1c15bf ("mmc: sdhci-f-sdh30: Replace with sdhci_pltfm")
Signed-off-by: Yangtao Li <frank.li@vivo.com>
Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20230727070051.17778-62-frank.li@vivo.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

net: fix the RTO timer retransmitting skb every 1ms if linear option is enabled

commit e4dd0d3a2f64b8bd8029ec70f52bdbebd0644408 upstream.

In the real workload, I encountered an issue which could cause the RTO
timer to retransmit the skb per 1ms with linear option enabled. The amount
of lost-retransmitted skbs can go up to 1000+ instantly.

The root cause is that if the icsk_rto happens to be zero in the 6th round
(which is the TCP_THIN_LINEAR_RETRIES value), then it will always be zero
due to the changed calculation method in tcp_retransmit_timer() as follows:

icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX);

Above line could be converted to
icsk->icsk_rto = min(0 << 1, TCP_RTO_MAX) = 0

Therefore, the timer expires so quickly without any doubt.

I read through the RFC 6298 and found that the RTO value can be rounded
up to a certain value, in Linux, say TCP_RTO_MIN as default, which is
regarded as the lower bound in this patch as suggested by Eric.

Fixes: 36e31b0af587 ("net: TCP thin linear timeouts")
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/nouveau/disp: fix use-after-free in error handling of nouveau_connector_create

commit 1b254b791d7b7dea6e8adc887fbbd51746d8bb27 upstream.

We can't simply free the connector after calling drm_connector_init on it.
We need to clean up the drm side first.

It might not fix all regressions from commit 2b5d1c29f6c4
("drm/nouveau/disp: PIOR DP uses GPIO for HPD, not PMGR AUX interrupts"),
but at least it fixes a memory corruption in error handling related to
that commit.

Link: https://lore.kernel.org/lkml/20230806213107.GFZNARG6moWpFuSJ9W@fat_crate.local/
Fixes: 95983aea8003 ("drm/nouveau/disp: add connector class")
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230814144933.3956959-1-kherbst@redhat.com
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

af_unix: Fix null-ptr-deref in unix_stream_sendpage().

Bing-Jhong Billy Jheng reported null-ptr-deref in unix_stream_sendpage()
with detailed analysis and a nice repro.

unix_stream_sendpage() tries to add data to the last skb in the peer's
recv queue without locking the queue.

If the peer's FD is passed to another socket and the socket's FD is
passed to the peer, there is a loop between them.  If we close both
sockets without receiving FD, the sockets will be cleaned up by garbage
collection.

The garbage collection iterates such sockets and unlinks skb with
FD from the socket's receive queue under the queue's lock.

So, there is a race where unix_stream_sendpage() could access an skb
locklessly that is being released by garbage collection, resulting in
use-after-free.

To avoid the issue, unix_stream_sendpage() must lock the peer's recv
queue.

Note the issue does not exist in 6.5+ thanks to the recent sendpage()
refactoring.

This patch is originally written by Linus Torvalds.

BUG: unable to handle page fault for address: ffff988004dd6870
PF: supervisor read access in kernel mode
PF: error_code(0x0000) - not-present page
PGD 0 P4D 0
PREEMPT SMP PTI
CPU: 4 PID: 297 Comm: garbage_uaf Not tainted 6.1.46 #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
RIP: 0010:kmem_cache_alloc_node+0xa2/0x1e0
Code: c0 0f 84 32 01 00 00 41 83 fd ff 74 10 48 8b 00 48 c1 e8 3a 41 39 c5 0f 85 1c 01 00 00 41 8b 44 24 28 49 8b 3c 24 48 8d 4a 40 <49> 8b 1c 06 4c 89 f0 65 48 0f c7 0f 0f 94 c0 84 c0 74 a1 41 8b 44
RSP: 0018:ffffc9000079fac0 EFLAGS: 00000246
RAX: 0000000000000070 RBX: 0000000000000005 RCX: 000000000001a284
RDX: 000000000001a244 RSI: 0000000000400cc0 RDI: 000000000002eee0
RBP: 0000000000400cc0 R08: 0000000000400cc0 R09: 0000000000000003
R10: 0000000000000001 R11: 0000000000000000 R12: ffff888003970f00
R13: 00000000ffffffff R14: ffff988004dd6800 R15: 00000000000000e8
FS:  00007f174d6f3600(0000) GS:ffff88807db00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffff988004dd6870 CR3: 00000000092be000 CR4: 00000000007506e0
PKRU: 55555554
Call Trace:
 <TASK>
 ? __die_body.cold+0x1a/0x1f
 ? page_fault_oops+0xa9/0x1e0
 ? fixup_exception+0x1d/0x310
 ? exc_page_fault+0xa8/0x150
 ? asm_exc_page_fault+0x22/0x30
 ? kmem_cache_alloc_node+0xa2/0x1e0
 ? __alloc_skb+0x16c/0x1e0
 __alloc_skb+0x16c/0x1e0
 alloc_skb_with_frags+0x48/0x1e0
 sock_alloc_send_pskb+0x234/0x270
 unix_stream_sendmsg+0x1f5/0x690
 sock_sendmsg+0x5d/0x60
 ____sys_sendmsg+0x210/0x260
 ___sys_sendmsg+0x83/0xd0
 ? kmem_cache_alloc+0xc6/0x1c0
 ? avc_disable+0x20/0x20
 ? percpu_counter_add_batch+0x53/0xc0
 ? alloc_empty_file+0x5d/0xb0
 ? alloc_file+0x91/0x170
 ? alloc_file_pseudo+0x94/0x100
 ? __fget_light+0x9f/0x120
 __sys_sendmsg+0x54/0xa0
 do_syscall_64+0x3b/0x90
 entry_SYSCALL_64_after_hwframe+0x69/0xd3
RIP: 0033:0x7f174d639a7d
Code: 28 89 54 24 1c 48 89 74 24 10 89 7c 24 08 e8 8a c1 f4 ff 8b 54 24 1c 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 2e 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 33 44 89 c7 48 89 44 24 08 e8 de c1 f4 ff 48
RSP: 002b:00007ffcb563ea50 EFLAGS: 00000293 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f174d639a7d
RDX: 0000000000000000 RSI: 00007ffcb563eab0 RDI: 0000000000000007
RBP: 00007ffcb563eb10 R08: 0000000000000000 R09: 00000000ffffffff
R10: 00000000004040a0 R11: 0000000000000293 R12: 00007ffcb563ec28
R13: 0000000000401398 R14: 0000000000403e00 R15: 00007f174d72c000
 </TASK>

Fixes: 869e7c62486e ("net: af_unix: implement stream sendpage support")
Reported-by: Bing-Jhong Billy Jheng <billy@starlabs.sg>
Reviewed-by: Bing-Jhong Billy Jheng <billy@starlabs.sg>
Co-developed-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amdgpu: keep irq count in amdgpu_irq_disable_all

commit 8ffd6f0442674f32c048ec8dffdbc5ec67829beb upstream.

This can clean up all irq warnings because of unbalanced
amdgpu_irq_get/put when unplugging/unbinding device, and leave
irq count decrease in each ip fini function.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amd/pm: skip the RLC stop when S0i3 suspend for SMU v13.0.4/11

commit 730d44e1fa306a20746ad4a85da550662aed9daa upstream.

For SMU v13.0.4/11, driver does not need to stop RLC for S0i3,
the firmwares will handle that properly.

Signed-off-by: Tim Huang <Tim.Huang@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64/ptrace: Ensure that SME is set up for target when writing SSVE state

commit 5d0a8d2fba50e9c07cde4aad7fba28c008b07a5b upstream.

When we use NT_ARM_SSVE to either enable streaming mode or change the
vector length for a process we do not currently do anything to ensure that
there is storage allocated for the SME specific register state.  If the
task had not previously used SME or we changed the vector length then
the task will not have had TIF_SME set or backing storage for ZA/ZT
allocated, resulting in inconsistent register sizes when saving state
and spurious traps which flush the newly set register state.

We should set TIF_SME to disable traps and ensure that storage is
allocated for ZA and ZT if it is not already allocated.  This requires
modifying sme_alloc() to make the flush of any existing register state
optional so we don't disturb existing state for ZA and ZT.

Fixes: e12310a0d30f ("arm64/sme: Implement ptrace support for streaming mode SVE registers")
Reported-by: David Spickett <David.Spickett@arm.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: <stable@vger.kernel.org> # 5.19.x
Link: https://lore.kernel.org/r/20230810-arm64-fix-ptrace-race-v1-1-a5361fad2bd6@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

netfilter: set default timeout to 3 secs for sctp shutdown send and recv state

commit 9bfab6d23a2865966a4f89a96536fbf23f83bc8c upstream.

In SCTP protocol, it is using the same timer (T2 timer) for SHUTDOWN and
SHUTDOWN_ACK retransmission. However in sctp conntrack the default timeout
value for SCTP_CONNTRACK_SHUTDOWN_ACK_SENT state is 3 secs while it's 300
msecs for SCTP_CONNTRACK_SHUTDOWN_SEND/RECV state.

As Paolo Valerio noticed, this might cause unwanted expiration of the ct
entry. In my test, with 1s tc netem delay set on the NAT path, after the
SHUTDOWN is sent, the sctp ct entry enters SCTP_CONNTRACK_SHUTDOWN_SEND
state. However, due to 300ms (too short) delay, when the SHUTDOWN_ACK is
sent back from the peer, the sctp ct entry has expired and been deleted,
and then the SHUTDOWN_ACK has to be dropped.

Also, it is confusing these two sysctl options always show 0 due to all
timeout values using sec as unit:

  net.netfilter.nf_conntrack_sctp_timeout_shutdown_recd = 0
  net.netfilter.nf_conntrack_sctp_timeout_shutdown_sent = 0

This patch fixes it by also using 3 secs for sctp shutdown send and recv
state in sctp conntrack, which is also RTO.initial value in SCTP protocol.

Note that the very short time value for SCTP_CONNTRACK_SHUTDOWN_SEND/RECV
was probably used for a rare scenario where SHUTDOWN is sent on 1st path
but SHUTDOWN_ACK is replied on 2nd path, then a new connection started
immediately on 1st path. So this patch also moves from SHUTDOWN_SEND/RECV
to CLOSE when receiving INIT in the ORIGINAL direction.

Fixes: 9fb9cbb1082d ("[NETFILTER]: Add nf_conntrack subsystem.")
Reported-by: Paolo Valerio <pvalerio@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

hugetlb: do not clear hugetlb dtor until allocating vmemmap

commit 32c877191e022b55fe3a374f3d7e9fb5741c514d upstream.

Patch series "Fix hugetlb free path race with memory errors".

In the discussion of Jiaqi Yan's series "Improve hugetlbfs read on
HWPOISON hugepages" the race window was discovered.
https://lore.kernel.org/linux-mm/20230616233447.GB7371@monkey/

Freeing a hugetlb page back to low level memory allocators is performed
in two steps.
1) Under hugetlb lock, remove page from hugetlb lists and clear destructor
2) Outside lock, allocate vmemmap if necessary and call low level free
Between these two steps, the hugetlb page will appear as a normal
compound page.  However, vmemmap for tail pages could be missing.
If a memory error occurs at this time, we could try to update page
flags non-existant page structs.

A much more detailed description is in the first patch.

The first patch addresses the race window.  However, it adds a
hugetlb_lock lock/unlock cycle to every vmemmap optimized hugetlb page
free operation.  This could lead to slowdowns if one is freeing a large
number of hugetlb pages.

The second path optimizes the update_and_free_pages_bulk routine to only
take the lock once in bulk operations.

The second patch is technically not a bug fix, but includes a Fixes tag
and Cc stable to avoid a performance regression.  It can be combined with
the first, but was done separately make reviewing easier.

This patch (of 2):

Freeing a hugetlb page and releasing base pages back to the underlying
allocator such as buddy or cma is performed in two steps:
- remove_hugetlb_folio() is called to remove the folio from hugetlb
  lists, get a ref on the page and remove hugetlb destructor.  This
  all must be done under the hugetlb lock.  After this call, the page
  can be treated as a normal compound page or a collection of base
  size pages.
- update_and_free_hugetlb_folio() is called to allocate vmemmap if
  needed and the free routine of the underlying allocator is called
  on the resulting page.  We can not hold the hugetlb lock here.

One issue with this scheme is that a memory error could occur between
these two steps.  In this case, the memory error handling code treats
the old hugetlb page as a normal compound page or collection of base
pages.  It will then try to SetPageHWPoison(page) on the page with an
error.  If the page with error is a tail page without vmemmap, a write
error will occur when trying to set the flag.

Address this issue by modifying remove_hugetlb_folio() and
update_and_free_hugetlb_folio() such that the hugetlb destructor is not
cleared until after allocating vmemmap.  Since clearing the destructor
requires holding the hugetlb lock, the clearing is done in
remove_hugetlb_folio() if the vmemmap is present.  This saves a
lock/unlock cycle.  Otherwise, destructor is cleared in
update_and_free_hugetlb_folio() after allocating vmemmap.

Note that this will leave hugetlb pages in a state where they are marked
free (by hugetlb specific page flag) and have a ref count.  This is not
a normal state.  The only code that would notice is the memory error
code, and it is set up to retry in such a case.

A subsequent patch will create a routine to do bulk processing of
vmemmap allocation.  This will eliminate a lock/unlock cycle for each
hugetlb page in the case where we are freeing a large number of pages.

Link: https://lkml.kernel.org/r/20230711220942.43706-1-mike.kravetz@oracle.com
Link: https://lkml.kernel.org/r/20230711220942.43706-2-mike.kravetz@oracle.com
Fixes: ad2fa3717b74 ("mm: hugetlb: alloc the vmemmap pages associated with each HugeTLB page")
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Reviewed-by: Muchun Song <songmuchun@bytedance.com>
Tested-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: James Houghton <jthoughton@google.com>
Cc: Jiaqi Yan <jiaqiyan@google.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amd/display: Implement workaround for writing to OTG_PIXEL_RATE_DIV register

commit 74fa4c81aadf418341f0d073c864ea7dca730a2e upstream.

[Why and How]
Current implementation requires FPGA builds to take a different
code path from DCN32 to write to OTG_PIXEL_RATE_DIV. Now that
we have a workaround to write to OTG_PIXEL_RATE_DIV register without
blanking display on hotplug on DCN32, we can allow the code paths for
FPGA to be exactly the same allowing for more consistent
testing.

Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Saaem Rizvi <SyedSaaem.Rizvi@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: "Limonciello, Mario" <mario.limonciello@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

sched/fair: Remove capacity inversion detection

commit a2e90611b9f425adbbfcdaa5b5e49958ddf6f61b upstream.

Remove the capacity inversion detection which is now handled by
util_fits_cpu() returning -1 when we need to continue to look for a
potential CPU with better performance.

