drm/xe/query: Increase timestamp width

Starting with Xe2 the timestamp is a full 64 bit counter, contrary to
the 36 bit that was available before. Although 36 should be sufficient
for any reasonable delta calculation (for Xe2, of about 30min), it's
surprising to userspace to get something truncated. Also if the
timestamp being compared to is coming from the GPU and the application
is not careful enough to apply the width there, a delta calculation
would be wrong.

Extend it to full 64-bits starting with Xe2.

v2: Expand width=64 to media gt, as it's just a wrong tagging in the
spec - empirical tests show it goes beyond 36 bits and match the engines
for the main gt

Bspec: 60411
Cc: Szymon Morek <szymon.morek@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241011035618.1057602-1-lucas.demarchi@intel.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
(cherry picked from commit 9d559cdcb21f42188d4c3ff3b4fe42b240f4af5d)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

drm/xe: Don't free job in TDR

Freeing job in TDR is not safe as TDR can pass the run_job thread
resulting in UAF. It is only safe for free job to naturally be called by
the scheduler. Rather free job in TDR, add to pending list.

Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2811
Cc: Matthew Auld <matthew.auld@intel.com>
Fixes: e275d61c5f3f ("drm/xe/guc: Handle timing out of signaled jobs gracefully")
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241003001657.3517883-3-matthew.brost@intel.com
(cherry picked from commit ea2f6a77d0c40d97f4a4dc93fee4afe15d94926d)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

drm/xe: Take job list lock in xe_sched_add_pending_job

A fragile micro optimization in xe_sched_add_pending_job relied on both
the GPU scheduler being stopped and fence signaling stopped to safely
add a job to the pending list without the job list lock in
xe_sched_add_pending_job. Remove this optimization and just take the job
list lock.

Fixes: 7ddb9403dd74 ("drm/xe: Sample ctx timestamp to determine if jobs have timed out")
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241003001657.3517883-2-matthew.brost@intel.com
(cherry picked from commit 90521df5fc43980e4575bd8c5b1cb62afe1a9f5f)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

drm/xe: fix unbalanced rpm put() with declare_wedged()

Technically the or_reset() means we call the action on failure, however
that would lead to unbalanced rpm put(). Move the get() earlier to fix
this. It should be extremely unlikely to ever trigger this in practice.

Fixes: 90936a0a4c54 ("drm/xe: Don't suspend device upon wedge")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Nirmoy Das <nirmoy.das@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241009084808.204432-4-matthew.auld@intel.com
(cherry picked from commit a187c1b0a800565a4db6372268692aff99df7f53)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

drm/xe: fix unbalanced rpm put() with fence_fini()

Currently we can call fence_fini() twice if something goes wrong when
sending the GuC CT for the tlb request, since we signal the fence and
return an error, leading to the caller also calling fini() on the error
path in the case of stack version of the flow, which leads to an extra
rpm put() which might later cause device to enter suspend when it
shouldn't. It looks like we can just drop the fini() call since the
fence signaller side will already call this for us.

There are known mysterious splats with device going to sleep even with
an rpm ref, and this could be one candidate.

v2 (Matt B):
  - Prefer warning if we detect double fini()

Fixes: f002702290fc ("drm/xe: Hold a PM ref when GT TLB invalidations are inflight")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Nirmoy Das <nirmoy.das@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241009084808.204432-3-matthew.auld@intel.com
(cherry picked from commit cfcbc0520d5055825f0647ab922b655688605183)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

drm/xe/xe2lpg: Extend Wa_15016589081 for xe2lpg

Add workaround (wa) 15016589081 which applies to Xe2_v3_LPG_MD.

Xe2_v3_LPG_MD is a Lunar Lake platform with GFX version: 20.04.
This wa is type: permanent, and hence is applicable on all steppings.

Signed-off-by: Aradhya Bhatia <aradhya.bhatia@intel.com>
Reviewed-by: Tejas Upadhyay <tejas.upadhyay@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241009065542.283151-1-aradhya.bhatia@intel.com
(cherry picked from commit 8fb1da9f9bfb02f710a7f826d50781b0b030cf53)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

blk-rq-qos: fix crash on rq_qos_wait vs. rq_qos_wake_function race

We're seeing crashes from rq_qos_wake_function that look like this:

  BUG: unable to handle page fault for address: ffffafe180a40084
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  PGD 100000067 P4D 100000067 PUD 10027c067 PMD 10115d067 PTE 0
  Oops: Oops: 0002 [#1] PREEMPT SMP PTI
  CPU: 17 UID: 0 PID: 0 Comm: swapper/17 Not tainted 6.12.0-rc3-00013-geca631b8fe80 #11
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
  RIP: 0010:_raw_spin_lock_irqsave+0x1d/0x40
  Code: 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f 1e fa 0f 1f 44 00 00 41 54 9c 41 5c fa 65 ff 05 62 97 30 4c 31 c0 ba 01 00 00 00 <f0> 0f b1 17 75 0a 4c 89 e0 41 5c c3 cc cc cc cc 89 c6 e8 2c 0b 00
  RSP: 0018:ffffafe180580ca0 EFLAGS: 00010046
  RAX: 0000000000000000 RBX: ffffafe180a3f7a8 RCX: 0000000000000011
  RDX: 0000000000000001 RSI: 0000000000000003 RDI: ffffafe180a40084
  RBP: 0000000000000000 R08: 00000000001e7240 R09: 0000000000000011
  R10: 0000000000000028 R11: 0000000000000888 R12: 0000000000000002
  R13: ffffafe180a40084 R14: 0000000000000000 R15: 0000000000000003
  FS:  0000000000000000(0000) GS:ffff9aaf1f280000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: ffffafe180a40084 CR3: 000000010e428002 CR4: 0000000000770ef0
  DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  PKRU: 55555554
  Call Trace:
   <IRQ>
   try_to_wake_up+0x5a/0x6a0
   rq_qos_wake_function+0x71/0x80
   __wake_up_common+0x75/0xa0
   __wake_up+0x36/0x60
   scale_up.part.0+0x50/0x110
   wb_timer_fn+0x227/0x450
   ...

So rq_qos_wake_function() calls wake_up_process(data->task), which calls
try_to_wake_up(), which faults in raw_spin_lock_irqsave(&p->pi_lock).

p comes from data->task, and data comes from the waitqueue entry, which
is stored on the waiter's stack in rq_qos_wait(). Analyzing the core
dump with drgn, I found that the waiter had already woken up and moved
on to a completely unrelated code path, clobbering what was previously
data->task. Meanwhile, the waker was passing the clobbered garbage in
data->task to wake_up_process(), leading to the crash.

What's happening is that in between rq_qos_wake_function() deleting the
waitqueue entry and calling wake_up_process(), rq_qos_wait() is finding
that it already got a token and returning. The race looks like this:

rq_qos_wait()                           rq_qos_wake_function()
==============================================================
prepare_to_wait_exclusive()
                                        data->got_token = true;
                                        list_del_init(&curr->entry);
if (data.got_token)
        break;
finish_wait(&rqw->wait, &data.wq);
  ^- returns immediately because
     list_empty_careful(&wq_entry->entry)
     is true
... return, go do something else ...
                                        wake_up_process(data->task)
                                          (NO LONGER VALID!)-^

Normally, finish_wait() is supposed to synchronize against the waker.
But, as noted above, it is returning immediately because the waitqueue
entry has already been removed from the waitqueue.

The bug is that rq_qos_wake_function() is accessing the waitqueue entry
AFTER deleting it. Note that autoremove_wake_function() wakes the waiter
and THEN deletes the waitqueue entry, which is the proper order.

Fix it by swapping the order. We also need to use
list_del_init_careful() to match the list_empty_careful() in
finish_wait().

Fixes: 38cfb5a45ee0 ("blk-wbt: improve waking of tasks")
Cc: stable@vger.kernel.org
Signed-off-by: Omar Sandoval <osandov@fb.com>
Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Link: https://lore.kernel.org/r/d3bee2463a67b1ee597211823bf7ad3721c26e41.1729014591.git.osandov@fb.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>

io_uring/rsrc: ignore dummy_ubuf for buffer cloning

For placeholder buffers, &dummy_ubuf is assigned which is a static
value. When buffers are attempted cloned, don't attempt to grab a
reference to it, as we both don't need it and it'll actively fail as
dummy_ubuf doesn't have a valid reference count setup.

Link: https://lore.kernel.org/io-uring/Zw8dkUzsxQ5LgAJL@ly-workstation/
Reported-by: Lai, Yi <yi1.lai@linux.intel.com>
Fixes: 7cc2a6eadcd7 ("io_uring: add IORING_REGISTER_COPY_BUFFERS method")
Signed-off-by: Jens Axboe <axboe@kernel.dk>

nilfs2: fix kernel bug due to missing clearing of buffer delay flag

Syzbot reported that after nilfs2 reads a corrupted file system image
and degrades to read-only, the BUG_ON check for the buffer delay flag
in submit_bh_wbc() may fail, causing a kernel bug.

This is because the buffer delay flag is not cleared when clearing the
buffer state flags to discard a page/folio or a buffer head. So, fix
this.

This became necessary when the use of nilfs2's own page clear routine
was expanded.  This state inconsistency does not occur if the buffer
is written normally by log writing.

Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Link: https://lore.kernel.org/r/20241015213300.7114-1-konishi.ryusuke@gmail.com
Fixes: 8c26c4e2694a ("nilfs2: fix issue with flush kernel thread after remount in RO mode because of driver's internal error or metadata corruption")
Reported-by: syzbot+985ada84bf055a575c07@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=985ada84bf055a575c07
Cc: stable@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>

drm/i915/dp_mst: Don't require DSC hblank quirk for a non-DSC compatible mode

If an MST branch device doesn't support DSC for a given mode, but the
MST link has enough BW for the mode, assume that the branch device does
support the mode using an uncompressed stream.

Fixes: 55eaef164174 ("drm/i915/dp_mst: Handle the Synaptics HBlank expansion quirk")
Cc: stable@vger.kernel.org # v6.8+
Reviewed-by: Suraj Kandpal <suraj.kandpal@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241009110135.1216498-2-imre.deak@intel.com
(cherry picked from commit 4e75c3e208a06ad6fd9b3517fb77337460d7c2b0)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/dp_mst: Handle error during DSC BW overhead/slice calculation

The MST branch device may not support the number of DSC slices a mode
requires, handle the error in this case.

Fixes: 4e0837a8d00a ("drm/i915/dp_mst: Account for FEC and DSC overhead during BW allocation")
Cc: stable@vger.kernel.org # v6.8+
Reviewed-by: Suraj Kandpal <suraj.kandpal@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241009110135.1216498-1-imre.deak@intel.com
(cherry picked from commit 802a69b6b8a0502a9e2309afec7e1b77f67874f2)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

fuse: update inode size after extending passthrough write

yangyun reported that libfuse test test_copy_file_range() copies zero
bytes from a newly written file when fuse passthrough is enabled.

The reason is that extending passthrough write is not updating the fuse
inode size and when vfs_copy_file_range() observes a zero size inode,
it returns without calling the filesystem copy_file_range() method.

Fix this by adjusting the fuse inode size after an extending passthrough
write.

This does not provide cache coherency of fuse inode attributes and
backing inode attributes, but it should prevent situations where fuse
inode size is too small, causing read/copy to be wrongly shortened.

Reported-by: yangyun <yangyun50@huawei.com>
Closes: https://github.com/libfuse/libfuse/issues/1048
Fixes: 57e1176e6086 ("fuse: implement read/write passthrough")
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>

fs: pass offset and result to backing_file end_write() callback

This is needed for extending fuse inode size after fuse passthrough write.

Suggested-by: Miklos Szeredi <miklos@szeredi.hu>
Link: https://lore.kernel.org/linux-fsdevel/CAJfpegs=cvZ_NYy6Q_D42XhYS=Sjj5poM1b5TzXzOVvX=R36aA@mail.gmail.com/
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>

s390: Update defconfigs

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

s390: Initialize psw mask in perf_arch_fetch_caller_regs()

Also initialize regs->psw.mask in perf_arch_fetch_caller_regs().
This way user_mode(regs) will return false, like it should.

It looks like all current users initialize regs to zero, so that this
doesn't fix a bug currently. However it is better to not rely on callers
to do this.

Fixes: 914d52e46490 ("s390: implement perf_arch_fetch_caller_regs")
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

s390/sclp_vt220: Convert newlines to CRLF instead of LFCR

According to the VT220 specification the possible character combinations
sent on RETURN are only CR or CRLF [0].

The Return key sends either a CR character (0/13) or a CR
character (0/13) and an LF character (0/10), depending on the
set/reset state of line feed/new line mode (LNM).

The sclp/vt220 driver however uses LFCR. This can confuse tools, for
example the kunit runner.

Link: https://vt100.net/docs/vt220-rm/chapter3.html#S3.2
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Cc: stable@vger.kernel.org
Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
Reviewed-by: Sven Schnelle <svens@linux.ibm.com>
Link: https://lore.kernel.org/r/20241014-s390-kunit-v1-2-941defa765a6@linutronix.de
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

s390/sclp: Deactivate sclp after all its users

On reboot the SCLP interface is deactivated through a reboot notifier.
This happens before other components using SCLP have the chance to run
their own reboot notifiers.
Two of those components are the SCLP console and tty drivers which try
to flush the last outstanding messages.
At that point the SCLP interface is already unusable and the messages
are discarded.

Execute sclp_deactivate() as late as possible to avoid this issue.

Fixes: 4ae46db99cd8 ("s390/consoles: improve panic notifiers reliability")
Cc: stable@vger.kernel.org
Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
Reviewed-by: Sven Schnelle <svens@linux.ibm.com>
Link: https://lore.kernel.org/r/20241014-s390-kunit-v1-1-941defa765a6@linutronix.de
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

s390/pkey_pckmo: Return with success for valid protected key types

The key_to_protkey handler function in module pkey_pckmo should return
with success on all known protected key types, including the new types
introduced by fd197556eef5 ("s390/pkey: Add AES xts and HMAC clear key
token support").

Fixes: fd197556eef5 ("s390/pkey: Add AES xts and HMAC clear key token support")
Signed-off-by: Holger Dengler <dengler@linux.ibm.com>
Reviewed-by: Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

ALSA: hda/conexant - Use cached pin control for Node 0x1d on HP EliteOne 1000 G2

The cached version avoids redundant commands to the codec, improving
stability and reducing unnecessary operations. This change ensures
better power management and reliable restoration of pin configurations,
especially after hibernation (S4) and other power transitions.

