linux-block.git
12 months agoMerge tag 'vfs-6.10-rc8.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs
Linus Torvalds [Thu, 11 Jul 2024 16:03:28 +0000 (09:03 -0700)]
Merge tag 'vfs-6.10-rc8.fixes' of git://git./linux/kernel/git/vfs/vfs

Pull vfs fixes from Christian Brauner:
 "cachefiles:

   - Export an existing and add a new cachefile helper to be used in
     filesystems to fix reference count bugs

   - Use the newly added fscache_ty_get_volume() helper to get a
     reference count on an fscache_volume to handle volumes that are
     about to be removed cleanly

   - After withdrawing a fscache_cache via FSCACHE_CACHE_IS_WITHDRAWN
     wait for all ongoing cookie lookups to complete and for the object
     count to reach zero

   - Propagate errors from vfs_getxattr() to avoid an infinite loop in
     cachefiles_check_volume_xattr() because it keeps seeing ESTALE

   - Don't send new requests when an object is dropped by raising
     CACHEFILES_ONDEMAND_OJBSTATE_DROPPING

   - Cancel all requests for an object that is about to be dropped

   - Wait for the ondemand_boject_worker to finish before dropping a
     cachefiles object to prevent use-after-free

   - Use cyclic allocation for message ids to better handle id recycling

   - Add missing lock protection when iterating through the xarray when
     polling

  netfs:

   - Use standard logging helpers for debug logging

  VFS:

   - Fix potential use-after-free in file locks during
     trace_posix_lock_inode(). The tracepoint could fire while another
     task raced it and freed the lock that was requested to be traced

   - Only increment the nr_dentry_negative counter for dentries that are
     present on the superblock LRU. Currently, DCACHE_LRU_LIST list is
     used to detect this case. However, the flag is also raised in
     combination with DCACHE_SHRINK_LIST to indicate that dentry->d_lru
     is used. So checking only DCACHE_LRU_LIST will lead to wrong
     nr_dentry_negative count. Fix the check to not count dentries that
     are on a shrink related list

  Misc:

   - hfsplus: fix an uninitialized value issue in copy_name

   - minix: fix minixfs_rename with HIGHMEM. It still uses kunmap() even
     though we switched it to kmap_local_page() a while ago"

* tag 'vfs-6.10-rc8.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
  minixfs: Fix minixfs_rename with HIGHMEM
  hfsplus: fix uninit-value in copy_name
  vfs: don't mod negative dentry count when on shrinker list
  filelock: fix potential use-after-free in posix_lock_inode
  cachefiles: add missing lock protection when polling
  cachefiles: cyclic allocation of msg_id to avoid reuse
  cachefiles: wait for ondemand_object_worker to finish when dropping object
  cachefiles: cancel all requests for the object that is being dropped
  cachefiles: stop sending new request when dropping object
  cachefiles: propagate errors from vfs_getxattr() to avoid infinite loop
  cachefiles: fix slab-use-after-free in cachefiles_withdraw_cookie()
  cachefiles: fix slab-use-after-free in fscache_withdraw_volume()
  netfs, fscache: export fscache_put_volume() and add fscache_try_get_volume()
  netfs: Switch debug logging to pr_debug()

12 months agoMerge tag 'mm-hotfixes-stable-2024-07-10-13-19' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Wed, 10 Jul 2024 21:59:41 +0000 (14:59 -0700)]
Merge tag 'mm-hotfixes-stable-2024-07-10-13-19' of git://git./linux/kernel/git/akpm/mm

Pull misc fixes from Andrew Morton:
 "21 hotfixes, 15 of which are cc:stable.

  No identifiable theme here - all are singleton patches, 19 are for MM"

* tag 'mm-hotfixes-stable-2024-07-10-13-19' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (21 commits)
  mm/hugetlb: fix kernel NULL pointer dereference when migrating hugetlb folio
  mm/hugetlb: fix potential race in __update_and_free_hugetlb_folio()
  filemap: replace pte_offset_map() with pte_offset_map_nolock()
  arch/xtensa: always_inline get_current() and current_thread_info()
  sched.h: always_inline alloc_tag_{save|restore} to fix modpost warnings
  MAINTAINERS: mailmap: update Lorenzo Stoakes's email address
  mm: fix crashes from deferred split racing folio migration
  lib/build_OID_registry: avoid non-destructive substitution for Perl < 5.13.2 compat
  mm: gup: stop abusing try_grab_folio
  nilfs2: fix kernel bug on rename operation of broken directory
  mm/hugetlb_vmemmap: fix race with speculative PFN walkers
  cachestat: do not flush stats in recency check
  mm/shmem: disable PMD-sized page cache if needed
  mm/filemap: skip to create PMD-sized page cache if needed
  mm/readahead: limit page cache size in page_cache_ra_order()
  mm/filemap: make MAX_PAGECACHE_ORDER acceptable to xarray
  mm/damon/core: merge regions aggressively when max_nr_regions is unmet
  Fix userfaultfd_api to return EINVAL as expected
  mm: vmalloc: check if a hash-index is in cpu_possible_mask
  mm: prevent derefencing NULL ptr in pfn_section_valid()
  ...

12 months agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Wed, 10 Jul 2024 21:47:35 +0000 (14:47 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "One core change that moves a disk start message to a location where it
  will only be printed once instead of twice plus a couple of error
  handling race fixes in the ufs driver"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: sd: Do not repeat the starting disk message
  scsi: ufs: core: Fix ufshcd_abort_one racing issue
  scsi: ufs: core: Fix ufshcd_clear_cmd racing issue

12 months agoMerge tag 'vfio-v6.10' of https://github.com/awilliam/linux-vfio
Linus Torvalds [Wed, 10 Jul 2024 19:00:43 +0000 (12:00 -0700)]
Merge tag 'vfio-v6.10' of https://github.com/awilliam/linux-vfio

Pull VFIO fix from Alex Williamson:

 - Recent stable backports are exposing a bug introduced in the v6.10
   development cycle where a counter value is uninitialized.  This leads
   to regressions in userspace drivers like QEMU where where the kernel
   might ask for an arbitrary buffer size or return out of memory itself
   based on a bogus value.  Zero initialize the counter.  (Yi Liu)

* tag 'vfio-v6.10' of https://github.com/awilliam/linux-vfio:
  vfio/pci: Init the count variable in collecting hot-reset devices

12 months agoMerge tag 'bcachefs-2024-07-10' of https://evilpiepirate.org/git/bcachefs
Linus Torvalds [Wed, 10 Jul 2024 18:50:16 +0000 (11:50 -0700)]
Merge tag 'bcachefs-2024-07-10' of https://evilpiepirate.org/git/bcachefs

Pull bcachefs fixes from Kent Overstreet:

 - Switch some asserts to WARN()

 - Fix a few "transaction not locked" asserts in the data read retry
   paths and backpointers gc

 - Fix a race that would cause the journal to get stuck on a flush
   commit

 - Add missing fsck checks for the fragmentation LRU

 - The usual assorted ssorted syzbot fixes

* tag 'bcachefs-2024-07-10' of https://evilpiepirate.org/git/bcachefs: (22 commits)
  bcachefs: Add missing bch2_trans_begin()
  bcachefs: Fix missing error check in journal_entry_btree_keys_validate()
  bcachefs: Warn on attempting a move with no replicas
  bcachefs: bch2_data_update_to_text()
  bcachefs: Log mount failure error code
  bcachefs: Fix undefined behaviour in eytzinger1_first()
  bcachefs: Mark bch_inode_info as SLAB_ACCOUNT
  bcachefs: Fix bch2_inode_insert() race path for tmpfiles
  closures: fix closure_sync + closure debugging
  bcachefs: Fix journal getting stuck on a flush commit
  bcachefs: io clock: run timer fns under clock lock
  bcachefs: Repair fragmentation_lru in alloc_write_key()
  bcachefs: add check for missing fragmentation in check_alloc_to_lru_ref()
  bcachefs: bch2_btree_write_buffer_maybe_flush()
  bcachefs: Add missing printbuf_tabstops_reset() calls
  bcachefs: Fix loop restart in bch2_btree_transactions_read()
  bcachefs: Fix bch2_read_retry_nodecode()
  bcachefs: Don't use the new_fs() bucket alloc path on an initialized fs
  bcachefs: Fix shift greater than integer size
  bcachefs: Change bch2_fs_journal_stop() BUG_ON() to warning
  ...

12 months agoMerge tag 'platform-drivers-x86-v6.10-6' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 10 Jul 2024 16:08:22 +0000 (09:08 -0700)]
Merge tag 'platform-drivers-x86-v6.10-6' of git://git./linux/kernel/git/pdx86/platform-drivers-x86

Pull x86 platform driver fix from Hans de Goede:
 "One-liner fix for a dmi_system_id array in the toshiba_acpi driver not
  being terminated properly.

  Something which somehow has escaped detection since being introduced
  in 2022 until now"

* tag 'platform-drivers-x86-v6.10-6' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:
  platform/x86: toshiba_acpi: Fix array out-of-bounds access

12 months agoMerge tag 'acpi-6.10-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Wed, 10 Jul 2024 16:05:22 +0000 (09:05 -0700)]
Merge tag 'acpi-6.10-rc8' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI fix from Rafael Wysocki:
 "Fix the sorting of _CST output data in the ACPI processor idle driver
  (Kuan-Wei Chiu)"

* tag 'acpi-6.10-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI: processor_idle: Fix invalid comparison with insertion sort for latency

12 months agoMerge tag 'pm-6.10-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Linus Torvalds [Wed, 10 Jul 2024 16:03:21 +0000 (09:03 -0700)]
Merge tag 'pm-6.10-rc8' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management fixes from Rafael Wysocki:
 "Fix two issues related to boost frequencies handling, one in the
  cpufreq core and one in the ACPI cpufreq driver (Mario Limonciello)"

* tag 'pm-6.10-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq: ACPI: Mark boost policy as enabled when setting boost
  cpufreq: Allow drivers to advertise boost enabled

12 months agoMerge tag 'thermal-6.10-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Wed, 10 Jul 2024 16:00:55 +0000 (09:00 -0700)]
Merge tag 'thermal-6.10-rc8' of git://git./linux/kernel/git/rafael/linux-pm

Pull thermal control fixes from Rafael Wysocki:
 "These fix a possible NULL pointer dereference in a thermal governor,
  fix up the handling of thermal zones enabled before their temperature
  can be determined and fix list sorting during thermal zone temperature
  updates.

  Specifics:

   - Prevent the Power Allocator thermal governor from dereferencing a
     NULL pointer if it is bound to a tripless thermal zone (Nícolas
     Prado)

   - Prevent thermal zones enabled too early from staying effectively
     dormant forever because their temperature cannot be determined
     initially (Rafael Wysocki)

   - Fix list sorting during thermal zone temperature updates to ensure
     the proper ordering of trip crossing notifications (Rafael
     Wysocki)"

* tag 'thermal-6.10-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  thermal: core: Fix list sorting in __thermal_zone_device_update()
  thermal: core: Call monitor_thermal_zone() if zone temperature is invalid
  thermal: gov_power_allocator: Return early in manage if trip_max is NULL

12 months agoMerge tag 'devicetree-fixes-for-6.10-2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 10 Jul 2024 15:58:50 +0000 (08:58 -0700)]
Merge tag 'devicetree-fixes-for-6.10-2' of git://git./linux/kernel/git/robh/linux

Pull devicetree fix from Rob Herring:

 - One fix for PASemi Nemo board interrupts

* tag 'devicetree-fixes-for-6.10-2' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  of/irq: Disable "interrupt-map" parsing for PASEMI Nemo

12 months agovfio/pci: Init the count variable in collecting hot-reset devices
Yi Liu [Wed, 10 Jul 2024 00:41:50 +0000 (17:41 -0700)]
vfio/pci: Init the count variable in collecting hot-reset devices

The count variable is used without initialization, it results in mistakes
in the device counting and crashes the userspace if the get hot reset info
path is triggered.

Fixes: f6944d4a0b87 ("vfio/pci: Collect hot-reset devices to local buffer")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=219010
Reported-by: Žilvinas Žaltiena <zaltys@natrix.lt>
Cc: Beld Zhang <beldzhang@gmail.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20240710004150.319105-1-yi.l.liu@intel.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
12 months agoplatform/x86: toshiba_acpi: Fix array out-of-bounds access
Armin Wolf [Tue, 9 Jul 2024 14:38:51 +0000 (16:38 +0200)]
platform/x86: toshiba_acpi: Fix array out-of-bounds access

In order to use toshiba_dmi_quirks[] together with the standard DMI
matching functions, it must be terminated by a empty entry.

Since this entry is missing, an array out-of-bounds access occurs
every time the quirk list is processed.

Fix this by adding the terminating empty entry.

Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202407091536.8b116b3d-lkp@intel.com
Fixes: 3cb1f40dfdc3 ("drivers/platform: toshiba_acpi: Call HCI_PANEL_POWER_ON on resume on some models")
Cc: stable@vger.kernel.org
Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Link: https://lore.kernel.org/r/20240709143851.10097-1-W_Armin@gmx.de
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
12 months agobcachefs: Add missing bch2_trans_begin()
Kent Overstreet [Fri, 5 Jul 2024 13:10:35 +0000 (09:10 -0400)]
bcachefs: Add missing bch2_trans_begin()

this fixes a 'transaction should be locked' error in backpointers fsck

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: Fix missing error check in journal_entry_btree_keys_validate()
Kent Overstreet [Fri, 5 Jul 2024 01:18:06 +0000 (21:18 -0400)]
bcachefs: Fix missing error check in journal_entry_btree_keys_validate()

Closes: https://syzkaller.appspot.com/bug?extid=8996d8f176cf946ef641
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: Warn on attempting a move with no replicas
Kent Overstreet [Wed, 3 Jul 2024 17:23:58 +0000 (13:23 -0400)]
bcachefs: Warn on attempting a move with no replicas

Instead of popping an assert in bch2_write(), WARN and print out some
debugging info.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: bch2_data_update_to_text()
Kent Overstreet [Wed, 3 Jul 2024 17:25:13 +0000 (13:25 -0400)]
bcachefs: bch2_data_update_to_text()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: Log mount failure error code
Kent Overstreet [Wed, 3 Jul 2024 16:58:34 +0000 (12:58 -0400)]
bcachefs: Log mount failure error code

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: Fix undefined behaviour in eytzinger1_first()
Kent Overstreet [Wed, 3 Jul 2024 15:56:19 +0000 (11:56 -0400)]
bcachefs: Fix undefined behaviour in eytzinger1_first()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: Mark bch_inode_info as SLAB_ACCOUNT
Youling Tang [Wed, 3 Jul 2024 07:09:55 +0000 (15:09 +0800)]
bcachefs: Mark bch_inode_info as SLAB_ACCOUNT

After commit 230e9fc28604 ("slab: add SLAB_ACCOUNT flag"), we need to mark
the inode cache as SLAB_ACCOUNT, similar to commit 5d097056c9a0 ("kmemcg:
account for certain kmem allocations to memcg")

Signed-off-by: Youling Tang <tangyouling@kylinos.cn>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: Fix bch2_inode_insert() race path for tmpfiles
Kent Overstreet [Mon, 1 Jul 2024 20:23:54 +0000 (16:23 -0400)]
bcachefs: Fix bch2_inode_insert() race path for tmpfiles

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agoclosures: fix closure_sync + closure debugging
Kent Overstreet [Sun, 30 Jun 2024 02:12:09 +0000 (22:12 -0400)]
closures: fix closure_sync + closure debugging

originally, stack closures were only used synchronously, and with the
original implementation of closure_sync() the ref never hit 0; thus,
closure_put_after_sub() assumes that if the ref hits 0 it's on the debug
list, in debug mode.

that's no longer true with the current implementation of closure_sync,
so we need a new magic so closure_debug_destroy() doesn't pop an assert.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agobcachefs: Fix journal getting stuck on a flush commit
Kent Overstreet [Sun, 30 Jun 2024 01:40:57 +0000 (21:40 -0400)]
bcachefs: Fix journal getting stuck on a flush commit

silly race

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
12 months agominixfs: Fix minixfs_rename with HIGHMEM
Matthew Wilcox (Oracle) [Tue, 9 Jul 2024 19:58:39 +0000 (20:58 +0100)]
minixfs: Fix minixfs_rename with HIGHMEM

minixfs now uses kmap_local_page(), so we can't call kunmap() to
undo it.  This one call was missed as part of the commit this fixes.

Fixes: 6628f69ee66a (minixfs: Use dir_put_page() in minix_unlink() and minix_rename())
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Link: https://lore.kernel.org/r/20240709195841.1986374-1-willy@infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
12 months agomm/hugetlb: fix kernel NULL pointer dereference when migrating hugetlb folio
Miaohe Lin [Tue, 9 Jul 2024 12:04:33 +0000 (20:04 +0800)]
mm/hugetlb: fix kernel NULL pointer dereference when migrating hugetlb folio

A kernel crash was observed when migrating hugetlb folio:

BUG: kernel NULL pointer dereference, address: 0000000000000008
PGD 0 P4D 0
Oops: Oops: 0002 [#1] PREEMPT SMP NOPTI
CPU: 0 PID: 3435 Comm: bash Not tainted 6.10.0-rc6-00450-g8578ca01f21f #66
RIP: 0010:__folio_undo_large_rmappable+0x70/0xb0
RSP: 0018:ffffb165c98a7b38 EFLAGS: 00000097
RAX: fffffbbc44528090 RBX: 0000000000000000 RCX: 0000000000000000
RDX: ffffa30e000a2800 RSI: 0000000000000246 RDI: ffffa3153ffffcc0
RBP: fffffbbc44528000 R08: 0000000000002371 R09: ffffffffbe4e5868
R10: 0000000000000001 R11: 0000000000000001 R12: ffffa3153ffffcc0
R13: fffffbbc44468000 R14: 0000000000000001 R15: 0000000000000001
FS:  00007f5b3a716740(0000) GS:ffffa3151fc00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000008 CR3: 000000010959a000 CR4: 00000000000006f0
Call Trace:
 <TASK>
 __folio_migrate_mapping+0x59e/0x950
 __migrate_folio.constprop.0+0x5f/0x120
 move_to_new_folio+0xfd/0x250
 migrate_pages+0x383/0xd70
 soft_offline_page+0x2ab/0x7f0
 soft_offline_page_store+0x52/0x90
 kernfs_fop_write_iter+0x12c/0x1d0
 vfs_write+0x380/0x540
 ksys_write+0x64/0xe0
 do_syscall_64+0xb9/0x1d0
 entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f5b3a514887
RSP: 002b:00007ffe138fce68 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 000000000000000c RCX: 00007f5b3a514887
RDX: 000000000000000c RSI: 0000556ab809ee10 RDI: 0000000000000001
RBP: 0000556ab809ee10 R08: 00007f5b3a5d1460 R09: 000000007fffffff
R10: 0000000000000000 R11: 0000000000000246 R12: 000000000000000c
R13: 00007f5b3a61b780 R14: 00007f5b3a617600 R15: 00007f5b3a616a00

It's because hugetlb folio is passed to __folio_undo_large_rmappable()
unexpectedly.  large_rmappable flag is imperceptibly set to hugetlb folio
since commit f6a8dd98a2ce ("hugetlb: convert alloc_buddy_hugetlb_folio to
use a folio").  Then commit be9581ea8c05 ("mm: fix crashes from deferred
split racing folio migration") makes folio_migrate_mapping() call
folio_undo_large_rmappable() triggering the bug.  Fix this issue by
clearing large_rmappable flag for hugetlb folios.  They don't need that
flag set anyway.

Link: https://lkml.kernel.org/r/20240709120433.4136700-1-linmiaohe@huawei.com
Fixes: f6a8dd98a2ce ("hugetlb: convert alloc_buddy_hugetlb_folio to use a folio")
Fixes: be9581ea8c05 ("mm: fix crashes from deferred split racing folio migration")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm/hugetlb: fix potential race in __update_and_free_hugetlb_folio()
Miaohe Lin [Mon, 8 Jul 2024 02:51:27 +0000 (10:51 +0800)]
mm/hugetlb: fix potential race in __update_and_free_hugetlb_folio()

There is a potential race between __update_and_free_hugetlb_folio() and
try_memory_failure_hugetlb():

 CPU1 CPU2
 __update_and_free_hugetlb_folio try_memory_failure_hugetlb
 folio_test_hugetlb
  -- It's still hugetlb folio.
  folio_clear_hugetlb_hwpoison
     spin_lock_irq(&hugetlb_lock);
   __get_huge_page_for_hwpoison
    folio_set_hugetlb_hwpoison
  spin_unlock_irq(&hugetlb_lock);
  spin_lock_irq(&hugetlb_lock);
  __folio_clear_hugetlb(folio);
   -- Hugetlb flag is cleared but too late.
  spin_unlock_irq(&hugetlb_lock);

When the above race occurs, raw error page info will be leaked.  Even
worse, raw error pages won't have hwpoisoned flag set and hit
pcplists/buddy.  Fix this issue by deferring
folio_clear_hugetlb_hwpoison() until __folio_clear_hugetlb() is done.  So
all raw error pages will have hwpoisoned flag set.

Link: https://lkml.kernel.org/r/20240708025127.107713-1-linmiaohe@huawei.com
Fixes: 32c877191e02 ("hugetlb: do not clear hugetlb dtor until allocating vmemmap")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Acked-by: Muchun Song <muchun.song@linux.dev>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agofilemap: replace pte_offset_map() with pte_offset_map_nolock()
ZhangPeng [Wed, 13 Mar 2024 01:29:13 +0000 (09:29 +0800)]
filemap: replace pte_offset_map() with pte_offset_map_nolock()

The vmf->ptl in filemap_fault_recheck_pte_none() is still set from
handle_pte_fault().  But at the same time, we did a pte_unmap(vmf->pte).
After a pte_unmap(vmf->pte) unmap and rcu_read_unlock(), the page table
may be racily changed and vmf->ptl maybe fails to protect the actual page
table.  Fix this by replacing pte_offset_map() with
pte_offset_map_nolock().

As David said, the PTL pointer might be stale so if we continue to use
it infilemap_fault_recheck_pte_none(), it might trigger UAF.  Also, if
the PTL fails, the issue fixed by commit 58f327f2ce80 ("filemap: avoid
unnecessary major faults in filemap_fault()") might reappear.

Link: https://lkml.kernel.org/r/20240313012913.2395414-1-zhangpeng362@huawei.com
Fixes: 58f327f2ce80 ("filemap: avoid unnecessary major faults in filemap_fault()")
Signed-off-by: ZhangPeng <zhangpeng362@huawei.com>
Suggested-by: David Hildenbrand <david@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Cc: "Huang, Ying" <ying.huang@intel.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Kefeng Wang <wangkefeng.wang@huawei.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Nanyong Sun <sunnanyong@huawei.com>
Cc: Yang Shi <shy828301@gmail.com>
Cc: Yin Fengwei <fengwei.yin@intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agoarch/xtensa: always_inline get_current() and current_thread_info()
Suren Baghdasaryan [Thu, 4 Jul 2024 13:25:06 +0000 (06:25 -0700)]
arch/xtensa: always_inline get_current() and current_thread_info()

Mark get_current() and current_thread_info() functions as always_inline to
fix the following modpost warning:

WARNING: modpost: vmlinux: section mismatch in reference: get_current+0xc (section: .text.unlikely) -> initcall_level_names (section: .init.data)

The warning happens when these functions are called from an __init
function and they don't get inlined (remain in the .text section) while
the value they return points into .init.data section.  Assuming
get_current() always returns a valid address, this situation can happen
only during init stage and accessing .init.data from .text section during
that stage should pose no issues.

Link: https://lkml.kernel.org/r/20240704132506.1011978-2-surenb@google.com
Fixes: 22d407b164ff ("lib: add allocation tagging support for memory allocation profiling")
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Chris Zankel <chris@zankel.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: kernel test robot <lkp@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agosched.h: always_inline alloc_tag_{save|restore} to fix modpost warnings
Suren Baghdasaryan [Thu, 4 Jul 2024 13:25:05 +0000 (06:25 -0700)]
sched.h: always_inline alloc_tag_{save|restore} to fix modpost warnings

Mark alloc_tag_{save|restore} as always_inline to fix the following
modpost warnings:

WARNING: modpost: vmlinux: section mismatch in reference: alloc_tag_save+0x1c (section: .text.unlikely) -> initcall_level_names (section: .init.data)
WARNING: modpost: vmlinux: section mismatch in reference: alloc_tag_restore+0x3c (section: .text.unlikely) -> initcall_level_names (section: .init.data)

The warnings happen when these functions are called from an __init
function and they don't get inlined (remain in the .text section) while
the value returned by get_current() points into .init.data section.
Assuming get_current() always returns a valid address, this situation can
happen only during init stage and accessing .init.data from .text section
during that stage should pose no issues.

Link: https://lkml.kernel.org/r/20240704132506.1011978-1-surenb@google.com
Fixes: 22d407b164ff ("lib: add allocation tagging support for memory allocation profiling")
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202407032306.gi9nZsBi-lkp@intel.com/
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Chris Zankel <chris@zankel.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agoMerge tag '6.10-rc6-smb3-server-fixes' of git://git.samba.org/ksmbd
Linus Torvalds [Tue, 9 Jul 2024 15:16:18 +0000 (08:16 -0700)]
Merge tag '6.10-rc6-smb3-server-fixes' of git://git.samba.org/ksmbd

Pull smb server fixes from Steve French:

 - fix access flags to address fuse incompatibility

 - fix device type returned by get filesystem info

* tag '6.10-rc6-smb3-server-fixes' of git://git.samba.org/ksmbd:
  ksmbd: discard write access to the directory open
  ksmbd: return FILE_DEVICE_DISK instead of super magic

12 months agoMerge tag 'linux_kselftest-fixes-6.10' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 9 Jul 2024 15:11:39 +0000 (08:11 -0700)]
Merge tag 'linux_kselftest-fixes-6.10' of git://git./linux/kernel/git/shuah/linux-kselftest

Pull kselftest fixes from Shuah Khan
 "Fixes to clang build failures to timerns, vDSO tests and fixes to vDSO
  makefile"

* tag 'linux_kselftest-fixes-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
  selftests/vDSO: remove duplicate compiler invocations from Makefile
  selftests/vDSO: remove partially duplicated "all:" target in Makefile
  selftests/vDSO: fix clang build errors and warnings
  selftest/timerns: fix clang build failures for abs() calls

12 months agos390/mm: Add NULL pointer check to crst_table_free() base_crst_free()
Heiko Carstens [Tue, 9 Jul 2024 06:50:56 +0000 (08:50 +0200)]
s390/mm: Add NULL pointer check to crst_table_free() base_crst_free()

crst_table_free() used to work with NULL pointers before the conversion
to ptdescs.  Since crst_table_free() can be called with a NULL pointer
(error handling in crst_table_upgrade() add an explicit check.