This ends up almost reverting patches below except for some comments:
commit da07d2f9c153 ("sched/fair: Fixes for capacity inversion detection")
commit aa69c36f31aa ("sched/fair: Consider capacity inversion in util_fits_cpu()")
commit 44c7b80bffc3 ("sched/fair: Detect capacity inversion")

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20230201143628.270912-3-vincent.guittot@linaro.org
Signed-off-by: Qais Yousef (Google) <qyousef@layalina.io>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

sched/fair: unlink misfit task from cpu overutilized

commit e5ed0550c04c5469ecdc1634d8aa18c8609590f0 upstream.

By taking into account uclamp_min, the 1:1 relation between task misfit
and cpu overutilized is no more true as a task with a small util_avg may
not fit a high capacity cpu because of uclamp_min constraint.

Add a new state in util_fits_cpu() to reflect the case that task would fit
a CPU except for the uclamp_min hint which is a performance requirement.

Use -1 to reflect that a CPU doesn't fit only because of uclamp_min so we
can use this new value to take additional action to select the best CPU
that doesn't match uclamp_min hint.

When util_fits_cpu() returns -1, we will continue to look for a possible
CPU with better performance, which replaces Capacity Inversion detection
with capacity_orig_of() - thermal_load_avg to detect a capacity inversion.

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-and-tested-by: Qais Yousef <qyousef@layalina.io>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
Link: https://lore.kernel.org/r/20230201143628.270912-2-vincent.guittot@linaro.org
Signed-off-by: Qais Yousef (Google) <qyousef@layalina.io>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

zsmalloc: allow only one active pool compaction context

commit d2658f2052c7db6ec0a79977205f8cf1cb9effc2 upstream.

zsmalloc pool can be compacted concurrently by many contexts,
e.g.

 cc1 handle_mm_fault()
      do_anonymous_page()
       __alloc_pages_slowpath()
        try_to_free_pages()
         do_try_to_free_pages(
          lru_gen_shrink_node()
           shrink_slab()
            do_shrink_slab()
             zs_shrinker_scan()
              zs_compact()

Pool compaction is currently (basically) single-threaded as
it is performed under pool->lock. Having multiple compaction
threads results in unnecessary contention, as each thread
competes for pool->lock. This, in turn, affects all zsmalloc
operations such as zs_malloc(), zs_map_object(), zs_free(), etc.

Introduce the pool->compaction_in_progress atomic variable,
which ensures that only one compaction context can run at a
time. This reduces overall pool->lock contention in (corner)
cases when many contexts attempt to shrink zspool simultaneously.

Link: https://lkml.kernel.org/r/20230418074639.1903197-1-senozhatsky@chromium.org
Fixes: c0547d0b6a4b ("zsmalloc: consolidate zs_pool's migrate_lock and size_class's locks")
Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Reviewed-by: Yosry Ahmed <yosryahmed@google.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amd/display: disable RCO for DCN314

commit 85e41f1ed5d94a26fe4e57003c399936d291ed70 upstream.

[Why]
RCO is causing error messages on some DCN314 systems

[How]
Force disable RCO for DCN314

Fixes: 17fbdbda9cc8 ("drm/amd/display: Enable dcn314 DPP RCO")
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Daniel Miess <daniel.miess@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ASoC: amd: vangogh: select CONFIG_SND_AMD_ACP_CONFIG

commit 812a05256d673b2b9c5db906775d1e6625ba4787 upstream.

The vangogh driver just gained a link time dependency that now causes
randconfig builds to fail:

x86_64-linux-ld: sound/soc/amd/vangogh/pci-acp5x.o: in function `snd_acp5x_probe':
pci-acp5x.c:(.text+0xbb): undefined reference to `snd_amd_acp_find_config'

Fixes: e89f45edb747e ("ASoC: amd: vangogh: Add check for acp config flags in vangogh platform")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230602124447.863476-1-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amdgpu/pm: fix throttle_status for other than MP1 11.0.7

commit 6a92761a86817ad15c9a562e2a809386237fae3e upstream.

Use the right metrics table version based on the firmware.

Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2720
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Umio Yasuno <coelacanth_dream@protonmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amdgpu: skip fence GFX interrupts disable/enable for S0ix

commit f1740b1ab2703b2a057da7cf33b03297e0381aa0 upstream.

GFX v11.0.1 reported fence fallback timer expired issue on
SDMA and GFX rings after S0ix resume. This is generated by
EOP interrupts are disabled when S0ix suspend but fails to
re-enable when resume because of the GFX is in GFXOFF.

[  203.349571] [drm] Fence fallback timer expired on ring sdma0
[  203.349572] [drm] Fence fallback timer expired on ring gfx_0.0.0
[  203.861635] [drm] Fence fallback timer expired on ring gfx_0.0.0

For S0ix, GFX is in GFXOFF state, avoid to touch the GFX registers
to configure the fence driver interrupts for rings that belong to GFX.
The interrupts configuration will be restored by GFXOFF exit.

Signed-off-by: Tim Huang <Tim.Huang@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amd: flush any delayed gfxoff on suspend entry

commit a7b7d9e8aee4f71b4c7151702fd74237b8cef989 upstream.

DCN 3.1.4 is reported to hang on s2idle entry if graphics activity
is happening during entry.  This is because GFXOFF was scheduled as
delayed but RLC gets disabled in s2idle entry sequence which will
hang GFX IP if not already in GFXOFF.

To help this problem, flush any delayed work for GFXOFF early in
s2idle entry sequence to ensure that it's off when RLC is changed.

commit 4b31b92b143f ("drm/amdgpu: complete gfxoff allow signal during
suspend without delay") modified power gating flow so that if called
in s0ix that it ensured that GFXOFF wasn't put in work queue but
instead processed immediately.

This is dead code due to commit 10cb67eb8a1b ("drm/amdgpu: skip
CG/PG for gfx during S0ix") because GFXOFF will now not be explicitly
called as part of the suspend entry code.  Remove that dead code.

Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/i915/sdvo: fix panel_type initialization

commit 2002eb6d3ea954dde9f8a223018d5335779937d0 upstream.

Commit 3f9ffce5765d ("drm/i915: Do panel VBT init early if the VBT
declares an explicit panel type") started using -1 as the value for
unset panel_type. It gets initialized in intel_panel_init_alloc(), but
the SDVO code never calls it.

Call intel_panel_init_alloc() to initialize the panel, including the
panel_type.

Reported-by: Tomi Leppänen <tomi@tomin.site>
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/8896
Fixes: 3f9ffce5765d ("drm/i915: Do panel VBT init early if the VBT declares an explicit panel type")
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: <stable@vger.kernel.org> # v6.1+
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Tested-by: Tomi Leppänen <tomi@tomin.site>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230803122706.838721-1-jani.nikula@intel.com
(cherry picked from commit 26e60294e8eacedc8ebb33405b2c375fd80e0900)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/qxl: fix UAF on handle creation

commit c611589b4259ed63b9b77be6872b1ce07ec0ac16 upstream.

qxl_mode_dumb_create() dereferences the qobj returned by
qxl_gem_object_create_with_handle(), but the handle is the only one
holding a reference to it.

A potential attacker could guess the returned handle value and closes it
between the return of qxl_gem_object_create_with_handle() and the qobj
usage, triggering a use-after-free scenario.

Reproducer:

int dri_fd =-1;
struct drm_mode_create_dumb arg = {0};

void gem_close(int handle);

void* trigger(void* ptr)
{
int ret;
arg.width = arg.height = 0x20;
arg.bpp = 32;
ret = ioctl(dri_fd, DRM_IOCTL_MODE_CREATE_DUMB, &arg);
if(ret)
{
perror("[*] DRM_IOCTL_MODE_CREATE_DUMB Failed");
exit(-1);
}
gem_close(arg.handle);
while(1) {
struct drm_mode_create_dumb args = {0};
args.width = args.height = 0x20;
args.bpp = 32;
ret = ioctl(dri_fd, DRM_IOCTL_MODE_CREATE_DUMB, &args);
if (ret) {
perror("[*] DRM_IOCTL_MODE_CREATE_DUMB Failed");
exit(-1);
}

printf("[*] DRM_IOCTL_MODE_CREATE_DUMB created, %d\n", args.handle);
gem_close(args.handle);
}
return NULL;
}

void gem_close(int handle)
{
struct drm_gem_close args;
args.handle = handle;
int ret = ioctl(dri_fd, DRM_IOCTL_GEM_CLOSE, &args); // gem close handle
if (!ret)
printf("gem close handle %d\n", args.handle);
}

int main(void)
{
dri_fd= open("/dev/dri/card0", O_RDWR);
printf("fd:%d\n", dri_fd);

if(dri_fd == -1)
return -1;

pthread_t tid1;

if(pthread_create(&tid1,NULL,trigger,NULL)){
perror("[*] thread_create tid1\n");
return -1;
}
while (1)
{
gem_close(arg.handle);
}
return 0;
}

This is a KASAN report:

==================================================================
BUG: KASAN: slab-use-after-free in qxl_mode_dumb_create+0x3c2/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:69
Write of size 1 at addr ffff88801136c240 by task poc/515

CPU: 1 PID: 515 Comm: poc Not tainted 6.3.0 #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
Call Trace:
<TASK>
__dump_stack linux/lib/dump_stack.c:88
dump_stack_lvl+0x48/0x70 linux/lib/dump_stack.c:106
print_address_description linux/mm/kasan/report.c:319
print_report+0xd2/0x660 linux/mm/kasan/report.c:430
kasan_report+0xd2/0x110 linux/mm/kasan/report.c:536
__asan_report_store1_noabort+0x17/0x30 linux/mm/kasan/report_generic.c:383
qxl_mode_dumb_create+0x3c2/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:69
drm_mode_create_dumb linux/drivers/gpu/drm/drm_dumb_buffers.c:96
drm_mode_create_dumb_ioctl+0x1f5/0x2d0 linux/drivers/gpu/drm/drm_dumb_buffers.c:102
drm_ioctl_kernel+0x21d/0x430 linux/drivers/gpu/drm/drm_ioctl.c:788
drm_ioctl+0x56f/0xcc0 linux/drivers/gpu/drm/drm_ioctl.c:891
vfs_ioctl linux/fs/ioctl.c:51
__do_sys_ioctl linux/fs/ioctl.c:870
__se_sys_ioctl linux/fs/ioctl.c:856
__x64_sys_ioctl+0x13d/0x1c0 linux/fs/ioctl.c:856
do_syscall_x64 linux/arch/x86/entry/common.c:50
do_syscall_64+0x5b/0x90 linux/arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc linux/arch/x86/entry/entry_64.S:120
RIP: 0033:0x7ff5004ff5f7
Code: 00 00 00 48 8b 05 99 c8 0d 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 69 c8 0d 00 f7 d8 64 89 01 48

RSP: 002b:00007ff500408ea8 EFLAGS: 00000286 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007ff5004ff5f7
RDX: 00007ff500408ec0 RSI: 00000000c02064b2 RDI: 0000000000000003
RBP: 00007ff500408ef0 R08: 0000000000000000 R09: 000000000000002a
R10: 0000000000000000 R11: 0000000000000286 R12: 00007fff1c6cdafe
R13: 00007fff1c6cdaff R14: 00007ff500408fc0 R15: 0000000000802000
</TASK>

Allocated by task 515:
kasan_save_stack+0x38/0x70 linux/mm/kasan/common.c:45
kasan_set_track+0x25/0x40 linux/mm/kasan/common.c:52
kasan_save_alloc_info+0x1e/0x40 linux/mm/kasan/generic.c:510
____kasan_kmalloc linux/mm/kasan/common.c:374
__kasan_kmalloc+0xc3/0xd0 linux/mm/kasan/common.c:383
kasan_kmalloc linux/./include/linux/kasan.h:196
kmalloc_trace+0x48/0xc0 linux/mm/slab_common.c:1066
kmalloc linux/./include/linux/slab.h:580
kzalloc linux/./include/linux/slab.h:720
qxl_bo_create+0x11a/0x610 linux/drivers/gpu/drm/qxl/qxl_object.c:124
qxl_gem_object_create+0xd9/0x360 linux/drivers/gpu/drm/qxl/qxl_gem.c:58
qxl_gem_object_create_with_handle+0xa1/0x180 linux/drivers/gpu/drm/qxl/qxl_gem.c:89
qxl_mode_dumb_create+0x1cd/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:63
drm_mode_create_dumb linux/drivers/gpu/drm/drm_dumb_buffers.c:96
drm_mode_create_dumb_ioctl+0x1f5/0x2d0 linux/drivers/gpu/drm/drm_dumb_buffers.c:102
drm_ioctl_kernel+0x21d/0x430 linux/drivers/gpu/drm/drm_ioctl.c:788
drm_ioctl+0x56f/0xcc0 linux/drivers/gpu/drm/drm_ioctl.c:891
vfs_ioctl linux/fs/ioctl.c:51
__do_sys_ioctl linux/fs/ioctl.c:870
__se_sys_ioctl linux/fs/ioctl.c:856
__x64_sys_ioctl+0x13d/0x1c0 linux/fs/ioctl.c:856
do_syscall_x64 linux/arch/x86/entry/common.c:50
do_syscall_64+0x5b/0x90 linux/arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc linux/arch/x86/entry/entry_64.S:120

Freed by task 515:
kasan_save_stack+0x38/0x70 linux/mm/kasan/common.c:45
kasan_set_track+0x25/0x40 linux/mm/kasan/common.c:52
kasan_save_free_info+0x2e/0x60 linux/mm/kasan/generic.c:521
____kasan_slab_free linux/mm/kasan/common.c:236
____kasan_slab_free+0x180/0x1f0 linux/mm/kasan/common.c:200
__kasan_slab_free+0x12/0x30 linux/mm/kasan/common.c:244
kasan_slab_free linux/./include/linux/kasan.h:162
slab_free_hook linux/mm/slub.c:1781
slab_free_freelist_hook+0xd2/0x1a0 linux/mm/slub.c:1807
slab_free linux/mm/slub.c:3787
__kmem_cache_free+0x196/0x2d0 linux/mm/slub.c:3800
kfree+0x78/0x120 linux/mm/slab_common.c:1019
qxl_ttm_bo_destroy+0x140/0x1a0 linux/drivers/gpu/drm/qxl/qxl_object.c:49
ttm_bo_release+0x678/0xa30 linux/drivers/gpu/drm/ttm/ttm_bo.c:381
kref_put linux/./include/linux/kref.h:65
ttm_bo_put+0x50/0x80 linux/drivers/gpu/drm/ttm/ttm_bo.c:393
qxl_gem_object_free+0x3e/0x60 linux/drivers/gpu/drm/qxl/qxl_gem.c:42
drm_gem_object_free+0x5c/0x90 linux/drivers/gpu/drm/drm_gem.c:974
kref_put linux/./include/linux/kref.h:65
__drm_gem_object_put linux/./include/drm/drm_gem.h:431
drm_gem_object_put linux/./include/drm/drm_gem.h:444
qxl_gem_object_create_with_handle+0x151/0x180 linux/drivers/gpu/drm/qxl/qxl_gem.c:100
qxl_mode_dumb_create+0x1cd/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:63
drm_mode_create_dumb linux/drivers/gpu/drm/drm_dumb_buffers.c:96
drm_mode_create_dumb_ioctl+0x1f5/0x2d0 linux/drivers/gpu/drm/drm_dumb_buffers.c:102
drm_ioctl_kernel+0x21d/0x430 linux/drivers/gpu/drm/drm_ioctl.c:788
drm_ioctl+0x56f/0xcc0 linux/drivers/gpu/drm/drm_ioctl.c:891
vfs_ioctl linux/fs/ioctl.c:51
__do_sys_ioctl linux/fs/ioctl.c:870
__se_sys_ioctl linux/fs/ioctl.c:856
__x64_sys_ioctl+0x13d/0x1c0 linux/fs/ioctl.c:856
do_syscall_x64 linux/arch/x86/entry/common.c:50
do_syscall_64+0x5b/0x90 linux/arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc linux/arch/x86/entry/entry_64.S:120

The buggy address belongs to the object at ffff88801136c000
which belongs to the cache kmalloc-1k of size 1024
The buggy address is located 576 bytes inside of
freed 1024-byte region [ffff88801136c000, ffff88801136c400)

The buggy address belongs to the physical page:
page:0000000089fc329b refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x11368
head:0000000089fc329b order:3 entire_mapcount:0 nr_pages_mapped:0 pincount:0
flags: 0xfffffc0010200(slab|head|node=0|zone=1|lastcpupid=0x1fffff)
raw: 000fffffc0010200 ffff888007841dc0 dead000000000122 0000000000000000
raw: 0000000000000000 0000000080100010 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
ffff88801136c100: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff88801136c180: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>ffff88801136c200: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
^
ffff88801136c280: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff88801136c300: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Disabling lock debugging due to kernel taint

Instead of returning a weak reference to the qxl_bo object, return the
created drm_gem_object and let the caller decrement the reference count
when it no longer needs it. As a convenience, if the caller is not
interested in the gobj object, it can pass NULL to the parameter and the
reference counting is descremented internally.