Fixes: 9988844c457f ("ALSA: hda/conexant - Fix audio routing for HP EliteOne 1000 G2")
Suggested-by: Kai-Heng Feng <kaihengf@nvidia.com>
Suggested-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
Link: https://patch.msgid.link/20241016080713.46801-1-kovalev@altlinux.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>

usb: gadget: f_uac2: fix return value for UAC2_ATTRIBUTE_STRING store

The configfs store callback should return the number of bytes consumed
not the total number of bytes we actually stored. These could differ if
for example the passed in string had a newline we did not store.

If the returned value does not match the number of bytes written the
writer might assume a failure or keep trying to write the remaining bytes.

For example the following command will hang trying to write the final
newline over and over again (tested on bash 2.05b):

  echo foo > function_name

Fixes: 993a44fa85c1 ("usb: gadget: f_uac2: allow changing interface name via configfs")
Cc: stable <stable@kernel.org>
Signed-off-by: Kevin Groeneveld <kgroeneveld@lenbrook.com>
Link: https://lore.kernel.org/r/20241006232637.4267-1-kgroeneveld@lenbrook.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: dwc3: core: Fix system suspend on TI AM62 platforms

Since commit 6d735722063a ("usb: dwc3: core: Prevent phy suspend during init"),
system suspend is broken on AM62 TI platforms.

Before that commit, both DWC3_GUSB3PIPECTL_SUSPHY and DWC3_GUSB2PHYCFG_SUSPHY
bits (hence forth called 2 SUSPHY bits) were being set during core
initialization and even during core re-initialization after a system
suspend/resume.

These bits are required to be set for system suspend/resume to work correctly
on AM62 platforms.

Since that commit, the 2 SUSPHY bits are not set for DEVICE/OTG mode if gadget
driver is not loaded and started.
For Host mode, the 2 SUSPHY bits are set before the first system suspend but
get cleared at system resume during core re-init and are never set again.

This patch resovles these two issues by ensuring the 2 SUSPHY bits are set
before system suspend and restored to the original state during system resume.

Cc: stable@vger.kernel.org # v6.9+
Fixes: 6d735722063a ("usb: dwc3: core: Prevent phy suspend during init")
Link: https://lore.kernel.org/all/1519dbe7-73b6-4afc-bfe3-23f4f75d772f@kernel.org/
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Acked-by: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
Tested-by: Markus Schneider-Pargmann <msp@baylibre.com>
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Link: https://lore.kernel.org/r/20241011-am62-lpm-usb-v3-1-562d445625b5@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xhci: tegra: fix checked USB2 port number

If USB virtualizatoin is enabled, USB2 ports are shared between all
Virtual Functions. The USB2 port number owned by an USB2 root hub in
a Virtual Function may be less than total USB2 phy number supported
by the Tegra XUSB controller.

Using total USB2 phy number as port number to check all PORTSC values
would cause invalid memory access.

[  116.923438] Unable to handle kernel paging request at virtual address 006c622f7665642f
...
[  117.213640] Call trace:
[  117.216783]  tegra_xusb_enter_elpg+0x23c/0x658
[  117.222021]  tegra_xusb_runtime_suspend+0x40/0x68
[  117.227260]  pm_generic_runtime_suspend+0x30/0x50
[  117.232847]  __rpm_callback+0x84/0x3c0
[  117.237038]  rpm_suspend+0x2dc/0x740
[  117.241229] pm_runtime_work+0xa0/0xb8
[  117.245769]  process_scheduled_works+0x24c/0x478
[  117.251007]  worker_thread+0x23c/0x328
[  117.255547]  kthread+0x104/0x1b0
[  117.259389]  ret_from_fork+0x10/0x20
[  117.263582] Code: 54000222 f9461ae8 f8747908 b4ffff48 (f9400100)

Cc: stable@vger.kernel.org # v6.3+
Fixes: a30951d31b25 ("xhci: tegra: USB2 pad power controls")
Signed-off-by: Henry Lin <henryl@nvidia.com>
Link: https://lore.kernel.org/r/20241014042134.27664-1-henryl@nvidia.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: dwc3: Wait for EndXfer completion before restoring GUSB2PHYCFG

DWC3 programming guide mentions that when operating in USB2.0 speeds,
if GUSB2PHYCFG[6] or GUSB2PHYCFG[8] is set, it must be cleared prior
to issuing commands and may be set again  after the command completes.
But currently while issuing EndXfer command without CmdIOC set, we
wait for 1ms after GUSB2PHYCFG is restored. This results in cases
where EndXfer command doesn't get completed and causes SMMU faults
since requests are unmapped afterwards. Hence restore GUSB2PHYCFG
after waiting for EndXfer command completion.

Cc: stable@vger.kernel.org
Fixes: 1d26ba0944d3 ("usb: dwc3: Wait unconditionally after issuing EndXfer command")
Signed-off-by: Prashanth K <quic_prashk@quicinc.com>
Acked-by: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
Link: https://lore.kernel.org/r/20240924093208.2524531-1-quic_prashk@quicinc.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: typec: qcom-pmic-typec: fix sink status being overwritten with RP_DEF

This line is overwriting the result of the above switch-case.

This fixes the tcpm driver getting stuck in a "Sink TX No Go" loop.

Fixes: a4422ff22142 ("usb: typec: qcom: Add Qualcomm PMIC Type-C driver")
Cc: stable <stable@kernel.org>
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Acked-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20241005144146.2345-1-jonathan@marek.ca
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: typec: altmode should keep reference to parent

The altmode device release refers to its parent device, but without keeping
a reference to it.

When registering the altmode, get a reference to the parent and put it in
the release function.

Before this fix, when using CONFIG_DEBUG_KOBJECT_RELEASE, we see issues
like this:

[   43.572860] kobject: 'port0.0' (ffff8880057ba008): kobject_release, parent 0000000000000000 (delayed 3000)
[   43.573532] kobject: 'port0.1' (ffff8880057bd008): kobject_release, parent 0000000000000000 (delayed 1000)
[   43.574407] kobject: 'port0' (ffff8880057b9008): kobject_release, parent 0000000000000000 (delayed 3000)
[   43.575059] kobject: 'port1.0' (ffff8880057ca008): kobject_release, parent 0000000000000000 (delayed 4000)
[   43.575908] kobject: 'port1.1' (ffff8880057c9008): kobject_release, parent 0000000000000000 (delayed 4000)
[   43.576908] kobject: 'typec' (ffff8880062dbc00): kobject_release, parent 0000000000000000 (delayed 4000)
[   43.577769] kobject: 'port1' (ffff8880057bf008): kobject_release, parent 0000000000000000 (delayed 3000)
[   46.612867] ==================================================================
[   46.613402] BUG: KASAN: slab-use-after-free in typec_altmode_release+0x38/0x129
[   46.614003] Read of size 8 at addr ffff8880057b9118 by task kworker/2:1/48
[   46.614538]
[   46.614668] CPU: 2 UID: 0 PID: 48 Comm: kworker/2:1 Not tainted 6.12.0-rc1-00138-gedbae730ad31 #535
[   46.615391] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.15.0-1 04/01/2014
[   46.616042] Workqueue: events kobject_delayed_cleanup
[   46.616446] Call Trace:
[   46.616648]  <TASK>
[   46.616820]  dump_stack_lvl+0x5b/0x7c
[   46.617112]  ? typec_altmode_release+0x38/0x129
[   46.617470]  print_report+0x14c/0x49e
[   46.617769]  ? rcu_read_unlock_sched+0x56/0x69
[   46.618117]  ? __virt_addr_valid+0x19a/0x1ab
[   46.618456]  ? kmem_cache_debug_flags+0xc/0x1d
[   46.618807]  ? typec_altmode_release+0x38/0x129
[   46.619161]  kasan_report+0x8d/0xb4
[   46.619447]  ? typec_altmode_release+0x38/0x129
[   46.619809]  ? process_scheduled_works+0x3cb/0x85f
[   46.620185]  typec_altmode_release+0x38/0x129
[   46.620537]  ? process_scheduled_works+0x3cb/0x85f
[   46.620907]  device_release+0xaf/0xf2
[   46.621206]  kobject_delayed_cleanup+0x13b/0x17a
[   46.621584]  process_scheduled_works+0x4f6/0x85f
[   46.621955]  ? __pfx_process_scheduled_works+0x10/0x10
[   46.622353]  ? hlock_class+0x31/0x9a
[   46.622647]  ? lock_acquired+0x361/0x3c3
[   46.622956]  ? move_linked_works+0x46/0x7d
[   46.623277]  worker_thread+0x1ce/0x291
[   46.623582]  ? __kthread_parkme+0xc8/0xdf
[   46.623900]  ? __pfx_worker_thread+0x10/0x10
[   46.624236]  kthread+0x17e/0x190
[   46.624501]  ? kthread+0xfb/0x190
[   46.624756]  ? __pfx_kthread+0x10/0x10
[   46.625015]  ret_from_fork+0x20/0x40
[   46.625268]  ? __pfx_kthread+0x10/0x10
[   46.625532]  ret_from_fork_asm+0x1a/0x30
[   46.625805]  </TASK>
[   46.625953]
[   46.626056] Allocated by task 678:
[   46.626287]  kasan_save_stack+0x24/0x44
[   46.626555]  kasan_save_track+0x14/0x2d
[   46.626811]  __kasan_kmalloc+0x3f/0x4d
[   46.627049]  __kmalloc_noprof+0x1bf/0x1f0
[   46.627362]  typec_register_port+0x23/0x491
[   46.627698]  cros_typec_probe+0x634/0xbb6
[   46.628026]  platform_probe+0x47/0x8c
[   46.628311]  really_probe+0x20a/0x47d
[   46.628605]  device_driver_attach+0x39/0x72
[   46.628940]  bind_store+0x87/0xd7
[   46.629213]  kernfs_fop_write_iter+0x1aa/0x218
[   46.629574]  vfs_write+0x1d6/0x29b
[   46.629856]  ksys_write+0xcd/0x13b
[   46.630128]  do_syscall_64+0xd4/0x139
[   46.630420]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[   46.630820]
[   46.630946] Freed by task 48:
[   46.631182]  kasan_save_stack+0x24/0x44
[   46.631493]  kasan_save_track+0x14/0x2d
[   46.631799]  kasan_save_free_info+0x3f/0x4d
[   46.632144]  __kasan_slab_free+0x37/0x45
[   46.632474]  kfree+0x1d4/0x252
[   46.632725]  device_release+0xaf/0xf2
[   46.633017]  kobject_delayed_cleanup+0x13b/0x17a
[   46.633388]  process_scheduled_works+0x4f6/0x85f
[   46.633764]  worker_thread+0x1ce/0x291
[   46.634065]  kthread+0x17e/0x190
[   46.634324]  ret_from_fork+0x20/0x40
[   46.634621]  ret_from_fork_asm+0x1a/0x30

Fixes: 8a37d87d72f0 ("usb: typec: Bus type for alternate modes")
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20241004123738.2964524-1-cascardo@igalia.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

MAINTAINERS: usb: raw-gadget: add bug tracker link

Add a link to the GitHub repository where Raw Gadget issues are managed.

Signed-off-by: Andrey Konovalov <andreyknvl@gmail.com>
Link: https://lore.kernel.org/r/20241012225853.118217-1-andrey.konovalov@linux.dev
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

MAINTAINERS: Add an entry for the LJCA drivers

Add a MAINTAINERS entry for the Intel La Jolla Cove Adapter (LJCA) set of
drivers.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20241011070414.3124-1-sakari.ailus@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

crypto: lib/mpi - Fix an "Uninitialized scalar variable" issue

The "err" variable may be returned without an initialized value.

Fixes: 8e3a67f2de87 ("crypto: lib/mpi - Add error checks to extension")
Signed-off-by: Qianqiang Liu <qianqiang.liu@163.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

cifs: Remove unused functions

cifs_ses_find_chan() has been unused since commit
f486ef8e2003 ("cifs: use the chans_need_reconnect bitmap for reconnect status")

cifs_read_page_from_socket() has been unused since commit
d08089f649a0 ("cifs: Change the I/O paths to use an iterator rather than a page list")

cifs_chan_in_reconnect() has been unused since commit
bc962159e8e3 ("cifs: avoid race conditions with parallel reconnects")

Remove them.

Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Signed-off-by: Steve French <stfrench@microsoft.com>

smb/client: Fix logically dead code

The if condition in collect_sample: can never be satisfied
because of a logical contradiction. The indicated dead code
may have performed some action; that action will never occur.

Fixes: 94ae8c3fee94 ("smb: client: compress: LZ77 code improvements cleanup")
Signed-off-by: Advait Dhamorikar <advaitdhamorikar@gmail.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

smb: client: fix OOBs when building SMB2_IOCTL request

When using encryption, either enforced by the server or when using
'seal' mount option, the client will squash all compound request buffers
down for encryption into a single iov in smb2_set_next_command().

