Pull ata fixes from Damien Le Moal:
- Fix error message output in the pata_arasan_cf driver (Minjie)
- Fix invalid error return in the pata_octeon_cf driver initialization
(Yingliang)
- Fix a compilation warning due to a missing static function
declaration in the pata_ns87415 driver (Arnd)
- Fix the condition evaluating when to fetch sense data for successful
completions, which should be done only when command duration limits
are being used (Niklas)
* tag 'ata-6.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/libata:
ata: libata-core: fix when to fetch sense data for successful commands
ata: pata_ns87415: mark ns87560_tf_read static
ata: pata_octeon_cf: fix error return code in octeon_cf_probe()
ata: pata_arasan_cf: Use dev_err_probe() instead dev_err() in data_xfer()
Pull hotfixes from Andrew Morton:
"11 hotfixes. Five are cc:stable and the remainder address post-6.4
issues or aren't considered serious enough to justify backporting"
* tag 'mm-hotfixes-stable-2023-07-28-15-52' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
mm/memory-failure: fix hardware poison check in unpoison_memory()
proc/vmcore: fix signedness bug in read_from_oldmem()
mailmap: update remaining active codeaurora.org email addresses
mm: lock VMA in dup_anon_vma() before setting ->anon_vma
mm: fix memory ordering for mm_lock_seq and vm_lock_seq
scripts/spelling.txt: remove 'thead' as a typo
mm/pagewalk: fix EFI_PGT_DUMP of espfix area
shmem: minor fixes to splice-read implementation
tmpfs: fix Documentation of noswap and huge mount options
Revert "um: Use swap() to make code cleaner"
mm/damon/core-test: initialise context before test in damon_test_set_attrs()
Pull thermal control fixes from Rafael Wysocki:
"Constify thermal_zone_device_register() parameters, which was omitted
by mistake, and fix a double free on thermal zone unregistration in
the generic DT thermal driver (Ahmad Fatoum)"
* tag 'thermal-6.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
thermal: of: fix double-free on unregistration
thermal: core: constify params in thermal_zone_device_register
Pull power management fixes from Rafael Wysocki:
"Fix the arming of wakeup IRQs in the generic wakeup IRQ code
(wakeirq), drop unused functions from it and fix up a driver using it
and trying to work around the IRQ arming issue in a questionable way
(Johan Hovold)"
* tag 'pm-6.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
serial: qcom-geni: drop bogus runtime pm state update
PM: sleep: wakeirq: drop unused enable helpers
PM: sleep: wakeirq: fix wake irq arming
Pull rdma fixes from Jason Gunthorpe:
"Several smaller driver fixes and a core RDMA CM regression fix:
- Fix improperly accepting flags from userspace in mlx4
- Add missing DMA barriers for irdma
- Fix two kcsan warnings in irdma
- Report the correct CQ op code to userspace in irdma
- Report the correct MW bind error code for irdma
- Load the destination address in RDMA CM to resolve a recent
regression
- Fix a QP regression in mthca
- Remove a race processing completions in bnxt_re resulting in a
crash
- Fix driver unloading races with interrupts and tasklets in bnxt_re
- Fix missing error unwind in rxe"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
RDMA/irdma: Report correct WC error
RDMA/irdma: Fix op_type reporting in CQEs
RDMA/rxe: Fix an error handling path in rxe_bind_mw()
RDMA/bnxt_re: Fix hang during driver unload
RDMA/bnxt_re: Prevent handling any completions after qp destroy
RDMA/mthca: Fix crash when polling CQ for shared QPs
RDMA/core: Update CMA destination address on rdma_resolve_addr
RDMA/irdma: Fix data race on CQP request done
RDMA/irdma: Fix data race on CQP completion stats
RDMA/irdma: Add missing read barriers
RDMA/mlx4: Make check for invalid flags stricter
Pull tpm fixes from Jarkko Sakkinen:
"I picked up three small scale updates that I think would improve the
quality of the release"
* tag 'tpmdd-v6.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/jarkko/linux-tpmdd:
tpm_tis: Explicitly check for error code
tpm: Switch i2c drivers back to use .probe()
security: keys: perform capable check only on privileged operations
Pull arm64 fixes from Catalin Marinas:
- A couple of SME updates for recent fixes (one of which went to
stable): reverting the flushing of the SME hardware state along with
the thread flushing and making sure we have the correct vector length
before reallocating.
- An ACPI/IORT fix to avoid skipping ID mappings whose "number of IDs"
is 0 (the spec reports the number of IDs in the mapping range minus
1).