The bug and the reproducer were originally found by the Zero Day Initiative project (ZDI-CAN-20940).

Link: https://www.zerodayinitiative.com/
Signed-off-by: Wander Lairson Costa <wander@redhat.com>
Cc: stable@vger.kernel.org
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230814165119.90847-1-wander@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mmc: block: Fix in_flight[issue_type] value error

commit 4b430d4ac99750ee2ae2f893f1055c7af1ec3dc5 upstream.

For a completed request, after the mmc_blk_mq_complete_rq(mq, req)
function is executed, the bitmap_tags corresponding to the
request will be cleared, that is, the request will be regarded as
idle. If the request is acquired by a different type of process at
this time, the issue_type of the request may change. It further
caused the value of mq->in_flight[issue_type] to be abnormal,
and a large number of requests could not be sent.

p1:       p2:
mmc_blk_mq_complete_rq
  blk_mq_free_request
      blk_mq_get_request
        blk_mq_rq_ctx_init
mmc_blk_mq_dec_in_flight
  mmc_issue_type(mq, req)

This strategy can ensure the consistency of issue_type
before and after executing mmc_blk_mq_complete_rq.

Fixes: 81196976ed94 ("mmc: block: Add blk-mq support")
Cc: stable@vger.kernel.org
Signed-off-by: Yibin Ding <yibin.ding@unisoc.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20230802023023.1318134-1-yunlong.xing@unisoc.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mmc: wbsd: fix double mmc_free_host() in wbsd_init()

commit d83035433701919ac6db15f7737cbf554c36c1a6 upstream.

mmc_free_host() has already be called in wbsd_free_mmc(),
remove the mmc_free_host() in error path in wbsd_init().

Fixes: dc5b9b50fc9d ("mmc: wbsd: fix return value check of mmc_add_host()")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20230807124443.3431366-1-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

blk-crypto: dynamically allocate fallback profile

commit c984ff1423ae9f70b1f28ce811856db0d9c99021 upstream.

blk_crypto_profile_init() calls lockdep_register_key(), which warns and
does not register if the provided memory is a static object.
blk-crypto-fallback currently has a static blk_crypto_profile and calls
blk_crypto_profile_init() thereupon, resulting in the warning and
failure to register.

Fortunately it is simple enough to use a dynamically allocated profile
and make lockdep function correctly.

Fixes: 2fb48d88e77f ("blk-crypto: use dynamic lock class for blk_crypto_profile::lock")
Cc: stable@vger.kernel.org
Signed-off-by: Sweet Tea Dorminy <sweettea-kernel@dorminy.me>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230817141615.15387-1-sweettea-kernel@dorminy.me
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arm64: dts: rockchip: Fix Wifi/Bluetooth on ROCK Pi 4 boards

commit ebceec271e552a2b05e47d8ef0597052b1a39449 upstream.

This patch fixes an issue affecting the Wifi/Bluetooth connectivity on
ROCK Pi 4 boards. Commit f471b1b2db08 ("arm64: dts: rockchip: Fix Bluetooth
on ROCK Pi 4 boards") introduced a problem with the clock configuration.
Specifically, the clock-names property of the sdio-pwrseq node was not
updated to 'lpo', causing the driver to wait indefinitely for the wrong clock
signal 'ext_clock' instead of the expected one 'lpo'. This prevented the proper
initialization of Wifi/Bluetooth chip on ROCK Pi 4 boards.

To address this, this patch updates the clock-names property of the
sdio-pwrseq node to "lpo" to align with the changes made to the bluetooth node.

This patch has been tested on ROCK Pi 4B.

Fixes: f471b1b2db08 ("arm64: dts: rockchip: Fix Bluetooth on ROCK Pi 4 boards")
Cc: stable@vger.kernel.org
Signed-off-by: Yogesh Hegde <yogi.kernel@gmail.com>
Link: https://lore.kernel.org/r/ZLbATQRjOl09aLAp@zephyrusG14
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

virtio-net: Zero max_tx_vq field for VIRTIO_NET_CTRL_MQ_HASH_CONFIG case

commit 2c507ce90e02cd78d00fd4b0fe26c8641873c13f upstream.

Kernel uses `struct virtio_net_ctrl_rss` to save command-specific-data
for both the VIRTIO_NET_CTRL_MQ_HASH_CONFIG and
VIRTIO_NET_CTRL_MQ_RSS_CONFIG commands.

According to the VirtIO standard, "Field reserved MUST contain zeroes.
It is defined to make the structure to match the layout of
virtio_net_rss_config structure, defined in 5.1.6.5.7.".

Yet for the VIRTIO_NET_CTRL_MQ_HASH_CONFIG command case, the `max_tx_vq`
field in struct virtio_net_ctrl_rss, which corresponds to the
`reserved` field in struct virtio_net_hash_config, is not zeroed,
thereby violating the VirtIO standard.

This patch solves this problem by zeroing this field in
virtnet_init_default_rss().

Cc: Andrew Melnychenko <andrew@daynix.com>
Cc: stable@vger.kernel.org
Fixes: c7114b1249fa ("drivers/net/virtio_net: Added basic RSS support.")
Signed-off-by: Hawkins Jiawei <yin31149@gmail.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20230810110405.25558-1-yin31149@gmail.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

cifs: Release folio lock on fscache read hit.

commit 69513dd669e243928f7450893190915a88f84a2b upstream.

Under the current code, when cifs_readpage_worker is called, the call
contract is that the callee should unlock the page. This is documented
in the read_folio section of Documentation/filesystems/vfs.rst as:

> The filesystem should unlock the folio once the read has completed,
> whether it was successful or not.

Without this change, when fscache is in use and cache hit occurs during
a read, the page lock is leaked, producing the following stack on
subsequent reads (via mmap) to the page:

$ cat /proc/3890/task/12864/stack
[<0>] folio_wait_bit_common+0x124/0x350
[<0>] filemap_read_folio+0xad/0xf0
[<0>] filemap_fault+0x8b1/0xab0
[<0>] __do_fault+0x39/0x150
[<0>] do_fault+0x25c/0x3e0
[<0>] __handle_mm_fault+0x6ca/0xc70
[<0>] handle_mm_fault+0xe9/0x350
[<0>] do_user_addr_fault+0x225/0x6c0
[<0>] exc_page_fault+0x84/0x1b0
[<0>] asm_exc_page_fault+0x27/0x30

This requires a reboot to resolve; it is a deadlock.

Note however that the call to cifs_readpage_from_fscache does mark the
page clean, but does not free the folio lock. This happens in
__cifs_readpage_from_fscache on success. Releasing the lock at that
point however is not appropriate as cifs_readahead also calls
cifs_readpage_from_fscache and *does* unconditionally release the lock
after its return. This change therefore effectively makes
cifs_readpage_worker work like cifs_readahead.

Signed-off-by: Russell Harmon <russ@har.mn>
Acked-by: Paulo Alcantara (SUSE) <pc@manguebit.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ALSA: usb-audio: Add support for Mythware XA001AU capture and playback interfaces.

commit 788449ae57f4273111b779bbcaad552b67f517d5 upstream.

This patch adds a USB quirk for Mythware XA001AU USB interface.

Signed-off-by: dengxiang <dengxiang@nfschina.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20230803024437.370069-1-dengxiang@nfschina.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: 8250: Fix oops for port->pm on uart_change_pm()

[ Upstream commit dfe2aeb226fd5e19b0ee795f4f6ed8bc494c1534 ]

Unloading a hardware specific 8250 driver can produce error "Unable to
handle kernel paging request at virtual address" about ten seconds after
unloading the driver. This happens on uart_hangup() calling
uart_change_pm().

Turns out commit 04e82793f068 ("serial: 8250: Reinit port->pm on port
specific driver unbind") was only a partial fix. If the hardware specific
driver has initialized port->pm function, we need to clear port->pm too.
Just reinitializing port->ops does not do this. Otherwise serial8250_pm()
will call port->pm() instead of serial8250_do_pm().

Fixes: 04e82793f068 ("serial: 8250: Reinit port->pm on port specific driver unbind")
Signed-off-by: Tony Lindgren <tony@atomide.com>
Link: https://lore.kernel.org/r/20230804131553.52927-1-tony@atomide.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

riscv: uaccess: Return the number of bytes effectively not copied

[ Upstream commit 4b05b993900dd3eba0fc83ef5c5ddc7d65d786c6 ]

It was reported that the riscv kernel hangs while executing the test
in [1].

Indeed, the test hangs when trying to write a buffer to a file. The
problem is that the riscv implementation of raw_copy_from_user() does not
return the correct number of bytes not written when an exception happens
and is fixed up, instead it always returns the initial size to copy,
even if some bytes were actually copied.

generic_perform_write() pre-faults the user pages and bails out if nothing
can be written, otherwise it will access the userspace buffer: here the
riscv implementation keeps returning it was not able to copy any byte
though the pre-faulting indicates otherwise. So generic_perform_write()
keeps retrying to access the user memory and ends up in an infinite
loop.

Note that before the commit mentioned in [1] that introduced this
regression, it worked because generic_perform_write() would bail out if
only one byte could not be written.

So fix this by returning the number of bytes effectively not written in
__asm_copy_[to|from]_user() and __clear_user(), as it is expected.

Link: https://lore.kernel.org/linux-riscv/20230309151841.bomov6hq3ybyp42a@debian/
Fixes: ebcbd75e3962 ("riscv: Fix the bug in memory access fixup code")
Reported-by: Bo YU <tsu.yubo@gmail.com>
Closes: https://lore.kernel.org/linux-riscv/20230309151841.bomov6hq3ybyp42a@debian/#t
Reported-by: Aurelien Jarno <aurelien@aurel32.net>
Closes: https://lore.kernel.org/linux-riscv/ZNOnCakhwIeue3yr@aurel32.net/
Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com>
Reviewed-by: Björn Töpel <bjorn@rivosinc.com>
Tested-by: Aurelien Jarno <aurelien@aurel32.net>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Link: https://lore.kernel.org/r/20230811150604.1621784-1-alexghiti@rivosinc.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ALSA: hda/realtek - Remodified 3k pull low procedure

[ Upstream commit 46cdff2369cbdf8d78081a22526e77bd1323f563 ]

Set spec->en_3kpull_low default to true.
Then fillback ALC236 and ALC257 to false.

Additional note: this addresses a regression caused by the previous
fix 69ea4c9d02b7 ("ALSA: hda/realtek - remove 3k pull low procedure").
The previous workaround was applied too widely without necessity,
which resulted in the pop noise at PM again.  This patch corrects the
condition and restores the old behavior for the devices that don't
suffer from the original problem.

Fixes: 69ea4c9d02b7 ("ALSA: hda/realtek - remove 3k pull low procedure")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217732
Link: https://lore.kernel.org/r/01e212a538fc407ca6edd10b81ff7b05@realtek.com
Signed-off-by: Kailang Yang <kailang@realtek.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

soc: aspeed: socinfo: Add kfree for kstrdup

[ Upstream commit 6e6d847a8ce18ab2fbec4f579f682486a82d2c6b ]

Add kfree() in the later error handling in order to avoid memory leak.

Fixes: e0218dca5787 ("soc: aspeed: Add soc info driver")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Link: https://lore.kernel.org/r/20230707021625.7727-1-jiasheng@iscas.ac.cn
Signed-off-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20230810123104.231167-1-joel@jms.id.au
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

soc: aspeed: uart-routing: Use __sysfs_match_string

[ Upstream commit e4ad279ae345413d900d791f2f618d0a1cd0d791 ]

The existing use of match_string() caused it to reject 'echo foo' due
to the implicitly appended newline, which was somewhat ergonomically
awkward and inconsistent with typical sysfs behavior.  Using the
__sysfs_* variant instead provides more convenient and consistent
linefeed-agnostic behavior.

Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
Fixes: c6807970c3bc ("soc: aspeed: Add UART routing support")
Reviewed-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20230628083735.19946-2-zev@bewilderbeest.net
Signed-off-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20230810122941.231085-1-joel@jms.id.au
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ALSA: hda/realtek: Add quirks for HP G11 Laptops

[ Upstream commit fb8cce69e5e56eedb35fc4d77b2f099860965859 ]

These HP G11 laptops use Realtek HDA codec combined with
2xCS35L41 Amplifiers using SPI or I2C with External Boost.

Laptop 103c8c26 has been removed as this has been replaced
by this new series of laptops.

Fixes: 3e10f6ca76c4 ("ALSA: hda/realtek: Add quirk for HP EliteBook G10 laptops")
Signed-off-by: Stefan Binding <sbinding@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230809142957.675933-2-sbinding@opensource.cirrus.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: meson: axg-tdm-formatter: fix channel slot allocation

[ Upstream commit c1f848f12103920ca165758aedb1c10904e193e1 ]

When the tdm lane mask is computed, the driver currently fills the 1st lane
before moving on to the next. If the stream has less channels than the
lanes can accommodate, slots will be disabled on the last lanes.

Unfortunately, the HW distribute channels in a different way. It distribute
channels in pair on each lanes before moving on the next slots.

This difference leads to problems if a device has an interface with more
than 1 lane and with more than 2 slots per lane.