SMB2_ioctl_init() allocates a small buffer (448 bytes) to hold the
SMB2_IOCTL request in the first iov, and if the user passes an input
buffer that is greater than 328 bytes, smb2_set_next_command() will
end up writing off the end of @rqst->iov[0].iov_base as shown below:

  mount.cifs //srv/share /mnt -o ...,seal
  ln -s $(perl -e "print('a')for 1..1024") /mnt/link

  BUG: KASAN: slab-out-of-bounds in
  smb2_set_next_command.cold+0x1d6/0x24c [cifs]
  Write of size 4116 at addr ffff8881148fcab8 by task ln/859

  CPU: 1 UID: 0 PID: 859 Comm: ln Not tainted 6.12.0-rc3 #1
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS
  1.16.3-2.fc40 04/01/2014
  Call Trace:
   <TASK>
   dump_stack_lvl+0x5d/0x80
   ? smb2_set_next_command.cold+0x1d6/0x24c [cifs]
   print_report+0x156/0x4d9
   ? smb2_set_next_command.cold+0x1d6/0x24c [cifs]
   ? __virt_addr_valid+0x145/0x310
   ? __phys_addr+0x46/0x90
   ? smb2_set_next_command.cold+0x1d6/0x24c [cifs]
   kasan_report+0xda/0x110
   ? smb2_set_next_command.cold+0x1d6/0x24c [cifs]
   kasan_check_range+0x10f/0x1f0
   __asan_memcpy+0x3c/0x60
   smb2_set_next_command.cold+0x1d6/0x24c [cifs]
   smb2_compound_op+0x238c/0x3840 [cifs]
   ? kasan_save_track+0x14/0x30
   ? kasan_save_free_info+0x3b/0x70
   ? vfs_symlink+0x1a1/0x2c0
   ? do_symlinkat+0x108/0x1c0
   ? __pfx_smb2_compound_op+0x10/0x10 [cifs]
   ? kmem_cache_free+0x118/0x3e0
   ? cifs_get_writable_path+0xeb/0x1a0 [cifs]
   smb2_get_reparse_inode+0x423/0x540 [cifs]
   ? __pfx_smb2_get_reparse_inode+0x10/0x10 [cifs]
   ? rcu_is_watching+0x20/0x50
   ? __kmalloc_noprof+0x37c/0x480
   ? smb2_create_reparse_symlink+0x257/0x490 [cifs]
   ? smb2_create_reparse_symlink+0x38f/0x490 [cifs]
   smb2_create_reparse_symlink+0x38f/0x490 [cifs]
   ? __pfx_smb2_create_reparse_symlink+0x10/0x10 [cifs]
   ? find_held_lock+0x8a/0xa0
   ? hlock_class+0x32/0xb0
   ? __build_path_from_dentry_optional_prefix+0x19d/0x2e0 [cifs]
   cifs_symlink+0x24f/0x960 [cifs]
   ? __pfx_make_vfsuid+0x10/0x10
   ? __pfx_cifs_symlink+0x10/0x10 [cifs]
   ? make_vfsgid+0x6b/0xc0
   ? generic_permission+0x96/0x2d0
   vfs_symlink+0x1a1/0x2c0
   do_symlinkat+0x108/0x1c0
   ? __pfx_do_symlinkat+0x10/0x10
   ? strncpy_from_user+0xaa/0x160
   __x64_sys_symlinkat+0xb9/0xf0
   do_syscall_64+0xbb/0x1d0
   entry_SYSCALL_64_after_hwframe+0x77/0x7f
  RIP: 0033:0x7f08d75c13bb

Reported-by: David Howells <dhowells@redhat.com>
Fixes: e77fe73c7e38 ("cifs: we can not use small padding iovs together with encryption")
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

smb: client: fix possible double free in smb2_set_ea()

Clang static checker(scan-build) warning：
fs/smb/client/smb2ops.c:1304:2: Attempt to free released memory.
 1304 |         kfree(ea);
      |         ^~~~~~~~~

There is a double free in such case:
'ea is initialized to NULL' -> 'first successful memory allocation for
ea' -> 'something failed, goto sea_exit' -> 'first memory release for ea'
-> 'goto replay_again' -> 'second goto sea_exit before allocate memory
for ea' -> 'second memory release for ea resulted in double free'.

Re-initialie 'ea' to NULL near to the replay_again label, it can fix this
double free problem.

Fixes: 4f1fffa23769 ("cifs: commands that are retried should have replay flag set")
Reviewed-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Su Hui <suhui@nfschina.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

cpufreq/amd-pstate: Use nominal perf for limits when boost is disabled

When boost has been disabled the limit for perf should be nominal perf not
the highest perf.  Using the latter to do calculations will lead to
incorrect values that are still above nominal.

Fixes: ad4caad58d91 ("cpufreq: amd-pstate: Merge amd_pstate_highest_perf_set() into amd_get_boost_ratio_numerator()")
Reported-by: Peter Jung <ptr1337@cachyos.org>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219348
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Tested-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com>
Link: https://lore.kernel.org/r/20241012174519.897-1-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>

Merge tag 'sched_ext-for-6.12-rc3-fixes' of git://git./linux/kernel/git/tj/sched_ext

Pull sched_ext fixes from Tejun Heo:

 - More issues reported in the enable/disable paths on large machines
   with many tasks due to scx_tasks_lock being held too long. Break up
   the task iterations

 - Remove ops.select_cpu() dependency in bypass mode so that a
   misbehaving implementation can't live-lock the machine by pushing all
   tasks to few CPUs in bypass mode

 - Other misc fixes

* tag 'sched_ext-for-6.12-rc3-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:
  sched_ext: Remove unnecessary cpu_relax()
  sched_ext: Don't hold scx_tasks_lock for too long
  sched_ext: Move scx_tasks_lock handling into scx_task_iter helpers
  sched_ext: bypass mode shouldn't depend on ops.select_cpu()
  sched_ext: Move scx_buildin_idle_enabled check to scx_bpf_select_cpu_dfl()
  sched_ext: Start schedulers with consistent p->scx.slice values
  Revert "sched_ext: Use shorter slice while bypassing"
  sched_ext: use correct function name in pick_task_scx() warning message
  selftests: sched_ext: Add sched_ext as proper selftest target

scsi: target: core: Fix null-ptr-deref in target_alloc_device()

There is a null-ptr-deref issue reported by KASAN:

BUG: KASAN: null-ptr-deref in target_alloc_device+0xbc4/0xbe0 [target_core_mod]
...
 kasan_report+0xb9/0xf0
 target_alloc_device+0xbc4/0xbe0 [target_core_mod]
 core_dev_setup_virtual_lun0+0xef/0x1f0 [target_core_mod]
 target_core_init_configfs+0x205/0x420 [target_core_mod]
 do_one_initcall+0xdd/0x4e0
...
 entry_SYSCALL_64_after_hwframe+0x76/0x7e

In target_alloc_device(), if allocing memory for dev queues fails, then
dev will be freed by dev->transport->free_device(), but dev->transport
is not initialized at that time, which will lead to a null pointer
reference problem.

Fixing this bug by freeing dev with hba->backend->ops->free_device().

Fixes: 1526d9f10c61 ("scsi: target: Make state_list per CPU")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Link: https://lore.kernel.org/r/20241011113444.40749-1-wanghai38@huawei.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Validate SAS port assignments

A sanity check on phy_mask was added in commit 3668651def2c ("scsi:
mpi3mr: Sanitise num_phys"). This causes warning messages when more than
64 phys are detected and devices connected to phys greater than 64 are
dropped.

The phy_mask bitmap is only needed for controller phys and not required
for expander phys. Controller phys can go up to a maximum of 64 and
therefore u64 is good enough to contain phy_mask bitmap.

To suppress those warnings and allow devices to be discovered as before
the offending commit, restrict the phy_mask setting and lowest phy
setting only to the controller phys.

Fixes: 3668651def2c ("scsi: mpi3mr: Sanitise num_phys")
Cc: stable@vger.kernel.org
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202410051943.Mp9o5DlF-lkp@intel.com/
Reported-by: Alexander Motin <mav@ixsystems.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
Link: https://lore.kernel.org/r/20241008074353.200379-1-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Set SDEV_OFFLINE when UFS is shut down

There is a history of deadlock if reboot is performed at the beginning
of booting. SDEV_QUIESCE was set for all LU's scsi_devices by UFS
shutdown, and at that time the audio driver was waiting on
blk_mq_submit_bio() holding a mutex_lock while reading the fw binary.
After that, a deadlock issue occurred while audio driver shutdown was
waiting for mutex_unlock of blk_mq_submit_bio(). To solve this, set
SDEV_OFFLINE for all LUs except WLUN, so that any I/O that comes down
after a UFS shutdown will return an error.

[   31.907781]I[0:      swapper/0:    0]        1        130705007       1651079834      11289729804                0 D(   2) 3 ffffff882e208000 *             init [device_shutdown]
[   31.907793]I[0:      swapper/0:    0] Mutex: 0xffffff8849a2b8b0: owner[0xffffff882e28cb00 kworker/6:0 :49]
[   31.907806]I[0:      swapper/0:    0] Call trace:
[   31.907810]I[0:      swapper/0:    0]  __switch_to+0x174/0x338
[   31.907819]I[0:      swapper/0:    0]  __schedule+0x5ec/0x9cc
[   31.907826]I[0:      swapper/0:    0]  schedule+0x7c/0xe8
[   31.907834]I[0:      swapper/0:    0]  schedule_preempt_disabled+0x24/0x40
[   31.907842]I[0:      swapper/0:    0]  __mutex_lock+0x408/0xdac
[   31.907849]I[0:      swapper/0:    0]  __mutex_lock_slowpath+0x14/0x24
[   31.907858]I[0:      swapper/0:    0]  mutex_lock+0x40/0xec
[   31.907866]I[0:      swapper/0:    0]  device_shutdown+0x108/0x280
[   31.907875]I[0:      swapper/0:    0]  kernel_restart+0x4c/0x11c
[   31.907883]I[0:      swapper/0:    0]  __arm64_sys_reboot+0x15c/0x280
[   31.907890]I[0:      swapper/0:    0]  invoke_syscall+0x70/0x158
[   31.907899]I[0:      swapper/0:    0]  el0_svc_common+0xb4/0xf4
[   31.907909]I[0:      swapper/0:    0]  do_el0_svc+0x2c/0xb0
[   31.907918]I[0:      swapper/0:    0]  el0_svc+0x34/0xe0
[   31.907928]I[0:      swapper/0:    0]  el0t_64_sync_handler+0x68/0xb4
[   31.907937]I[0:      swapper/0:    0]  el0t_64_sync+0x1a0/0x1a4

[   31.908774]I[0:      swapper/0:    0]       49                0         11960702      11236868007                0 D(   2) 6 ffffff882e28cb00 *      kworker/6:0 [__bio_queue_enter]
[   31.908783]I[0:      swapper/0:    0] Call trace:
[   31.908788]I[0:      swapper/0:    0]  __switch_to+0x174/0x338
[   31.908796]I[0:      swapper/0:    0]  __schedule+0x5ec/0x9cc
[   31.908803]I[0:      swapper/0:    0]  schedule+0x7c/0xe8
[   31.908811]I[0:      swapper/0:    0]  __bio_queue_enter+0xb8/0x178
[   31.908818]I[0:      swapper/0:    0]  blk_mq_submit_bio+0x194/0x67c
[   31.908827]I[0:      swapper/0:    0]  __submit_bio+0xb8/0x19c

Fixes: b294ff3e3449 ("scsi: ufs: core: Enable power management for wlun")
Cc: stable@vger.kernel.org
Signed-off-by: Seunghwan Baek <sh8267.baek@samsung.com>
Link: https://lore.kernel.org/r/20240829093913.6282-2-sh8267.baek@samsung.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Requeue aborted request

After the SQ cleanup fix, the CQ will receive a response with the
corresponding tag marked as OCS: ABORTED. To align with the behavior of
Legacy SDB mode, the handling of OCS: ABORTED has been changed to match
that of OCS_INVALID_COMMAND_STATUS (SDB), with both returning a SCSI
result of DID_REQUEUE.

Furthermore, the workaround implemented before the SQ cleanup fix can be
removed.

Fixes: ab248643d3d6 ("scsi: ufs: core: Add error handling for MCQ mode")
Cc: stable@vger.kernel.org
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Link: https://lore.kernel.org/r/20241001091917.6917-3-peter.wang@mediatek.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Fix the issue of ICU failure

When setting the ICU bit without using read-modify-write, SQRTCy will
restart SQ again and receive an RTC return error code 2 (Failure - SQ
not stopped).

Additionally, the error log has been modified so that this type of error
can be observed.

Fixes: ab248643d3d6 ("scsi: ufs: core: Add error handling for MCQ mode")
Cc: stable@vger.kernel.org
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Link: https://lore.kernel.org/r/20241001091917.6917-2-peter.wang@mediatek.com
Reviewed-by: Bao D. Nguyen <quic_nguyenb@quicinc.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

net: dsa: vsc73xx: fix reception from VLAN-unaware bridges

Similar to the situation described for sja1105 in commit 1f9fc48fd302
("net: dsa: sja1105: fix reception from VLAN-unaware bridges"), the
vsc73xx driver uses tag_8021q and doesn't need the ds->untag_bridge_pvid
request. In fact, this option breaks packet reception.

The ds->untag_bridge_pvid option strips VLANs from packets received on
VLAN-unaware bridge ports. But those VLANs should already be stripped
by tag_vsc73xx_8021q.c as part of vsc73xx_rcv() - they are not VLANs in
VLAN-unaware mode, but DSA tags. Thus, dsa_software_vlan_untag() tries
to untag a VLAN that doesn't exist, corrupting the packet.

Fixes: 93e4649efa96 ("net: dsa: provide a software untagging function on RX for VLAN-aware bridges")
Tested-by: Pawel Dembicki <paweldembicki@gmail.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://patch.msgid.link/20241014153041.1110364-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ravb: Only advertise Rx/Tx timestamps if hardware supports it

Recent work moving the reporting of Rx software timestamps to the core
[1] highlighted an issue where hardware time stamping was advertised
for the platforms where it is not supported.

Fix this by covering advertising support for hardware timestamps only if
the hardware supports it. Due to the Tx implementation in RAVB software
Tx timestamping is also only considered if the hardware supports
hardware timestamps. This should be addressed in future, but this fix
only reflects what the driver currently implements.