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
ACPI/IORT: Remove erroneous id_count check in iort_node_get_rmr_info()
arm64/sme: Set new vector length before reallocating
arm64/fpsimd: Don't flush SME register hardware state along with thread
Pull xen fixes from Juergen Gross:
- A fix for a performance problem in QubesOS, adding a way to drain the
queue of grants experiencing delayed unmaps faster
- A patch enabling the use of static event channels from user mode,
which was omitted when introducing supporting static event channels
- A fix for a problem where Xen related code didn't check properly for
running in a Xen environment, resulting in a WARN splat
* tag 'for-linus-6.5a-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: speed up grant-table reclaim
xen/evtchn: Introduce new IOCTL to bind static evtchn
xenbus: check xen_domain in xenbus_probe_initcall
recv_data either returns the number of received bytes, or a negative value
representing an error code. Adding the return value directly to the total
number of received bytes therefore looks a little weird, since it might add
a negative error code to a sum of bytes.
The following check for size < expected usually makes the function return
ETIME in that case, so it does not cause too many problems in practice. But
to make the code look cleaner and because the caller might still be
interested in the original error code, explicitly check for the presence of
an error code and pass that through.
Cc: stable@vger.kernel.org
Fixes: cb5354253a ("[PATCH] tpm: spacing cleanups 2")
Signed-off-by: Alexander Steffen <Alexander.Steffen@infineon.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
After commit b8a1a4cd5a ("i2c: Provide a temporary .probe_new()
call-back type"), all drivers being converted to .probe_new() and then
03c835f498 ("i2c: Switch .probe() to not take an id parameter")
convert back to (the new) .probe() to be able to eventually drop
.probe_new() from struct i2c_driver.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
If the current task fails the check for the queried capability via
`capable(CAP_SYS_ADMIN)` LSMs like SELinux generate a denial message.
Issuing such denial messages unnecessarily can lead to a policy author
granting more privileges to a subject than needed to silence them.
Reorder CAP_SYS_ADMIN checks after the check whether the operation is
actually privileged.
Signed-off-by: Christian Göttsche <cgzones@googlemail.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Pull ceph fixes from Ilya Dryomov:
"A patch to reduce the potential for erroneous RBD exclusive lock
blocklisting (fencing) with a couple of prerequisites and a fixup to
prevent metrics from being sent to the MDS even just once after that
has been disabled by the user. All marked for stable"
* tag 'ceph-for-6.5-rc4' of https://github.com/ceph/ceph-client:
rbd: retrieve and check lock owner twice before blocklisting
rbd: harden get_lock_owner_info() a bit
rbd: make get_lock_owner_info() return a single locker or NULL
ceph: never send metrics if disable_send_metrics is set
Pull 9p fixes from Eric Van Hensbergen:
"Misc set of fixes for 9p.
Most of these clean up warnings we've gotten out of compilation tools,
but several of them were from inspection while hunting down a couple
of regressions.
The most important one is 75b396821c ("fs/9p: remove unnecessary and
overrestrictive check") which caused a regression for some folks by
restricting mmap in any case where writeback caches weren't enabled.
Most of the other bugs caught via inspection were type mismatches"
* tag '9p-fixes-6.5-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
fs/9p: Remove unused extern declaration
9p: remove dead stores (variable set again without being read)
9p: virtio: skip incrementing unused variable
9p: virtio: make sure 'offs' is initialized in zc_request
9p: virtio: fix unlikely null pointer deref in handle_rerror
9p: fix ignored return value in v9fs_dir_release
fs/9p: remove unnecessary invalidate_inode_pages2
fs/9p: fix type mismatch in file cache mode helper
fs/9p: fix typo in comparison logic for cache mode
fs/9p: remove unnecessary and overrestrictive check
fs/9p: Fix a datatype used with V9FS_DIRECT_IO
Pull block fixes from Jens Axboe:
"A few fixes that should go into the current kernel release, mainly:
- Set of fixes for dasd (Stefan)
- Handle interruptible waits returning because of a signal for ublk
(Ming)"
* tag 'block-6.5-2023-07-28' of git://git.kernel.dk/linux:
ublk: return -EINTR if breaking from waiting for existed users in DEL_DEV
ublk: fail to recover device if queue setup is interrupted
ublk: fail to start device if queue setup is interrupted
block: Fix a source code comment in include/uapi/linux/blkzoned.h
s390/dasd: print copy pair message only for the correct error
s390/dasd: fix hanging device after request requeue
s390/dasd: use correct number of retries for ERP requests
s390/dasd: fix hanging device after quiesce/resume
Pull io_uring fix from Jens Axboe:
"Just a single tweak to a patch from last week, to avoid having idle
cqring waits be attributed as iowait"
* tag 'io_uring-6.5-2023-07-28' of git://git.kernel.dk/linux:
io_uring: gate iowait schedule on having pending requests
Pull iommufd fixes from Jason Gunthorpe:
"Two user triggerable problems:
- Syzkaller found a way to trigger a WARN_ON and leak memory by
racing destroy with other actions
- There is still a bug in the "batch carry" stuff that gets invoked
for complex cases with accesses and unmapping of huge pages. The
test suite found this (triggers rarely)"
* tag 'for-linus-iommufd' of git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd:
iommufd: Set end correctly when doing batch carry
iommufd: IOMMUFD_DESTROY should not increase the refcount
Pull device mapper fixes from Mike Snitzer:
- Fix double free on memory allocation failure in DM integrity target's
integrity_recalc()
- Fix locking in DM raid target's raid_ctr() and around call to
md_stop()
- Fix DM cache target's cleaner policy to always allow work to be
queued for writeback; even if cache isn't idle.