For example: a playback interface with 2 lanes and 4 slots each (total 8
slots - zero based numbering)
- Playing a 8ch stream:
  - All slots activated by the driver
  - channel #2 will be played on lane #1 - slot #0 following HW placement
- Playing a 4ch stream:
  - Lane #1 disabled by the driver
  - channel #2 will be played on lane #0 - slot #2

This behaviour is obviously not desirable.

Change the way slots are activated on the TDM lanes to follow what the HW
does and make sure each channel always get mapped to the same slot/lane.

Fixes: 1a11d88f499c ("ASoC: meson: add tdm formatter base driver")
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20230809171931.1244502-1-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: rt5665: add missed regulator_bulk_disable

[ Upstream commit c163108e706909570f8aa9aa5bcf6806e2b4c98c ]

The driver forgets to call regulator_bulk_disable()

Add the missed call to fix it.

Fixes: 33ada14a26c8 ("ASoC: add rt5665 codec driver")
Signed-off-by: Zhang Shurong <zhang_shurong@foxmail.com>
Link: https://lore.kernel.org/r/tencent_A560D01E3E0A00A85A12F137E4B5205B3508@qq.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: imx93: Fix anatop node size

[ Upstream commit 78e869dd8b2ba19765ac9b05cdea3e432d1dc188 ]

Although the memory map of i.MX93 reference manual rev. 2 claims that
analog top has start address of 0x44480000 and end address of 0x4448ffff,
this overlaps with TMU memory area starting at 0x44482000, as stated in
section 73.6.1.
As PLL configuration registers start at addresses up to 0x44481400, as used
by clk-imx93, reduce the anatop size to 0x2000, so exclude the TMU area
but keep all PLL registers inside.

Fixes: ec8b5b5058ea ("arm64: dts: freescale: Add i.MX93 dtsi support")
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: imx: Set default tuning step for imx6sx usdhc

[ Upstream commit 0a2b96e42a0284c4fc03022236f656a085ca714a ]

If the tuning step is not set, the tuning step is set to 1.
For some sd cards, the following Tuning timeout will occur.

Tuning failed, falling back to fixed sampling clock

So set the default tuning step. This refers to the NXP vendor's
commit below:

https://github.com/nxp-imx/linux-imx/blob/lf-6.1.y/
arch/arm/boot/dts/imx6sx.dtsi#L1108-L1109

Fixes: 1e336aa0c025 ("mmc: sdhci-esdhc-imx: correct the tuning start tap and step setting")
Signed-off-by: Xiaolei Wang <xiaolei.wang@windriver.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: imx8mm: Drop CSI1 PHY reference clock configuration

[ Upstream commit f02b53375e8f14b4c27a14f6e4fb6e89914fdc29 ]

The CSI1 PHY reference clock is limited to 125 MHz according to:
i.MX 8M Mini Applications Processor Reference Manual, Rev. 3, 11/2020
Table 5-1. Clock Root Table (continued) / page 307
Slice Index n = 123 .

Currently the IMX8MM_CLK_CSI1_PHY_REF clock is configured to be
fed directly from 1 GHz PLL2 , which overclocks them. Instead, drop
the configuration altogether, which defaults the clock to 24 MHz REF
clock input, which for the PHY reference clock is just fine.

Based on a patch from Marek Vasut for the imx8mn.

Fixes: e523b7c54c05 ("arm64: dts: imx8mm: Add CSI nodes")
Signed-off-by: Fabio Estevam <festevam@denx.de>
Reviewed-by: Marek Vasut <marex@denx.de>
Reviewed-by: Marco Felsch <m.felsch@pengutronix.de>
Reviewed-by: Adam Ford <aford173@gmail.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: imx6: phytec: fix RTC interrupt level

[ Upstream commit 762b700982a1e0f562184363f19860c3b9bdd0bf ]

RTC interrupt level should be set to "LOW". This was revealed by the
introduction of commit:

  f181987ef477 ("rtc: m41t80: use IRQ flags obtained from fwnode")

which changed the way IRQ type is obtained.

Signed-off-by: Andrej Picej <andrej.picej@norik.com>
Reviewed-by: Stefan Riedmüller <s.riedmueller@phytec.de>
Fixes: 800d595151bb ("ARM: dts: imx6: Add initial support for phyBOARD-Mira")
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: imx: align LED node names with dtschema

[ Upstream commit 4b0d1f2738899dbcc7a026d826373530019aa31b ]

The node names should be generic and DT schema expects certain pattern:

  imx50-kobo-aura.dtb: gpio-leds: 'on' does not match any of the regexes: '(^led-[0-9a-f]$|led)', 'pinctrl-[0-9]+'
  imx6dl-yapp4-draco.dtb: led-controller@30: 'chan@0', 'chan@1', 'chan@2' do not match any of the regexes: '^led@[0-8]$', '^multi-led@[0-8]$', 'pinctrl-[0-9]+'

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Stable-dep-of: 762b700982a1 ("ARM: dts: imx6: phytec: fix RTC interrupt level")
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: rockchip: Disable HS400 for eMMC on ROCK 4C+

[ Upstream commit 2bd1d2dd808c60532283e9cf05110bf1bf2f9079 ]

There is some instablity with some eMMC modules on ROCK Pi 4 SBCs running
in HS400 mode. This ends up resulting in some block errors after a while
or after a "heavy" operation utilising the eMMC (e.g. resizing a
filesystem). An example of these errors is as follows:

    [  289.171014] mmc1: running CQE recovery
    [  290.048972] mmc1: running CQE recovery
    [  290.054834] mmc1: running CQE recovery
    [  290.060817] mmc1: running CQE recovery
    [  290.061337] blk_update_request: I/O error, dev mmcblk1, sector 1411072 op 0x1:(WRITE) flags 0x800 phys_seg 36 prio class 0
    [  290.061370] EXT4-fs warning (device mmcblk1p1): ext4_end_bio:348: I/O error 10 writing to inode 29547 starting block 176466)
    [  290.061484] Buffer I/O error on device mmcblk1p1, logical block 172288
    [  290.061531] Buffer I/O error on device mmcblk1p1, logical block 172289
    [  290.061551] Buffer I/O error on device mmcblk1p1, logical block 172290
    [  290.061574] Buffer I/O error on device mmcblk1p1, logical block 172291
    [  290.061592] Buffer I/O error on device mmcblk1p1, logical block 172292
    [  290.061615] Buffer I/O error on device mmcblk1p1, logical block 172293
    [  290.061632] Buffer I/O error on device mmcblk1p1, logical block 172294
    [  290.061654] Buffer I/O error on device mmcblk1p1, logical block 172295
    [  290.061673] Buffer I/O error on device mmcblk1p1, logical block 172296
    [  290.061695] Buffer I/O error on device mmcblk1p1, logical block 172297

Disabling the Command Queue seems to stop the CQE recovery from running,
but doesn't seem to improve the I/O errors. Until this can be investigated
further, disable HS400 mode on the ROCK Pi 4 SBCs to at least stop I/O
errors from occurring.

Fixes: 246450344dad ("arm64: dts: rockchip: rk3399: Radxa ROCK 4C+")
Signed-off-by: Christopher Obbard <chris.obbard@collabora.com>
Link: https://lore.kernel.org/r/20230705144255.115299-3-chris.obbard@collabora.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: rockchip: Disable HS400 for eMMC on ROCK Pi 4

[ Upstream commit cee572756aa2cb46e959e9797ad4b730b78a050b ]

There is some instablity with some eMMC modules on ROCK Pi 4 SBCs running
in HS400 mode. This ends up resulting in some block errors after a while
or after a "heavy" operation utilising the eMMC (e.g. resizing a
filesystem). An example of these errors is as follows:

    [  289.171014] mmc1: running CQE recovery
    [  290.048972] mmc1: running CQE recovery
    [  290.054834] mmc1: running CQE recovery
    [  290.060817] mmc1: running CQE recovery
    [  290.061337] blk_update_request: I/O error, dev mmcblk1, sector 1411072 op 0x1:(WRITE) flags 0x800 phys_seg 36 prio class 0
    [  290.061370] EXT4-fs warning (device mmcblk1p1): ext4_end_bio:348: I/O error 10 writing to inode 29547 starting block 176466)
    [  290.061484] Buffer I/O error on device mmcblk1p1, logical block 172288
    [  290.061531] Buffer I/O error on device mmcblk1p1, logical block 172289
    [  290.061551] Buffer I/O error on device mmcblk1p1, logical block 172290
    [  290.061574] Buffer I/O error on device mmcblk1p1, logical block 172291
    [  290.061592] Buffer I/O error on device mmcblk1p1, logical block 172292
    [  290.061615] Buffer I/O error on device mmcblk1p1, logical block 172293
    [  290.061632] Buffer I/O error on device mmcblk1p1, logical block 172294
    [  290.061654] Buffer I/O error on device mmcblk1p1, logical block 172295
    [  290.061673] Buffer I/O error on device mmcblk1p1, logical block 172296
    [  290.061695] Buffer I/O error on device mmcblk1p1, logical block 172297

Disabling the Command Queue seems to stop the CQE recovery from running,
but doesn't seem to improve the I/O errors. Until this can be investigated
further, disable HS400 mode on the ROCK Pi 4 SBCs to at least stop I/O
errors from occurring.

While we are here, set the eMMC maximum clock frequency to 1.5MHz to
follow the ROCK 4C+.

Fixes: 1b5715c602fd ("arm64: dts: rockchip: add ROCK Pi 4 DTS support")
Signed-off-by: Christopher Obbard <chris.obbard@collabora.com>
Tested-By: Folker Schwesinger <dev@folker-schwesinger.de>
Link: https://lore.kernel.org/r/20230705144255.115299-2-chris.obbard@collabora.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: qcom: qrb5165-rb5: fix thermal zone conflict

[ Upstream commit 798f1df86e5709b7b6aedf493cc04c7fedbf544a ]

The commit 3a786086c6f8 ("arm64: dts: qcom: Add missing "-thermal"
suffix for thermal zones") renamed the thermal zone in the pm8150l.dtsi
file to comply with the schema. However this resulted in a clash with
the RB5 board file, which already contained the pm8150l-thermal zone for
the on-board sensor. This resulted in the board file definition
overriding the thermal zone defined in the PMIC include file (and thus
the on-die PMIC temp alarm was not probing at all).

Rename the thermal zone in qcom/qrb5165-rb5.dts to remove this override.

Fixes: 3a786086c6f8 ("arm64: dts: qcom: Add missing "-thermal" suffix for thermal zones")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Link: https://lore.kernel.org/r/20230613131224.666668-1-dmitry.baryshkov@linaro.org
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bus: ti-sysc: Flush posted write on enable before reset

[ Upstream commit 34539b442b3bc7d5bf10164750302b60b91f18a7 ]

The am335x devices started producing boot errors for resetting musb module
in because of subtle timing changes:

Unhandled fault: external abort on non-linefetch (0x1008)
...
sysc_poll_reset_sysconfig from sysc_reset+0x109/0x12
sysc_reset from sysc_probe+0xa99/0xeb0
...

The fix is to flush posted write after enable before reset during
probe. Note that some devices also need to specify the delay after enable
with ti,sysc-delay-us, but this is not needed for musb on am335x based on
my tests.

Reported-by: kernelci.org bot <bot@kernelci.org>
Closes: https://storage.kernelci.org/next/master/next-20230614/arm/multi_v7_defconfig+CONFIG_THUMB2_KERNEL=y/gcc-10/lab-cip/baseline-beaglebone-black.html
Fixes: 596e7955692b ("bus: ti-sysc: Add support for software reset")
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ice: Block switchdev mode when ADQ is active and vice versa

[ Upstream commit 43d00e102d9ecbe2635d7e3f2e14d2e90183d6af ]

ADQ and switchdev are not supported simultaneously. Enabling both at the
same time can result in nullptr dereference.

To prevent this, check if ADQ is active when changing devlink mode to
switchdev mode, and check if switchdev is active when enabling ADQ.

Fixes: fbc7b27af0f9 ("ice: enable ndo_setup_tc support for mqprio_qdisc")
Signed-off-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Tested-by: Sujai Buvaneswaran <sujai.buvaneswaran@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230816193405.1307580-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

qede: fix firmware halt over suspend and resume

[ Upstream commit 2eb9625a3a32251ecea470cd576659a3a03b4e59 ]

While performing certain power-off sequences, PCI drivers are
called to suspend and resume their underlying devices through
PCI PM (power management) interface. However this NIC hardware
does not support PCI PM suspend/resume operations so system wide
suspend/resume leads to bad MFW (management firmware) state which
causes various follow-up errors in driver when communicating with
the device/firmware afterwards.

To fix this driver implements PCI PM suspend handler to indicate
unsupported operation to the PCI subsystem explicitly, thus avoiding
system to go into suspended/standby mode.

Without this fix device/firmware does not recover unless system
is power cycled.

Fixes: 2950219d87b0 ("qede: Add basic network device support")
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Alok Prasad <palok@marvell.com>
Reviewed-by: John Meneghini <jmeneghi@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230816150711.59035-1-manishc@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: do not allow gso_size to be set to GSO_BY_FRAGS

[ Upstream commit b616be6b97688f2f2bd7c4a47ab32f27f94fb2a9 ]

One missing check in virtio_net_hdr_to_skb() allowed
syzbot to crash kernels again [1]

Do not allow gso_size to be set to GSO_BY_FRAGS (0xffff),
because this magic value is used by the kernel.