1. Commit 277901ee3a26 ("ravb: Remove setting of RX software timestamp")

Fixes: 7e09a052dc4e ("ravb: Exclude gPTP feature support for RZ/G2L")
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com>
Tested-by: Paul Barker <paul.barker.ct@bp.renesas.com>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Link: https://patch.msgid.link/20241014124343.3875285-1-niklas.soderlund+renesas@ragnatech.se
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: microchip: vcap api: Fix memory leaks in vcap_api_encode_rule_test()

Commit a3c1e45156ad ("net: microchip: vcap: Fix use-after-free error in
kunit test") fixed the use-after-free error, but introduced below
memory leaks by removing necessary vcap_free_rule(), add it to fix it.

unreferenced object 0xffffff80ca58b700 (size 192):
  comm "kunit_try_catch", pid 1215, jiffies 4294898264
  hex dump (first 32 bytes):
    00 12 7a 00 05 00 00 00 0a 00 00 00 64 00 00 00  ..z.........d...
    00 00 00 00 00 00 00 00 00 04 0b cc 80 ff ff ff  ................
  backtrace (crc 9c09c3fe):
    [<0000000052a0be73>] kmemleak_alloc+0x34/0x40
    [<0000000043605459>] __kmalloc_cache_noprof+0x26c/0x2f4
    [<0000000040a01b8d>] vcap_alloc_rule+0x3cc/0x9c4
    [<000000003fe86110>] vcap_api_encode_rule_test+0x1ac/0x16b0
    [<00000000b3595fc4>] kunit_try_run_case+0x13c/0x3ac
    [<0000000010f5d2bf>] kunit_generic_run_threadfn_adapter+0x80/0xec
    [<00000000c5d82c9a>] kthread+0x2e8/0x374
    [<00000000f4287308>] ret_from_fork+0x10/0x20
unreferenced object 0xffffff80cc0b0400 (size 64):
  comm "kunit_try_catch", pid 1215, jiffies 4294898265
  hex dump (first 32 bytes):
    80 04 0b cc 80 ff ff ff 18 b7 58 ca 80 ff ff ff  ..........X.....
    39 00 00 00 02 00 00 00 06 05 04 03 02 01 ff ff  9...............
  backtrace (crc daf014e9):
    [<0000000052a0be73>] kmemleak_alloc+0x34/0x40
    [<0000000043605459>] __kmalloc_cache_noprof+0x26c/0x2f4
    [<000000000ff63fd4>] vcap_rule_add_key+0x2cc/0x528
    [<00000000dfdb1e81>] vcap_api_encode_rule_test+0x224/0x16b0
    [<00000000b3595fc4>] kunit_try_run_case+0x13c/0x3ac
    [<0000000010f5d2bf>] kunit_generic_run_threadfn_adapter+0x80/0xec
    [<00000000c5d82c9a>] kthread+0x2e8/0x374
    [<00000000f4287308>] ret_from_fork+0x10/0x20
unreferenced object 0xffffff80cc0b0700 (size 64):
  comm "kunit_try_catch", pid 1215, jiffies 4294898265
  hex dump (first 32 bytes):
    80 07 0b cc 80 ff ff ff 28 b7 58 ca 80 ff ff ff  ........(.X.....
    3c 00 00 00 00 00 00 00 01 2f 03 b3 ec ff ff ff  <......../......
  backtrace (crc 8d877792):
    [<0000000052a0be73>] kmemleak_alloc+0x34/0x40
    [<0000000043605459>] __kmalloc_cache_noprof+0x26c/0x2f4
    [<000000006eadfab7>] vcap_rule_add_action+0x2d0/0x52c
    [<00000000323475d1>] vcap_api_encode_rule_test+0x4d4/0x16b0
    [<00000000b3595fc4>] kunit_try_run_case+0x13c/0x3ac
    [<0000000010f5d2bf>] kunit_generic_run_threadfn_adapter+0x80/0xec
    [<00000000c5d82c9a>] kthread+0x2e8/0x374
    [<00000000f4287308>] ret_from_fork+0x10/0x20
unreferenced object 0xffffff80cc0b0900 (size 64):
  comm "kunit_try_catch", pid 1215, jiffies 4294898266
  hex dump (first 32 bytes):
    80 09 0b cc 80 ff ff ff 80 06 0b cc 80 ff ff ff  ................
    7d 00 00 00 01 00 00 00 00 00 00 00 ff 00 00 00  }...............
  backtrace (crc 34181e56):
    [<0000000052a0be73>] kmemleak_alloc+0x34/0x40
    [<0000000043605459>] __kmalloc_cache_noprof+0x26c/0x2f4
    [<000000000ff63fd4>] vcap_rule_add_key+0x2cc/0x528
    [<00000000991e3564>] vcap_val_rule+0xcf0/0x13e8
    [<00000000fc9868e5>] vcap_api_encode_rule_test+0x678/0x16b0
    [<00000000b3595fc4>] kunit_try_run_case+0x13c/0x3ac
    [<0000000010f5d2bf>] kunit_generic_run_threadfn_adapter+0x80/0xec
    [<00000000c5d82c9a>] kthread+0x2e8/0x374
    [<00000000f4287308>] ret_from_fork+0x10/0x20
unreferenced object 0xffffff80cc0b0980 (size 64):
  comm "kunit_try_catch", pid 1215, jiffies 4294898266
  hex dump (first 32 bytes):
    18 b7 58 ca 80 ff ff ff 00 09 0b cc 80 ff ff ff  ..X.............
    67 00 00 00 00 00 00 00 01 01 74 88 c0 ff ff ff  g.........t.....
  backtrace (crc 275fd9be):
    [<0000000052a0be73>] kmemleak_alloc+0x34/0x40
    [<0000000043605459>] __kmalloc_cache_noprof+0x26c/0x2f4
    [<000000000ff63fd4>] vcap_rule_add_key+0x2cc/0x528
    [<000000001396a1a2>] test_add_def_fields+0xb0/0x100
    [<000000006e7621f0>] vcap_val_rule+0xa98/0x13e8
    [<00000000fc9868e5>] vcap_api_encode_rule_test+0x678/0x16b0
    [<00000000b3595fc4>] kunit_try_run_case+0x13c/0x3ac
    [<0000000010f5d2bf>] kunit_generic_run_threadfn_adapter+0x80/0xec
    [<00000000c5d82c9a>] kthread+0x2e8/0x374
    [<00000000f4287308>] ret_from_fork+0x10/0x20
......

Cc: stable@vger.kernel.org
Fixes: a3c1e45156ad ("net: microchip: vcap: Fix use-after-free error in kunit test")
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Jens Emil Schulz Østergaard <jensemil.schulzostergaard@microchip.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Link: https://patch.msgid.link/20241014121922.1280583-1-ruanjinjie@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-phy-mdio-bcm-unimac-add-bcm6846-variant'

Linus Walleij says:

====================
net: phy: mdio-bcm-unimac: Add BCM6846 variant

As pointed out by Florian:
https://lore.kernel.org/linux-devicetree/b542b2e8-115c-4234-a464-e73aa6bece5c@broadcom.com/

The BCM6846 has a few extra registers and cannot reuse the
compatible string from other variants of the Unimac
MDIO block: we need to be able to tell them apart.
====================

Link: https://patch.msgid.link/20241012-bcm6846-mdio-v1-0-c703ca83e962@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: phy: mdio-bcm-unimac: Add BCM6846 support

Add Unimac mdio compatible string for the special BCM6846
variant.

This variant has a few extra registers compared to other
versions.

Suggested-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://lore.kernel.org/linux-devicetree/b542b2e8-115c-4234-a464-e73aa6bece5c@broadcom.com/
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://patch.msgid.link/20241012-bcm6846-mdio-v1-2-c703ca83e962@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: net: brcm,unimac-mdio: Add bcm6846-mdio

The MDIO block in the BCM6846 is not identical to any of the
previous versions, but has extended registers not present in
the other variants. For this reason we need to use a new
compatible especially for this SoC.

Suggested-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://lore.kernel.org/linux-devicetree/b542b2e8-115c-4234-a464-e73aa6bece5c@broadcom.com/
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://patch.msgid.link/20241012-bcm6846-mdio-v1-1-c703ca83e962@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

udp: Compute L4 checksum as usual when not segmenting the skb

If:

  1) the user requested USO, but
  2) there is not enough payload for GSO to kick in, and
  3) the egress device doesn't offer checksum offload, then

we want to compute the L4 checksum in software early on.

In the case when we are not taking the GSO path, but it has been requested,
the software checksum fallback in skb_segment doesn't get a chance to
compute the full checksum, if the egress device can't do it. As a result we
end up sending UDP datagrams with only a partial checksum filled in, which
the peer will discard.

Fixes: 10154dbded6d ("udp: Allow GSO transmit from devices with no checksum offload")
Reported-by: Ivan Babrou <ivan@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20241011-uso-swcsum-fixup-v2-1-6e1ddc199af9@cloudflare.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

genetlink: hold RCU in genlmsg_mcast()

While running net selftests with CONFIG_PROVE_RCU_LIST=y I saw
one lockdep splat [1].

genlmsg_mcast() uses for_each_net_rcu(), and must therefore hold RCU.

Instead of letting all callers guard genlmsg_multicast_allns()
with a rcu_read_lock()/rcu_read_unlock() pair, do it in genlmsg_mcast().

This also means the @flags parameter is useless, we need to always use
GFP_ATOMIC.

[1]
[10882.424136] =============================
[10882.424166] WARNING: suspicious RCU usage
[10882.424309] 6.12.0-rc2-virtme #1156 Not tainted
[10882.424400] -----------------------------
[10882.424423] net/netlink/genetlink.c:1940 RCU-list traversed in non-reader section!!
[10882.424469]
other info that might help us debug this:

[10882.424500]
rcu_scheduler_active = 2, debug_locks = 1
[10882.424744] 2 locks held by ip/15677:
[10882.424791] #0: ffffffffb6b491b0 (cb_lock){++++}-{3:3}, at: genl_rcv (net/netlink/genetlink.c:1219)
[10882.426334] #1: ffffffffb6b49248 (genl_mutex){+.+.}-{3:3}, at: genl_rcv_msg (net/netlink/genetlink.c:61 net/netlink/genetlink.c:57 net/netlink/genetlink.c:1209)
[10882.426465]
stack backtrace:
[10882.426805] CPU: 14 UID: 0 PID: 15677 Comm: ip Not tainted 6.12.0-rc2-virtme #1156
[10882.426919] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014
[10882.427046] Call Trace:
[10882.427131]  <TASK>
[10882.427244] dump_stack_lvl (lib/dump_stack.c:123)
[10882.427335] lockdep_rcu_suspicious (kernel/locking/lockdep.c:6822)
[10882.427387] genlmsg_multicast_allns (net/netlink/genetlink.c:1940 (discriminator 7) net/netlink/genetlink.c:1977 (discriminator 7))
[10882.427436] l2tp_tunnel_notify.constprop.0 (net/l2tp/l2tp_netlink.c:119) l2tp_netlink
[10882.427683] l2tp_nl_cmd_tunnel_create (net/l2tp/l2tp_netlink.c:253) l2tp_netlink
[10882.427748] genl_family_rcv_msg_doit (net/netlink/genetlink.c:1115)
[10882.427834] genl_rcv_msg (net/netlink/genetlink.c:1195 net/netlink/genetlink.c:1210)
[10882.427877] ? __pfx_l2tp_nl_cmd_tunnel_create (net/l2tp/l2tp_netlink.c:186) l2tp_netlink
[10882.427927] ? __pfx_genl_rcv_msg (net/netlink/genetlink.c:1201)
[10882.427959] netlink_rcv_skb (net/netlink/af_netlink.c:2551)
[10882.428069] genl_rcv (net/netlink/genetlink.c:1220)
[10882.428095] netlink_unicast (net/netlink/af_netlink.c:1332 net/netlink/af_netlink.c:1357)
[10882.428140] netlink_sendmsg (net/netlink/af_netlink.c:1901)
[10882.428210] ____sys_sendmsg (net/socket.c:729 (discriminator 1) net/socket.c:744 (discriminator 1) net/socket.c:2607 (discriminator 1))

Fixes: 33f72e6f0c67 ("l2tp : multicast notification to the registered listeners")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: James Chapman <jchapman@katalix.com>
Cc: Tom Parkin <tparkin@katalix.com>
Cc: Johannes Berg <johannes.berg@intel.com>
Link: https://patch.msgid.link/20241011171217.3166614-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: mv88e6xxx: Fix the max_vid definition for the MV88E6361

According to the Marvell datasheet the 88E6361 has two VTU pages
(4k VIDs per page) so the max_vid should be 8191, not 4095.

In the current implementation mv88e6xxx_vtu_walk() gives unexpected
results because of this error. I verified that mv88e6xxx_vtu_walk()
works correctly on the MV88E6361 with this patch in place.

Fixes: 12899f299803 ("net: dsa: mv88e6xxx: enable support for 88E6361 switch")
Signed-off-by: Peter Rashleigh <peter@rashleigh.ca>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20241014204342.5852-1-peter@rashleigh.ca
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tcp/dccp: Don't use timer_pending() in reqsk_queue_unlink().

Martin KaFai Lau reported use-after-free [0] in reqsk_timer_handler().

  """
  We are seeing a use-after-free from a bpf prog attached to
  trace_tcp_retransmit_synack. The program passes the req->sk to the
  bpf_sk_storage_get_tracing kernel helper which does check for null
  before using it.
  """

The commit 83fccfc3940c ("inet: fix potential deadlock in
reqsk_queue_unlink()") added timer_pending() in reqsk_queue_unlink() not
to call del_timer_sync() from reqsk_timer_handler(), but it introduced a
small race window.

Before the timer is called, expire_timers() calls detach_timer(timer, true)
to clear timer->entry.pprev and marks it as not pending.

If reqsk_queue_unlink() checks timer_pending() just after expire_timers()
calls detach_timer(), TCP will miss del_timer_sync(); the reqsk timer will
continue running and send multiple SYN+ACKs until it expires.

The reported UAF could happen if req->sk is close()d earlier than the timer
expiration, which is 63s by default.

The scenario would be

  1. inet_csk_complete_hashdance() calls inet_csk_reqsk_queue_drop(),
     but del_timer_sync() is missed

  2. reqsk timer is executed and scheduled again

  3. req->sk is accept()ed and reqsk_put() decrements rsk_refcnt, but
     reqsk timer still has another one, and inet_csk_accept() does not
     clear req->sk for non-TFO sockets

  4. sk is close()d

  5. reqsk timer is executed again, and BPF touches req->sk

Let's not use timer_pending() by passing the caller context to
__inet_csk_reqsk_queue_drop().

Note that reqsk timer is pinned, so the issue does not happen in most
use cases. [1]

[0]
BUG: KFENCE: use-after-free read in bpf_sk_storage_get_tracing+0x2e/0x1b0

Use-after-free read at 0x00000000a891fb3a (in kfence-#1):
bpf_sk_storage_get_tracing+0x2e/0x1b0
bpf_prog_5ea3e95db6da0438_tcp_retransmit_synack+0x1d20/0x1dda
bpf_trace_run2+0x4c/0xc0
tcp_rtx_synack+0xf9/0x100
reqsk_timer_handler+0xda/0x3d0
run_timer_softirq+0x292/0x8a0
irq_exit_rcu+0xf5/0x320
sysvec_apic_timer_interrupt+0x6d/0x80
asm_sysvec_apic_timer_interrupt+0x16/0x20
intel_idle_irq+0x5a/0xa0
cpuidle_enter_state+0x94/0x273
cpu_startup_entry+0x15e/0x260
start_secondary+0x8a/0x90
secondary_startup_64_no_verify+0xfa/0xfb

kfence-#1: 0x00000000a72cc7b6-0x00000000d97616d9, size=2376, cache=TCPv6

allocated by task 0 on cpu 9 at 260507.901592s:
sk_prot_alloc+0x35/0x140
sk_clone_lock+0x1f/0x3f0
inet_csk_clone_lock+0x15/0x160
tcp_create_openreq_child+0x1f/0x410
tcp_v6_syn_recv_sock+0x1da/0x700
tcp_check_req+0x1fb/0x510
tcp_v6_rcv+0x98b/0x1420
ipv6_list_rcv+0x2258/0x26e0
napi_complete_done+0x5b1/0x2990
mlx5e_napi_poll+0x2ae/0x8d0
net_rx_action+0x13e/0x590
irq_exit_rcu+0xf5/0x320
common_interrupt+0x80/0x90
asm_common_interrupt+0x22/0x40
cpuidle_enter_state+0xfb/0x273
cpu_startup_entry+0x15e/0x260
start_secondary+0x8a/0x90
secondary_startup_64_no_verify+0xfa/0xfb

freed by task 0 on cpu 9 at 260507.927527s:
rcu_core_si+0x4ff/0xf10
irq_exit_rcu+0xf5/0x320
sysvec_apic_timer_interrupt+0x6d/0x80
asm_sysvec_apic_timer_interrupt+0x16/0x20
cpuidle_enter_state+0xfb/0x273
cpu_startup_entry+0x15e/0x260
start_secondary+0x8a/0x90
secondary_startup_64_no_verify+0xfa/0xfb

Fixes: 83fccfc3940c ("inet: fix potential deadlock in reqsk_queue_unlink()")
Reported-by: Martin KaFai Lau <martin.lau@kernel.org>
Closes: https://lore.kernel.org/netdev/eb6684d0-ffd9-4bdc-9196-33f690c25824@linux.dev/
Link: https://lore.kernel.org/netdev/b55e2ca0-42f2-4b7c-b445-6ffd87ca74a0@linux.dev/
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20241014223312.4254-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

drm/msm/a6xx+: Insert a fence wait before SMMU table update

The CP_SMMU_TABLE_UPDATE _should_ be waiting for idle, but on some
devices (x1-85, possibly others), it seems to pass that barrier while
there are still things in the event completion FIFO waiting to be
written back to memory.