* tag 'for-6.5/dm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
dm cache policy smq: ensure IO doesn't prevent cleaner policy progress
dm raid: protect md_stop() with 'reconfig_mutex'
dm raid: clean up four equivalent goto tags in raid_ctr()
dm raid: fix missing reconfig_mutex unlock in raid_ctr() error paths
dm integrity: fix double free on memory allocation failure
Pull sound fixes from Takashi Iwai:
"A collection of device-specific small fixes such as ASoC Realtek codec
fixes for PM issues, ASoC nau8821 quirk additions, and usual HD- and
USB-audio quirks"
* tag 'sound-6.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ALSA: hda/realtek: Support ASUS G713PV laptop
ALSA: usb-audio: Update for native DSD support quirks
ALSA: usb-audio: Add quirk for Microsoft Modern Wireless Headset
ALSA: hda/relatek: Enable Mute LED on HP 250 G8
ASoC: atmel: Fix the 8K sample parameter in I2SC master
ASoC: rt711-sdca: fix for JD event handling in ClockStop Mode0
ASoC: rt711: fix for JD event handling in ClockStop Mode0
ASoC: rt722-sdca: fix for JD event handling in ClockStop Mode0
ASoC: rt712-sdca: fix for JD event handling in ClockStop Mode0
ASoc: codecs: ES8316: Fix DMIC config
ASoC: rt5682-sdw: fix for JD event handling in ClockStop Mode0
ASoC: wm8904: Fill the cache for WM8904_ADC_TEST_0 register
ASoC: nau8821: Add DMI quirk mechanism for active-high jack-detect
ASoC: da7219: Check for failure reading AAD IRQ events
ASoC: da7219: Flush pending AAD IRQ when suspending
ALSA: seq: remove redundant unsigned comparison to zero
ASoC: fsl_spdif: Silence output on stop
Pull drm fixes from Dave Airlie:
"Regular scheduled fixes, msm and amdgpu leading the way, with some
i915 and a single misc fbdev, all seems fine.
fbdev:
- remove unused function
amdgpu:
- gfxhub partition fix
- Fix error handling in psp_sw_init()
- SMU13 fix
- DCN 3.1 fix
- DCN 3.2 fix
- Fix for display PHY programming sequence
- DP MST error handling fix
- GFX 9.4.3 fix
amdkfd:
- GFX11 trap handling fix
i915:
- Use shmem for dpt objects
- Fix an error handling path in igt_write_huge()
msm:
- display:
- Fix to correct the UBWC programming for decoder version 4.3 seen
on SM8550
- Add the missing flush and fetch bits for DMA4 and DMA5 SSPPs.
- Fix to drop the unused dpu_core_perf_data_bus_id enum from the
code
- Drop the unused dsi_phy_14nm_17mA_regulators from QCM 2290 DSI
cfg.
- gpu:
- Fix warn splat for newer devices without revn
- Remove name/revn for a690.. we shouldn't be populating these for
newer devices, for consistency, but it slipped through review
- Fix a6xx gpu snapshot BINDLESS_DATA size (was listed in bytes
instead of dwords, causing AHB faults on a6xx gen4/a660-family)
- Disallow submit with fence id 0"
* tag 'drm-fixes-2023-07-28' of git://anongit.freedesktop.org/drm/drm: (22 commits)
drm/msm: Disallow submit with fence id 0
drm/amdgpu: Restore HQD persistent state register
drm/amd/display: Unlock on error path in dm_handle_mst_sideband_msg_ready_event()
drm/amd/display: Exit idle optimizations before attempt to access PHY
drm/amd/display: Don't apply FIFO resync W/A if rdivider = 0
drm/amd/display: Guard DCN31 PHYD32CLK logic against chip family
drm/amd/smu: use AverageGfxclkFrequency* to replace previous GFX Curr Clock
drm/amd: Fix an error handling mistake in psp_sw_init()
drm/amdgpu: Fix infinite loop in gfxhub_v1_2_xcc_gart_enable (v2)
drm/amdkfd: fix trap handling work around for debugging
drm/fb-helper: Remove unused inline function drm_fb_helper_defio_init()
drm/i915: Fix an error handling path in igt_write_huge()
drm/i915/dpt: Use shmem for dpt objects
drm/msm: Fix hw_fence error path cleanup
drm/msm: Fix IS_ERR_OR_NULL() vs NULL check in a5xx_submit_in_rb()
drm/msm/adreno: Fix snapshot BINDLESS_DATA size
drm/msm/a690: Remove revn and name
drm/msm/adreno: Fix warn splat for devices without revn
drm/msm/dsi: Drop unused regulators from QCM2290 14nm DSI PHY config
drm/msm/dpu: drop enum dpu_core_perf_data_bus_id
...