Also add the same check to base_crst_free() for consistency reasons.

In real life this should not happen, since order two GFP_KERNEL
allocations will not fail, unless FAIL_PAGE_ALLOC is enabled and used.

Reported-by: Yunseong Kim <yskelg@gmail.com>
Fixes: 6326c26c1514 ("s390: convert various pgalloc functions to use ptdescs")
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 months agoof/irq: Disable "interrupt-map" parsing for PASEMI Nemo
Marc Zyngier [Tue, 2 Jul 2024 21:42:46 +0000 (22:42 +0100)]
of/irq: Disable "interrupt-map" parsing for PASEMI Nemo

Once again, we've broken PASEMI Nemo boards with its incomplete
"interrupt-map" translations. Commit 935df1bd40d4 ("of/irq: Factor out
parsing of interrupt-map parent phandle+args from of_irq_parse_raw()")
changed the behavior resulting in the existing work-around not taking
effect. Rework the work-around to just skip parsing "interrupt-map" up
front by using the of_irq_imap_abusers list.

Fixes: 935df1bd40d4 ("of/irq: Factor out parsing of interrupt-map parent phandle+args from of_irq_parse_raw()")
Reported-by: Christian Zigotzky <chzigotzky@xenosoft.de>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/86ed8ba2sp.wl-maz@kernel.org
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
12 months agoMerge tag 'perf-tools-fixes-for-v6.10-2024-07-08' of git://git.kernel.org/pub/scm...
Linus Torvalds [Mon, 8 Jul 2024 21:08:43 +0000 (14:08 -0700)]
Merge tag 'perf-tools-fixes-for-v6.10-2024-07-08' of git://git./linux/kernel/git/perf/perf-tools

Pull perf tools fixes from Namhyung Kim:
 "Fix performance issue for v6.10

  These address the performance issues reported by Matt, Namhyung and
  Linus. Recently perf changed the processing of the comm string and DSO
  using sorted arrays but this caused it to sort the array whenever
  adding a new entry.

  This caused a performance issue and the fix is to enhance the sorting
  by finding the insertion point in the sorted array and to shift
  righthand side using memmove()"

* tag 'perf-tools-fixes-for-v6.10-2024-07-08' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools:
  perf dsos: When adding a dso into sorted dsos maintain the sort order
  perf comm str: Avoid sort during insert

12 months agothermal: core: Fix list sorting in __thermal_zone_device_update()
Rafael J. Wysocki [Mon, 8 Jul 2024 15:16:00 +0000 (17:16 +0200)]
thermal: core: Fix list sorting in __thermal_zone_device_update()

The order in which lists are sorted in __thermal_zone_device_update()
is reverse with respect to what it should be due to a mistake in
thermal_trip_notify_cmp().

Fix it and observe that it is not necessary to sort the lists in
different orders.  They can both be sorted in ascending order if
way_down_list is walked in reverse order which allows the code to
be slightly more straightforward (and less prone to silly mistakes).

Fixes: 7454f2c42cce ("thermal: core: Sort trip point crossing notifications by temperature")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://patch.msgid.link/12481676.O9o76ZdvQC@rjwysocki.net
12 months agoperf dsos: When adding a dso into sorted dsos maintain the sort order
Ian Rogers [Wed, 3 Jul 2024 17:21:17 +0000 (10:21 -0700)]
perf dsos: When adding a dso into sorted dsos maintain the sort order

dsos__add would add at the end of the dso array possibly requiring a
later find to re-sort the array. Patterns of find then add were
becoming O(n*log n) due to the sorts. Change the add routine to be
O(n) rather than O(1) but to maintain the sorted-ness of the dsos
array so that later finds don't need the O(n*log n) sort.

Fixes: 3f4ac23a9908 ("perf dsos: Switch backing storage to array from rbtree/list")
Reported-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Steinar Gunderson <sesse@google.com>
Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Cc: Matt Fleming <matt@readmodwrite.com>
Link: https://lore.kernel.org/r/20240703172117.810918-3-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
12 months agoperf comm str: Avoid sort during insert
Ian Rogers [Wed, 3 Jul 2024 17:21:16 +0000 (10:21 -0700)]
perf comm str: Avoid sort during insert

The array is sorted, so just move the elements and insert in order.

Fixes: 13ca628716c6 ("perf comm: Add reference count checking to 'struct comm_str'")
Reported-by: Matt Fleming <matt@readmodwrite.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Matt Fleming <matt@readmodwrite.com>
Cc: Steinar Gunderson <sesse@google.com>
Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Link: https://lore.kernel.org/r/20240703172117.810918-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
12 months agoLinux 6.10-rc7 v6.10-rc7
Linus Torvalds [Sun, 7 Jul 2024 21:23:46 +0000 (14:23 -0700)]
Linux 6.10-rc7

12 months agoMerge tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 7 Jul 2024 17:59:38 +0000 (10:59 -0700)]
Merge tag 'clk-fixes-for-linus' of git://git./linux/kernel/git/clk/linux

Pull clk fixes from Stephen Boyd:
 "A set of clk fixes for the Qualcomm, Mediatek, and Allwinner drivers:

   - Fix the Qualcomm Stromer Plus PLL set_rate() clk_op to explicitly
     set the alpha enable bit and not set bits that don't exist

   - Mark Qualcomm IPQ9574 crypto clks as voted to avoid stuck clk
     warnings

   - Fix the parent of some PLLs on Qualcomm sm6530 so their rate is
     correct

   - Fix the min/max rate clamping logic in the Allwinner driver that
     got broken in v6.9

   - Limit runtime PM enabling in the Mediatek driver to only
     mt8183-mfgcfg so that system wide resume doesn't break on other
     Mediatek SoCs"

* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
  clk: mediatek: mt8183: Only enable runtime PM on mt8183-mfgcfg
  clk: sunxi-ng: common: Don't call hw_to_ccu_common on hw without common
  clk: qcom: gcc-ipq9574: Add BRANCH_HALT_VOTED flag
  clk: qcom: apss-ipq-pll: remove 'config_ctl_hi_val' from Stromer pll configs
  clk: qcom: clk-alpha-pll: set ALPHA_EN bit for Stromer Plus PLLs
  clk: qcom: gcc-sm6350: Fix gpll6* & gpll7 parents

12 months agoMerge tag 'powerpc-6.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sun, 7 Jul 2024 01:31:24 +0000 (18:31 -0700)]
Merge tag 'powerpc-6.10-4' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:

 - Fix unnecessary copy to 0 when kernel is booted at address 0

 - Fix usercopy crash when dumping dtl via debugfs

 - Avoid possible crash when PCI hotplug races with error handling

 - Fix kexec crash caused by scv being disabled before other CPUs
   call-in

 - Fix powerpc selftests build with USERCFLAGS set

Thanks to Anjali K, Ganesh Goudar, Gautam Menghani, Jinglin Wen,
Nicholas Piggin, Sourabh Jain, Srikar Dronamraju, and Vishal Chourasia.

* tag 'powerpc-6.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  selftests/powerpc: Fix build with USERCFLAGS set
  powerpc/pseries: Fix scv instruction crash with kexec
  powerpc/eeh: avoid possible crash when edev->pdev changes
  powerpc/pseries: Whitelist dtl slub object for copying to userspace
  powerpc/64s: Fix unnecessary copy to 0 when kernel is booted at address 0

12 months agoMerge tag '6.10-rc6-smb3-client-fix' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sat, 6 Jul 2024 23:16:58 +0000 (16:16 -0700)]
Merge tag '6.10-rc6-smb3-client-fix' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client fix from Steve French:
 "Fix for smb3 readahead performance regression"

* tag '6.10-rc6-smb3-client-fix' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: Fix read-performance regression by dropping readahead expansion

12 months agoMAINTAINERS: mailmap: update Lorenzo Stoakes's email address
Lorenzo Stoakes [Wed, 3 Jul 2024 09:27:04 +0000 (10:27 +0100)]
MAINTAINERS: mailmap: update Lorenzo Stoakes's email address

Now working at Oracle.

Link: https://lkml.kernel.org/r/20240703092704.11571-1-lorenzo.stoakes@oracle.com
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm: fix crashes from deferred split racing folio migration
Hugh Dickins [Tue, 2 Jul 2024 07:40:55 +0000 (00:40 -0700)]
mm: fix crashes from deferred split racing folio migration

Even on 6.10-rc6, I've been seeing elusive "Bad page state"s (often on
flags when freeing, yet the flags shown are not bad: PG_locked had been
set and cleared??), and VM_BUG_ON_PAGE(page_ref_count(page) == 0)s from
deferred_split_scan()'s folio_put(), and a variety of other BUG and WARN
symptoms implying double free by deferred split and large folio migration.

6.7 commit 9bcef5973e31 ("mm: memcg: fix split queue list crash when large
folio migration") was right to fix the memcg-dependent locking broken in
85ce2c517ade ("memcontrol: only transfer the memcg data for migration"),
but missed a subtlety of deferred_split_scan(): it moves folios to its own
local list to work on them without split_queue_lock, during which time
folio->_deferred_list is not empty, but even the "right" lock does nothing
to secure the folio and the list it is on.

Fortunately, deferred_split_scan() is careful to use folio_try_get(): so
folio_migrate_mapping() can avoid the race by folio_undo_large_rmappable()
while the old folio's reference count is temporarily frozen to 0 - adding
such a freeze in the !mapping case too (originally, folio lock and
unmapping and no swap cache left an anon folio unreachable, so no freezing
was needed there: but the deferred split queue offers a way to reach it).

Link: https://lkml.kernel.org/r/29c83d1a-11ca-b6c9-f92e-6ccb322af510@google.com
Fixes: 9bcef5973e31 ("mm: memcg: fix split queue list crash when large folio migration")
Signed-off-by: Hugh Dickins <hughd@google.com>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Barry Song <baohua@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Kefeng Wang <wangkefeng.wang@huawei.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Nhat Pham <nphamcs@gmail.com>
Cc: Yang Shi <shy828301@gmail.com>
Cc: Zi Yan <ziy@nvidia.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agolib/build_OID_registry: avoid non-destructive substitution for Perl < 5.13.2 compat
Paul Menzel [Mon, 1 Jul 2024 15:58:01 +0000 (17:58 +0200)]
lib/build_OID_registry: avoid non-destructive substitution for Perl < 5.13.2 compat

On a system with Perl 5.12.1, commit 5ef6dc08cfde
("lib/build_OID_registry: don't mention the full path of the script in
output") causes the build to fail with the error below.

     Bareword found where operator expected at ./lib/build_OID_registry line 41, near "s#^\Q$abs_srctree/\E##r"
     syntax error at ./lib/build_OID_registry line 41, near "s#^\Q$abs_srctree/\E##r"
     Execution of ./lib/build_OID_registry aborted due to compilation errors.
     make[3]: *** [lib/Makefile:352: lib/oid_registry_data.c] Error 255

Ahmad Fatoum analyzed that non-destructive substitution is only supported since
Perl 5.13.2. Instead of dropping `r` and having the side effect of modifying
`$0`, introduce a dedicated variable to support older Perl versions.