Work around that by adding a fence wait before context switch.  The
CP_EVENT_WRITE that writes the fence is the last write from a submit,
so seeing this value hit memory is a reliable indication that it is
safe to proceed with the context switch.

v2: Only emit CP_WAIT_TIMESTAMP on a7xx, as it is not supported on a6xx.
    Conversely, I've not been able to reproduce this issue on a6xx, so
    hopefully it is limited to a7xx, or perhaps just certain a7xx
    devices.

Fixes: af66706accdf ("drm/msm/a6xx: Add skeleton A7xx support")
Closes: https://gitlab.freedesktop.org/drm/msm/-/issues/63
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>

net: bcmasp: fix potential memory leak in bcmasp_xmit()

The bcmasp_xmit() returns NETDEV_TX_OK without freeing skb
in case of mapping fails, add dev_kfree_skb() to fix it.

Fixes: 490cb412007d ("net: bcmasp: Add support for ASP2.0 Ethernet controller")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://patch.msgid.link/20241014145901.48940-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

powerpc/powernv: Free name on error in opal_event_init()

In opal_event_init() if request_irq() fails name is not freed, leading
to a memory leak. The code only runs at boot time, there's no way for a
user to trigger it, so there's no security impact.

Fix the leak by freeing name in the error path.

Reported-by: 2639161967 <2639161967@qq.com>
Closes: https://lore.kernel.org/linuxppc-dev/87wmjp3wig.fsf@mail.lhotse
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://patch.msgid.link/20240920093520.67997-1-mpe@ellerman.id.au

drm/msm/dpu: don't always program merge_3d block

Only program the merge_3d block for the video phys encoder when the 3d
blend mode is not NONE

Fixes: 3e79527a33a8 ("drm/msm/dpu: enable merge_3d support on sm8150/sm8250")
Suggested-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/619095/
Link: https://lore.kernel.org/r/20241009-merge3d-fix-v1-1-0d0b6f5c244e@quicinc.com
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>

drm/msm/dpu: Don't always set merge_3d pending flush

Don't set the merge_3d pending flush bits if the mode_3d is
BLEND_3D_NONE.

Always flushing merge_3d can cause timeout issues when there are
multiple commits with concurrent writeback enabled.

This is because the video phys enc waits for the hw_ctl flush register
to be completely cleared [1] in its wait_for_commit_done(), but the WB
encoder always sets the merge_3d pending flush during each commit
regardless of if the merge_3d is actually active.

This means that the hw_ctl flush register will never be 0 when there are
multiple CWB commits and the video phys enc will hit vblank timeout
errors after the first CWB commit.

[1] commit fe9df3f50c39 ("drm/msm/dpu: add real wait_for_commit_done()")

Fixes: 3e79527a33a8 ("drm/msm/dpu: enable merge_3d support on sm8150/sm8250")
Fixes: d7d0e73f7de3 ("drm/msm/dpu: introduce the dpu_encoder_phys_* for writeback")
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/619092/
Link: https://lore.kernel.org/r/20241009-mode3d-fix-v1-1-c0258354fadc@quicinc.com
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>

irqchip/renesas-rzg2l: Fix missing put_device

rzg2l_irqc_common_init() calls of_find_device_by_node(), but the
corresponding put_device() call is missing.  This also gets reported by
make coccicheck.

Make use of the cleanup interfaces from cleanup.h to call into
__free_put_device(), which in turn calls into put_device when leaving
function rzg2l_irqc_common_init() and variable "dev" goes out of scope.

To prevent that the device is put on successful completion, assign NULL to
"dev" to prevent __free_put_device() from calling into put_device() within
the successful path.

"make coccicheck" will still complain about missing put_device() calls,
but those are false positives now.

Fixes: 3fed09559cd8 ("irqchip: Add RZ/G2L IA55 Interrupt Controller driver")
Signed-off-by: Fabrizio Castro <fabrizio.castro.jz@renesas.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20241011172003.1242841-1-fabrizio.castro.jz@renesas.com

irqchip/riscv-intc: Fix SMP=n boot with ACPI

When CONFIG_SMP is disabled, the static array rintc_acpi_data with size
NR_CPUS is not sufficient to hold all RINTC structures passed from the
firmware.

All RINTC structures are required to configure IMSIC/APLIC/PLIC properly
irrespective of SMP in the OS. So, allocate dynamic memory based on the
number of RINTC structures in MADT to fix this issue.

Fixes: f8619b66bdb1 ("irqchip/riscv-intc: Add ACPI support for AIA")
Reported-by: Björn Töpel <bjorn@kernel.org>
Signed-off-by: Sunil V L <sunilvl@ventanamicro.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Alexandre Ghiti <alexghiti@rivosinc.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/all/20241014065739.656959-1-sunilvl@ventanamicro.com
Closes: https://github.com/linux-riscv/linux-riscv/actions/runs/11280997511/job/31375229012

Merge tag 'scmi-fixes-6.12' of https://git./linux/kernel/git/sudeep.holla/linux into arm/fixes

Arm SCMI fixes for v6.12

Couple of fixes to address the issues found and reported on Broadcom
STB platforms following the recent refactor of all the SCMI transports
as standalone drivers.

One of the issue is that the effective timeout value is much less than
the intended value due to the way mailbox messages are queues in the
mailbox framework. Since we block or serialise the shmem access anyway,
there is no point in utilizing mailbox queues. The issue is fixed with
exclusive lock on the channel when sending the message.

The other issues is actually non-issue for upstream, but the workaround
is just changing the link order of the transport drivers which enables
Broadcom STB platforms to run both upstream and custom downstream kernel
without any device tree changes. So pushing this to help them test upstream
seamlessly as it has no practical or theoretical impact for others.

There is also a fix to address possible double freeing of the name string
in scmi_debugfs_common_cleanup() when devm_add_action_or_reset() fails.

* tag 'scmi-fixes-6.12' of https://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux:
  firmware: arm_scmi: Queue in scmi layer for mailbox implementation
  firmware: arm_scmi: Give SMC transport precedence over mailbox
  firmware: arm_scmi: Fix the double free in scmi_debugfs_common_setup()

Link: https://lore.kernel.org/r/20241015185128.1000604-1-sudeep.holla@arm.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'ffa-fixes-6.12' of https://git./linux/kernel/git/sudeep.holla/linux into arm/fixes

Arm FF-A fixes for v6.12

Couple of fixes to avoid string-fortify warnings in export_uuid()
and memcpy() from the recently added functions to support
FFA_MSG_SEND_DIRECT_REQ2 and FFA_MSG_SEND_DIRECT_RESP2.

* tag 'ffa-fixes-6.12' of https://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux:
  firmware: arm_ffa: Avoid string-fortify warning caused by memcpy()
  firmware: arm_ffa: Avoid string-fortify warning in export_uuid()

Link: https://lore.kernel.org/r/20241015185037.1000435-1-sudeep.holla@arm.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'mvebu-fixes-6.12-1' of https://git./linux/kernel/git/gclement/mvebu into arm/fixes

mvebu fixes for 6.12 (part 1)

Fix cp0 mdio pin numbers on SolidRun CN9130 SoM

* tag 'mvebu-fixes-6.12-1' of https://git.kernel.org/pub/scm/linux/kernel/git/gclement/mvebu:
  arm64: dts: marvell: cn9130-sr-som: fix cp0 mdio pin numbers

Link: https://lore.kernel.org/r/87ldyud25o.fsf@BLaptop.bootlin.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'reset-fixes-for-v6.12' of git://git.pengutronix.de/pza/linux into arm/fixes

Reset controller fixes for v6.12

Fix a NULL pointer dereference in reset-starfive-jh71x0 and replace two
accidental commas at line endings with semicolons in reset-npcm.

* tag 'reset-fixes-for-v6.12' of git://git.pengutronix.de/pza/linux:
  reset: starfive: jh71x0: Fix accessing the empty member on JH7110 SoC
  reset: npcm: convert comma to semicolon

Link: https://lore.kernel.org/r/20240930165733.1541936-1-p.zabel@pengutronix.de
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

net: systemport: fix potential memory leak in bcm_sysport_xmit()

The bcm_sysport_xmit() returns NETDEV_TX_OK without freeing skb
in case of dma_map_single() fails, add dev_kfree_skb() to fix it.

Fixes: 80105befdb4b ("net: systemport: add Broadcom SYSTEMPORT Ethernet MAC driver")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Link: https://patch.msgid.link/20241014145115.44977-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge tag 'trace-ringbuffer-v6.12-rc3' of git://git./linux/kernel/git/trace/linux-trace

Pull ring-buffer fixes from Steven Rostedt:

 - Fix ref counter of buffers assigned at boot up

   A tracing instance can be created from the kernel command line. If it
   maps to memory, it is considered permanent and should not be deleted,
   or bad things can happen. If it is not mapped to memory, then the
   user is fine to delete it via rmdir from the instances directory. But
   the ref counts assumed 0 was free to remove and greater than zero was
   not. But this was not the case. When an instance is created, it
   should have the reference of 1, and if it should not be removed, it
   must be greater than 1. The boot up code set normal instances with a
   ref count of 0, which could get removed if something accessed it and
   then released it. And memory mapped instances had a ref count of 1
   which meant it could be deleted, and bad things happen. Keep normal
   instances ref count as 1, and set memory mapped instances ref count
   to 2.

 - Protect sub buffer size (order) updates from other modifications

   When a ring buffer is changing the size of its sub-buffers, no other
   operations should be performed on the ring buffer. That includes
   reading it. But the locking only grabbed the buffer->mutex that keeps
   some operations from touching the ring buffer. It also must hold the
   cpu_buffer->reader_lock as well when updates happen as other paths
   use that to do some operations on the ring buffer.

* tag 'trace-ringbuffer-v6.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  ring-buffer: Fix reader locking when changing the sub buffer order
  ring-buffer: Fix refcount setting of boot mapped buffers

Merge branch 'fix-truncation-bug-in-coerce_reg_to_size_sx-and-extend-selftests'

Dimitar Kanaliev says:

====================
Fix truncation bug in coerce_reg_to_size_sx and extend selftests.

This patch series addresses a truncation bug in the eBPF verifier function
coerce_reg_to_size_sx(). The issue was caused by the incorrect ordering
of assignments between 32-bit and 64-bit min/max values, leading to
improper truncation when updating the register state. This issue has been
reported previously by Zac Ecob[1] , but was not followed up on.

The first patch fixes the assignment order in coerce_reg_to_size_sx()
to ensure correct truncation. The subsequent patches add selftests for
coerce_{reg,subreg}_to_size_sx.

Changelog:
v1 -> v2:
 - Moved selftests inside the conditional check for cpuv4

[1] (https://lore.kernel.org/bpf/h3qKLDEO6m9nhif0eAQX4fVrqdO0D_OPb0y5HfMK9jBePEKK33wQ3K-bqSVnr0hiZdFZtSJOsbNkcEQGpv_yJk61PAAiO8fUkgMRSO-lB50=@protonmail.com/)
====================

Link: https://lore.kernel.org/r/20241014121155.92887-1-dimitar.kanaliev@siteground.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

selftests/bpf: Add test for sign extension in coerce_subreg_to_size_sx()

Add a test for unsigned ranges after signed extension instruction. This
case isn't currently covered by existing tests in verifier_movsx.c.

Acked-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Signed-off-by: Dimitar Kanaliev <dimitar.kanaliev@siteground.com>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20241014121155.92887-4-dimitar.kanaliev@siteground.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

selftests/bpf: Add test for truncation after sign extension in coerce_reg_to_size_sx()

Add test that checks whether unsigned ranges deduced by the verifier for
sign extension instruction is correct. Without previous patch that
fixes truncation in coerce_reg_to_size_sx() this test fails.

Acked-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Signed-off-by: Dimitar Kanaliev <dimitar.kanaliev@siteground.com>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20241014121155.92887-3-dimitar.kanaliev@siteground.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

bpf: Fix truncation bug in coerce_reg_to_size_sx()

coerce_reg_to_size_sx() updates the register state after a sign-extension
operation. However, there's a bug in the assignment order of the unsigned
min/max values, leading to incorrect truncation:

  0: (85) call bpf_get_prandom_u32#7    ; R0_w=scalar()
  1: (57) r0 &= 1                       ; R0_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=1,var_off=(0x0; 0x1))
  2: (07) r0 += 254                     ; R0_w=scalar(smin=umin=smin32=umin32=254,smax=umax=smax32=umax32=255,var_off=(0xfe; 0x1))
  3: (bf) r0 = (s8)r0                   ; R0_w=scalar(smin=smin32=-2,smax=smax32=-1,umin=umin32=0xfffffffe,umax=0xffffffff,var_off=(0xfffffffffffffffe; 0x1))

In the current implementation, the unsigned 32-bit min/max values
(u32_min_value and u32_max_value) are assigned directly from the 64-bit
signed min/max values (s64_min and s64_max):

  reg->umin_value = reg->u32_min_value = s64_min;
  reg->umax_value = reg->u32_max_value = s64_max;

Due to the chain assigmnent, this is equivalent to:

  reg->u32_min_value = s64_min;  // Unintended truncation
  reg->umin_value = reg->u32_min_value;
  reg->u32_max_value = s64_max;  // Unintended truncation
  reg->umax_value = reg->u32_max_value;

Fixes: 1f9a1ea821ff ("bpf: Support new sign-extension load insns")
Reported-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Reported-by: Zac Ecob <zacecob@protonmail.com>
Signed-off-by: Dimitar Kanaliev <dimitar.kanaliev@siteground.com>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Reviewed-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Link: https://lore.kernel.org/r/20241014121155.92887-2-dimitar.kanaliev@siteground.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Merge tag 'bcachefs-2024-10-14' of git://evilpiepirate.org/bcachefs

Pull bcachefs fixes from Kent Overstreet:

 - New metadata version inode_has_child_snapshots

   This fixes bugs with handling of unlinked inodes + snapshots, in
   particular when an inode is reattached after taking a snapshot;
   deleted inodes now get correctly cleaned up across snapshots.