Pull cxl fixes from Vishal Verma:
- Update MAINTAINERS for cxl
- A few static analysis fixes
- Fix a Kconfig dependency for CONFIG_FW_LOADER
* tag 'cxl-fixes-6.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl:
tools/testing/cxl: Remove unused SZ_512G macro
cxl/acpi: Return 'rc' instead of '0' in cxl_parse_cfmws()
cxl/acpi: Fix a use-after-free in cxl_parse_cfmws()
cxl: Update MAINTAINERS
cxl/mem: Fix a double shift bug
cxl: fix CONFIG_FW_LOADER dependency
This reverts commit 9e46e4dcd9.
kbuild reports a warning in memblock_remove_region() because of a false
positive caused by partial reset of the memblock state.
Doing the full reset will remove the false positives, but will allow
late use of memblock_free() to go unnoticed, so it is better to revert
the offending commit.
WARNING: CPU: 0 PID: 1 at mm/memblock.c:352 memblock_remove_region (kbuild/src/x86_64/mm/memblock.c:352 (discriminator 1))
Modules linked in:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 6.5.0-rc3-00001-g9e46e4dcd9d6 #2
RIP: 0010:memblock_remove_region (kbuild/src/x86_64/mm/memblock.c:352 (discriminator 1))
Call Trace:
memblock_discard (kbuild/src/x86_64/mm/memblock.c:383)
page_alloc_init_late (kbuild/src/x86_64/include/linux/find.h:208 kbuild/src/x86_64/include/linux/nodemask.h:266 kbuild/src/x86_64/mm/mm_init.c:2405)
kernel_init_freeable (kbuild/src/x86_64/init/main.c:1325 kbuild/src/x86_64/init/main.c:1546)
kernel_init (kbuild/src/x86_64/init/main.c:1439)
ret_from_fork (kbuild/src/x86_64/arch/x86/kernel/process.c:145)
ret_from_fork_asm (kbuild/src/x86_64/arch/x86/entry/entry_64.S:298)
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202307271656.447aa17e-oliver.sang@intel.com
Signed-off-by: "Mike Rapoport (IBM)" <rppt@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mbind() calls down into vma_replace_policy() without taking the per-VMA
locks, replaces the VMA's vma->vm_policy pointer, and frees the old
policy. That's bad; a concurrent page fault might still be using the
old policy (in vma_alloc_folio()), resulting in use-after-free.
Normally this will manifest as a use-after-free read first, but it can
result in memory corruption, including because vma_alloc_folio() can
call mpol_cond_put() on the freed policy, which conditionally changes
the policy's refcount member.
This bug is specific to CONFIG_NUMA, but it does also affect non-NUMA
systems as long as the kernel was built with CONFIG_NUMA.
Signed-off-by: Jann Horn <jannh@google.com>
Reviewed-by: Suren Baghdasaryan <surenb@google.com>
Fixes: 5e31275cc9 ("mm: add per-VMA lock and helper functions to control it")
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The condition to fetch sense data was supposed to be:
ATA_SENSE set AND either
1) Command was NCQ and ATA_DFLAG_CDL_ENABLED flag set (flag
ATA_DFLAG_CDL_ENABLED will only be set if the Successful NCQ command
sense data supported bit is set); or
2) Command was non-NCQ and regular sense data reporting is enabled.
However the check in 2) accidentally had the negation at the wrong place,
causing it to try to fetch sense data if it was a non-NCQ command _or_
if regular sense data reporting was _not_ enabled.
Fix this by removing the extra parentheses that should not be there,
such that only the correct return (ata_is_ncq()) is negated.
Fixes: 18bd7718b5 ("scsi: ata: libata: Handle completion of CDL commands using policy 0xD")
Reported-by: Borislav Petkov <bp@alien8.de>
Closes: https://lore.kernel.org/linux-ide/20230722155621.GIZLv8JbURKzHtKvQE@fat_crate.local/
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Tested-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Fixes for v6.5-rc4
Display:
+ Fix to correct the UBWC programming for decoder version 4.3 seen
on SM8550
+ Add the missing flush and fetch bits for DMA4 and DMA5 SSPPs.
+ Fix to drop the unused dpu_core_perf_data_bus_id enum from the code
+ Drop the unused dsi_phy_14nm_17mA_regulators from QCM 2290 DSI cfg.