[1]
general protection fault, probably for non-canonical address 0xdffffc000000000e: 0000 [#1] PREEMPT SMP KASAN
KASAN: null-ptr-deref in range [0x0000000000000070-0x0000000000000077]
CPU: 0 PID: 5039 Comm: syz-executor401 Not tainted 6.5.0-rc5-next-20230809-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/26/2023
RIP: 0010:skb_segment+0x1a52/0x3ef0 net/core/skbuff.c:4500
Code: 00 00 00 e9 ab eb ff ff e8 6b 96 5d f9 48 8b 84 24 00 01 00 00 48 8d 78 70 48 b8 00 00 00 00 00 fc ff df 48 89 fa 48 c1 ea 03 <0f> b6 04 02 84 c0 74 08 3c 03 0f 8e ea 21 00 00 48 8b 84 24 00 01
RSP: 0018:ffffc90003d3f1c8 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: 000000000001fffe RCX: 0000000000000000
RDX: 000000000000000e RSI: ffffffff882a3115 RDI: 0000000000000070
RBP: ffffc90003d3f378 R08: 0000000000000005 R09: 000000000000ffff
R10: 000000000000ffff R11: 5ee4a93e456187d6 R12: 000000000001ffc6
R13: dffffc0000000000 R14: 0000000000000008 R15: 000000000000ffff
FS: 00005555563f2380(0000) GS:ffff8880b9800000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000020020000 CR3: 000000001626d000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
udp6_ufo_fragment+0x9d2/0xd50 net/ipv6/udp_offload.c:109
ipv6_gso_segment+0x5c4/0x17b0 net/ipv6/ip6_offload.c:120
skb_mac_gso_segment+0x292/0x610 net/core/gso.c:53
__skb_gso_segment+0x339/0x710 net/core/gso.c:124
skb_gso_segment include/net/gso.h:83 [inline]
validate_xmit_skb+0x3a5/0xf10 net/core/dev.c:3625
__dev_queue_xmit+0x8f0/0x3d60 net/core/dev.c:4329
dev_queue_xmit include/linux/netdevice.h:3082 [inline]
packet_xmit+0x257/0x380 net/packet/af_packet.c:276
packet_snd net/packet/af_packet.c:3087 [inline]
packet_sendmsg+0x24c7/0x5570 net/packet/af_packet.c:3119
sock_sendmsg_nosec net/socket.c:727 [inline]
sock_sendmsg+0xd9/0x180 net/socket.c:750
____sys_sendmsg+0x6ac/0x940 net/socket.c:2496
___sys_sendmsg+0x135/0x1d0 net/socket.c:2550
__sys_sendmsg+0x117/0x1e0 net/socket.c:2579
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x38/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7ff27cdb34d9

Fixes: 3953c46c3ac7 ("sk_buff: allow segmenting based on frag sizes")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Xin Long <lucien.xin@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Link: https://lore.kernel.org/r/20230816142158.1779798-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

sock: Fix misuse of sk_under_memory_pressure()

[ Upstream commit 2d0c88e84e483982067a82073f6125490ddf3614 ]

The status of global socket memory pressure is updated when:

  a) __sk_mem_raise_allocated():

enter: sk_memory_allocated(sk) >  sysctl_mem[1]
leave: sk_memory_allocated(sk) <= sysctl_mem[0]

  b) __sk_mem_reduce_allocated():

leave: sk_under_memory_pressure(sk) &&
sk_memory_allocated(sk) < sysctl_mem[0]

So the conditions of leaving global pressure are inconstant, which
may lead to the situation that one pressured net-memcg prevents the
global pressure from being cleared when there is indeed no global
pressure, thus the global constrains are still in effect unexpectedly
on the other sockets.

This patch fixes this by ignoring the net-memcg's pressure when
deciding whether should leave global memory pressure.

Fixes: e1aab161e013 ("socket: initial cgroup code.")
Signed-off-by: Abel Wu <wuyun.abel@bytedance.com>
Acked-by: Shakeel Butt <shakeelb@google.com>
Link: https://lore.kernel.org/r/20230816091226.1542-1-wuyun.abel@bytedance.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

sfc: don't unregister flow_indr if it was never registered

[ Upstream commit fa165e1949976704500a442faeef8d9596faee76 ]

In efx_init_tc(), move the setting of efx->tc->up after the
 flow_indr_dev_register() call, so that if it fails, efx_fini_tc()
 won't call flow_indr_dev_unregister().

Fixes: 5b2e12d51bd8 ("sfc: bind indirect blocks for TC offload on EF100")
Suggested-by: Pieter Jansen van Vuuren <pieter.jansen-van-vuuren@amd.com>
Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Edward Cree <ecree.xilinx@gmail.com>
Link: https://lore.kernel.org/r/a81284d7013aba74005277bd81104e4cfbea3f6f.1692114888.git.ecree.xilinx@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: dsa: mv88e6xxx: Wait for EEPROM done before HW reset

[ Upstream commit 23d775f12dcd23d052a4927195f15e970e27ab26 ]

If the switch is reset during active EEPROM transactions, as in
just after an SoC reset after power up, the I2C bus transaction
may be cut short leaving the EEPROM internal I2C state machine
in the wrong state.  When the switch is reset again, the bad
state machine state may result in data being read from the wrong
memory location causing the switch to enter unexpected mode
rendering it inoperational.

Fixes: a3dcb3e7e70c ("net: dsa: mv88e6xxx: Wait for EEPROM done after HW reset")
Signed-off-by: Alfred Lee <l00g33k@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20230815001323.24739-1-l00g33k@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

i40e: fix misleading debug logs

[ Upstream commit 2f2beb8874cb0844e84ad26e990f05f4f13ff63f ]

Change "write" into the actual "read" word.
Change parameters description.

Fixes: 7073f46e443e ("i40e: Add AQ commands for NVM Update for X722")
Signed-off-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Andrii Staikov <andrii.staikov@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iavf: fix FDIR rule fields masks validation

[ Upstream commit 751969e5b1196821ef78f0aa664a8a97c92c9057 ]

Return an error if a field's mask is neither full nor empty. When a mask
is only partial the field is not being used for rule programming but it
gives a wrong impression it is used. Fix by returning an error on any
partial mask to make it clear they are not supported.
The ip_ver assignment is moved earlier in code to allow using it in
iavf_validate_fdir_fltr_masks.

Fixes: 527691bf0682 ("iavf: Support IPv4 Flow Director filters")
Fixes: e90cbc257a6f ("iavf: Support IPv6 Flow Director filters")
Signed-off-by: Piotr Gardocki <piotrx.gardocki@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: openvswitch: reject negative ifindex

[ Upstream commit a552bfa16bab4ce901ee721346a28c4e483f4066 ]

Recent changes in net-next (commit 759ab1edb56c ("net: store netdevs
in an xarray")) refactored the handling of pre-assigned ifindexes
and let syzbot surface a latent problem in ovs. ovs does not validate
ifindex, making it possible to create netdev ports with negative
ifindex values. It's easy to repro with YNL:

$ ./cli.py --spec netlink/specs/ovs_datapath.yaml \
         --do new \
 --json '{"upcall-pid": 1, "name":"my-dp"}'
$ ./cli.py --spec netlink/specs/ovs_vport.yaml \
 --do new \
 --json '{"upcall-pid": "00000001", "name": "some-port0", "dp-ifindex":3,"ifindex":4294901760,"type":2}'

$ ip link show
-65536: some-port0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
    link/ether 7a:48:21:ad:0b:fb brd ff:ff:ff:ff:ff:ff
...

Validate the inputs. Now the second command correctly returns:

$ ./cli.py --spec netlink/specs/ovs_vport.yaml \
 --do new \
 --json '{"upcall-pid": "00000001", "name": "some-port0", "dp-ifindex":3,"ifindex":4294901760,"type":2}'

lib.ynl.NlError: Netlink error: Numerical result out of range
nl_len = 108 (92) nl_flags = 0x300 nl_type = 2
error: -34 extack: {'msg': 'integer out of range', 'unknown': [[type:4 len:36] b'\x0c\x00\x02\x00\x00\x00\x00\x00\x00\x00\x00\x00\x0c\x00\x03\x00\xff\xff\xff\x7f\x00\x00\x00\x00\x08\x00\x01\x00\x08\x00\x00\x00'], 'bad-attr': '.ifindex'}

Accept 0 since it used to be silently ignored.

Fixes: 54c4ef34c4b6 ("openvswitch: allow specifying ifindex of new interfaces")
Reported-by: syzbot+7456b5dcf65111553320@syzkaller.appspotmail.com
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Aaron Conole <aconole@redhat.com>
Link: https://lore.kernel.org/r/20230814203840.2908710-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

team: Fix incorrect deletion of ETH_P_8021AD protocol vid from slaves

[ Upstream commit dafcbce07136d799edc4c67f04f9fd69ff1eac1f ]

Similar to commit 01f4fd270870 ("bonding: Fix incorrect deletion of
ETH_P_8021AD protocol vid from slaves"), we can trigger BUG_ON(!vlan_info)
in unregister_vlan_dev() with the following testcase:

  # ip netns add ns1
  # ip netns exec ns1 ip link add team1 type team
  # ip netns exec ns1 ip link add team_slave type veth peer veth2
  # ip netns exec ns1 ip link set team_slave master team1
  # ip netns exec ns1 ip link add link team_slave name team_slave.10 type vlan id 10 protocol 802.1ad
  # ip netns exec ns1 ip link add link team1 name team1.10 type vlan id 10 protocol 802.1ad
  # ip netns exec ns1 ip link set team_slave nomaster
  # ip netns del ns1

Add S-VLAN tag related features support to team driver. So the team driver
will always propagate the VLAN info to its slaves.

Fixes: 8ad227ff89a7 ("net: vlan: add 802.1ad support")
Suggested-by: Ido Schimmel <idosch@idosch.org>
Signed-off-by: Ziyang Xuan <william.xuanziyang@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230814032301.2804971-1-william.xuanziyang@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: phy: broadcom: stub c45 read/write for 54810

[ Upstream commit 096516d092d54604d590827d05b1022c8f326639 ]

The 54810 does not support c45. The mmd_phy_indirect accesses return
arbirtary values leading to odd behavior like saying it supports EEE
when it doesn't. We also see that reading/writing these non-existent
MMD registers leads to phy instability in some cases.

Fixes: b14995ac2527 ("net: phy: broadcom: Add BCM54810 PHY entry")
Signed-off-by: Justin Chen <justin.chen@broadcom.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://lore.kernel.org/r/1691901708-28650-1-git-send-email-justin.chen@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: nft_dynset: disallow object maps

[ Upstream commit 23185c6aed1ffb8fc44087880ba2767aba493779 ]

Do not allow to insert elements from datapath to objects maps.

Fixes: 8aeff920dcc9 ("netfilter: nf_tables: add stateful object reference to set elements")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ipvs: fix racy memcpy in proc_do_sync_threshold

[ Upstream commit 5310760af1d4fbea1452bfc77db5f9a680f7ae47 ]

When two threads run proc_do_sync_threshold() in parallel,
data races could happen between the two memcpy():

Thread-1 Thread-2
memcpy(val, valp, sizeof(val));
memcpy(valp, val, sizeof(val));

This race might mess up the (struct ctl_table *) table->data,
so we add a mutex lock to serialize them.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Link: https://lore.kernel.org/netdev/B6988E90-0A1E-4B85-BF26-2DAF6D482433@gmail.com/
Signed-off-by: Sishuai Gong <sishuai.system@gmail.com>
Acked-by: Simon Horman <horms@kernel.org>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: nf_tables: deactivate catchall elements in next generation

[ Upstream commit 90e5b3462efa37b8bba82d7c4e63683856e188af ]

When flushing, individual set elements are disabled in the next
generation via the ->flush callback.

Catchall elements are not disabled.  This is incorrect and may lead to
double-deactivations of catchall elements which then results in memory
leaks:

WARNING: CPU: 1 PID: 3300 at include/net/netfilter/nf_tables.h:1172 nft_map_deactivate+0x549/0x730
CPU: 1 PID: 3300 Comm: nft Not tainted 6.5.0-rc5+ #60
RIP: 0010:nft_map_deactivate+0x549/0x730
 [..]
 ? nft_map_deactivate+0x549/0x730
 nf_tables_delset+0xb66/0xeb0

(the warn is due to nft_use_dec() detecting underflow).

Fixes: aaa31047a6d2 ("netfilter: nftables: add catch-all set element support")
Reported-by: lonial con <kongln9170@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: nf_tables: fix false-positive lockdep splat

[ Upstream commit b9f052dc68f69dac89fe1e24693354c033daa091 ]

->abort invocation may cause splat on debug kernels:

WARNING: suspicious RCU usage
net/netfilter/nft_set_pipapo.c:1697 suspicious rcu_dereference_check() usage!
[..]
rcu_scheduler_active = 2, debug_locks = 1
1 lock held by nft/133554: [..] (nft_net->commit_mutex){+.+.}-{3:3}, at: nf_tables_valid_genid
[..]
 lockdep_rcu_suspicious+0x1ad/0x260
 nft_pipapo_abort+0x145/0x180
 __nf_tables_abort+0x5359/0x63d0
 nf_tables_abort+0x24/0x40
 nfnetlink_rcv+0x1a0a/0x22c0
 netlink_unicast+0x73c/0x900
 netlink_sendmsg+0x7f0/0xc20
 ____sys_sendmsg+0x48d/0x760

Transaction mutex is held, so parallel updates are not possible.
Switch to _protected and check mutex is held for lockdep enabled builds.

Fixes: 212ed75dc5fb ("netfilter: nf_tables: integrate pipapo into commit protocol")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

octeon_ep: cancel tx_timeout_task later in remove sequence

[ Upstream commit 28458c80006bb4e993a09fc094094a8578cad292 ]

tx_timeout_task is canceled too early when removing the driver. Nothing
prevents .ndo_tx_timeout from triggering and queuing the work again.

Better cancel it after the netdev is unregistered.
It's harmless for octep_tx_timeout_task to run in the window between the
unregistration and cancelation, because it checks netif_running.

Fixes: 862cd659a6fb ("octeon_ep: Add driver framework and device initialization")
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Link: https://lore.kernel.org/r/20230810150114.107765-3-mschmidt@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: macb: In ZynqMP resume always configure PS GTR for non-wakeup source

[ Upstream commit 6c461e394d11a981c662cc16cebfb05b602e23ba ]

On Zynq UltraScale+ MPSoC ubuntu platform when systemctl issues suspend,
network manager bring down the interface and goes into suspend. When it
wakes up it again enables the interface.

This leads to xilinx-psgtr "PLL lock timeout" on interface bringup, as
the power management controller power down the entire FPD (including
SERDES) if none of the FPD devices are in use and serdes is not
initialized on resume.

$ sudo rtcwake -m no -s 120 -v
$ sudo systemctl suspend  <this does ifconfig eth1 down>
$ ifconfig eth1 up
xilinx-psgtr fd400000.phy: lane 0 (type 10, protocol 5): PLL lock timeout
phy phy-fd400000.phy.0: phy poweron failed --> -110

macb driver is called in this way:
1. macb_close: Stop network interface. In this function, it
   reset MACB IP and disables PHY and network interface.

2. macb_suspend: It is called in kernel suspend flow. But because
   network interface has been disabled(netif_running(ndev) is
   false), it does nothing and returns directly;

3. System goes into suspend state. Some time later, system is
   waken up by RTC wakeup device;

4. macb_resume: It does nothing because network interface has
   been disabled;

5. macb_open: It is called to enable network interface again. ethernet
   interface is initialized in this API but serdes which is power-off
   by PMUFW during FPD-off suspend is not initialized again and so
   we hit GT PLL lock issue on open.

To resolve this PLL timeout issue always do PS GTR initialization
when ethernet device is configured as non-wakeup source.

Fixes: f22bd29ba19a ("net: macb: Fix ZynqMP SGMII non-wakeup source resume failure")
Fixes: 8b73fa3ae02b ("net: macb: Added ZynqMP-specific initialization")
Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@amd.com>
Link: https://lore.kernel.org/r/1691414091-2260697-1-git-send-email-radhey.shyam.pandey@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/panel: simple: Fix AUO G121EAN01 panel timings according to the docs

[ Upstream commit e8470c0a7bcaa82f78ad34282d662dd7bd9630c2 ]

Commit 03e909acd95a ("drm/panel: simple: Add support for AUO G121EAN01.4
panel") added support for this panel model, but the timings it implements
are very different from what the datasheet describes. I checked both the
G121EAN01.0 datasheet from [0] and the G121EAN01.4 one from [1] and they
all have the same timings: for example the LVDS clock typical value is 74.4
MHz, not 66.7 MHz as implemented.