 - Disk accounting rewrite fixes
     - validation fixes for when a device has been removed
     - fix journal replay failing with "journal_reclaim_would_deadlock"

 - Some more small fixes for erasure coding + device removal

 - Assorted small syzbot fixes

* tag 'bcachefs-2024-10-14' of git://evilpiepirate.org/bcachefs: (27 commits)
  bcachefs: Fix sysfs warning in fstests generic/730,731
  bcachefs: Handle race between stripe reuse, invalidate_stripe_to_dev
  bcachefs: Fix kasan splat in new_stripe_alloc_buckets()
  bcachefs: Add missing validation for bch_stripe.csum_granularity_bits
  bcachefs: Fix missing bounds checks in bch2_alloc_read()
  bcachefs: fix uaf in bch2_dio_write_done()
  bcachefs: Improve check_snapshot_exists()
  bcachefs: Fix bkey_nocow_lock()
  bcachefs: Fix accounting replay flags
  bcachefs: Fix invalid shift in member_to_text()
  bcachefs: Fix bch2_have_enough_devs() for BCH_SB_MEMBER_INVALID
  bcachefs: __wait_for_freeing_inode: Switch to wait_bit_queue_entry
  bcachefs: Check if stuck in journal_res_get()
  closures: Add closure_wait_event_timeout()
  bcachefs: Fix state lock involved deadlock
  bcachefs: Fix NULL pointer dereference in bch2_opt_to_text
  bcachefs: Release transaction before wake up
  bcachefs: add check for btree id against max in try read node
  bcachefs: Disk accounting device validation fixes
  bcachefs: bch2_inode_or_descendents_is_open()
  ...

net: ethernet: rtsn: fix potential memory leak in rtsn_start_xmit()

The rtsn_start_xmit() returns NETDEV_TX_OK without freeing skb
in case of skb->len being too long, add dev_kfree_skb_any() to fix it.

Fixes: b0d3969d2b4d ("net: ethernet: rtsn: Add support for Renesas Ethernet-TSN")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241014144250.38802-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: xilinx: axienet: fix potential memory leak in axienet_start_xmit()

The axienet_start_xmit() returns NETDEV_TX_OK without freeing skb
in case of dma_map_single() fails, add dev_kfree_skb_any() to fix it.

Fixes: 71791dc8bdea ("net: axienet: Check for DMA mapping errors")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Reviewed-by: Radhey Shyam Pandey <radhey.shyam.pandey@amd.com>
Link: https://patch.msgid.link/20241014143704.31938-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'mptcp-prevent-mpc-handshake-on-port-based-signal-endpoints'

Matthieu Baerts says:

====================
mptcp: prevent MPC handshake on port-based signal endpoints

MPTCP connection requests toward a listening socket created by the
in-kernel PM for a port based signal endpoint will never be accepted,
they need to be explicitly rejected.

- Patch 1: Explicitly reject such requests. A fix for >= v5.12.

- Patch 2: Cover this case in the MPTCP selftests to avoid regressions.

Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
v1: https://lore.kernel.org/20240908180620.822579-1-xiyou.wangcong@gmail.com

Link: https://lore.kernel.org/a5289a0d-2557-40b8-9575-6f1a0bbf06e4@redhat.com
====================

Link: https://patch.msgid.link/20241014-net-mptcp-mpc-port-endp-v2-0-7faea8e6b6ae@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

selftests: mptcp: join: test for prohibited MPC to port-based endp

Explicitly verify that MPC connection attempts towards a port-based
signal endpoint fail with a reset.

Note that this new test is a bit different from the other ones, not
using 'run_tests'. It is then needed to add the capture capability, and
the picking the right port which have been extracted into three new
helpers. The info about the capture can also be printed from a single
point, which simplifies the exit paths in do_transfer().

The 'Fixes' tag here below is the same as the one from the previous
commit: this patch here is not fixing anything wrong in the selftests,
but it validates the previous fix for an issue introduced by this commit
ID.

Fixes: 1729cf186d8a ("mptcp: create the listening socket for new port")
Cc: stable@vger.kernel.org
Co-developed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20241014-net-mptcp-mpc-port-endp-v2-2-7faea8e6b6ae@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: prevent MPC handshake on port-based signal endpoints

Syzkaller reported a lockdep splat:

  ============================================
  WARNING: possible recursive locking detected
  6.11.0-rc6-syzkaller-00019-g67784a74e258 #0 Not tainted
  --------------------------------------------
  syz-executor364/5113 is trying to acquire lock:
  ffff8880449f1958 (k-slock-AF_INET){+.-.}-{2:2}, at: spin_lock include/linux/spinlock.h:351 [inline]
  ffff8880449f1958 (k-slock-AF_INET){+.-.}-{2:2}, at: sk_clone_lock+0x2cd/0xf40 net/core/sock.c:2328

  but task is already holding lock:
  ffff88803fe3cb58 (k-slock-AF_INET){+.-.}-{2:2}, at: spin_lock include/linux/spinlock.h:351 [inline]
  ffff88803fe3cb58 (k-slock-AF_INET){+.-.}-{2:2}, at: sk_clone_lock+0x2cd/0xf40 net/core/sock.c:2328

  other info that might help us debug this:
   Possible unsafe locking scenario:

         CPU0
         ----
    lock(k-slock-AF_INET);
    lock(k-slock-AF_INET);

   *** DEADLOCK ***

   May be due to missing lock nesting notation

  7 locks held by syz-executor364/5113:
   #0: ffff8880449f0e18 (sk_lock-AF_INET){+.+.}-{0:0}, at: lock_sock include/net/sock.h:1607 [inline]
   #0: ffff8880449f0e18 (sk_lock-AF_INET){+.+.}-{0:0}, at: mptcp_sendmsg+0x153/0x1b10 net/mptcp/protocol.c:1806
   #1: ffff88803fe39ad8 (k-sk_lock-AF_INET){+.+.}-{0:0}, at: lock_sock include/net/sock.h:1607 [inline]
   #1: ffff88803fe39ad8 (k-sk_lock-AF_INET){+.+.}-{0:0}, at: mptcp_sendmsg_fastopen+0x11f/0x530 net/mptcp/protocol.c:1727
   #2: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: rcu_lock_acquire include/linux/rcupdate.h:326 [inline]
   #2: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: rcu_read_lock include/linux/rcupdate.h:838 [inline]
   #2: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: __ip_queue_xmit+0x5f/0x1b80 net/ipv4/ip_output.c:470
   #3: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: rcu_lock_acquire include/linux/rcupdate.h:326 [inline]
   #3: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: rcu_read_lock include/linux/rcupdate.h:838 [inline]
   #3: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: ip_finish_output2+0x45f/0x1390 net/ipv4/ip_output.c:228
   #4: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: local_lock_acquire include/linux/local_lock_internal.h:29 [inline]
   #4: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: process_backlog+0x33b/0x15b0 net/core/dev.c:6104
   #5: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: rcu_lock_acquire include/linux/rcupdate.h:326 [inline]
   #5: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: rcu_read_lock include/linux/rcupdate.h:838 [inline]
   #5: ffffffff8e938320 (rcu_read_lock){....}-{1:2}, at: ip_local_deliver_finish+0x230/0x5f0 net/ipv4/ip_input.c:232
   #6: ffff88803fe3cb58 (k-slock-AF_INET){+.-.}-{2:2}, at: spin_lock include/linux/spinlock.h:351 [inline]
   #6: ffff88803fe3cb58 (k-slock-AF_INET){+.-.}-{2:2}, at: sk_clone_lock+0x2cd/0xf40 net/core/sock.c:2328

  stack backtrace:
  CPU: 0 UID: 0 PID: 5113 Comm: syz-executor364 Not tainted 6.11.0-rc6-syzkaller-00019-g67784a74e258 #0
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
  Call Trace:
   <IRQ>
   __dump_stack lib/dump_stack.c:93 [inline]
   dump_stack_lvl+0x241/0x360 lib/dump_stack.c:119
   check_deadlock kernel/locking/lockdep.c:3061 [inline]
   validate_chain+0x15d3/0x5900 kernel/locking/lockdep.c:3855
   __lock_acquire+0x137a/0x2040 kernel/locking/lockdep.c:5142
   lock_acquire+0x1ed/0x550 kernel/locking/lockdep.c:5759
   __raw_spin_lock include/linux/spinlock_api_smp.h:133 [inline]
   _raw_spin_lock+0x2e/0x40 kernel/locking/spinlock.c:154
   spin_lock include/linux/spinlock.h:351 [inline]
   sk_clone_lock+0x2cd/0xf40 net/core/sock.c:2328
   mptcp_sk_clone_init+0x32/0x13c0 net/mptcp/protocol.c:3279
   subflow_syn_recv_sock+0x931/0x1920 net/mptcp/subflow.c:874
   tcp_check_req+0xfe4/0x1a20 net/ipv4/tcp_minisocks.c:853
   tcp_v4_rcv+0x1c3e/0x37f0 net/ipv4/tcp_ipv4.c:2267
   ip_protocol_deliver_rcu+0x22e/0x440 net/ipv4/ip_input.c:205
   ip_local_deliver_finish+0x341/0x5f0 net/ipv4/ip_input.c:233
   NF_HOOK+0x3a4/0x450 include/linux/netfilter.h:314
   NF_HOOK+0x3a4/0x450 include/linux/netfilter.h:314
   __netif_receive_skb_one_core net/core/dev.c:5661 [inline]
   __netif_receive_skb+0x2bf/0x650 net/core/dev.c:5775
   process_backlog+0x662/0x15b0 net/core/dev.c:6108
   __napi_poll+0xcb/0x490 net/core/dev.c:6772
   napi_poll net/core/dev.c:6841 [inline]
   net_rx_action+0x89b/0x1240 net/core/dev.c:6963
   handle_softirqs+0x2c4/0x970 kernel/softirq.c:554
   do_softirq+0x11b/0x1e0 kernel/softirq.c:455
   </IRQ>
   <TASK>
   __local_bh_enable_ip+0x1bb/0x200 kernel/softirq.c:382
   local_bh_enable include/linux/bottom_half.h:33 [inline]
   rcu_read_unlock_bh include/linux/rcupdate.h:908 [inline]
   __dev_queue_xmit+0x1763/0x3e90 net/core/dev.c:4450
   dev_queue_xmit include/linux/netdevice.h:3105 [inline]
   neigh_hh_output include/net/neighbour.h:526 [inline]
   neigh_output include/net/neighbour.h:540 [inline]
   ip_finish_output2+0xd41/0x1390 net/ipv4/ip_output.c:235
   ip_local_out net/ipv4/ip_output.c:129 [inline]
   __ip_queue_xmit+0x118c/0x1b80 net/ipv4/ip_output.c:535
   __tcp_transmit_skb+0x2544/0x3b30 net/ipv4/tcp_output.c:1466
   tcp_rcv_synsent_state_process net/ipv4/tcp_input.c:6542 [inline]
   tcp_rcv_state_process+0x2c32/0x4570 net/ipv4/tcp_input.c:6729
   tcp_v4_do_rcv+0x77d/0xc70 net/ipv4/tcp_ipv4.c:1934
   sk_backlog_rcv include/net/sock.h:1111 [inline]
   __release_sock+0x214/0x350 net/core/sock.c:3004
   release_sock+0x61/0x1f0 net/core/sock.c:3558
   mptcp_sendmsg_fastopen+0x1ad/0x530 net/mptcp/protocol.c:1733
   mptcp_sendmsg+0x1884/0x1b10 net/mptcp/protocol.c:1812
   sock_sendmsg_nosec net/socket.c:730 [inline]
   __sock_sendmsg+0x1a6/0x270 net/socket.c:745
   ____sys_sendmsg+0x525/0x7d0 net/socket.c:2597
   ___sys_sendmsg net/socket.c:2651 [inline]
   __sys_sendmmsg+0x3b2/0x740 net/socket.c:2737
   __do_sys_sendmmsg net/socket.c:2766 [inline]
   __se_sys_sendmmsg net/socket.c:2763 [inline]
   __x64_sys_sendmmsg+0xa0/0xb0 net/socket.c:2763
   do_syscall_x64 arch/x86/entry/common.c:52 [inline]
   do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83
   entry_SYSCALL_64_after_hwframe+0x77/0x7f
  RIP: 0033:0x7f04fb13a6b9
  Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 01 1a 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
  RSP: 002b:00007ffd651f42d8 EFLAGS: 00000246 ORIG_RAX: 0000000000000133
  RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007f04fb13a6b9
  RDX: 0000000000000001 RSI: 0000000020000d00 RDI: 0000000000000004
  RBP: 00007ffd651f4310 R08: 0000000000000001 R09: 0000000000000001
  R10: 0000000020000080 R11: 0000000000000246 R12: 00000000000f4240
  R13: 00007f04fb187449 R14: 00007ffd651f42f4 R15: 00007ffd651f4300
   </TASK>

As noted by Cong Wang, the splat is false positive, but the code
path leading to the report is an unexpected one: a client is
attempting an MPC handshake towards the in-kernel listener created
by the in-kernel PM for a port based signal endpoint.

Such connection will be never accepted; many of them can make the
listener queue full and preventing the creation of MPJ subflow via
such listener - its intended role.

Explicitly detect this scenario at initial-syn time and drop the
incoming MPC request.