GPU:
+ Fix warn splat for newer devices without revn
+ Remove name/revn for a690.. we shouldn't be populating these for
newer devices, for consistency, but it slipped through review
+ Fix a6xx gpu snapshot BINDLESS_DATA size (was listed in bytes
instead of dwords, causing AHB faults on a6xx gen4/a660-family)
+ Disallow submit with fence id 0
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rob Clark <robdclark@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/CAF6AEGs9MwCSfiyv8i7yWAsJKYEzCDyzaTx=ujX80Y23rZd9RA@mail.gmail.com
The global function triggers a warning because of the missing prototype
drivers/ata/pata_ns87415.c:263:6: warning: no previous prototype for 'ns87560_tf_read' [-Wmissing-prototypes]
263 | void ns87560_tf_read(struct ata_port *ap, struct ata_taskfile *tf)
There are no other references to this, so just make it static.
Fixes: c4b5b7b6c4 ("pata_ns87415: Initial cut at 87415/87560 IDE support")
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Reviewed-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
When VMAs are merged, dup_anon_vma() is called with `dst` pointing to the
VMA that is being expanded to cover the area previously occupied by
another VMA. This currently happens while `dst` is not write-locked.
This means that, in the `src->anon_vma && !dst->anon_vma` case, as soon as
the assignment `dst->anon_vma = src->anon_vma` has happened, concurrent
page faults can happen on `dst` under the per-VMA lock. This is already
icky in itself, since such page faults can now install pages into `dst`
that are attached to an `anon_vma` that is not yet tied back to the
`anon_vma` with an `anon_vma_chain`. But if `anon_vma_clone()` fails due
to an out-of-memory error, things get much worse: `anon_vma_clone()` then
reverts `dst->anon_vma` back to NULL, and `dst` remains completely
unconnected to the `anon_vma`, even though we can have pages in the area
covered by `dst` that point to the `anon_vma`.
This means the `anon_vma` of such pages can be freed while the pages are
still mapped into userspace, which leads to UAF when a helper like
folio_lock_anon_vma_read() tries to look up the anon_vma of such a page.
This theoretically is a security bug, but I believe it is really hard to
actually trigger as an unprivileged user because it requires that you can
make an order-0 GFP_KERNEL allocation fail, and the page allocator tries
pretty hard to prevent that.
I think doing the vma_start_write() call inside dup_anon_vma() is the most
straightforward fix for now.
For a kernel-assisted reproducer, see the notes section of the patch mail.
Link: https://lkml.kernel.org/r/20230721034643.616851-1-jannh@google.com
Fixes: 5e31275cc9 ("mm: add per-VMA lock and helper functions to control it")
Signed-off-by: Jann Horn <jannh@google.com>
Reviewed-by: Suren Baghdasaryan <surenb@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
mm->mm_lock_seq effectively functions as a read/write lock; therefore it
must be used with acquire/release semantics.
A specific example is the interaction between userfaultfd_register() and
lock_vma_under_rcu().
userfaultfd_register() does the following from the point where it changes
a VMA's flags to the point where concurrent readers are permitted again
(in a simple scenario where only a single private VMA is accessed and no
merging/splitting is involved):
userfaultfd_register
userfaultfd_set_vm_flags
vm_flags_reset
vma_start_write
down_write(&vma->vm_lock->lock)
vma->vm_lock_seq = mm_lock_seq [marks VMA as busy]
up_write(&vma->vm_lock->lock)
vm_flags_init
[sets VM_UFFD_* in __vm_flags]
vma->vm_userfaultfd_ctx.ctx = ctx
mmap_write_unlock
vma_end_write_all
WRITE_ONCE(mm->mm_lock_seq, mm->mm_lock_seq + 1) [unlocks VMA]
There are no memory barriers in between the __vm_flags update and the
mm->mm_lock_seq update that unlocks the VMA, so the unlock can be
reordered to above the `vm_flags_init()` call, which means from the
perspective of a concurrent reader, a VMA can be marked as a userfaultfd
VMA while it is not VMA-locked. That's bad, we definitely need a
store-release for the unlock operation.
The non-atomic write to vma->vm_lock_seq in vma_start_write() is mostly
fine because all accesses to vma->vm_lock_seq that matter are always
protected by the VMA lock. There is a racy read in vma_start_read()
though that can tolerate false-positives, so we should be using
WRITE_ONCE() to keep things tidy and data-race-free (including for KCSAN).
On the other side, lock_vma_under_rcu() works as follows in the relevant
region for locking and userfaultfd check:
lock_vma_under_rcu
vma_start_read
vma->vm_lock_seq == READ_ONCE(vma->vm_mm->mm_lock_seq) [early bailout]
down_read_trylock(&vma->vm_lock->lock)
vma->vm_lock_seq == READ_ONCE(vma->vm_mm->mm_lock_seq) [main check]
userfaultfd_armed
checks vma->vm_flags & __VM_UFFD_FLAGS
Here, the interesting aspect is how far down the mm->mm_lock_seq read can
be reordered - if this read is reordered down below the vma->vm_flags
access, this could cause lock_vma_under_rcu() to partly operate on
information that was read while the VMA was supposed to be locked. To
prevent this kind of downwards bleeding of the mm->mm_lock_seq read, we
need to read it with a load-acquire.