Replace the timings with the ones from the documentation. These timings
have been tested and the clock frequencies verified with an oscilloscope to
ensure they are correct.

Also use struct display_timing instead of struct drm_display_mode in order
to also specify the minimum and maximum values.

[0] https://embedded.avnet.com/product/g121ean01-0/
[1] https://embedded.avnet.com/product/g121ean01-4/

Fixes: 03e909acd95a ("drm/panel: simple: Add support for AUO G121EAN01.4 panel")
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20230804151239.835216-1-luca.ceresoli@bootlin.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests: mirror_gre_changes: Tighten up the TTL test match

[ Upstream commit 855067defa36b1f9effad8c219d9a85b655cf500 ]

This test verifies whether the encapsulated packets have the correct
configured TTL. It does so by sending ICMP packets through the test
topology and mirroring them to a gretap netdevice. On a busy host
however, more than just the test ICMP packets may end up flowing
through the topology, get mirrored, and counted. This leads to
potential spurious failures as the test observes much more mirrored
packets than the sent test packets, and assumes a bug.

Fix this by tightening up the mirror action match. Change it from
matchall to a flower classifier matching on ICMP packets specifically.

Fixes: 45315673e0c5 ("selftests: forwarding: Test changes in mirror-to-gretap")
Signed-off-by: Petr Machata <petrm@nvidia.com>
Tested-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: phy: fix IRQ-based wake-on-lan over hibernate / power off

[ Upstream commit cc941e548bffc01b5816b4edc5cb432a137a58b3 ]

Uwe reports:
"Most PHYs signal WoL using an interrupt. So disabling interrupts [at
shutdown] breaks WoL at least on PHYs covered by the marvell driver."

Discussing with Ioana, the problem which was trying to be solved was:
"The board in question is a LS1021ATSN which has two AR8031 PHYs that
share an interrupt line. In case only one of the PHYs is probed and
there are pending interrupts on the PHY#2 an IRQ storm will happen
since there is no entity to clear the interrupt from PHY#2's registers.
PHY#1's driver will get stuck in .handle_interrupt() indefinitely."

Further confirmation that "the two AR8031 PHYs are on the same MDIO
bus."

With WoL using interrupts to wake the system, in such a case, the
system will begin booting with an asserted interrupt. Thus, we need to
cope with an interrupt asserted during boot.

Solve this instead by disabling interrupts during PHY probe. This will
ensure in Ioana's situation that both PHYs of the same type sharing an
interrupt line on a common MDIO bus will have their interrupt outputs
disabled when the driver probes the device, but before we hook in any
interrupt handlers - thus avoiding the interrupt storm.

A better fix would be for platform firmware to disable the interrupting
devices at source during boot, before control is handed to the kernel.

Fixes: e2f016cf7751 ("net: phy: add a shutdown procedure")
Link: 20230804071757.383971-1-u.kleine-koenig@pengutronix.de
Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: pcs: Add missing put_device call in miic_create

[ Upstream commit 829c6524d6729d05a82575dbcc16f99be5ee843d ]

The reference of pdev->dev is taken by of_find_device_by_node, so
it should be released when not need anymore.

Fixes: 7dc54d3b8d91 ("net: pcs: add Renesas MII converter driver")
Signed-off-by: Xiang Yang <xiangyang3@huawei.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

virtio-net: set queues after driver_ok

[ Upstream commit 51b813176f098ff61bd2833f627f5319ead098a5 ]

Commit 25266128fe16 ("virtio-net: fix race between set queues and
probe") tries to fix the race between set queues and probe by calling
_virtnet_set_queues() before DRIVER_OK is set. This violates virtio
spec. Fixing this by setting queues after virtio_device_ready().

Note that rtnl needs to be held for userspace requests to change the
number of queues. So we are serialized in this way.

Fixes: 25266128fe16 ("virtio-net: fix race between set queues and probe")
Reported-by: Dragos Tatulea <dtatulea@nvidia.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

virtio_net: notify MAC address change on device initialization

[ Upstream commit 9f62d221a4b0aa6a8d2a18053a0ca349c025297c ]

In virtnet_probe(), if the device doesn't provide a MAC address the
driver assigns a random one.
As we modify the MAC address we need to notify the device to allow it
to update all the related information.

The problem can be seen with vDPA and mlx5_vdpa driver as it doesn't
assign a MAC address by default. The virtio_net device uses a random
MAC address (we can see it with "ip link"), but we can't ping a net
namespace from another one using the virtio-vdpa device because the
new MAC address has not been provided to the hardware:
RX packets are dropped since they don't go through the receive filters,
TX packets go through unaffected.

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 51b813176f09 ("virtio-net: set queues after driver_ok")
Signed-off-by: Sasha Levin <sashal@kernel.org>

xfrm: add forgotten nla_policy for XFRMA_MTIMER_THRESH

[ Upstream commit 5e2424708da7207087934c5c75211e8584d553a0 ]

The previous commit 4e484b3e969b ("xfrm: rate limit SA mapping change
message to user space") added one additional attribute named
XFRMA_MTIMER_THRESH and described its type at compat_policy
(net/xfrm/xfrm_compat.c).

However, the author forgot to also describe the nla_policy at
xfrma_policy (net/xfrm/xfrm_user.c). Hence, this suppose NLA_U32 (4
bytes) value can be faked as empty (0 bytes) by a malicious user, which
leads to 4 bytes overflow read and heap information leak when parsing
nlattrs.

To exploit this, one malicious user can spray the SLUB objects and then
leverage this 4 bytes OOB read to leak the heap data into
x->mapping_maxage (see xfrm_update_ae_params(...)), and leak it to
userspace via copy_to_user_state_extra(...).

The above bug is assigned CVE-2023-3773. To fix it, this commit just
completes the nla_policy description for XFRMA_MTIMER_THRESH, which
enforces the length check and avoids such OOB read.

Fixes: 4e484b3e969b ("xfrm: rate limit SA mapping change message to user space")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

xfrm: add NULL check in xfrm_update_ae_params

[ Upstream commit 00374d9b6d9f932802b55181be9831aa948e5b7c ]

Normally, x->replay_esn and x->preplay_esn should be allocated at
xfrm_alloc_replay_state_esn(...) in xfrm_state_construct(...), hence the
xfrm_update_ae_params(...) is okay to update them. However, the current
implementation of xfrm_new_ae(...) allows a malicious user to directly
dereference a NULL pointer and crash the kernel like below.

BUG: kernel NULL pointer dereference, address: 0000000000000000
PGD 8253067 P4D 8253067 PUD 8e0e067 PMD 0
Oops: 0002 [#1] PREEMPT SMP KASAN NOPTI
CPU: 0 PID: 98 Comm: poc.npd Not tainted 6.4.0-rc7-00072-gdad9774deaf1 #8
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.o4
RIP: 0010:memcpy_orig+0xad/0x140
Code: e8 4c 89 5f e0 48 8d 7f e0 73 d2 83 c2 20 48 29 d6 48 29 d7 83 fa 10 72 34 4c 8b 06 4c 8b 4e 08 c
RSP: 0018:ffff888008f57658 EFLAGS: 00000202
RAX: 0000000000000000 RBX: ffff888008bd0000 RCX: ffffffff8238e571
RDX: 0000000000000018 RSI: ffff888007f64844 RDI: 0000000000000000
RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: ffff888008f57818
R13: ffff888007f64aa4 R14: 0000000000000000 R15: 0000000000000000
FS:  00000000014013c0(0000) GS:ffff88806d600000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000000 CR3: 00000000054d8000 CR4: 00000000000006f0
Call Trace:
 <TASK>
 ? __die+0x1f/0x70
 ? page_fault_oops+0x1e8/0x500
 ? __pfx_is_prefetch.constprop.0+0x10/0x10
 ? __pfx_page_fault_oops+0x10/0x10
 ? _raw_spin_unlock_irqrestore+0x11/0x40
 ? fixup_exception+0x36/0x460
 ? _raw_spin_unlock_irqrestore+0x11/0x40
 ? exc_page_fault+0x5e/0xc0
 ? asm_exc_page_fault+0x26/0x30
 ? xfrm_update_ae_params+0xd1/0x260
 ? memcpy_orig+0xad/0x140
 ? __pfx__raw_spin_lock_bh+0x10/0x10
 xfrm_update_ae_params+0xe7/0x260
 xfrm_new_ae+0x298/0x4e0
 ? __pfx_xfrm_new_ae+0x10/0x10
 ? __pfx_xfrm_new_ae+0x10/0x10
 xfrm_user_rcv_msg+0x25a/0x410
 ? __pfx_xfrm_user_rcv_msg+0x10/0x10
 ? __alloc_skb+0xcf/0x210
 ? stack_trace_save+0x90/0xd0
 ? filter_irq_stacks+0x1c/0x70
 ? __stack_depot_save+0x39/0x4e0
 ? __kasan_slab_free+0x10a/0x190
 ? kmem_cache_free+0x9c/0x340
 ? netlink_recvmsg+0x23c/0x660
 ? sock_recvmsg+0xeb/0xf0
 ? __sys_recvfrom+0x13c/0x1f0
 ? __x64_sys_recvfrom+0x71/0x90
 ? do_syscall_64+0x3f/0x90
 ? entry_SYSCALL_64_after_hwframe+0x72/0xdc
 ? copyout+0x3e/0x50
 netlink_rcv_skb+0xd6/0x210
 ? __pfx_xfrm_user_rcv_msg+0x10/0x10
 ? __pfx_netlink_rcv_skb+0x10/0x10
 ? __pfx_sock_has_perm+0x10/0x10
 ? mutex_lock+0x8d/0xe0
 ? __pfx_mutex_lock+0x10/0x10
 xfrm_netlink_rcv+0x44/0x50
 netlink_unicast+0x36f/0x4c0
 ? __pfx_netlink_unicast+0x10/0x10
 ? netlink_recvmsg+0x500/0x660
 netlink_sendmsg+0x3b7/0x700

This Null-ptr-deref bug is assigned CVE-2023-3772. And this commit
adds additional NULL check in xfrm_update_ae_params to fix the NPD.

Fixes: d8647b79c3b7 ("xfrm: Add user interface for esn and big anti-replay windows")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ip_vti: fix potential slab-use-after-free in decode_session6

[ Upstream commit 6018a266279b1a75143c7c0804dd08a5fc4c3e0b ]

When ip_vti device is set to the qdisc of the sfb type, the cb field
of the sent skb may be modified during enqueuing. Then,
slab-use-after-free may occur when ip_vti device sends IPv6 packets.
As commit f855691975bb ("xfrm6: Fix the nexthdr offset in
_decode_session6.") showed, xfrm_decode_session was originally intended
only for the receive path. IP6CB(skb)->nhoff is not set during
transmission. Therefore, set the cb field in the skb to 0 before
sending packets.

Fixes: f855691975bb ("xfrm6: Fix the nexthdr offset in _decode_session6.")
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ip6_vti: fix slab-use-after-free in decode_session6

[ Upstream commit 9fd41f1ba638938c9a1195d09bc6fa3be2712f25 ]

When ipv6_vti device is set to the qdisc of the sfb type, the cb field
of the sent skb may be modified during enqueuing. Then,
slab-use-after-free may occur when ipv6_vti device sends IPv6 packets.

The stack information is as follows:
BUG: KASAN: slab-use-after-free in decode_session6+0x103f/0x1890
Read of size 1 at addr ffff88802e08edc2 by task swapper/0/0
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.4.0-next-20230707-00001-g84e2cad7f979 #410
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-1.fc33 04/01/2014
Call Trace:
<IRQ>
dump_stack_lvl+0xd9/0x150
print_address_description.constprop.0+0x2c/0x3c0
kasan_report+0x11d/0x130
decode_session6+0x103f/0x1890
__xfrm_decode_session+0x54/0xb0
vti6_tnl_xmit+0x3e6/0x1ee0
dev_hard_start_xmit+0x187/0x700
sch_direct_xmit+0x1a3/0xc30
__qdisc_run+0x510/0x17a0
__dev_queue_xmit+0x2215/0x3b10
neigh_connected_output+0x3c2/0x550
ip6_finish_output2+0x55a/0x1550
ip6_finish_output+0x6b9/0x1270
ip6_output+0x1f1/0x540
ndisc_send_skb+0xa63/0x1890
ndisc_send_rs+0x132/0x6f0
addrconf_rs_timer+0x3f1/0x870
call_timer_fn+0x1a0/0x580
expire_timers+0x29b/0x4b0
run_timer_softirq+0x326/0x910
__do_softirq+0x1d4/0x905
irq_exit_rcu+0xb7/0x120
sysvec_apic_timer_interrupt+0x97/0xc0
</IRQ>
Allocated by task 9176:
kasan_save_stack+0x22/0x40
kasan_set_track+0x25/0x30
__kasan_slab_alloc+0x7f/0x90
kmem_cache_alloc_node+0x1cd/0x410
kmalloc_reserve+0x165/0x270
__alloc_skb+0x129/0x330
netlink_sendmsg+0x9b1/0xe30
sock_sendmsg+0xde/0x190
____sys_sendmsg+0x739/0x920
___sys_sendmsg+0x110/0x1b0
__sys_sendmsg+0xf7/0x1c0
do_syscall_64+0x39/0xb0
entry_SYSCALL_64_after_hwframe+0x63/0xcd
Freed by task 9176:
kasan_save_stack+0x22/0x40
kasan_set_track+0x25/0x30
kasan_save_free_info+0x2b/0x40
____kasan_slab_free+0x160/0x1c0
slab_free_freelist_hook+0x11b/0x220
kmem_cache_free+0xf0/0x490
skb_free_head+0x17f/0x1b0
skb_release_data+0x59c/0x850
consume_skb+0xd2/0x170
netlink_unicast+0x54f/0x7f0
netlink_sendmsg+0x926/0xe30
sock_sendmsg+0xde/0x190
____sys_sendmsg+0x739/0x920
___sys_sendmsg+0x110/0x1b0
__sys_sendmsg+0xf7/0x1c0
do_syscall_64+0x39/0xb0
entry_SYSCALL_64_after_hwframe+0x63/0xcd
The buggy address belongs to the object at ffff88802e08ed00
which belongs to the cache skbuff_small_head of size 640
The buggy address is located 194 bytes inside of
freed 640-byte region [ffff88802e08ed00, ffff88802e08ef80)

As commit f855691975bb ("xfrm6: Fix the nexthdr offset in
_decode_session6.") showed, xfrm_decode_session was originally intended
only for the receive path. IP6CB(skb)->nhoff is not set during
transmission. Therefore, set the cb field in the skb to 0 before
sending packets.

Fixes: f855691975bb ("xfrm6: Fix the nexthdr offset in _decode_session6.")
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

xfrm: fix slab-use-after-free in decode_session6

[ Upstream commit 53223f2ed1ef5c90dad814daaaefea4e68a933c8 ]

When the xfrm device is set to the qdisc of the sfb type, the cb field
of the sent skb may be modified during enqueuing. Then,
slab-use-after-free may occur when the xfrm device sends IPv6 packets.