Fixes: 1729cf186d8a ("mptcp: create the listening socket for new port")
Cc: stable@vger.kernel.org
Reported-by: syzbot+f4aacdfef2c6a6529c3e@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=f4aacdfef2c6a6529c3e
Cc: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20241014-net-mptcp-mpc-port-endp-v2-1-7faea8e6b6ae@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/smc: Fix searching in list of known pnetids in smc_pnet_add_pnetid

pnetid of pi (not newly allocated pe) should be compared

Fixes: e888a2e8337c ("net/smc: introduce list of pnetids for Ethernet devices")
Reviewed-by: D. Wythe <alibuda@linux.alibaba.com>
Reviewed-by: Wen Gu <guwen@linux.alibaba.com>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Link: https://patch.msgid.link/20241014115321.33234-1-lirongqing@baidu.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: macb: Avoid 20s boot delay by skipping MDIO bus registration for fixed-link PHY

A boot delay was introduced by commit 79540d133ed6 ("net: macb: Fix
handling of fixed-link node"). This delay was caused by the call to
`mdiobus_register()` in cases where a fixed-link PHY was present. The
MDIO bus registration triggered unnecessary PHY address scans, leading
to a 20-second delay due to attempts to detect Clause 45 (C45)
compatible PHYs, despite no MDIO bus being attached.

The commit 79540d133ed6 ("net: macb: Fix handling of fixed-link node")
was originally introduced to fix a regression caused by commit
7897b071ac3b4 ("net: macb: convert to phylink"), which caused the driver
to misinterpret fixed-link nodes as PHY nodes. This resulted in warnings
like:
mdio_bus f0028000.ethernet-ffffffff: fixed-link has invalid PHY address
mdio_bus f0028000.ethernet-ffffffff: scan phy fixed-link at address 0
...
mdio_bus f0028000.ethernet-ffffffff: scan phy fixed-link at address 31

This patch reworks the logic to avoid registering and allocation of the
MDIO bus when:
  - The device tree contains a fixed-link node.
  - There is no "mdio" child node in the device tree.

If a child node named "mdio" exists, the MDIO bus will be registered to
support PHYs  attached to the MACB's MDIO bus. Otherwise, with only a
fixed-link, the MDIO bus is skipped.

Tested on a sama5d35 based system with a ksz8863 switch attached to
macb0.

Fixes: 79540d133ed6 ("net: macb: Fix handling of fixed-link node")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Cc: stable@vger.kernel.org
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20241013052916.3115142-1-o.rempel@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: aeroflex: fix potential memory leak in greth_start_xmit_gbit()

The greth_start_xmit_gbit() returns NETDEV_TX_OK without freeing skb
in case of skb->len being too long, add dev_kfree_skb() to fix it.

Fixes: d4c41139df6e ("net: Add Aeroflex Gaisler 10/100/1G Ethernet MAC driver")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Reviewed-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Link: https://patch.msgid.link/20241012110434.49265-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdevsim: use cond_resched() in nsim_dev_trap_report_work()

I am still seeing many syzbot reports hinting that syzbot
might fool nsim_dev_trap_report_work() with hundreds of ports [1]

Lets use cond_resched(), and system_unbound_wq
instead of implicit system_wq.

[1]
INFO: task syz-executor:20633 blocked for more than 143 seconds.
      Not tainted 6.12.0-rc2-syzkaller-00205-g1d227fcc7222 #0
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
task:syz-executor    state:D stack:25856 pid:20633 tgid:20633 ppid:1      flags:0x00004006
...
NMI backtrace for cpu 1
CPU: 1 UID: 0 PID: 16760 Comm: kworker/1:0 Not tainted 6.12.0-rc2-syzkaller-00205-g1d227fcc7222 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/13/2024
Workqueue: events nsim_dev_trap_report_work
 RIP: 0010:__sanitizer_cov_trace_pc+0x0/0x70 kernel/kcov.c:210
Code: 89 fb e8 23 00 00 00 48 8b 3d 04 fb 9c 0c 48 89 de 5b e9 c3 c7 5d 00 0f 1f 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 <f3> 0f 1e fa 48 8b 04 24 65 48 8b 0c 25 c0 d7 03 00 65 8b 15 60 f0
RSP: 0018:ffffc90000a187e8 EFLAGS: 00000246
RAX: 0000000000000100 RBX: ffffc90000a188e0 RCX: ffff888027d3bc00
RDX: ffff888027d3bc00 RSI: 0000000000000000 RDI: 0000000000000000
RBP: ffff88804a2e6000 R08: ffffffff8a4bc495 R09: ffffffff89da3577
R10: 0000000000000004 R11: ffffffff8a4bc2b0 R12: dffffc0000000000
R13: ffff88806573b503 R14: dffffc0000000000 R15: ffff8880663cca00
FS:  0000000000000000(0000) GS:ffff8880b8700000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007fc90a747f98 CR3: 000000000e734000 CR4: 00000000003526f0
DR0: 0000000000000000 DR1: 000000000000002b DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Call Trace:
 <NMI>
 </NMI>
 <TASK>
  __local_bh_enable_ip+0x1bb/0x200 kernel/softirq.c:382
  spin_unlock_bh include/linux/spinlock.h:396 [inline]
  nsim_dev_trap_report drivers/net/netdevsim/dev.c:820 [inline]
  nsim_dev_trap_report_work+0x75d/0xaa0 drivers/net/netdevsim/dev.c:850
  process_one_work kernel/workqueue.c:3229 [inline]
  process_scheduled_works+0xa63/0x1850 kernel/workqueue.c:3310
  worker_thread+0x870/0xd30 kernel/workqueue.c:3391
  kthread+0x2f0/0x390 kernel/kthread.c:389
  ret_from_fork+0x4b/0x80 arch/x86/kernel/process.c:147
  ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
 </TASK>

Fixes: ba5e1272142d ("netdevsim: avoid potential loop in nsim_dev_trap_report_work()")
Reported-by: syzbot+d383dc9579a76f56c251@syzkaller.appspotmail.com
Reported-by: syzbot+c596faae21a68bf7afd0@syzkaller.appspotmail.com
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20241012094230.3893510-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

macsec: don't increment counters for an unrelated SA

On RX, we shouldn't be incrementing the stats for an arbitrary SA in
case the actual SA hasn't been set up. Those counters are intended to
track packets for their respective AN when the SA isn't currently
configured. Due to the way MACsec is implemented, we don't keep
counters unless the SA is configured, so we can't track those packets,
and those counters will remain at 0.

The RXSC's stats keeps track of those packets without telling us which
AN they belonged to. We could add counters for non-existent SAs, and
then find a way to integrate them in the dump to userspace, but I
don't think it's worth the effort.

Fixes: 91ec9bd57f35 ("macsec: Fix traffic counters/statistics")
Reported-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Link: https://patch.msgid.link/f5ac92aaa5b89343232615f4c03f9f95042c6aa0.1728657709.git.sd@queasysnail.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

gpu: host1x: Set up device DMA parameters

In order to store device DMA parameters, the DMA framework depends on
the device's dma_parms field to point at a valid memory location. Add
backing storage for this in struct host1x_memory_context and point to
it.

Reported-by: Jonathan Hunter <jonathanh@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240916133320.368620-1-thierry.reding@gmail.com
(cherry picked from commit b4ad4ef374d66cc8df3188bb1ddb65bce5fc9e50)
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/amdgpu/swsmu: Only force workload setup on init

Needed to set the workload type at init time so that
we can apply the navi3x margin optimization.

Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3618
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/3131
Fixes: c50fe289ed72 ("drm/amdgpu/swsmu: always force a state reprogram on init")
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 580ad7cbd4b7be8d2cb5ab5c1fca6bb76045eb0e)
Cc: stable@vger.kernel.org

drm/radeon: Fix encoder->possible_clones

Include the encoder itself in its possible_clones bitmask.
In the past nothing validated that drivers were populating
possible_clones correctly, but that changed in commit
74d2aacbe840 ("drm: Validate encoder->possible_clones").
Looks like radeon never got the memo and is still not
following the rules 100% correctly.

This results in some warnings during driver initialization:
Bogus possible_clones: [ENCODER:46:TV-46] possible_clones=0x4 (full encoder mask=0x7)
WARNING: CPU: 0 PID: 170 at drivers/gpu/drm/drm_mode_config.c:615 drm_mode_config_validate+0x113/0x39c
...

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Fixes: 74d2aacbe840 ("drm: Validate encoder->possible_clones")
Reported-by: Erhard Furtner <erhard_f@mailbox.org>
Closes: https://lore.kernel.org/dri-devel/20241009000321.418e4294@yea/
Tested-by: Erhard Furtner <erhard_f@mailbox.org>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 3b6e7d40649c0d75572039aff9d0911864c689db)
Cc: stable@vger.kernel.org

drm/amdgpu/smu13: always apply the powersave optimization

It can avoid margin issues in some very demanding applications.

Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3618
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/3131
Fixes: c50fe289ed72 ("drm/amdgpu/swsmu: always force a state reprogram on init")
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 62f38b4ccaa6aa063ca781d80b10aacd39dc5c76)
Cc: stable@vger.kernel.org

drm/amdkfd: Accounting pdd vram_usage for svm

Process device data pdd->vram_usage is read by rocm-smi via sysfs, this
is currently missing the svm_bo usage accounting, so "rocm-smi
--showpids" per process VRAM usage report is incorrect.

Add pdd->vram_usage accounting when svm_bo allocation and release,
change to atomic64_t type because it is updated outside process mutex
now.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 98c0b0efcc11f2a5ddf3ce33af1e48eedf808b04)

drm/amd/amdgpu: Fix double unlock in amdgpu_mes_add_ring

This patch addresses a double unlock issue in the amdgpu_mes_add_ring
function. The mutex was being unlocked twice under certain error
conditions, which could lead to undefined behavior.

The fix ensures that the mutex is unlocked only once before jumping to
the clean_up_memory label. The unlock operation is moved to just before
the goto statement within the conditional block that checks the return
value of amdgpu_ring_init. This prevents the second unlock attempt after
the clean_up_memory label, which is no longer necessary as the mutex is
already unlocked by this point in the code flow.

This change resolves the potential double unlock and maintains the
correct mutex handling throughout the function.

Fixes below:
Commit d0c423b64765 ("drm/amdgpu/mes: use ring for kernel queue
submission"), leads to the following Smatch static checker warning:

drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c:1240 amdgpu_mes_add_ring()
warn: double unlock '&adev->mes.mutex_hidden' (orig line 1213)

drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
    1143 int amdgpu_mes_add_ring(struct amdgpu_device *adev, int gang_id,
    1144                         int queue_type, int idx,
    1145                         struct amdgpu_mes_ctx_data *ctx_data,
    1146                         struct amdgpu_ring **out)
    1147 {
    1148         struct amdgpu_ring *ring;
    1149         struct amdgpu_mes_gang *gang;
    1150         struct amdgpu_mes_queue_properties qprops = {0};
    1151         int r, queue_id, pasid;
    1152
    1153         /*
    1154          * Avoid taking any other locks under MES lock to avoid circular
    1155          * lock dependencies.
    1156          */
    1157         amdgpu_mes_lock(&adev->mes);
    1158         gang = idr_find(&adev->mes.gang_id_idr, gang_id);
    1159         if (!gang) {
    1160                 DRM_ERROR("gang id %d doesn't exist\n", gang_id);
    1161                 amdgpu_mes_unlock(&adev->mes);
    1162                 return -EINVAL;
    1163         }
    1164         pasid = gang->process->pasid;
    1165
    1166         ring = kzalloc(sizeof(struct amdgpu_ring), GFP_KERNEL);
    1167         if (!ring) {
    1168                 amdgpu_mes_unlock(&adev->mes);
    1169                 return -ENOMEM;
    1170         }
    1171
    1172         ring->ring_obj = NULL;
    1173         ring->use_doorbell = true;
    1174         ring->is_mes_queue = true;
    1175         ring->mes_ctx = ctx_data;
    1176         ring->idx = idx;
    1177         ring->no_scheduler = true;
    1178
    1179         if (queue_type == AMDGPU_RING_TYPE_COMPUTE) {
    1180                 int offset = offsetof(struct amdgpu_mes_ctx_meta_data,
    1181                                       compute[ring->idx].mec_hpd);
    1182                 ring->eop_gpu_addr =
    1183                         amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset);
    1184         }
    1185
    1186         switch (queue_type) {
    1187         case AMDGPU_RING_TYPE_GFX:
    1188                 ring->funcs = adev->gfx.gfx_ring[0].funcs;
    1189                 ring->me = adev->gfx.gfx_ring[0].me;
    1190                 ring->pipe = adev->gfx.gfx_ring[0].pipe;
    1191                 break;
    1192         case AMDGPU_RING_TYPE_COMPUTE:
    1193                 ring->funcs = adev->gfx.compute_ring[0].funcs;
    1194                 ring->me = adev->gfx.compute_ring[0].me;
    1195                 ring->pipe = adev->gfx.compute_ring[0].pipe;
    1196                 break;
    1197         case AMDGPU_RING_TYPE_SDMA:
    1198                 ring->funcs = adev->sdma.instance[0].ring.funcs;
    1199                 break;
    1200         default:
    1201                 BUG();
    1202         }
    1203
    1204         r = amdgpu_ring_init(adev, ring, 1024, NULL, 0,
    1205                              AMDGPU_RING_PRIO_DEFAULT, NULL);
    1206         if (r)
    1207                 goto clean_up_memory;
    1208
    1209         amdgpu_mes_ring_to_queue_props(adev, ring, &qprops);
    1210
    1211         dma_fence_wait(gang->process->vm->last_update, false);
    1212         dma_fence_wait(ctx_data->meta_data_va->last_pt_update, false);
    1213         amdgpu_mes_unlock(&adev->mes);
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

    1214
    1215         r = amdgpu_mes_add_hw_queue(adev, gang_id, &qprops, &queue_id);
    1216         if (r)
    1217                 goto clean_up_ring;
                         ^^^^^^^^^^^^^^^^^^

    1218
    1219         ring->hw_queue_id = queue_id;
    1220         ring->doorbell_index = qprops.doorbell_off;
    1221
    1222         if (queue_type == AMDGPU_RING_TYPE_GFX)
    1223                 sprintf(ring->name, "gfx_%d.%d.%d", pasid, gang_id, queue_id);
    1224         else if (queue_type == AMDGPU_RING_TYPE_COMPUTE)
    1225                 sprintf(ring->name, "compute_%d.%d.%d", pasid, gang_id,
    1226                         queue_id);
    1227         else if (queue_type == AMDGPU_RING_TYPE_SDMA)
    1228                 sprintf(ring->name, "sdma_%d.%d.%d", pasid, gang_id,
    1229                         queue_id);
    1230         else
    1231                 BUG();
    1232
    1233         *out = ring;
    1234         return 0;
    1235
    1236 clean_up_ring:
    1237         amdgpu_ring_fini(ring);
    1238 clean_up_memory:
    1239         kfree(ring);
--> 1240         amdgpu_mes_unlock(&adev->mes);
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

    1241         return r;
    1242 }

Fixes: d0c423b64765 ("drm/amdgpu/mes: use ring for kernel queue submission")
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Hawking Zhang <Hawking.Zhang@amd.com>
Suggested-by: Jack Xiao <Jack.Xiao@amd.com>
Reported by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Jack Xiao <Jack.Xiao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit bfaf1883605fd0c0dbabacd67ed49708470d5ea4)

drm/amdgpu/mes: fix issue of writing to the same log buffer from 2 MES pipes

With Unified MES enabled in gfx12, need separate event log buffer for the
2 MES pipes to avoid data overwrite.