Link: https://lkml.kernel.org/r/20240702223512.8329-2-pmenzel@molgen.mpg.de
Link: https://lkml.kernel.org/r/20240701155802.75152-1-pmenzel@molgen.mpg.de
Fixes: 5ef6dc08cfde ("lib/build_OID_registry: don't mention the full path of the script in output")
Link: https://lore.kernel.org/all/259f7a87-2692-480e-9073-1c1c35b52f67@molgen.mpg.de/
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Suggested-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Cc: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Nicolas Schier <nicolas@fjasle.eu>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm: gup: stop abusing try_grab_folio
Yang Shi [Fri, 28 Jun 2024 19:14:58 +0000 (12:14 -0700)]
mm: gup: stop abusing try_grab_folio

A kernel warning was reported when pinning folio in CMA memory when
launching SEV virtual machine.  The splat looks like:

[  464.325306] WARNING: CPU: 13 PID: 6734 at mm/gup.c:1313 __get_user_pages+0x423/0x520
[  464.325464] CPU: 13 PID: 6734 Comm: qemu-kvm Kdump: loaded Not tainted 6.6.33+ #6
[  464.325477] RIP: 0010:__get_user_pages+0x423/0x520
[  464.325515] Call Trace:
[  464.325520]  <TASK>
[  464.325523]  ? __get_user_pages+0x423/0x520
[  464.325528]  ? __warn+0x81/0x130
[  464.325536]  ? __get_user_pages+0x423/0x520
[  464.325541]  ? report_bug+0x171/0x1a0
[  464.325549]  ? handle_bug+0x3c/0x70
[  464.325554]  ? exc_invalid_op+0x17/0x70
[  464.325558]  ? asm_exc_invalid_op+0x1a/0x20
[  464.325567]  ? __get_user_pages+0x423/0x520
[  464.325575]  __gup_longterm_locked+0x212/0x7a0
[  464.325583]  internal_get_user_pages_fast+0xfb/0x190
[  464.325590]  pin_user_pages_fast+0x47/0x60
[  464.325598]  sev_pin_memory+0xca/0x170 [kvm_amd]
[  464.325616]  sev_mem_enc_register_region+0x81/0x130 [kvm_amd]

Per the analysis done by yangge, when starting the SEV virtual machine, it
will call pin_user_pages_fast(..., FOLL_LONGTERM, ...) to pin the memory.
But the page is in CMA area, so fast GUP will fail then fallback to the
slow path due to the longterm pinnalbe check in try_grab_folio().

The slow path will try to pin the pages then migrate them out of CMA area.
But the slow path also uses try_grab_folio() to pin the page, it will
also fail due to the same check then the above warning is triggered.

In addition, the try_grab_folio() is supposed to be used in fast path and
it elevates folio refcount by using add ref unless zero.  We are guaranteed
to have at least one stable reference in slow path, so the simple atomic add
could be used.  The performance difference should be trivial, but the
misuse may be confusing and misleading.

Redefined try_grab_folio() to try_grab_folio_fast(), and try_grab_page()
to try_grab_folio(), and use them in the proper paths.  This solves both
the abuse and the kernel warning.

The proper naming makes their usecase more clear and should prevent from
abusing in the future.

peterx said:

: The user will see the pin fails, for gpu-slow it further triggers the WARN
: right below that failure (as in the original report):
:
:         folio = try_grab_folio(page, page_increm - 1,
:                                 foll_flags);
:         if (WARN_ON_ONCE(!folio)) { <------------------------ here
:                 /*
:                         * Release the 1st page ref if the
:                         * folio is problematic, fail hard.
:                         */
:                 gup_put_folio(page_folio(page), 1,
:                                 foll_flags);
:                 ret = -EFAULT;
:                 goto out;
:         }

[1] https://lore.kernel.org/linux-mm/1719478388-31917-1-git-send-email-yangge1116@126.com/

[shy828301@gmail.com: fix implicit declaration of function try_grab_folio_fast]
Link: https://lkml.kernel.org/r/CAHbLzkowMSso-4Nufc9hcMehQsK9PNz3OSu-+eniU-2Mm-xjhA@mail.gmail.com
Link: https://lkml.kernel.org/r/20240628191458.2605553-1-yang@os.amperecomputing.com
Fixes: 57edfcfd3419 ("mm/gup: accelerate thp gup even for "pages != NULL"")
Signed-off-by: Yang Shi <yang@os.amperecomputing.com>
Reported-by: yangge <yangge1116@126.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: <stable@vger.kernel.org> [6.6+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agoMerge tag 'i2c-for-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sat, 6 Jul 2024 16:51:00 +0000 (09:51 -0700)]
Merge tag 'i2c-for-6.10-rc7' of git://git./linux/kernel/git/wsa/linux

Pull i2c fix from Wolfram Sang:
 "An i2c driver fix"

* tag 'i2c-for-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: pnx: Fix potential deadlock warning from del_timer_sync() call in isr

12 months agoselftests/powerpc: Fix build with USERCFLAGS set
Michael Ellerman [Sat, 6 Jul 2024 12:08:33 +0000 (22:08 +1000)]
selftests/powerpc: Fix build with USERCFLAGS set

Currently building the powerpc selftests with USERCFLAGS set to anything
causes the build to break:

  $ make -C tools/testing/selftests/powerpc V=1 USERCFLAGS=-Wno-error
  ...
  gcc -Wno-error    cache_shape.c ...
  cache_shape.c:18:10: fatal error: utils.h: No such file or directory
     18 | #include "utils.h"
        |          ^~~~~~~~~
  compilation terminated.

This happens because the USERCFLAGS are added to CFLAGS in lib.mk, which
causes the check of CFLAGS in powerpc/flags.mk to skip setting CFLAGS at
all, resulting in none of the usual CFLAGS being passed. That can
be seen in the output above, the only flag passed to the compiler is
-Wno-error.

Fix it by dropping the conditional setting of CFLAGS in flags.mk.
Instead always set CFLAGS, but also append USERCFLAGS if they are set.

Note that appending to CFLAGS (with +=) wouldn't work, because flags.mk
is included by multiple Makefiles (to support partial builds), causing
CFLAGS to be appended to multiple times. Additionally that would place
the USERCFLAGS prior to the standard CFLAGS, meaning the USERCFLAGS
couldn't override the standard flags. Being able to override the
standard flags is desirable, for example for adding -Wno-error.

With the fix in place, the CFLAGS are set correctly, including the
USERCFLAGS:

  $ make -C tools/testing/selftests/powerpc V=1 USERCFLAGS=-Wno-error
  ...
  gcc -std=gnu99 -O2 -Wall -Werror -DGIT_VERSION='"v6.10-rc2-7-gdea17e7e56c3"'
  -I/home/michael/linux/tools/testing/selftests/powerpc/include -Wno-error
  cache_shape.c ...

Fixes: 5553a79387e9 ("selftests/powerpc: Add flags.mk to support pmu buildable")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20240706120833.909853-1-mpe@ellerman.id.au
12 months agohfsplus: fix uninit-value in copy_name
Edward Adam Davis [Tue, 21 May 2024 05:21:46 +0000 (13:21 +0800)]
hfsplus: fix uninit-value in copy_name

[syzbot reported]
BUG: KMSAN: uninit-value in sized_strscpy+0xc4/0x160
 sized_strscpy+0xc4/0x160
 copy_name+0x2af/0x320 fs/hfsplus/xattr.c:411
 hfsplus_listxattr+0x11e9/0x1a50 fs/hfsplus/xattr.c:750
 vfs_listxattr fs/xattr.c:493 [inline]
 listxattr+0x1f3/0x6b0 fs/xattr.c:840
 path_listxattr fs/xattr.c:864 [inline]
 __do_sys_listxattr fs/xattr.c:876 [inline]
 __se_sys_listxattr fs/xattr.c:873 [inline]
 __x64_sys_listxattr+0x16b/0x2f0 fs/xattr.c:873
 x64_sys_call+0x2ba0/0x3b50 arch/x86/include/generated/asm/syscalls_64.h:195
 do_syscall_x64 arch/x86/entry/common.c:52 [inline]
 do_syscall_64+0xcf/0x1e0 arch/x86/entry/common.c:83
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

Uninit was created at:
 slab_post_alloc_hook mm/slub.c:3877 [inline]
 slab_alloc_node mm/slub.c:3918 [inline]
 kmalloc_trace+0x57b/0xbe0 mm/slub.c:4065
 kmalloc include/linux/slab.h:628 [inline]
 hfsplus_listxattr+0x4cc/0x1a50 fs/hfsplus/xattr.c:699
 vfs_listxattr fs/xattr.c:493 [inline]
 listxattr+0x1f3/0x6b0 fs/xattr.c:840
 path_listxattr fs/xattr.c:864 [inline]
 __do_sys_listxattr fs/xattr.c:876 [inline]
 __se_sys_listxattr fs/xattr.c:873 [inline]
 __x64_sys_listxattr+0x16b/0x2f0 fs/xattr.c:873
 x64_sys_call+0x2ba0/0x3b50 arch/x86/include/generated/asm/syscalls_64.h:195
 do_syscall_x64 arch/x86/entry/common.c:52 [inline]
 do_syscall_64+0xcf/0x1e0 arch/x86/entry/common.c:83
 entry_SYSCALL_64_after_hwframe+0x77/0x7f
[Fix]
When allocating memory to strbuf, initialize memory to 0.

Reported-and-tested-by: syzbot+efde959319469ff8d4d7@syzkaller.appspotmail.com
Signed-off-by: Edward Adam Davis <eadavis@qq.com>
Link: https://lore.kernel.org/r/tencent_8BBB6433BC9E1C1B7B4BDF1BF52574BA8808@qq.com
Reported-and-tested-by: syzbot+01ade747b16e9c8030e0@syzkaller.appspotmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
12 months agoMerge tag 'integrity-v6.10-fix' of ssh://ra.kernel.org/pub/scm/linux/kernel/git/zohar...
Linus Torvalds [Fri, 5 Jul 2024 23:21:54 +0000 (16:21 -0700)]
Merge tag 'integrity-v6.10-fix' of ssh://ra./linux/kernel/git/zohar/linux-integrity

Pull integrity fix from Mimi Zohar:
 "A single bug fix to properly remove all of the securityfs IMA
  measurement lists"

* tag 'integrity-v6.10-fix' of ssh://ra.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity:
  ima: fix wrong zero-assignment during securityfs dentry remove

12 months agoselftests/vDSO: remove duplicate compiler invocations from Makefile
John Hubbard [Fri, 5 Jul 2024 19:57:36 +0000 (09:57 -1000)]
selftests/vDSO: remove duplicate compiler invocations from Makefile

The Makefile open-codes compiler invocations that ../lib.mk already
provides.

Avoid this by using a Make feature that allows setting per-target
variables, which in this case are: CFLAGS and LDFLAGS. This approach
generates the exact same compiler invocations as before, but removes all
of the code duplication, along with the quirky mangled variable names.
So now the Makefile is smaller, less unusual, and easier to read.

The new dependencies are listed after including lib.mk, in order to
let lib.mk provide the first target ("all:"), and are grouped together
with their respective source file dependencies, for visual clarity.

Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
12 months agoselftests/vDSO: remove partially duplicated "all:" target in Makefile
John Hubbard [Fri, 5 Jul 2024 19:57:35 +0000 (09:57 -1000)]
selftests/vDSO: remove partially duplicated "all:" target in Makefile

There were a couple of errors here:

1. TEST_GEN_PROGS was incorrectly prepending $(OUTPUT) to each program
to be built. However, lib.mk already does that because it assumes "bare"
program names are passed in, so this ended up creating
$(OUTPUT)/$(OUTPUT)/file.c, which of course won't work as intended.

2. lib.mk was included before TEST_GEN_PROGS was set, which led to
lib.mk's "all:" target not seeing anything to rebuild.

So nothing worked, which caused the author to force things by creating
an "all:" target locally--while still including ../lib.mk.

Fix all of this by including ../lib.mk at the right place, and removing
the $(OUTPUT) prefix to the programs to be built, and removing the
duplicate "all:" target.

Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
12 months agoselftests/vDSO: fix clang build errors and warnings
John Hubbard [Fri, 5 Jul 2024 19:57:34 +0000 (09:57 -1000)]
selftests/vDSO: fix clang build errors and warnings

When building with clang, via:

    make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and
allows these tests to run and pass.

1. Fix linker error (undefined reference to memcpy) by providing a local
   version of memcpy.

2. clang complains about using this form:

    if (g = h & 0xf0000000)

...so factor out the assignment into a separate step.

3. The code is passing a signed const char* to elf_hash(), which expects
   a const unsigned char *. There are several callers, so fix this at
   the source by allowing the function to accept a signed argument, and
   then converting to unsigned operations, once inside the function.

4. clang doesn't have __attribute__((externally_visible)) and generates
   a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem
   to require that attribute in order to build, run and pass tests here,
   so remove it.