The stack information is as follows:
BUG: KASAN: slab-use-after-free in decode_session6+0x103f/0x1890
Read of size 1 at addr ffff8881111458ef by task swapper/3/0
CPU: 3 PID: 0 Comm: swapper/3 Not tainted 6.4.0-next-20230707 #409
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-1.fc33 04/01/2014
Call Trace:
<IRQ>
dump_stack_lvl+0xd9/0x150
print_address_description.constprop.0+0x2c/0x3c0
kasan_report+0x11d/0x130
decode_session6+0x103f/0x1890
__xfrm_decode_session+0x54/0xb0
xfrmi_xmit+0x173/0x1ca0
dev_hard_start_xmit+0x187/0x700
sch_direct_xmit+0x1a3/0xc30
__qdisc_run+0x510/0x17a0
__dev_queue_xmit+0x2215/0x3b10
neigh_connected_output+0x3c2/0x550
ip6_finish_output2+0x55a/0x1550
ip6_finish_output+0x6b9/0x1270
ip6_output+0x1f1/0x540
ndisc_send_skb+0xa63/0x1890
ndisc_send_rs+0x132/0x6f0
addrconf_rs_timer+0x3f1/0x870
call_timer_fn+0x1a0/0x580
expire_timers+0x29b/0x4b0
run_timer_softirq+0x326/0x910
__do_softirq+0x1d4/0x905
irq_exit_rcu+0xb7/0x120
sysvec_apic_timer_interrupt+0x97/0xc0
</IRQ>
<TASK>
asm_sysvec_apic_timer_interrupt+0x1a/0x20
RIP: 0010:intel_idle_hlt+0x23/0x30
Code: 1f 84 00 00 00 00 00 f3 0f 1e fa 41 54 41 89 d4 0f 1f 44 00 00 66 90 0f 1f 44 00 00 0f 00 2d c4 9f ab 00 0f 1f 44 00 00 fb f4 <fa> 44 89 e0 41 5c c3 66 0f 1f 44 00 00 f3 0f 1e fa 41 54 41 89 d4
RSP: 0018:ffffc90000197d78 EFLAGS: 00000246
RAX: 00000000000a83c3 RBX: ffffe8ffffd09c50 RCX: ffffffff8a22d8e5
RDX: 0000000000000001 RSI: ffffffff8d3f8080 RDI: ffffe8ffffd09c50
RBP: ffffffff8d3f8080 R08: 0000000000000001 R09: ffffed1026ba6d9d
R10: ffff888135d36ceb R11: 0000000000000001 R12: 0000000000000001
R13: ffffffff8d3f8100 R14: 0000000000000001 R15: 0000000000000000
cpuidle_enter_state+0xd3/0x6f0
cpuidle_enter+0x4e/0xa0
do_idle+0x2fe/0x3c0
cpu_startup_entry+0x18/0x20
start_secondary+0x200/0x290
secondary_startup_64_no_verify+0x167/0x16b
</TASK>
Allocated by task 939:
kasan_save_stack+0x22/0x40
kasan_set_track+0x25/0x30
__kasan_slab_alloc+0x7f/0x90
kmem_cache_alloc_node+0x1cd/0x410
kmalloc_reserve+0x165/0x270
__alloc_skb+0x129/0x330
inet6_ifa_notify+0x118/0x230
__ipv6_ifa_notify+0x177/0xbe0
addrconf_dad_completed+0x133/0xe00
addrconf_dad_work+0x764/0x1390
process_one_work+0xa32/0x16f0
worker_thread+0x67d/0x10c0
kthread+0x344/0x440
ret_from_fork+0x1f/0x30
The buggy address belongs to the object at ffff888111145800
which belongs to the cache skbuff_small_head of size 640
The buggy address is located 239 bytes inside of
freed 640-byte region [ffff888111145800, ffff888111145a80)

As commit f855691975bb ("xfrm6: Fix the nexthdr offset in
_decode_session6.") showed, xfrm_decode_session was originally intended
only for the receive path. IP6CB(skb)->nhoff is not set during
transmission. Therefore, set the cb field in the skb to 0 before
sending packets.

Fixes: f855691975bb ("xfrm6: Fix the nexthdr offset in _decode_session6.")
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: xfrm: Amend XFRMA_SEC_CTX nla_policy structure

[ Upstream commit d1e0e61d617ba17aa516db707aa871387566bbf7 ]

According to all consumers code of attrs[XFRMA_SEC_CTX], like

* verify_sec_ctx_len(), convert to xfrm_user_sec_ctx*
* xfrm_state_construct(), call security_xfrm_state_alloc whose prototype
is int security_xfrm_state_alloc(.., struct xfrm_user_sec_ctx *sec_ctx);
* copy_from_user_sec_ctx(), convert to xfrm_user_sec_ctx *
...

It seems that the expected parsing result for XFRMA_SEC_CTX should be
structure xfrm_user_sec_ctx, and the current xfrm_sec_ctx is confusing
and misleading (Luckily, they happen to have same size 8 bytes).

This commit amend the policy structure to xfrm_user_sec_ctx to avoid
ambiguity.

Fixes: cf5cb79f6946 ("[XFRM] netlink: Establish an attribute policy")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: af_key: fix sadb_x_filter validation

[ Upstream commit 75065a8929069bc93181848818e23f147a73f83a ]

When running xfrm_state_walk_init(), the xfrm_address_filter being used
is okay to have a splen/dplen that equals to sizeof(xfrm_address_t)<<3.
This commit replaces >= to > to make sure the boundary checking is
correct.

Fixes: 37bd22420f85 ("af_key: pfkey_dump needs parameter validation")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: xfrm: Fix xfrm_address_filter OOB read

[ Upstream commit dfa73c17d55b921e1d4e154976de35317e43a93a ]

We found below OOB crash:

[   44.211730] ==================================================================
[   44.212045] BUG: KASAN: slab-out-of-bounds in memcmp+0x8b/0xb0
[   44.212045] Read of size 8 at addr ffff88800870f320 by task poc.xfrm/97
[   44.212045]
[   44.212045] CPU: 0 PID: 97 Comm: poc.xfrm Not tainted 6.4.0-rc7-00072-gdad9774deaf1-dirty #4
[   44.212045] Call Trace:
[   44.212045]  <TASK>
[   44.212045]  dump_stack_lvl+0x37/0x50
[   44.212045]  print_report+0xcc/0x620
[   44.212045]  ? __virt_addr_valid+0xf3/0x170
[   44.212045]  ? memcmp+0x8b/0xb0
[   44.212045]  kasan_report+0xb2/0xe0
[   44.212045]  ? memcmp+0x8b/0xb0
[   44.212045]  kasan_check_range+0x39/0x1c0
[   44.212045]  memcmp+0x8b/0xb0
[   44.212045]  xfrm_state_walk+0x21c/0x420
[   44.212045]  ? __pfx_dump_one_state+0x10/0x10
[   44.212045]  xfrm_dump_sa+0x1e2/0x290
[   44.212045]  ? __pfx_xfrm_dump_sa+0x10/0x10
[   44.212045]  ? __kernel_text_address+0xd/0x40
[   44.212045]  ? kasan_unpoison+0x27/0x60
[   44.212045]  ? mutex_lock+0x60/0xe0
[   44.212045]  ? __pfx_mutex_lock+0x10/0x10
[   44.212045]  ? kasan_save_stack+0x22/0x50
[   44.212045]  netlink_dump+0x322/0x6c0
[   44.212045]  ? __pfx_netlink_dump+0x10/0x10
[   44.212045]  ? mutex_unlock+0x7f/0xd0
[   44.212045]  ? __pfx_mutex_unlock+0x10/0x10
[   44.212045]  __netlink_dump_start+0x353/0x430
[   44.212045]  xfrm_user_rcv_msg+0x3a4/0x410
[   44.212045]  ? __pfx__raw_spin_lock_irqsave+0x10/0x10
[   44.212045]  ? __pfx_xfrm_user_rcv_msg+0x10/0x10
[   44.212045]  ? __pfx_xfrm_dump_sa+0x10/0x10
[   44.212045]  ? __pfx_xfrm_dump_sa_done+0x10/0x10
[   44.212045]  ? __stack_depot_save+0x382/0x4e0
[   44.212045]  ? filter_irq_stacks+0x1c/0x70
[   44.212045]  ? kasan_save_stack+0x32/0x50
[   44.212045]  ? kasan_save_stack+0x22/0x50
[   44.212045]  ? kasan_set_track+0x25/0x30
[   44.212045]  ? __kasan_slab_alloc+0x59/0x70
[   44.212045]  ? kmem_cache_alloc_node+0xf7/0x260
[   44.212045]  ? kmalloc_reserve+0xab/0x120
[   44.212045]  ? __alloc_skb+0xcf/0x210
[   44.212045]  ? netlink_sendmsg+0x509/0x700
[   44.212045]  ? sock_sendmsg+0xde/0xe0
[   44.212045]  ? __sys_sendto+0x18d/0x230
[   44.212045]  ? __x64_sys_sendto+0x71/0x90
[   44.212045]  ? do_syscall_64+0x3f/0x90
[   44.212045]  ? entry_SYSCALL_64_after_hwframe+0x72/0xdc
[   44.212045]  ? netlink_sendmsg+0x509/0x700
[   44.212045]  ? sock_sendmsg+0xde/0xe0
[   44.212045]  ? __sys_sendto+0x18d/0x230
[   44.212045]  ? __x64_sys_sendto+0x71/0x90
[   44.212045]  ? do_syscall_64+0x3f/0x90
[   44.212045]  ? entry_SYSCALL_64_after_hwframe+0x72/0xdc
[   44.212045]  ? kasan_save_stack+0x22/0x50
[   44.212045]  ? kasan_set_track+0x25/0x30
[   44.212045]  ? kasan_save_free_info+0x2e/0x50
[   44.212045]  ? __kasan_slab_free+0x10a/0x190
[   44.212045]  ? kmem_cache_free+0x9c/0x340
[   44.212045]  ? netlink_recvmsg+0x23c/0x660
[   44.212045]  ? sock_recvmsg+0xeb/0xf0
[   44.212045]  ? __sys_recvfrom+0x13c/0x1f0
[   44.212045]  ? __x64_sys_recvfrom+0x71/0x90
[   44.212045]  ? do_syscall_64+0x3f/0x90
[   44.212045]  ? entry_SYSCALL_64_after_hwframe+0x72/0xdc
[   44.212045]  ? copyout+0x3e/0x50
[   44.212045]  netlink_rcv_skb+0xd6/0x210
[   44.212045]  ? __pfx_xfrm_user_rcv_msg+0x10/0x10
[   44.212045]  ? __pfx_netlink_rcv_skb+0x10/0x10
[   44.212045]  ? __pfx_sock_has_perm+0x10/0x10
[   44.212045]  ? mutex_lock+0x8d/0xe0
[   44.212045]  ? __pfx_mutex_lock+0x10/0x10
[   44.212045]  xfrm_netlink_rcv+0x44/0x50
[   44.212045]  netlink_unicast+0x36f/0x4c0
[   44.212045]  ? __pfx_netlink_unicast+0x10/0x10
[   44.212045]  ? netlink_recvmsg+0x500/0x660
[   44.212045]  netlink_sendmsg+0x3b7/0x700
[   44.212045]  ? __pfx_netlink_sendmsg+0x10/0x10
[   44.212045]  ? __pfx_netlink_sendmsg+0x10/0x10
[   44.212045]  sock_sendmsg+0xde/0xe0
[   44.212045]  __sys_sendto+0x18d/0x230
[   44.212045]  ? __pfx___sys_sendto+0x10/0x10
[   44.212045]  ? rcu_core+0x44a/0xe10
[   44.212045]  ? __rseq_handle_notify_resume+0x45b/0x740
[   44.212045]  ? _raw_spin_lock_irq+0x81/0xe0
[   44.212045]  ? __pfx___rseq_handle_notify_resume+0x10/0x10
[   44.212045]  ? __pfx_restore_fpregs_from_fpstate+0x10/0x10
[   44.212045]  ? __pfx_blkcg_maybe_throttle_current+0x10/0x10
[   44.212045]  ? __pfx_task_work_run+0x10/0x10
[   44.212045]  __x64_sys_sendto+0x71/0x90
[   44.212045]  do_syscall_64+0x3f/0x90
[   44.212045]  entry_SYSCALL_64_after_hwframe+0x72/0xdc
[   44.212045] RIP: 0033:0x44b7da
[   44.212045] RSP: 002b:00007ffdc8838548 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
[   44.212045] RAX: ffffffffffffffda RBX: 00007ffdc8839978 RCX: 000000000044b7da
[   44.212045] RDX: 0000000000000038 RSI: 00007ffdc8838770 RDI: 0000000000000003
[   44.212045] RBP: 00007ffdc88385b0 R08: 00007ffdc883858c R09: 000000000000000c
[   44.212045] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000001
[   44.212045] R13: 00007ffdc8839968 R14: 00000000004c37d0 R15: 0000000000000001
[   44.212045]  </TASK>
[   44.212045]
[   44.212045] Allocated by task 97:
[   44.212045]  kasan_save_stack+0x22/0x50
[   44.212045]  kasan_set_track+0x25/0x30
[   44.212045]  __kasan_kmalloc+0x7f/0x90
[   44.212045]  __kmalloc_node_track_caller+0x5b/0x140
[   44.212045]  kmemdup+0x21/0x50
[   44.212045]  xfrm_dump_sa+0x17d/0x290
[   44.212045]  netlink_dump+0x322/0x6c0
[   44.212045]  __netlink_dump_start+0x353/0x430
[   44.212045]  xfrm_user_rcv_msg+0x3a4/0x410
[   44.212045]  netlink_rcv_skb+0xd6/0x210
[   44.212045]  xfrm_netlink_rcv+0x44/0x50
[   44.212045]  netlink_unicast+0x36f/0x4c0
[   44.212045]  netlink_sendmsg+0x3b7/0x700
[   44.212045]  sock_sendmsg+0xde/0xe0
[   44.212045]  __sys_sendto+0x18d/0x230
[   44.212045]  __x64_sys_sendto+0x71/0x90
[   44.212045]  do_syscall_64+0x3f/0x90
[   44.212045]  entry_SYSCALL_64_after_hwframe+0x72/0xdc
[   44.212045]
[   44.212045] The buggy address belongs to the object at ffff88800870f300
[   44.212045]  which belongs to the cache kmalloc-64 of size 64
[   44.212045] The buggy address is located 32 bytes inside of
[   44.212045]  allocated 36-byte region [ffff88800870f300, ffff88800870f324)
[   44.212045]
[   44.212045] The buggy address belongs to the physical page:
[   44.212045] page:00000000e4de16ee refcount:1 mapcount:0 mapping:000000000 ...
[   44.212045] flags: 0x100000000000200(slab|node=0|zone=1)
[   44.212045] page_type: 0xffffffff()
[   44.212045] raw: 0100000000000200 ffff888004c41640 dead000000000122 0000000000000000
[   44.212045] raw: 0000000000000000 0000000080200020 00000001ffffffff 0000000000000000
[   44.212045] page dumped because: kasan: bad access detected
[   44.212045]
[   44.212045] Memory state around the buggy address:
[   44.212045]  ffff88800870f200: fa fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
[   44.212045]  ffff88800870f280: 00 00 00 00 00 fc fc fc fc fc fc fc fc fc fc fc
[   44.212045] >ffff88800870f300: 00 00 00 00 04 fc fc fc fc fc fc fc fc fc fc fc
[   44.212045]                                ^
[   44.212045]  ffff88800870f380: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   44.212045]  ffff88800870f400: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   44.212045] ==================================================================

By investigating the code, we find the root cause of this OOB is the lack
of checks in xfrm_dump_sa(). The buggy code allows a malicious user to pass
arbitrary value of filter->splen/dplen. Hence, with crafted xfrm states,
the attacker can achieve 8 bytes heap OOB read, which causes info leak.

  if (attrs[XFRMA_ADDRESS_FILTER]) {
    filter = kmemdup(nla_data(attrs[XFRMA_ADDRESS_FILTER]),
        sizeof(*filter), GFP_KERNEL);
    if (filter == NULL)
      return -ENOMEM;
    // NO MORE CHECKS HERE !!!
  }

This patch fixes the OOB by adding necessary boundary checks, just like
the code in pfkey_dump() function.