Signed-off-by: Michael Chen <michael.chen@amd.com>
Reviewed-by: Jack Xiao <Jack.Xiao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 144df260f3daab42c4611021f929b3342de516e5)
Cc: stable@vger.kernel.org # 6.11.x

drm/amdgpu: prevent BO_HANDLES error from being overwritten

Before this patch, if multiple BO_HANDLES chunks were submitted,
the error -EINVAL would be correctly set but could be overwritten
by the return value from amdgpu_cs_p1_bo_handles(). This patch
ensures that if there are multiple BO_HANDLES, we stop.

Fixes: fec5f8e8c6bc ("drm/amdgpu: disallow multiple BO_HANDLES chunks in one submit")
Signed-off-by: Mohammed Anees <pvmohammedanees2003@gmail.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 40f2cd98828f454bdc5006ad3d94330a5ea164b7)
Cc: stable@vger.kernel.org

drm/amdgpu: enable enforce_isolation sysfs node on VFs

It should be enabled on both bare metal and VFs.

Fixes: e189be9b2e38 ("drm/amdgpu: Add enforce_isolation sysfs attribute")
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Cc: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
(cherry picked from commit dc8847b054fd6679866ed4ee861e069e54c10799)

nvme-multipath: defer partition scanning

We need to suppress the partition scan from occuring within the
controller's scan_work context. If a path error occurs here, the IO will
wait until a path becomes available or all paths are torn down, but that
action also occurs within scan_work, so it would deadlock. Defer the
partion scan to a different context that does not block scan_work.

Reported-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>

ring-buffer: Fix reader locking when changing the sub buffer order

The function ring_buffer_subbuf_order_set() updates each
ring_buffer_per_cpu and installs new sub buffers that match the requested
page order. This operation may be invoked concurrently with readers that
rely on some of the modified data, such as the head bit (RB_PAGE_HEAD), or
the ring_buffer_per_cpu.pages and reader_page pointers. However, no
exclusive access is acquired by ring_buffer_subbuf_order_set(). Modifying
the mentioned data while a reader also operates on them can then result in
incorrect memory access and various crashes.

Fix the problem by taking the reader_lock when updating a specific
ring_buffer_per_cpu in ring_buffer_subbuf_order_set().

Link: https://lore.kernel.org/linux-trace-kernel/20240715145141.5528-1-petr.pavlu@suse.com/
Link: https://lore.kernel.org/linux-trace-kernel/20241010195849.2f77cc3f@gandalf.local.home/
Link: https://lore.kernel.org/linux-trace-kernel/20241011112850.17212b25@gandalf.local.home/
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://lore.kernel.org/20241015112440.26987-1-petr.pavlu@suse.com
Fixes: 8e7b58c27b3c ("ring-buffer: Just update the subbuffers when changing their allocation order")
Signed-off-by: Petr Pavlu <petr.pavlu@suse.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>

io_uring/sqpoll: close race on waiting for sqring entries

When an application uses SQPOLL, it must wait for the SQPOLL thread to
consume SQE entries, if it fails to get an sqe when calling
io_uring_get_sqe(). It can do so by calling io_uring_enter(2) with the
flag value of IORING_ENTER_SQ_WAIT. In liburing, this is generally done
with io_uring_sqring_wait(). There's a natural expectation that once
this call returns, a new SQE entry can be retrieved, filled out, and
submitted. However, the kernel uses the cached sq head to determine if
the SQRING is full or not. If the SQPOLL thread is currently in the
process of submitting SQE entries, it may have updated the cached sq
head, but not yet committed it to the SQ ring. Hence the kernel may find
that there are SQE entries ready to be consumed, and return successfully
to the application. If the SQPOLL thread hasn't yet committed the SQ
ring entries by the time the application returns to userspace and
attempts to get a new SQE, it will fail getting a new SQE.

Fix this by having io_sqring_full() always use the user visible SQ ring
head entry, rather than the internally cached one.

Cc: stable@vger.kernel.org # 5.10+
Link: https://github.com/axboe/liburing/discussions/1267
Reported-by: Benedek Thaler <thaler@thaler.hu>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

gpu: host1x: Fix boot regression for Tegra

Commit 4c27ac45e622 ("gpu: host1x: Request syncpoint IRQs only during
probe") caused a boot regression for the Tegra186 device. Following this
update the function host1x_intr_init() now calls
host1x_hw_intr_disable_all_syncpt_intrs() during probe. However,
host1x_intr_init() is called before runtime power-management is enabled
for Host1x and the function host1x_hw_intr_disable_all_syncpt_intrs() is
accessing hardware registers. So if the Host1x hardware is not enabled
prior to probing then the device will now hang on attempting to access
the registers. So far this is only observed on Tegra186, but potentially
could be seen on other devices.

Fix this by moving the call to the function host1x_intr_init() in probe
to after enabling the runtime power-management in the probe and update
the failure path in probe as necessary.

Fixes: 4c27ac45e622 ("gpu: host1x: Request syncpoint IRQs only during probe")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240925160504.60221-1-jonathanh@nvidia.com
(cherry picked from commit dc56f8428e5f34418f3243a60cec13166efe4fdb)
Signed-off-by: Thierry Reding <treding@nvidia.com>

firmware: arm_ffa: Avoid string-fortify warning caused by memcpy()

Copying from a 144 byte structure arm_smccc_1_2_regs at an offset of 32
into an 112 byte struct ffa_send_direct_data2 causes a compile-time warning:

 | In file included from drivers/firmware/arm_ffa/driver.c:25:
 | In function 'fortify_memcpy_chk',
 |    inlined from 'ffa_msg_send_direct_req2' at drivers/firmware/arm_ffa/driver.c:504:3:
 | include/linux/fortify-string.h:580:4: warning: call to '__read_overflow2_field'
 |  declared with 'warning' attribute: detected read beyond size of field
 |  (2nd parameter); maybe use struct_group()? [-Wattribute-warning]
 |                         __read_overflow2_field(q_size_field, size);

Fix it by not passing a plain buffer to memcpy() to avoid the overflow
warning.

Fixes: aaef3bc98129 ("firmware: arm_ffa: Add support for FFA_MSG_SEND_DIRECT_{REQ,RESP}2")
Signed-off-by: Gavin Shan <gshan@redhat.com>
Message-Id: <20241014004724.991353-1-gshan@redhat.com>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>

x86/apic: Always explicitly disarm TSC-deadline timer

New processors have become pickier about the local APIC timer state
before entering low power modes. These low power modes are used (for
example) when you close your laptop lid and suspend. If you put your
laptop in a bag and it is not in this low power mode, it is likely
to get quite toasty while it quickly sucks the battery dry.

The problem boils down to some CPUs' inability to power down until the
CPU recognizes that the local APIC timer is shut down. The current
kernel code works in one-shot and periodic modes but does not work for
deadline mode. Deadline mode has been the supported and preferred mode
on Intel CPUs for over a decade and uses an MSR to drive the timer
instead of an APIC register.

Disable the TSC Deadline timer in lapic_timer_shutdown() by writing to
MSR_IA32_TSC_DEADLINE when in TSC-deadline mode. Also avoid writing
to the initial-count register (APIC_TMICT) which is ignored in
TSC-deadline mode.

Note: The APIC_LVTT|=APIC_LVT_MASKED operation should theoretically be
enough to tell the hardware that the timer will not fire in any of the
timer modes. But mitigating AMD erratum 411[1] also requires clearing
out APIC_TMICT. Solely setting APIC_LVT_MASKED is also ineffective in
practice on Intel Lunar Lake systems, which is the motivation for this
change.

1. 411 Processor May Exit Message-Triggered C1E State Without an Interrupt if Local APIC Timer Reaches Zero - https://www.amd.com/content/dam/amd/en/documents/archived-tech-docs/revision-guides/41322_10h_Rev_Gd.pdf

Fixes: 279f1461432c ("x86: apic: Use tsc deadline for oneshot when available")
Suggested-by: Dave Hansen <dave.hansen@intel.com>
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Tested-by: Todd Brandt <todd.e.brandt@intel.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/all/20241015061522.25288-1-rui.zhang%40intel.com

octeontx2-af: Fix potential integer overflows on integer shifts

The left shift int 32 bit integer constants 1 is evaluated using 32 bit
arithmetic and then assigned to a 64 bit unsigned integer. In the case
where the shift is 32 or more this can lead to an overflow. Avoid this
by shifting using the BIT_ULL macro instead.

Fixes: 019aba04f08c ("octeontx2-af: Modify SMQ flush sequence to drop packets")
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: https://patch.msgid.link/20241010154519.768785-1-colin.i.king@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

kasan: Disable Software Tag-Based KASAN with GCC

Syzbot reports a KASAN failure early during boot on arm64 when building
with GCC 12.2.0 and using the Software Tag-Based KASAN mode:

  | BUG: KASAN: invalid-access in smp_build_mpidr_hash arch/arm64/kernel/setup.c:133 [inline]
  | BUG: KASAN: invalid-access in setup_arch+0x984/0xd60 arch/arm64/kernel/setup.c:356
  | Write of size 4 at addr 03ff800086867e00 by task swapper/0
  | Pointer tag: [03], memory tag: [fe]

Initial triage indicates that the report is a false positive and a
thorough investigation of the crash by Mark Rutland revealed the root
cause to be a bug in GCC:

  > When GCC is passed `-fsanitize=hwaddress` or
  > `-fsanitize=kernel-hwaddress` it ignores
  > `__attribute__((no_sanitize_address))`, and instruments functions
  > we require are not instrumented.
  >
  > [...]
  >
  > All versions [of GCC] I tried were broken, from 11.3.0 to 14.2.0
  > inclusive.
  >
  > I think we have to disable KASAN_SW_TAGS with GCC until this is
  > fixed

Disable Software Tag-Based KASAN when building with GCC by making
CC_HAS_KASAN_SW_TAGS depend on !CC_IS_GCC.

Cc: Andrey Konovalov <andreyknvl@gmail.com>
Suggested-by: Mark Rutland <mark.rutland@arm.com>
Reported-by: syzbot+908886656a02769af987@syzkaller.appspotmail.com
Link: https://lore.kernel.org/r/000000000000f362e80620e27859@google.com
Link: https://lore.kernel.org/r/ZvFGwKfoC4yVjN_X@J2N7QTR9R3
Link: https://bugzilla.kernel.org/show_bug.cgi?id=218854
Reviewed-by: Andrey Konovalov <andreyknvl@gmail.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/r/20241014161100.18034-1-will@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>

net: stmmac: dwmac-tegra: Fix link bring-up sequence

The Tegra MGBE driver sometimes fails to initialize, reporting the
following error, and as a result, it is unable to acquire an IP
address with DHCP:

 tegra-mgbe 6800000.ethernet: timeout waiting for link to become ready

As per the recommendation from the Tegra hardware design team, fix this
issue by:
- clearing the PHY_RDY bit before setting the CDR_RESET bit and then
setting PHY_RDY bit before clearing CDR_RESET bit. This ensures valid
data is present at UPHY RX inputs before starting the CDR lock.
- adding the required delays when bringing up the UPHY lane. Note we
need to use delays here because there is no alternative, such as
polling, for these cases. Using the usleep_range() instead of ndelay()
as sleeping is preferred over busy wait loop.

Without this change we would see link failures on boot sometimes as
often as 1 in 5 boots. With this fix we have not observed any failures
in over 1000 boots.

Fixes: d8ca113724e7 ("net: stmmac: tegra: Add MGBE support")
Signed-off-by: Paritosh Dixit <paritoshd@nvidia.com>
Link: https://patch.msgid.link/20241010142908.602712-1-paritoshd@nvidia.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

xfs: punch delalloc extents from the COW fork for COW writes

When ->iomap_end is called on a short write to the COW fork it needs to
punch stale delalloc data from the COW fork and not the data fork.

Ensure that IOMAP_F_NEW is set for new COW fork allocations in
xfs_buffered_write_iomap_begin, and then use the IOMAP_F_SHARED flag
in xfs_buffered_write_delalloc_punch to decide which fork to punch.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>

xfs: set IOMAP_F_SHARED for all COW fork allocations

Change to always set xfs_buffered_write_iomap_begin for COW fork
allocations even if they don't overlap existing data fork extents,
which will allow the iomap_end callback to detect if it has to punch
stale delalloc blocks from the COW fork instead of the data fork.  It
also means we sample the sequence counter for both the data and the COW
fork when writing to the COW fork, which ensures we properly revalidate
when only COW fork changes happens.

This is essentially a revert of commit 72a048c1056a ("xfs: only set
IOMAP_F_SHARED when providing a srcmap to a write"). This is fine because
the problem that the commit fixed has now been dealt with in iomap by
only looking at the actual srcmap and not the fallback to the write
iomap.

Note that the direct I/O path was never changed and has always set
IOMAP_F_SHARED for all COW fork allocations.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>

xfs: share more code in xfs_buffered_write_iomap_begin

Introduce a local iomap_flags variable so that the code allocating new
delalloc blocks in the data fork can fall through to the found_imap
label and reuse the code to unlock and fill the iomap.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>

xfs: support the COW fork in xfs_bmap_punch_delalloc_range

xfs_buffered_write_iomap_begin can also create delallocate reservations
that need cleaning up, prepare for that by adding support for the COW
fork in xfs_bmap_punch_delalloc_range.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>

xfs: IOMAP_ZERO and IOMAP_UNSHARE already hold invalidate_lock

All XFS callers of iomap_zero_range and iomap_file_unshare already hold
invalidate_lock, so we can't take it again in
iomap_file_buffered_write_punch_delalloc.

Use the passed in flags argument to detect if we're called from a zero
or unshare operation and don't take the lock again in this case.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>