Reviewed-by: Carlos Llamas <cmllamas@google.com>
Reviewed-by: Edward Liaw <edliaw@google.com>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Tested-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
12 months agoMerge tag 'pci-v6.10-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci
Linus Torvalds [Fri, 5 Jul 2024 19:33:00 +0000 (12:33 -0700)]
Merge tag 'pci-v6.10-fixes-2' of git://git./linux/kernel/git/pci/pci

Pull pci update from Bjorn Helgaas:

 - Update MAINTAINERS and CREDITS to credit Gustavo Pimentel with the
   Synopsys DesignWare eDMA driver and reflect that he is no longer at
   Synopsys and isn't in a position to maintain the DesignWare xData
   traffic generator (Bjorn Helgaas)

* tag 'pci-v6.10-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci:
  CREDITS: Add Synopsys DesignWare eDMA driver for Gustavo Pimentel
  MAINTAINERS: Orphan Synopsys DesignWare xData traffic generator

12 months agoMerge tag 'riscv-for-linus-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 5 Jul 2024 19:22:51 +0000 (12:22 -0700)]
Merge tag 'riscv-for-linus-6.10-rc7' of git://git./linux/kernel/git/riscv/linux

Pull RISC-V fixes from Palmer Dabbelt:

 - A fix for the CMODX example in the recently added icache flushing
   prctl()

 - A fix to the perf driver to avoid corrupting event data on counter
   overflows when external overflow handlers are in use

 - A fix to clear all hardware performance monitor events on boot, to
   avoid dangling events firmware or previously booted kernels from
   triggering spuriously

 - A fix to the perf event probing logic to avoid erroneously reporting
   the presence of unimplemented counters. This also prevents some
   implemented counters from being reported

 - A build fix for the vector sigreturn selftest on clang

 - A fix to ftrace, which now requires the previously optional index
   argument to ftrace_graph_ret_addr()

 - A fix to avoid deadlocking if kexec crash handling triggers in an
   interrupt context

* tag 'riscv-for-linus-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
  riscv: kexec: Avoid deadlock in kexec crash path
  riscv: stacktrace: fix usage of ftrace_graph_ret_addr()
  riscv: selftests: Fix vsetivli args for clang
  perf: RISC-V: Check standard event availability
  drivers/perf: riscv: Reset the counter to hpmevent mapping while starting cpus
  drivers/perf: riscv: Do not update the event data if uptodate
  documentation: Fix riscv cmodx example

12 months agoselftest/timerns: fix clang build failures for abs() calls
John Hubbard [Thu, 4 Jul 2024 02:52:47 +0000 (19:52 -0700)]
selftest/timerns: fix clang build failures for abs() calls

When building with clang, via:

    make LLVM=1 -C tools/testing/selftests

...clang warns about mismatches between the expected and required
integer length being supplied to abs(3).

Fix this by using the correct variant of abs(3): labs(3) or llabs(3), in
these cases.

Reviewed-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Acked-by: Andrei Vagin <avagin@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
12 months agoMerge tag 'drm-fixes-2024-07-05' of https://gitlab.freedesktop.org/drm/kernel
Linus Torvalds [Fri, 5 Jul 2024 18:53:40 +0000 (11:53 -0700)]
Merge tag 'drm-fixes-2024-07-05' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Daniel Vetter:
 "Just small fixes all over here, all quiet as it should.

  drivers:

   - amd: mostly amdgpu display fixes + radeon vm NULL deref fix

   - xe: migration error handling + typoed register name in gt setup

   - i915: usb-c fix to shut up warnings on MTL+

   - panthor: fix sync-only jobs + ioctl validation fix to not EINVAL
     wrongly

   - panel quirks

   - nouveau: NULL deref in get_modes

  drm core:

   - fbdev big endian fix for the dma memory backed variant

  drivers/firmware:

   - fix sysfb refcounting"

* tag 'drm-fixes-2024-07-05' of https://gitlab.freedesktop.org/drm/kernel:
  drm/xe/mcr: Avoid clobbering DSS steering
  drm/xe: fix error handling in xe_migrate_update_pgtables
  drm/ttm: Always take the bo delayed cleanup path for imported bos
  drm/fbdev-generic: Fix framebuffer on big endian devices
  drm/panthor: Fix sync-only jobs
  drm/panthor: Don't check the array stride on empty uobj arrays
  drm/amdgpu/atomfirmware: silence UBSAN warning
  drm/radeon: check bo_va->bo is non-NULL before using it
  drm/amd/display: Fix array-index-out-of-bounds in dml2/FCLKChangeSupport
  drm/amd/display: Update efficiency bandwidth for dcn351
  drm/amd/display: Fix refresh rate range for some panel
  drm/amd/display: Account for cursor prefetch BW in DML1 mode support
  drm/amd/display: Add refresh rate range check
  drm/amd/display: Reset freesync config before update new state
  drm: panel-orientation-quirks: Add labels for both Valve Steam Deck revisions
  drm: panel-orientation-quirks: Add quirk for Valve Galileo
  drm/i915/display: For MTL+ platforms skip mg dp programming
  drm/nouveau: fix null pointer dereference in nouveau_connector_get_modes
  firmware: sysfb: Fix reference count of sysfb parent device

12 months agoMerge tag 'gpio-fixes-for-v6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 5 Jul 2024 18:39:30 +0000 (11:39 -0700)]
Merge tag 'gpio-fixes-for-v6.10-rc7' of git://git./linux/kernel/git/brgl/linux

Pull gpio fixes from Bartosz Golaszewski:
 "Two OF lookup quirks and one fix for an issue in the generic gpio-mmio
  driver:

   - add two OF lookup quirks for TSC2005 and MIPS Lantiq

   - don't try to figure out bgpio_bits from the 'ngpios' property in
     gpio-mmio"

* tag 'gpio-fixes-for-v6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
  gpiolib: of: add polarity quirk for TSC2005
  gpio: mmio: do not calculate bgpio_bits via "ngpios"
  gpiolib: of: fix lookup quirk for MIPS Lantiq

12 months agoMerge tag 'tpmdd-next-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 5 Jul 2024 18:30:57 +0000 (11:30 -0700)]
Merge tag 'tpmdd-next-6.10-rc7' of git://git./linux/kernel/git/jarkko/linux-tpmdd

Pull TPM fixes from Jarkko Sakkinen:
 "This contains the fixes for !chip->auth condition, preventing the
  breakage of:
   - tpm_ftpm_tee.c
   - tpm_i2c_nuvoton.c
   - tpm_ibmvtpm.c
   - tpm_tis_i2c_cr50.c
   - tpm_vtpm_proxy.c

  All drivers will continue to work as they did in 6.9, except a single
  warning (dev_warn() not WARN()) is printed to klog only to inform that
  authenticated sessions are not enabled"

* tag 'tpmdd-next-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/jarkko/linux-tpmdd:
  tpm: Address !chip->auth in tpm_buf_append_hmac_session*()
  tpm: Address !chip->auth in tpm_buf_append_name()
  tpm: Address !chip->auth in tpm2_*_auth_session()

12 months agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Fri, 5 Jul 2024 18:23:30 +0000 (11:23 -0700)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm fix from Paolo Bonzini:

 - s390: fix support for z16 systems

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: s390: fix LPSWEY handling

12 months agovfs: don't mod negative dentry count when on shrinker list
Brian Foster [Wed, 3 Jul 2024 12:13:01 +0000 (08:13 -0400)]
vfs: don't mod negative dentry count when on shrinker list

The nr_dentry_negative counter is intended to only account negative
dentries that are present on the superblock LRU. Therefore, the LRU
add, remove and isolate helpers modify the counter based on whether
the dentry is negative, but the shrinker list related helpers do not
modify the counter, and the paths that change a dentry between
positive and negative only do so if DCACHE_LRU_LIST is set.

The problem with this is that a dentry on a shrinker list still has
DCACHE_LRU_LIST set to indicate ->d_lru is in use. The additional
DCACHE_SHRINK_LIST flag denotes whether the dentry is on LRU or a
shrink related list. Therefore if a relevant operation (i.e. unlink)
occurs while a dentry is present on a shrinker list, and the
associated codepath only checks for DCACHE_LRU_LIST, then it is
technically possible to modify the negative dentry count for a
dentry that is off the LRU. Since the shrinker list related helpers
do not modify the negative dentry count (because non-LRU dentries
should not be included in the count) when the dentry is ultimately
removed from the shrinker list, this can cause the negative dentry
count to become permanently inaccurate.

This problem can be reproduced via a heavy file create/unlink vs.
drop_caches workload. On an 80xcpu system, I start 80 tasks each
running a 1k file create/delete loop, and one task spinning on
drop_caches. After 10 minutes or so of runtime, the idle/clean cache
negative dentry count increases from somewhere in the range of 5-10
entries to several hundred (and increasingly grows beyond
nr_dentry_unused).

Tweak the logic in the paths that turn a dentry negative or positive
to filter out the case where the dentry is present on a shrink
related list. This allows the above workload to maintain an accurate
negative dentry count.

Fixes: af0c9af1b3f6 ("fs/dcache: Track & report number of negative dentries")
Signed-off-by: Brian Foster <bfoster@redhat.com>
Link: https://lore.kernel.org/r/20240703121301.247680-1-bfoster@redhat.com
Acked-by: Ian Kent <ikent@redhat.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Waiman Long <longman@redhat.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
12 months agofilelock: fix potential use-after-free in posix_lock_inode
Jeff Layton [Tue, 2 Jul 2024 22:44:48 +0000 (18:44 -0400)]
filelock: fix potential use-after-free in posix_lock_inode

Light Hsieh reported a KASAN UAF warning in trace_posix_lock_inode().
The request pointer had been changed earlier to point to a lock entry
that was added to the inode's list. However, before the tracepoint could
fire, another task raced in and freed that lock.

Fix this by moving the tracepoint inside the spinlock, which should
ensure that this doesn't happen.

Fixes: 74f6f5912693 ("locks: fix KASAN: use-after-free in trace_event_raw_event_filelock_lock")
Link: https://lore.kernel.org/linux-fsdevel/724ffb0a2962e912ea62bb0515deadf39c325112.camel@kernel.org/
Reported-by: Light Hsieh (謝明燈) <Light.Hsieh@mediatek.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Link: https://lore.kernel.org/r/20240702-filelock-6-10-v1-1-96e766aadc98@kernel.org
Reviewed-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
12 months agoMerge patch series "cachefiles: random bugfixes"
Christian Brauner [Wed, 3 Jul 2024 08:36:18 +0000 (10:36 +0200)]
Merge patch series "cachefiles: random bugfixes"

libaokun@huaweicloud.com <libaokun@huaweicloud.com> says:

This is the third version of this patch series, in which another patch set
is subsumed into this one to avoid confusing the two patch sets.
(https://patchwork.kernel.org/project/linux-fsdevel/list/?series=854914)

We've been testing ondemand mode for cachefiles since January, and we're
almost done. We hit a lot of issues during the testing period, and this
patch series fixes some of the issues. The patches have passed internal
testing without regression.

The following is a brief overview of the patches, see the patches for
more details.

Patch 1-2: Add fscache_try_get_volume() helper function to avoid
fscache_volume use-after-free on cache withdrawal.

Patch 3: Fix cachefiles_lookup_cookie() and cachefiles_withdraw_cache()
concurrency causing cachefiles_volume use-after-free.

Patch 4: Propagate error codes returned by vfs_getxattr() to avoid
endless loops.

Patch 5-7: A read request waiting for reopen could be closed maliciously
before the reopen worker is executing or waiting to be scheduled. So
ondemand_object_worker() may be called after the info and object and even
the cache have been freed and trigger use-after-free. So use
cancel_work_sync() in cachefiles_ondemand_clean_object() to cancel the
reopen worker or wait for it to finish. Since it makes no sense to wait
for the daemon to complete the reopen request, to avoid this pointless
operation blocking cancel_work_sync(), Patch 1 avoids request generation
by the DROPPING state when the request has not been sent, and Patch 2
flushes the requests of the current object before cancel_work_sync().

Patch 8: Cyclic allocation of msg_id to avoid msg_id reuse misleading
the daemon to cause hung.

Patch 9: Hold xas_lock during polling to avoid dereferencing reqs causing
use-after-free. This issue was triggered frequently in our tests, and we
found that anolis 5.10 had fixed it. So to avoid failing the test, this
patch is pushed upstream as well.

Baokun Li (7):
  netfs, fscache: export fscache_put_volume() and add
    fscache_try_get_volume()
  cachefiles: fix slab-use-after-free in fscache_withdraw_volume()
  cachefiles: fix slab-use-after-free in cachefiles_withdraw_cookie()
  cachefiles: propagate errors from vfs_getxattr() to avoid infinite
    loop
  cachefiles: stop sending new request when dropping object
  cachefiles: cancel all requests for the object that is being dropped
  cachefiles: cyclic allocation of msg_id to avoid reuse

Hou Tao (1):
  cachefiles: wait for ondemand_object_worker to finish when dropping
    object

Jingbo Xu (1):
  cachefiles: add missing lock protection when polling

 fs/cachefiles/cache.c          | 45 ++++++++++++++++++++++++++++-
 fs/cachefiles/daemon.c         |  4 +--
 fs/cachefiles/internal.h       |  3 ++
 fs/cachefiles/ondemand.c       | 52 ++++++++++++++++++++++++++++++----
 fs/cachefiles/volume.c         |  1 -
 fs/cachefiles/xattr.c          |  5 +++-
 fs/netfs/fscache_volume.c      | 14 +++++++++
 fs/netfs/internal.h            |  2 --
 include/linux/fscache-cache.h  |  6 ++++
 include/trace/events/fscache.h |  4 +++
 10 files changed, 123 insertions(+), 13 deletions(-)

Link: https://lore.kernel.org/r/20240628062930.2467993-1-libaokun@huaweicloud.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
12 months agoksmbd: discard write access to the directory open
Hobin Woo [Fri, 5 Jul 2024 03:27:25 +0000 (12:27 +0900)]
ksmbd: discard write access to the directory open

may_open() does not allow a directory to be opened with the write access.
However, some writing flags set by client result in adding write access
on server, making ksmbd incompatible with FUSE file system. Simply, let's
discard the write access when opening a directory.

list_add corruption. next is NULL.
------------[ cut here ]------------
kernel BUG at lib/list_debug.c:26!
pc : __list_add_valid+0x88/0xbc
lr : __list_add_valid+0x88/0xbc
Call trace:
__list_add_valid+0x88/0xbc
fuse_finish_open+0x11c/0x170
fuse_open_common+0x284/0x5e8
fuse_dir_open+0x14/0x24
do_dentry_open+0x2a4/0x4e0
dentry_open+0x50/0x80
smb2_open+0xbe4/0x15a4
handle_ksmbd_work+0x478/0x5ec
process_one_work+0x1b4/0x448
worker_thread+0x25c/0x430
kthread+0x104/0x1d4
ret_from_fork+0x10/0x20

Cc: stable@vger.kernel.org
Signed-off-by: Yoonho Shin <yoonho.shin@samsung.com>
Signed-off-by: Hobin Woo <hobin.woo@samsung.com>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
12 months agoMerge tag 'i2c-host-fixes-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel...
Wolfram Sang [Fri, 5 Jul 2024 14:08:55 +0000 (16:08 +0200)]
Merge tag 'i2c-host-fixes-6.10-rc7' of git://git./linux/kernel/git/andi.shyti/linux into i2c/for-current

This tag includes a nice fix in the PNX driver that has been
pending for a long time. Piotr has replaced a potential lock in
the interrupt context with a more efficient and straightforward
handling of the timeout signaling.