Some of the comment wording is based on suggestions by Suren.
BACKPORT WARNING: One of the functions changed by this patch (which I've
written against Linus' tree) is vma_try_start_write(), but this function
no longer exists in mm/mm-everything. I don't know whether the merged
version of this patch will be ordered before or after the patch that
removes vma_try_start_write(). If you're backporting this patch to a tree
with vma_try_start_write(), make sure this patch changes that function.
Link: https://lkml.kernel.org/r/20230721225107.942336-1-jannh@google.com
Fixes: 5e31275cc9 ("mm: add per-VMA lock and helper functions to control it")
Signed-off-by: Jann Horn <jannh@google.com>
Reviewed-by: Suren Baghdasaryan <surenb@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Booting x86_64 with CONFIG_EFI_PGT_DUMP=y shows messages of the form
"mm/pgtable-generic.c:53: bad pmd (____ptrval____)(8000000100077061)".
EFI_PGT_DUMP dumps all of efi_mm, including the espfix area, which is set
up with pmd entries which fit the pmd_bad() check: so 0d940a9b27 warns
and clears those entries, which would ruin running Win16 binaries.
The failing pte_offset_map() stopped such a kernel from even booting,
until a few commits later be872f83bf changed the pagewalk to tolerate
that: but it needs to be even more careful, to not spoil those entries.
I might have preferred to change init_espfix_ap() not to use "bad" pmd
entries; or to leave them out of the efi_mm dump. But there is great
value in staying away from there, and a pagewalk check of address against
TASK_SIZE may protect from other such aberrations too.
Link: https://lkml.kernel.org/r/22bca736-4cab-9ee5-6a52-73a3b2bbe865@google.com
Closes: https://lore.kernel.org/linux-mm/CABXGCsN3JqXckWO=V7p=FhPU1tK03RE1w9UE6xL5Y86SMk209w@mail.gmail.com/
Fixes: 0d940a9b27 ("mm/pgtable: allow pte_offset_map[_lock]() to fail")
Fixes: be872f83bf ("mm/pagewalk: walk_pte_range() allow for pte_offset_map()")
Signed-off-by: Hugh Dickins <hughd@google.com>
Reported-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Tested-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Cc: Bagas Sanjaya <bagasdotme@gmail.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
HWPoison: my reading of folio_test_hwpoison() is that it only tests the
head page of a large folio, whereas splice_folio_into_pipe() will splice
as much of the folio as it can: so for safety we should also check the
has_hwpoisoned flag, set if any of the folio's pages are hwpoisoned.
(Perhaps that ugliness can be improved at the mm end later.)
The call to splice_zeropage_into_pipe() risked overrunning past EOF: ask
it for "part" not "len".
Link: https://lkml.kernel.org/r/32c72c9c-72a8-115f-407d-f0148f368@google.com
Fixes: bd194b1871 ("shmem: Implement splice-read")
Signed-off-by: Hugh Dickins <hughd@google.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The noswap mount option is surely not one of the three options for sizing:
move its description down.
The huge= mount option does not accept numeric values: those are just in
an internal enum. Delete those numbers, and follow the manpage text more
closely (but there's not yet any fadvise() or fcntl() which applies here).
/sys/kernel/mm/transparent_hugepage/shmem_enabled is hard to describe, and
barely relevant to mounting a tmpfs: just refer to transhuge.rst (while
still using the words deny and force, to help as informal reminders).
[rdunlap@infradead.org: fixup Docs table for huge mount options]
Link: https://lkml.kernel.org/r/20230725052333.26857-1-rdunlap@infradead.org
Link: https://lkml.kernel.org/r/986cb0bf-9780-354-9bb-4bf57aadbab@google.com
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Fixes: d0f5a85442 ("shmem: update documentation")
Fixes: 2c6efe9cf2 ("shmem: add support to ignore swap")
Reviewed-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Christian Brauner <brauner@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Pull networking fixes from Paolo Abeni:
"Including fixes from can, netfilter.
Current release - regressions:
- core: fix splice_to_socket() for O_NONBLOCK socket
- af_unix: fix fortify_panic() in unix_bind_bsd().
- can: raw: fix lockdep issue in raw_release()
Previous releases - regressions:
- tcp: reduce chance of collisions in inet6_hashfn().