Fixes: d3623099d350 ("ipsec: add support of limited SA dump")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

i2c: designware: Handle invalid SMBus block data response length value

commit 69f035c480d76f12bf061148ccfd578e1099e5fc upstream.

In the I2C_FUNC_SMBUS_BLOCK_DATA case, the invalid length byte value
(outside of 1-32) of the SMBus block data response from the Slave device
is not correctly handled by the I2C Designware driver.

In case IC_EMPTYFIFO_HOLD_MASTER_EN==1, which cannot be detected
from the registers, the Master can be disabled only if the STOP bit
is set. Without STOP bit set, the Master remains active, holding the bus
until receiving a block data response length. This hangs the bus and
is unrecoverable.

Avoid this by issuing another dump read to reach the stop condition when
an invalid length byte is received.

Cc: stable@vger.kernel.org
Signed-off-by: Tam Nguyen <tamnguyenchi@os.amperecomputing.com>
Acked-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Link: https://lore.kernel.org/r/20230726080001.337353-3-tamnguyenchi@os.amperecomputing.com
Reviewed-by: Andi Shyti <andi.shyti@kernel.org>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

i2c: designware: Correct length byte validation logic

commit 49d4db3953cb9004ff94efc0c176e026c820af5a upstream.

Commit 0daede80f870 ("i2c: designware: Convert driver to using regmap API")
changes the logic to validate the whole 32-bit return value of
DW_IC_DATA_CMD register instead of 8-bit LSB without reason.

Later, commit f53f15ba5a85 ("i2c: designware: Get right data length"),
introduced partial fix but not enough because the "tmp > 0" still test
tmp as 32-bit value and is wrong in case the IC_DATA_CMD[11] is set.

Revert the logic to just before commit 0daede80f870
("i2c: designware: Convert driver to using regmap API").

Fixes: f53f15ba5a85 ("i2c: designware: Get right data length")
Fixes: 0daede80f870 ("i2c: designware: Convert driver to using regmap API")
Cc: stable@vger.kernel.org
Signed-off-by: Tam Nguyen <tamnguyenchi@os.amperecomputing.com>
Signed-off-by: Quan Nguyen <quan@os.amperecomputing.com>
Acked-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Link: https://lore.kernel.org/r/20230726080001.337353-2-tamnguyenchi@os.amperecomputing.com
Reviewed-by: Andi Shyti <andi.shyti@kernel.org>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

btrfs: fix BUG_ON condition in btrfs_cancel_balance

commit 29eefa6d0d07e185f7bfe9576f91e6dba98189c2 upstream.

Pausing and canceling balance can race to interrupt balance lead to BUG_ON
panic in btrfs_cancel_balance. The BUG_ON condition in btrfs_cancel_balance
does not take this race scenario into account.

However, the race condition has no other side effects. We can fix that.

Reproducing it with panic trace like this:

  kernel BUG at fs/btrfs/volumes.c:4618!
  RIP: 0010:btrfs_cancel_balance+0x5cf/0x6a0
  Call Trace:
   <TASK>
   ? do_nanosleep+0x60/0x120
   ? hrtimer_nanosleep+0xb7/0x1a0
   ? sched_core_clone_cookie+0x70/0x70
   btrfs_ioctl_balance_ctl+0x55/0x70
   btrfs_ioctl+0xa46/0xd20
   __x64_sys_ioctl+0x7d/0xa0
   do_syscall_64+0x38/0x80
   entry_SYSCALL_64_after_hwframe+0x63/0xcd

  Race scenario as follows:
  > mutex_unlock(&fs_info->balance_mutex);
  > --------------------
  > .......issue pause and cancel req in another thread
  > --------------------
  > ret = __btrfs_balance(fs_info);
  >
  > mutex_lock(&fs_info->balance_mutex);
  > if (ret == -ECANCELED && atomic_read(&fs_info->balance_pause_req)) {
  >         btrfs_info(fs_info, "balance: paused");
  >         btrfs_exclop_balance(fs_info, BTRFS_EXCLOP_BALANCE_PAUSED);
  > }

CC: stable@vger.kernel.org # 4.19+
Signed-off-by: xiaoshoukui <xiaoshoukui@ruijie.com.cn>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

btrfs: fix incorrect splitting in btrfs_drop_extent_map_range

commit c962098ca4af146f2625ed64399926a098752c9c upstream.

In production we were seeing a variety of WARN_ON()'s in the extent_map
code, specifically in btrfs_drop_extent_map_range() when we have to call
add_extent_mapping() for our second split.

Consider the following extent map layout

PINNED
[0 16K)  [32K, 48K)

and then we call btrfs_drop_extent_map_range for [0, 36K), with
skip_pinned == true.  The initial loop will have

start = 0
end = 36K
len = 36K

we will find the [0, 16k) extent, but since we are pinned we will skip
it, which has this code

start = em_end;
if (end != (u64)-1)
len = start + len - em_end;

em_end here is 16K, so now the values are

start = 16K
len = 16K + 36K - 16K = 36K

len should instead be 20K.  This is a problem when we find the next
extent at [32K, 48K), we need to split this extent to leave [36K, 48k),
however the code for the split looks like this

split->start = start + len;
split->len = em_end - (start + len);

In this case we have

em_end = 48K
split->start = 16K + 36K       // this should be 16K + 20K
split->len = 48K - (16K + 36K) // this overflows as 16K + 36K is 52K

and now we have an invalid extent_map in the tree that potentially
overlaps other entries in the extent map.  Even in the non-overlapping
case we will have split->start set improperly, which will cause problems
with any block related calculations.

We don't actually need len in this loop, we can simply use end as our
end point, and only adjust start up when we find a pinned extent we need
to skip.

Adjust the logic to do this, which keeps us from inserting an invalid
extent map.

We only skip_pinned in the relocation case, so this is relatively rare,
except in the case where you are running relocation a lot, which can
happen with auto relocation on.

Fixes: 55ef68990029 ("Btrfs: Fix btrfs_drop_extent_cache for skip pinned case")
CC: stable@vger.kernel.org # 4.14+
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

tty: serial: fsl_lpuart: Clear the error flags by writing 1 for lpuart32 platforms

commit 282069845af388b08d622ad192b831dcd0549c62 upstream.

Do not read the data register to clear the error flags for lpuart32
platforms, the additional read may cause the receive FIFO underflow
since the DMA has already read the data register.
Actually all lpuart32 platforms support write 1 to clear those error
bits, let's use this method to better clear the error flags.

Fixes: 42b68768e51b ("serial: fsl_lpuart: DMA support for 32-bit variant")
Cc: stable <stable@kernel.org>
Signed-off-by: Sherry Sun <sherry.sun@nxp.com>
Link: https://lore.kernel.org/r/20230801022304.24251-1-sherry.sun@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

tty: n_gsm: fix the UAF caused by race condition in gsm_cleanup_mux

commit 3c4f8333b582487a2d1e02171f1465531cde53e3 upstream.

In commit 9b9c8195f3f0 ("tty: n_gsm: fix UAF in gsm_cleanup_mux"), the UAF
problem is not completely fixed. There is a race condition in
gsm_cleanup_mux(), which caused this UAF.

The UAF problem is triggered by the following race:
task[5046]                     task[5054]
-----------------------        -----------------------
gsm_cleanup_mux();
dlci = gsm->dlci[0];
mutex_lock(&gsm->mutex);
                               gsm_cleanup_mux();
       dlci = gsm->dlci[0]; //Didn't take the lock
gsm_dlci_release(gsm->dlci[i]);
gsm->dlci[i] = NULL;
mutex_unlock(&gsm->mutex);
                               mutex_lock(&gsm->mutex);
       dlci->dead = true; //UAF

Fix it by assigning values after mutex_lock().

Link: https://syzkaller.appspot.com/text?tag=CrashReport&x=176188b5a80000
Cc: stable <stable@kernel.org>
Fixes: 9b9c8195f3f0 ("tty: n_gsm: fix UAF in gsm_cleanup_mux")
Fixes: aa371e96f05d ("tty: n_gsm: fix restart handling via CLD command")
Signed-off-by: Yi Yang <yiyang13@huawei.com>
Co-developed-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Signed-off-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Link: https://lore.kernel.org/r/20230811031121.153237-1-yiyang13@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

vdpa: Enable strict validation for netlinks ops

commit f46c1e1620c6bbc9aad5693082efd1b80822e97c upstream.

The previous patches added the missing nla policies that were required for
validation to work.

Now strict validation on netlink ops can be enabled. This patch does it.

Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Cc: stable@vger.kernel.org
Message-Id: <20230727175757.73988-9-dtatulea@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

vdpa: Add max vqp attr to vdpa_nl_policy for nlattr length check

commit 5d6ba607d6cb5c58a4ddf33381e18c83dbb4098f upstream.

The vdpa_nl_policy structure is used to validate the nlattr when parsing
the incoming nlmsg. It will ensure the attribute being described produces
a valid nlattr pointer in info->attrs before entering into each handler
in vdpa_nl_ops.

That is to say, the missing part in vdpa_nl_policy may lead to illegal
nlattr after parsing, which could lead to OOB read just like CVE-2023-3773.

This patch adds the missing nla_policy for vdpa max vqp attr to avoid
such bugs.

Fixes: ad69dd0bf26b ("vdpa: Introduce query of device config layout")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Cc: stable@vger.kernel.org
Message-Id: <20230727175757.73988-7-dtatulea@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

vdpa: Add queue index attr to vdpa_nl_policy for nlattr length check

commit b3003e1b54e057f5f3124e437b80c3bef26ed3fe upstream.

The vdpa_nl_policy structure is used to validate the nlattr when parsing
the incoming nlmsg. It will ensure the attribute being described produces
a valid nlattr pointer in info->attrs before entering into each handler
in vdpa_nl_ops.

That is to say, the missing part in vdpa_nl_policy may lead to illegal
nlattr after parsing, which could lead to OOB read just like CVE-2023-3773.

This patch adds the missing nla_policy for vdpa queue index attr to avoid
such bugs.

Fixes: 13b00b135665 ("vdpa: Add support for querying vendor statistics")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Cc: stable@vger.kernelorg
Message-Id: <20230727175757.73988-5-dtatulea@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

vdpa: Add features attr to vdpa_nl_policy for nlattr length check

commit 79c8651587504ba263d2fd67fd4406240fb21f69 upstream.

The vdpa_nl_policy structure is used to validate the nlattr when parsing
the incoming nlmsg. It will ensure the attribute being described produces
a valid nlattr pointer in info->attrs before entering into each handler
in vdpa_nl_ops.

That is to say, the missing part in vdpa_nl_policy may lead to illegal
nlattr after parsing, which could lead to OOB read just like CVE-2023-3773.

This patch adds the missing nla_policy for vdpa features attr to avoid
such bugs.

Fixes: 90fea5a800c3 ("vdpa: device feature provisioning")
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Cc: stable@vger.kernel.org
Message-Id: <20230727175757.73988-3-dtatulea@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

powerpc/rtas_flash: allow user copy to flash block cache objects

commit 4f3175979e62de3b929bfa54a0db4b87d36257a7 upstream.

With hardened usercopy enabled (CONFIG_HARDENED_USERCOPY=y), using the
/proc/powerpc/rtas/firmware_update interface to prepare a system
firmware update yields a BUG():

  kernel BUG at mm/usercopy.c:102!
  Oops: Exception in kernel mode, sig: 5 [#1]
  LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries
  Modules linked in:
  CPU: 0 PID: 2232 Comm: dd Not tainted 6.5.0-rc3+ #2
  Hardware name: IBM,8408-E8E POWER8E (raw) 0x4b0201 0xf000004 of:IBM,FW860.50 (SV860_146) hv:phyp pSeries
  NIP:  c0000000005991d0 LR: c0000000005991cc CTR: 0000000000000000
  REGS: c0000000148c76a0 TRAP: 0700   Not tainted  (6.5.0-rc3+)
  MSR:  8000000000029033 <SF,EE,ME,IR,DR,RI,LE>  CR: 24002242  XER: 0000000c
  CFAR: c0000000001fbd34 IRQMASK: 0
  [ ... GPRs omitted ... ]
  NIP usercopy_abort+0xa0/0xb0
  LR  usercopy_abort+0x9c/0xb0
  Call Trace:
    usercopy_abort+0x9c/0xb0 (unreliable)
    __check_heap_object+0x1b4/0x1d0
    __check_object_size+0x2d0/0x380
    rtas_flash_write+0xe4/0x250
    proc_reg_write+0xfc/0x160
    vfs_write+0xfc/0x4e0
    ksys_write+0x90/0x160
    system_call_exception+0x178/0x320
    system_call_common+0x160/0x2c4

The blocks of the firmware image are copied directly from user memory
to objects allocated from flash_block_cache, so flash_block_cache must
be created using kmem_cache_create_usercopy() to mark it safe for user
access.

Fixes: 6d07d1cd300f ("usercopy: Restrict non-usercopy caches to size 0")
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
[mpe: Trim and indent oops]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230810-rtas-flash-vs-hardened-usercopy-v2-1-dcf63793a938@linux.ibm.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

fbdev: mmp: fix value check in mmphw_probe()

commit 0872b2c0abc0e84ac82472959c8e14e35277549c upstream.

in mmphw_probe(), check the return value of clk_prepare_enable()
and return the error code if clk_prepare_enable() returns an
unexpected value.

Fixes: d63028c38905 ("video: mmp display controller support")
Signed-off-by: Yuanjun Gong <ruc_gongyuanjun@163.com>
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>