12 months agoMerge tag 'amd-drm-fixes-6.10-2024-07-03' of https://gitlab.freedesktop.org/agd5f...
Daniel Vetter [Fri, 5 Jul 2024 10:54:13 +0000 (12:54 +0200)]
Merge tag 'amd-drm-fixes-6.10-2024-07-03' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-6.10-2024-07-03:

amdgpu:
- Freesync fixes
- DML1 bandwidth fix
- DCN 3.5 fixes
- DML2 fix
- Silence an UBSAN warning

radeon:
- GPUVM fix

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240703184723.1981997-1-alexander.deucher@amd.com
12 months agogpiolib: of: add polarity quirk for TSC2005
Dmitry Torokhov [Wed, 3 Jul 2024 18:26:09 +0000 (11:26 -0700)]
gpiolib: of: add polarity quirk for TSC2005

DTS for Nokia N900 incorrectly specifies "active high" polarity for
the reset line, while the chip documentation actually specifies it as
"active low".  In the past the driver fudged gpiod API and inverted
the logic internally, but it was changed in d0d89493bff8.

Fixes: d0d89493bff8 ("Input: tsc2004/5 - switch to using generic device properties")
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/ZoWXwYtwgJIxi-hD@google.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
12 months agoMerge tag 'kvm-s390-master-6.10-1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Paolo Bonzini [Fri, 5 Jul 2024 08:45:53 +0000 (04:45 -0400)]
Merge tag 'kvm-s390-master-6.10-1' of git://git./linux/kernel/git/kvms390/linux into HEAD

KVM: s390: Fix z16 support

The z16 support might fail with the lpswey instruction. Provide a
handler.

12 months agoscsi: sd: Do not repeat the starting disk message
Damien Le Moal [Mon, 1 Jul 2024 21:53:26 +0000 (06:53 +0900)]
scsi: sd: Do not repeat the starting disk message

The SCSI disk message "Starting disk" to signal resuming of a suspended
disk is printed in both sd_resume() and sd_resume_common() which results
in this message being printed twice when resuming from e.g. autosuspend:

$ echo 5000 > /sys/block/sda/device/power/autosuspend_delay_ms
$ echo auto > /sys/block/sda/device/power/control

[ 4962.438293] sd 0:0:0:0: [sda] Synchronizing SCSI cache
[ 4962.501121] sd 0:0:0:0: [sda] Stopping disk

$ echo on > /sys/block/sda/device/power/control

[ 4972.805851] sd 0:0:0:0: [sda] Starting disk
[ 4980.558806] sd 0:0:0:0: [sda] Starting disk

Fix this double print by removing the call to sd_printk() from sd_resume()
and moving the call to sd_printk() in sd_resume_common() earlier in the
function, before the check using sd_do_start_stop().  Doing so, the message
is printed once regardless if sd_resume_common() actually executes
sd_start_stop_device() (i.e. SCSI device case) or not (libsas and libata
managed ATA devices case).

Fixes: 0c76106cb975 ("scsi: sd: Fix TCG OPAL unlock on system resume")
Cc: stable@vger.kernel.org
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Link: https://lore.kernel.org/r/20240701215326.128067-1-dlemoal@kernel.org
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: John Garry <john.g.garry@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
12 months agoscsi: ufs: core: Fix ufshcd_abort_one racing issue
Peter Wang [Fri, 28 Jun 2024 07:00:30 +0000 (15:00 +0800)]
scsi: ufs: core: Fix ufshcd_abort_one racing issue

When ufshcd_abort_one is racing with the completion ISR, the completed tag
of the request's mq_hctx pointer will be set to NULL by ISR.  Return
success when request is completed by ISR because ufshcd_abort_one does not
need to do anything.

The racing flow is:

Thread A
ufshcd_err_handler step 1
...
ufshcd_abort_one
ufshcd_try_to_abort_task
ufshcd_cmd_inflight(true) step 3
ufshcd_mcq_req_to_hwq
blk_mq_unique_tag
rq->mq_hctx->queue_num step 5

Thread B
ufs_mtk_mcq_intr(cq complete ISR) step 2
scsi_done
...
__blk_mq_free_request
rq->mq_hctx = NULL; step 4

Below is KE back trace.
  ufshcd_try_to_abort_task: cmd at tag 41 not pending in the device.
  ufshcd_try_to_abort_task: cmd at tag=41 is cleared.
  Aborting tag 41 / CDB 0x28 succeeded
  Unable to handle kernel NULL pointer dereference at virtual address 0000000000000194
  pc : [0xffffffddd7a79bf8] blk_mq_unique_tag+0x8/0x14
  lr : [0xffffffddd6155b84] ufshcd_mcq_req_to_hwq+0x1c/0x40 [ufs_mediatek_mod_ise]
   do_mem_abort+0x58/0x118
   el1_abort+0x3c/0x5c
   el1h_64_sync_handler+0x54/0x90
   el1h_64_sync+0x68/0x6c
   blk_mq_unique_tag+0x8/0x14
   ufshcd_err_handler+0xae4/0xfa8 [ufs_mediatek_mod_ise]
   process_one_work+0x208/0x4fc
   worker_thread+0x228/0x438
   kthread+0x104/0x1d4
   ret_from_fork+0x10/0x20

Fixes: 93e6c0e19d5b ("scsi: ufs: core: Clear cmd if abort succeeds in MCQ mode")
Suggested-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Link: https://lore.kernel.org/r/20240628070030.30929-3-peter.wang@mediatek.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
12 months agoscsi: ufs: core: Fix ufshcd_clear_cmd racing issue
Peter Wang [Fri, 28 Jun 2024 07:00:29 +0000 (15:00 +0800)]
scsi: ufs: core: Fix ufshcd_clear_cmd racing issue

When ufshcd_clear_cmd is racing with the completion ISR, the completed tag
of the request's mq_hctx pointer will be set to NULL by the ISR.  And
ufshcd_clear_cmd's call to ufshcd_mcq_req_to_hwq will get NULL pointer KE.
Return success when the request is completed by ISR because sq does not
need cleanup.

The racing flow is:

Thread A
ufshcd_err_handler step 1
ufshcd_try_to_abort_task
ufshcd_cmd_inflight(true) step 3
ufshcd_clear_cmd
...
ufshcd_mcq_req_to_hwq
blk_mq_unique_tag
rq->mq_hctx->queue_num step 5

Thread B
ufs_mtk_mcq_intr(cq complete ISR) step 2
scsi_done
...
__blk_mq_free_request
rq->mq_hctx = NULL; step 4

Below is KE back trace:

  ufshcd_try_to_abort_task: cmd pending in the device. tag = 6
  Unable to handle kernel NULL pointer dereference at virtual address 0000000000000194
   pc : [0xffffffd589679bf8] blk_mq_unique_tag+0x8/0x14
   lr : [0xffffffd5862f95b4] ufshcd_mcq_sq_cleanup+0x6c/0x1cc [ufs_mediatek_mod_ise]
   Workqueue: ufs_eh_wq_0 ufshcd_err_handler [ufs_mediatek_mod_ise]
   Call trace:
    dump_backtrace+0xf8/0x148
    show_stack+0x18/0x24
    dump_stack_lvl+0x60/0x7c
    dump_stack+0x18/0x3c
    mrdump_common_die+0x24c/0x398 [mrdump]
    ipanic_die+0x20/0x34 [mrdump]
    notify_die+0x80/0xd8
    die+0x94/0x2b8
    __do_kernel_fault+0x264/0x298
    do_page_fault+0xa4/0x4b8
    do_translation_fault+0x38/0x54
    do_mem_abort+0x58/0x118
    el1_abort+0x3c/0x5c
    el1h_64_sync_handler+0x54/0x90
    el1h_64_sync+0x68/0x6c
    blk_mq_unique_tag+0x8/0x14
    ufshcd_clear_cmd+0x34/0x118 [ufs_mediatek_mod_ise]
    ufshcd_try_to_abort_task+0x2c8/0x5b4 [ufs_mediatek_mod_ise]
    ufshcd_err_handler+0xa7c/0xfa8 [ufs_mediatek_mod_ise]
    process_one_work+0x208/0x4fc
    worker_thread+0x228/0x438
    kthread+0x104/0x1d4
    ret_from_fork+0x10/0x20

Fixes: 8d7290348992 ("scsi: ufs: mcq: Add supporting functions for MCQ abort")
Suggested-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Link: https://lore.kernel.org/r/20240628070030.30929-2-peter.wang@mediatek.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
12 months agotpm: Address !chip->auth in tpm_buf_append_hmac_session*()
Jarkko Sakkinen [Wed, 3 Jul 2024 15:47:46 +0000 (18:47 +0300)]
tpm: Address !chip->auth in tpm_buf_append_hmac_session*()

Unless tpm_chip_bootstrap() was called by the driver, !chip->auth can
cause a null derefence in tpm_buf_hmac_session*().  Thus, address
!chip->auth in tpm_buf_hmac_session*() and remove the fallback
implementation for !TCG_TPM2_HMAC.

Cc: stable@vger.kernel.org # v6.9+
Reported-by: Stefan Berger <stefanb@linux.ibm.com>
Closes: https://lore.kernel.org/linux-integrity/20240617193408.1234365-1-stefanb@linux.ibm.com/
Fixes: 1085b8276bb4 ("tpm: Add the rest of the session HMAC API")
Tested-by: Michael Ellerman <mpe@ellerman.id.au> # ppc
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
12 months agotpm: Address !chip->auth in tpm_buf_append_name()
Jarkko Sakkinen [Wed, 3 Jul 2024 15:33:14 +0000 (18:33 +0300)]
tpm: Address !chip->auth in tpm_buf_append_name()

Unless tpm_chip_bootstrap() was called by the driver, !chip->auth can
cause a null derefence in tpm_buf_append_name().  Thus, address
!chip->auth in tpm_buf_append_name() and remove the fallback
implementation for !TCG_TPM2_HMAC.

Cc: stable@vger.kernel.org # v6.10+
Reported-by: Stefan Berger <stefanb@linux.ibm.com>
Closes: https://lore.kernel.org/linux-integrity/20240617193408.1234365-1-stefanb@linux.ibm.com/
Fixes: d0a25bb961e6 ("tpm: Add HMAC session name/handle append")
Tested-by: Michael Ellerman <mpe@ellerman.id.au> # ppc
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
12 months agotpm: Address !chip->auth in tpm2_*_auth_session()
Jarkko Sakkinen [Wed, 3 Jul 2024 16:39:27 +0000 (19:39 +0300)]
tpm: Address !chip->auth in tpm2_*_auth_session()

Unless tpm_chip_bootstrap() was called by the driver, !chip->auth can cause
a null derefence in tpm2_*_auth_session(). Thus, address !chip->auth in
tpm2_*_auth_session().

Cc: stable@vger.kernel.org # v6.9+
Reported-by: Stefan Berger <stefanb@linux.ibm.com>
Closes: https://lore.kernel.org/linux-integrity/20240617193408.1234365-1-stefanb@linux.ibm.com/
Fixes: 699e3efd6c64 ("tpm: Add HMAC session start and end functions")
Tested-by: Michael Ellerman <mpe@ellerman.id.au> # ppc
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
12 months agoMerge tag 'for-6.10-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Thu, 4 Jul 2024 17:27:37 +0000 (10:27 -0700)]
Merge tag 'for-6.10-rc6-tag' of git://git./linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:

 - fix folio refcounting when releasing them (encoded write, dummy
   extent buffer)

 - fix out of bounds read when checking qgroup inherit data

 - fix how configurable chunk size is handled in zoned mode

 - in the ref-verify tool, fix uninitialized return value when checking
   extent owner ref and simple quota are not enabled

* tag 'for-6.10-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: fix folio refcount in __alloc_dummy_extent_buffer()
  btrfs: fix folio refcount in btrfs_do_encoded_write()
  btrfs: fix uninitialized return value in the ref-verify tool
  btrfs: always do the basic checks for btrfs_qgroup_inherit structure
  btrfs: zoned: fix calc_available_free_space() for zoned mode

12 months agoMerge tag 'net-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 4 Jul 2024 17:11:12 +0000 (10:11 -0700)]
Merge tag 'net-6.10-rc7' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from bluetooth, wireless and netfilter.