- netfilter: skip immediate deactivate in _PREPARE_ERROR
- tipc: stop tipc crypto on failure in tipc_node_create
- eth: igc: fix kernel panic during ndo_tx_timeout callback
- eth: iavf: fix potential deadlock on allocation failure
Previous releases - always broken:
- ipv6: fix bug where deleting a mngtmpaddr can create a new
temporary address
- eth: ice: fix memory management in ice_ethtool_fdir.c
- eth: hns3: fix the imp capability bit cannot exceed 32 bits issue
- eth: vxlan: calculate correct header length for GPE
- eth: stmmac: apply redundant write work around on 4.xx too"
* tag 'net-6.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (49 commits)
tipc: stop tipc crypto on failure in tipc_node_create
af_unix: Terminate sun_path when bind()ing pathname socket.
tipc: check return value of pskb_trim()
benet: fix return value check in be_lancer_xmit_workarounds()
virtio-net: fix race between set queues and probe
net/sched: mqprio: Add length check for TCA_MQPRIO_{MAX/MIN}_RATE64
splice, net: Fix splice_to_socket() for O_NONBLOCK socket
net: fec: tx processing does not call XDP APIs if budget is 0
mptcp: more accurate NL event generation
selftests: mptcp: join: only check for ip6tables if needed
tools: ynl-gen: fix parse multi-attr enum attribute
tools: ynl-gen: fix enum index in _decode_enum(..)
netfilter: nf_tables: disallow rule addition to bound chain via NFTA_RULE_CHAIN_ID
netfilter: nf_tables: skip immediate deactivate in _PREPARE_ERROR
netfilter: nft_set_rbtree: fix overlap expiration walk
igc: Fix Kernel Panic during ndo_tx_timeout callback
net: dsa: qca8k: fix mdb add/del case with 0 VID
net: dsa: qca8k: fix broken search_and_del
net: dsa: qca8k: fix search_and_insert wrong handling of new rule
net: dsa: qca8k: enable use_single_write for qca8xxx
...
Pull soundwire fixes from Vinod Koul:
- Core fix for enumeration completion
- Qualcomm driver fix to update status
- AMD driver fix for probe error check
* tag 'soundwire-6.5-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire:
soundwire: amd: Fix a check for errors in probe()
soundwire: qcom: update status correctly with mask
soundwire: fix enumeration completion
Pull phy fixes from Vinod Koul:
- Out of bound fix for hisilicon phy
- Qualcomm synopsis femto phy for keeping clock enabled during suspend
and enabling ref clocks
- Mediatek driver fixes for upper limit test and error code
* tag 'phy-fixes-6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy:
phy: hisilicon: Fix an out of bounds check in hisi_inno_phy_probe()
phy: qcom-snps-femto-v2: use qcom_snps_hsphy_suspend/resume error code
phy: qcom-snps-femto-v2: properly enable ref clock
phy: qcom-snps-femto-v2: keep cfg_ahb_clk enabled during runtime suspend
phy: mediatek: hdmi: mt8195: fix prediv bad upper limit test
phy: phy-mtk-dp: Fix an error code in probe()
Pull btrfs fixes from David Sterba:
- fix accounting of global block reserve size when block group tree is
enabled
- the async discard has been enabled in 6.2 unconditionally, but for
zoned mode it does not make that much sense to do it asynchronously
as the zones are reset as needed
- error handling and proper error value propagation fixes
* tag 'for-6.5-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
btrfs: check for commit error at btrfs_attach_transaction_barrier()
btrfs: check if the transaction was aborted at btrfs_wait_for_commit()
btrfs: remove BUG_ON()'s in add_new_free_space()
btrfs: account block group tree when calculating global reserve size
btrfs: zoned: do not enable async discard
Pull memblock fix from Mike Rapoport:
"A call to memblock_free() or memblock_phys_free() issued after
memblock data is discarded will result in use after free in
memblock_isolate_range().
Avoid those issues by making sure that memblock_discard points
memblock.reserved.regions back at the static buffer"
* tag 'fixes-2023-07-27' of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock:
mm,memblock: reset memblock.reserved to system init state to prevent UAF
lock_vma_under_rcu() tries to guarantee that __anon_vma_prepare() can't
be called in the VMA-locked page fault path by ensuring that
vma->anon_vma is set.
However, this check happens before the VMA is locked, which means a
concurrent move_vma() can concurrently call unlink_anon_vmas(), which
disassociates the VMA's anon_vma.
This means we can get UAF in the following scenario:
THREAD 1 THREAD 2
======== ========
<page fault>
lock_vma_under_rcu()
rcu_read_lock()
mas_walk()
check vma->anon_vma
mremap() syscall
move_vma()
vma_start_write()
unlink_anon_vmas()
<syscall end>
handle_mm_fault()
__handle_mm_fault()
handle_pte_fault()
do_pte_missing()
do_anonymous_page()
anon_vma_prepare()
__anon_vma_prepare()
find_mergeable_anon_vma()
mas_walk() [looks up VMA X]
munmap() syscall (deletes VMA X)
reusable_anon_vma() [called on freed VMA X]
This is a security bug if you can hit it, although an attacker would
have to win two races at once where the first race window is only a few
instructions wide.