  There's one fix for power management with Intel's e1000e here,
  Thorsten tells us there's another problem that started in v6.9. We're
  trying to wrap that up but I don't think it's blocking.

  Current release - new code bugs:

   - wifi: mac80211: disable softirqs for queued frame handling

   - af_unix: fix uninit-value in __unix_walk_scc(), with the new
     garbage collection algo

  Previous releases - regressions:

   - Bluetooth:
      - qca: fix BT enable failure for QCA6390 after warm reboot
      - add quirk to ignore reserved PHY bits in LE Extended Adv Report,
        abused by some Broadcom controllers found on Apple machines

   - wifi: wilc1000: fix ies_len type in connect path

  Previous releases - always broken:

   - tcp: fix DSACK undo in fast recovery to call tcp_try_to_open(),
     avoid premature timeouts

   - net: make sure skb_datagram_iter maps fragments page by page, in
     case we somehow get compound highmem mixed in

   - eth: bnx2x: fix multiple UBSAN array-index-out-of-bounds when more
     queues are used

  Misc:

   - MAINTAINERS: Remembering Larry Finger"

* tag 'net-6.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (62 commits)
  bnxt_en: Fix the resource check condition for RSS contexts
  mlxsw: core_linecards: Fix double memory deallocation in case of invalid INI file
  inet_diag: Initialize pad field in struct inet_diag_req_v2
  tcp: Don't flag tcp_sk(sk)->rx_opt.saw_unknown for TCP AO.
  selftests: make order checking verbose in msg_zerocopy selftest
  selftests: fix OOM in msg_zerocopy selftest
  ice: use proper macro for testing bit
  ice: Reject pin requests with unsupported flags
  ice: Don't process extts if PTP is disabled
  ice: Fix improper extts handling
  selftest: af_unix: Add test case for backtrack after finalising SCC.
  af_unix: Fix uninit-value in __unix_walk_scc()
  bonding: Fix out-of-bounds read in bond_option_arp_ip_targets_set()
  net: rswitch: Avoid use-after-free in rswitch_poll()
  netfilter: nf_tables: unconditionally flush pending work before notifier
  wifi: iwlwifi: mvm: check vif for NULL/ERR_PTR before dereference
  wifi: iwlwifi: mvm: avoid link lookup in statistics
  wifi: iwlwifi: mvm: don't wake up rx_sync_waitq upon RFKILL
  wifi: iwlwifi: properly set WIPHY_FLAG_SUPPORTS_EXT_KEK_KCK
  wifi: wilc1000: fix ies_len type in connect path
  ...

12 months agothermal: core: Call monitor_thermal_zone() if zone temperature is invalid
Rafael J. Wysocki [Fri, 28 Jun 2024 12:10:03 +0000 (14:10 +0200)]
thermal: core: Call monitor_thermal_zone() if zone temperature is invalid

Commit 202aa0d4bb53 ("thermal: core: Do not call handle_thermal_trip()
if zone temperature is invalid") caused __thermal_zone_device_update()
to return early if the current thermal zone temperature was invalid.

This was done to avoid running handle_thermal_trip() and governor
callbacks in that case which led to confusion.  However, it went too
far because monitor_thermal_zone() still needs to be called even when
the zone temperature is invalid to ensure that it will be updated
eventually in case thermal polling is enabled and the driver has no
other means to notify the core of zone temperature changes (for example,
it does not register an interrupt handler or ACPI notifier).

Also if the .set_trips() zone callback is expected to set up monitoring
interrupts for a thermal zone, it has to be provided with valid
boundaries and that can only happen if the zone temperature is known.

Accordingly, to ensure that __thermal_zone_device_update() will
run again after a failing zone temperature check, make it call
monitor_thermal_zone() regardless of whether or not the zone
temperature is valid and make the latter schedule a thermal zone
temperature update if the zone temperature is invalid even if
polling is not enabled for the thermal zone.

Fixes: 202aa0d4bb53 ("thermal: core: Do not call handle_thermal_trip() if zone temperature is invalid")
Reported-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Tested-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://patch.msgid.link/2764814.mvXUDI8C0e@rjwysocki.net
[ rjw: Changed THERMAL_RECHECK_DELAY_MS to 250 ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
12 months agoMerge tag 's390-6.10-8' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Thu, 4 Jul 2024 16:46:15 +0000 (09:46 -0700)]
Merge tag 's390-6.10-8' of git://git./linux/kernel/git/s390/linux

Pull s390 fixes from Heiko Carstens:

 - Fix and add physical to virtual address translations in dasd and
   virtio_ccw drivers. For virtio_ccw this is just a minimal fix.
   More code cleanup will follow.

 - Small defconfig updates

* tag 's390-6.10-8' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/dasd: Fix invalid dereferencing of indirect CCW data pointer
  s390/vfio_ccw: Fix target addresses of TIC CCWs
  s390: Update defconfigs

12 months agoMerge tag 'platform-drivers-x86-v6.10-5' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 4 Jul 2024 16:36:42 +0000 (09:36 -0700)]
Merge tag 'platform-drivers-x86-v6.10-5' of git://git./linux/kernel/git/pdx86/platform-drivers-x86

Pull x86 platform driver fix from Hans de Goede:

 - Fix regression in toshiba_acpi introduced in 6.10-rc1

* tag 'platform-drivers-x86-v6.10-5' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:
  platform/x86: toshiba_acpi: Fix quickstart quirk handling

12 months agoMerge tag 'kselftest-fix-2024-07-04' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 4 Jul 2024 16:29:42 +0000 (09:29 -0700)]
Merge tag 'kselftest-fix-2024-07-04' of git://git./linux/kernel/git/mic/linux

Pull Kselftest fix from Mickaël Salaün:
 "Fix Kselftests timeout.

  We can't use CLONE_VFORK, since that blocks the parent - and thus the
  timeout handling - until the child exits or execve's.

  Go back to using plain fork()"

* tag 'kselftest-fix-2024-07-04' of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux:
  selftests/harness: Fix tests timeout and race condition

12 months agoMerge tag 'mm-hotfixes-stable-2024-07-03-22-23' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Thu, 4 Jul 2024 16:13:02 +0000 (09:13 -0700)]
Merge tag 'mm-hotfixes-stable-2024-07-03-22-23' of git://git./linux/kernel/git/akpm/mm

Pull misc fixes from, Andrew Morton:
 "6 hotfies, all cc:stable. Some fixes for longstanding nilfs2 issues
  and three unrelated MM fixes"

* tag 'mm-hotfixes-stable-2024-07-03-22-23' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
  nilfs2: fix incorrect inode allocation from reserved inodes
  nilfs2: add missing check for inode numbers on directory entries
  nilfs2: fix inode number range checks
  mm: avoid overflows in dirty throttling logic
  Revert "mm/writeback: fix possible divide-by-zero in wb_dirty_limits(), again"
  mm: optimize the redundant loop of mm_update_owner_next()

12 months agoMerge tag 'drm-misc-fixes-2024-07-04' of https://gitlab.freedesktop.org/drm/misc...
Daniel Vetter [Thu, 4 Jul 2024 14:48:02 +0000 (16:48 +0200)]
Merge tag 'drm-misc-fixes-2024-07-04' of https://gitlab.freedesktop.org/drm/misc/kernel into drm-fixes

drm-misc-fixes for v6.10-rc7:
- Add panel quirks.
- Firmware sysfb refcount fix.
- Another null pointer mode deref fix for nouveau.
- Panthor sync and uobj fixes.
- Fix fbdev regression since v6.7.
- Delay free imported bo in ttm to fix lockdep splat.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ffba0c63-2798-40b6-948d-361cd3b14e9f@linux.intel.com
12 months agoMerge tag 'drm-xe-fixes-2024-07-04' of https://gitlab.freedesktop.org/drm/xe/kernel...
Daniel Vetter [Thu, 4 Jul 2024 14:44:16 +0000 (16:44 +0200)]
Merge tag 'drm-xe-fixes-2024-07-04' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes

Driver Changes:
- One copy/paste mistake fix.
- One error path fix causing an error pointer dereference.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Thomas Hellstrom <thomas.hellstrom@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ZoZ-wD66lgjiNh72@fedora
12 months agobnxt_en: Fix the resource check condition for RSS contexts
Pavan Chebbi [Wed, 3 Jul 2024 18:01:12 +0000 (11:01 -0700)]
bnxt_en: Fix the resource check condition for RSS contexts

While creating a new RSS context, bnxt_rfs_capable() currently
makes a strict check to see if the required VNICs are already
available.  If the current VNICs are not what is required,
either too many or not enough, it will call the firmware to
reserve the exact number required.

There is a bug in the firmware when the driver tries to
relinquish some reserved VNICs and RSS contexts.  It will
cause the default VNIC to lose its RSS configuration and
cause receive packets to be placed incorrectly.

Workaround this problem by skipping the resource reduction.
The driver will not reduce the VNIC and RSS context reservations
when a context is deleted.  The resources will be available for
use when new contexts are created later.

Potentially, this workaround can cause us to run out of VNIC
and RSS contexts if there are a lot of VF functions creating
and deleting RSS contexts.  In the future, we will conditionally
disable this workaround when the firmware fix is available.

Fixes: 438ba39b25fe ("bnxt_en: Improve RSS context reservation infrastructure")
Reported-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/netdev/20240625010210.2002310-1-kuba@kernel.org/
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20240703180112.78590-1-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agomlxsw: core_linecards: Fix double memory deallocation in case of invalid INI file
Aleksandr Mishin [Wed, 3 Jul 2024 20:32:51 +0000 (23:32 +0300)]
mlxsw: core_linecards: Fix double memory deallocation in case of invalid INI file

In case of invalid INI file mlxsw_linecard_types_init() deallocates memory
but doesn't reset pointer to NULL and returns 0. In case of any error
occurred after mlxsw_linecard_types_init() call, mlxsw_linecards_init()
calls mlxsw_linecard_types_fini() which performs memory deallocation again.

Add pointer reset to NULL.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: b217127e5e4e ("mlxsw: core_linecards: Add line card objects and implement provisioning")
Signed-off-by: Aleksandr Mishin <amishin@t-argos.ru>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Michal Kubiak <michal.kubiak@intel.com>
Link: https://patch.msgid.link/20240703203251.8871-1-amishin@t-argos.ru
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoMerge tag 'wireless-2024-07-04' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Thu, 4 Jul 2024 14:31:54 +0000 (07:31 -0700)]
Merge tag 'wireless-2024-07-04' of git://git./linux/kernel/git/wireless/wireless

Kalle Valo says:

====================
wireless fixes for v6.10

Hopefully the last fixes for v6.10. Fix a regression in wilc1000
where bitrate Information Elements longer than 255 bytes were broken.
Few fixes also to mac80211 and iwlwifi.

* tag 'wireless-2024-07-04' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless:
  wifi: iwlwifi: mvm: check vif for NULL/ERR_PTR before dereference
  wifi: iwlwifi: mvm: avoid link lookup in statistics
  wifi: iwlwifi: mvm: don't wake up rx_sync_waitq upon RFKILL
  wifi: iwlwifi: properly set WIPHY_FLAG_SUPPORTS_EXT_KEK_KCK
  wifi: wilc1000: fix ies_len type in connect path
  wifi: mac80211: fix BSS_CHANGED_UNSOL_BCAST_PROBE_RESP
====================

Link: https://patch.msgid.link/20240704111431.11DEDC3277B@smtp.kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoMerge tag 'drm-intel-fixes-2024-07-02' of https://gitlab.freedesktop.org/drm/i915...
Daniel Vetter [Thu, 4 Jul 2024 14:14:17 +0000 (16:14 +0200)]
Merge tag 'drm-intel-fixes-2024-07-02' of https://gitlab.freedesktop.org/drm/i915/kernel into drm-fixes

drm/i915 fixes for v6.10-rc7:
- Skip unnecessary MG programming, avoiding warnings (Imre)

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/87frss9ozs.fsf@intel.com
12 months agoMerge tag 'nf-24-07-04' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Paolo Abeni [Thu, 4 Jul 2024 13:31:26 +0000 (15:31 +0200)]
Merge tag 'nf-24-07-04' of git://git./linux/kernel/git/netfilter/nf

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

The following batch contains a oneliner patch to inconditionally flush
workqueue containing stale objects to be released, syzbot managed to
trigger UaF. Patch from Florian Westphal.

netfilter pull request 24-07-04

* tag 'nf-24-07-04' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: nf_tables: unconditionally flush pending work before notifier
====================

Link: https://patch.msgid.link/20240703223304.1455-1-pablo@netfilter.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
12 months agoinet_diag: Initialize pad field in struct inet_diag_req_v2
Shigeru Yoshida [Wed, 3 Jul 2024 09:16:49 +0000 (18:16 +0900)]
inet_diag: Initialize pad field in struct inet_diag_req_v2

KMSAN reported uninit-value access in raw_lookup() [1]. Diag for raw
sockets uses the pad field in struct inet_diag_req_v2 for the
underlying protocol. This field corresponds to the sdiag_raw_protocol
field in struct inet_diag_req_raw.

inet_diag_get_exact_compat() converts inet_diag_req to
inet_diag_req_v2, but leaves the pad field uninitialized. So the issue
occurs when raw_lookup() accesses the sdiag_raw_protocol field.