This patch is based on some previous discussion with Linus Torvalds on
the security list.
Cc: stable@vger.kernel.org
Fixes: 5e31275cc9 ("mm: add per-VMA lock and helper functions to control it")
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Industrial processor i3255 supports temperatures -40 deg celcius
to 105 deg Celcius. The current implementation of k10temp_read_temp
rounds off any negative temperatures to '0'. To fix this,
the following changes have been made.
A flag 'disp_negative' is added to struct k10temp_data to support
AMD i3255 processors. Flag 'disp_negative' is set if 3255 processor
is found during k10temp_probe. Flag 'disp_negative' is used to
determine whether to round off negative temperatures to '0' in
k10temp_read_temp.
Signed-off-by: Baskaran Kannan <Baski.Kannan@amd.com>
Link: https://lore.kernel.org/r/20230727162159.1056136-1-Baski.Kannan@amd.com
Fixes: aef17ca127 ("hwmon: (k10temp) Only apply temperature offset if result is positive")
Cc: stable@vger.kernel.org
[groeck: Fixed multi-line comment]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
syzkaller found a race where IOMMUFD_DESTROY increments the refcount:
obj = iommufd_get_object(ucmd->ictx, cmd->id, IOMMUFD_OBJ_ANY);
if (IS_ERR(obj))
return PTR_ERR(obj);
iommufd_ref_to_users(obj);
/* See iommufd_ref_to_users() */
if (!iommufd_object_destroy_user(ucmd->ictx, obj))
As part of the sequence to join the two existing primitives together.
Allowing the refcount the be elevated without holding the destroy_rwsem
violates the assumption that all temporary refcount elevations are
protected by destroy_rwsem. Racing IOMMUFD_DESTROY with
iommufd_object_destroy_user() will cause spurious failures:
WARNING: CPU: 0 PID: 3076 at drivers/iommu/iommufd/device.c:477 iommufd_access_destroy+0x18/0x20 drivers/iommu/iommufd/device.c:478
Modules linked in:
CPU: 0 PID: 3076 Comm: syz-executor.0 Not tainted 6.3.0-rc1-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/03/2023
RIP: 0010:iommufd_access_destroy+0x18/0x20 drivers/iommu/iommufd/device.c:477
Code: e8 3d 4e 00 00 84 c0 74 01 c3 0f 0b c3 0f 1f 44 00 00 f3 0f 1e fa 48 89 fe 48 8b bf a8 00 00 00 e8 1d 4e 00 00 84 c0 74 01 c3 <0f> 0b c3 0f 1f 44 00 00 41 57 41 56 41 55 4c 8d ae d0 00 00 00 41
RSP: 0018:ffffc90003067e08 EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff888109ea0300 RCX: 0000000000000000
RDX: 0000000000000001 RSI: 0000000000000000 RDI: 00000000ffffffff
RBP: 0000000000000004 R08: 0000000000000000 R09: ffff88810bbb3500
R10: ffff88810bbb3e48 R11: 0000000000000000 R12: ffffc90003067e88
R13: ffffc90003067ea8 R14: ffff888101249800 R15: 00000000fffffffe
FS: 00007ff7254fe6c0(0000) GS:ffff888237c00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000555557262da8 CR3: 000000010a6fd000 CR4: 0000000000350ef0
Call Trace:
<TASK>
iommufd_test_create_access drivers/iommu/iommufd/selftest.c:596 [inline]
iommufd_test+0x71c/0xcf0 drivers/iommu/iommufd/selftest.c:813
iommufd_fops_ioctl+0x10f/0x1b0 drivers/iommu/iommufd/main.c:337
vfs_ioctl fs/ioctl.c:51 [inline]
__do_sys_ioctl fs/ioctl.c:870 [inline]
__se_sys_ioctl fs/ioctl.c:856 [inline]
__x64_sys_ioctl+0x84/0xc0 fs/ioctl.c:856
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x38/0x80 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
The solution is to not increment the refcount on the IOMMUFD_DESTROY path
at all. Instead use the xa_lock to serialize everything. The refcount
check == 1 and xa_erase can be done under a single critical region. This
avoids the need for any refcount incrementing.
It has the downside that if userspace races destroy with other operations
it will get an EBUSY instead of waiting, but this is kind of racing is
already dangerous.
Fixes: 2ff4bed7fe ("iommufd: File descriptor, context, kconfig and makefiles")
Link: https://lore.kernel.org/r/2-v1-85aacb2af554+bc-iommufd_syz3_jgg@nvidia.com
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reported-by: syzbot+7574ebfe589049630608@syzkaller.appspotmail.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