Fix this by initializing the pad field in
inet_diag_get_exact_compat(). Also, do the same fix in
inet_diag_dump_compat() to avoid the similar issue in the future.

[1]
BUG: KMSAN: uninit-value in raw_lookup net/ipv4/raw_diag.c:49 [inline]
BUG: KMSAN: uninit-value in raw_sock_get+0x657/0x800 net/ipv4/raw_diag.c:71
 raw_lookup net/ipv4/raw_diag.c:49 [inline]
 raw_sock_get+0x657/0x800 net/ipv4/raw_diag.c:71
 raw_diag_dump_one+0xa1/0x660 net/ipv4/raw_diag.c:99
 inet_diag_cmd_exact+0x7d9/0x980
 inet_diag_get_exact_compat net/ipv4/inet_diag.c:1404 [inline]
 inet_diag_rcv_msg_compat+0x469/0x530 net/ipv4/inet_diag.c:1426
 sock_diag_rcv_msg+0x23d/0x740 net/core/sock_diag.c:282
 netlink_rcv_skb+0x537/0x670 net/netlink/af_netlink.c:2564
 sock_diag_rcv+0x35/0x40 net/core/sock_diag.c:297
 netlink_unicast_kernel net/netlink/af_netlink.c:1335 [inline]
 netlink_unicast+0xe74/0x1240 net/netlink/af_netlink.c:1361
 netlink_sendmsg+0x10c6/0x1260 net/netlink/af_netlink.c:1905
 sock_sendmsg_nosec net/socket.c:730 [inline]
 __sock_sendmsg+0x332/0x3d0 net/socket.c:745
 ____sys_sendmsg+0x7f0/0xb70 net/socket.c:2585
 ___sys_sendmsg+0x271/0x3b0 net/socket.c:2639
 __sys_sendmsg net/socket.c:2668 [inline]
 __do_sys_sendmsg net/socket.c:2677 [inline]
 __se_sys_sendmsg net/socket.c:2675 [inline]
 __x64_sys_sendmsg+0x27e/0x4a0 net/socket.c:2675
 x64_sys_call+0x135e/0x3ce0 arch/x86/include/generated/asm/syscalls_64.h:47
 do_syscall_x64 arch/x86/entry/common.c:52 [inline]
 do_syscall_64+0xd9/0x1e0 arch/x86/entry/common.c:83
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

Uninit was stored to memory at:
 raw_sock_get+0x650/0x800 net/ipv4/raw_diag.c:71
 raw_diag_dump_one+0xa1/0x660 net/ipv4/raw_diag.c:99
 inet_diag_cmd_exact+0x7d9/0x980
 inet_diag_get_exact_compat net/ipv4/inet_diag.c:1404 [inline]
 inet_diag_rcv_msg_compat+0x469/0x530 net/ipv4/inet_diag.c:1426
 sock_diag_rcv_msg+0x23d/0x740 net/core/sock_diag.c:282
 netlink_rcv_skb+0x537/0x670 net/netlink/af_netlink.c:2564
 sock_diag_rcv+0x35/0x40 net/core/sock_diag.c:297
 netlink_unicast_kernel net/netlink/af_netlink.c:1335 [inline]
 netlink_unicast+0xe74/0x1240 net/netlink/af_netlink.c:1361
 netlink_sendmsg+0x10c6/0x1260 net/netlink/af_netlink.c:1905
 sock_sendmsg_nosec net/socket.c:730 [inline]
 __sock_sendmsg+0x332/0x3d0 net/socket.c:745
 ____sys_sendmsg+0x7f0/0xb70 net/socket.c:2585
 ___sys_sendmsg+0x271/0x3b0 net/socket.c:2639
 __sys_sendmsg net/socket.c:2668 [inline]
 __do_sys_sendmsg net/socket.c:2677 [inline]
 __se_sys_sendmsg net/socket.c:2675 [inline]
 __x64_sys_sendmsg+0x27e/0x4a0 net/socket.c:2675
 x64_sys_call+0x135e/0x3ce0 arch/x86/include/generated/asm/syscalls_64.h:47
 do_syscall_x64 arch/x86/entry/common.c:52 [inline]
 do_syscall_64+0xd9/0x1e0 arch/x86/entry/common.c:83
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

Local variable req.i created at:
 inet_diag_get_exact_compat net/ipv4/inet_diag.c:1396 [inline]
 inet_diag_rcv_msg_compat+0x2a6/0x530 net/ipv4/inet_diag.c:1426
 sock_diag_rcv_msg+0x23d/0x740 net/core/sock_diag.c:282

CPU: 1 PID: 8888 Comm: syz-executor.6 Not tainted 6.10.0-rc4-00217-g35bb670d65fc #32
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-2.fc40 04/01/2014

Fixes: 432490f9d455 ("net: ip, diag -- Add diag interface for raw sockets")
Reported-by: syzkaller <syzkaller@googlegroups.com>
Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20240703091649.111773-1-syoshida@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
12 months agothermal: gov_power_allocator: Return early in manage if trip_max is NULL
Nícolas F. R. A. Prado [Tue, 2 Jul 2024 21:24:56 +0000 (17:24 -0400)]
thermal: gov_power_allocator: Return early in manage if trip_max is NULL

Commit da781936e7c3 ("thermal: gov_power_allocator: Allow binding
without trip points") allowed the governor to bind even when trip_max
is NULL. This allows a NULL pointer dereference to happen in the manage
callback.

Add an early return to prevent it, since the governor is expected to not do
anything in this case.

Fixes: da781936e7c3 ("thermal: gov_power_allocator: Allow binding without trip points")
Signed-off-by: Nícolas F. R. A. Prado <nfraprado@collabora.com>
Link: https://patch.msgid.link/20240702-power-allocator-null-trip-max-v1-1-47a60dc55414@collabora.com
Cc: All applicable <stable@vger.kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
12 months agotcp: Don't flag tcp_sk(sk)->rx_opt.saw_unknown for TCP AO.
Kuniyuki Iwashima [Wed, 3 Jul 2024 03:35:08 +0000 (20:35 -0700)]
tcp: Don't flag tcp_sk(sk)->rx_opt.saw_unknown for TCP AO.

When we process segments with TCP AO, we don't check it in
tcp_parse_options().  Thus, opt_rx->saw_unknown is set to 1,
which unconditionally triggers the BPF TCP option parser.

Let's avoid the unnecessary BPF invocation.

Fixes: 0a3a809089eb ("net/tcp: Verify inbound TCP-AO signed segments")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Dmitry Safonov <0x7f454c46@gmail.com>
Link: https://patch.msgid.link/20240703033508.6321-1-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
12 months agodrm/xe/mcr: Avoid clobbering DSS steering
Matt Roper [Wed, 26 Jun 2024 21:05:37 +0000 (14:05 -0700)]
drm/xe/mcr: Avoid clobbering DSS steering

A couple copy/paste mistakes in the code that selects steering targets
for OADDRM and INSTANCE0 unintentionally clobbered the steering target
for DSS ranges in some cases.

The OADDRM/INSTANCE0 values were also not assigned as intended, although
that mistake wound up being harmless since the desired values for those
specific ranges were '0' which the kzalloc of the GT structure should
have already taken care of implicitly.

Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240626210536.1620176-2-matthew.d.roper@intel.com
(cherry picked from commit 4f82ac6102788112e599a6074d2c1f2afce923df)
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
12 months agodrm/xe: fix error handling in xe_migrate_update_pgtables
Matthew Auld [Thu, 20 Jun 2024 10:20:26 +0000 (11:20 +0100)]
drm/xe: fix error handling in xe_migrate_update_pgtables

Don't call drm_suballoc_free with sa_bo pointing to PTR_ERR.

References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2120
Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: <stable@vger.kernel.org> # v6.8+
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240620102025.127699-2-matthew.auld@intel.com
(cherry picked from commit ce6b63336f79ec5f3996de65f452330e395f99ae)
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
12 months agodrm/ttm: Always take the bo delayed cleanup path for imported bos
Thomas Hellström [Fri, 28 Jun 2024 15:38:48 +0000 (17:38 +0200)]
drm/ttm: Always take the bo delayed cleanup path for imported bos

Bos can be put with multiple unrelated dma-resv locks held. But
imported bos attempt to grab the bo dma-resv during dma-buf detach
that typically happens during cleanup. That leads to lockde splats
similar to the below and a potential ABBA deadlock.

Fix this by always taking the delayed workqueue cleanup path for
imported bos.

Requesting stable fixes from when the Xe driver was introduced,
since its usage of drm_exec and wide vm dma_resvs appear to be
the first reliable trigger of this.

[22982.116427] ============================================
[22982.116428] WARNING: possible recursive locking detected
[22982.116429] 6.10.0-rc2+ #10 Tainted: G     U  W
[22982.116430] --------------------------------------------
[22982.116430] glxgears:sh0/5785 is trying to acquire lock:
[22982.116431] ffff8c2bafa539a8 (reservation_ww_class_mutex){+.+.}-{3:3}, at: dma_buf_detach+0x3b/0xf0
[22982.116438]
               but task is already holding lock:
[22982.116438] ffff8c2d9aba6da8 (reservation_ww_class_mutex){+.+.}-{3:3}, at: drm_exec_lock_obj+0x49/0x2b0 [drm_exec]
[22982.116442]
               other info that might help us debug this:
[22982.116442]  Possible unsafe locking scenario:

[22982.116443]        CPU0
[22982.116444]        ----
[22982.116444]   lock(reservation_ww_class_mutex);
[22982.116445]   lock(reservation_ww_class_mutex);
[22982.116447]
                *** DEADLOCK ***

[22982.116447]  May be due to missing lock nesting notation

[22982.116448] 5 locks held by glxgears:sh0/5785:
[22982.116449]  #0: ffff8c2d9aba58c8 (&xef->vm.lock){+.+.}-{3:3}, at: xe_file_close+0xde/0x1c0 [xe]
[22982.116507]  #1: ffff8c2e28cc8480 (&vm->lock){++++}-{3:3}, at: xe_vm_close_and_put+0x161/0x9b0 [xe]
[22982.116578]  #2: ffff8c2e31982970 (&val->lock){.+.+}-{3:3}, at: xe_validation_ctx_init+0x6d/0x70 [xe]
[22982.116647]  #3: ffffacdc469478a8 (reservation_ww_class_acquire){+.+.}-{0:0}, at: xe_vma_destroy_unlocked+0x7f/0xe0 [xe]
[22982.116716]  #4: ffff8c2d9aba6da8 (reservation_ww_class_mutex){+.+.}-{3:3}, at: drm_exec_lock_obj+0x49/0x2b0 [drm_exec]
[22982.116719]
               stack backtrace:
[22982.116720] CPU: 8 PID: 5785 Comm: glxgears:sh0 Tainted: G     U  W          6.10.0-rc2+ #10
[22982.116721] Hardware name: ASUS System Product Name/PRIME B560M-A AC, BIOS 2001 02/01/2023
[22982.116723] Call Trace:
[22982.116724]  <TASK>
[22982.116725]  dump_stack_lvl+0x77/0xb0
[22982.116727]  __lock_acquire+0x1232/0x2160
[22982.116730]  lock_acquire+0xcb/0x2d0
[22982.116732]  ? dma_buf_detach+0x3b/0xf0
[22982.116734]  ? __lock_acquire+0x417/0x2160
[22982.116736]  __ww_mutex_lock.constprop.0+0xd0/0x13b0
[22982.116738]  ? dma_buf_detach+0x3b/0xf0
[22982.116741]  ? dma_buf_detach+0x3b/0xf0
[22982.116743]  ? ww_mutex_lock+0x2b/0x90
[22982.116745]  ww_mutex_lock+0x2b/0x90
[22982.116747]  dma_buf_detach+0x3b/0xf0
[22982.116749]  drm_prime_gem_destroy+0x2f/0x40 [drm]
[22982.116775]  xe_ttm_bo_destroy+0x32/0x220 [xe]
[22982.116818]  ? __mutex_unlock_slowpath+0x3a/0x290
[22982.116821]  drm_exec_unlock_all+0xa1/0xd0 [drm_exec]
[22982.116823]  drm_exec_fini+0x12/0xb0 [drm_exec]
[22982.116824]  xe_validation_ctx_fini+0x15/0x40 [xe]
[22982.116892]  xe_vma_destroy_unlocked+0xb1/0xe0 [xe]
[22982.116959]  xe_vm_close_and_put+0x41a/0x9b0 [xe]
[22982.117025]  ? xa_find+0xe3/0x1e0
[22982.117028]  xe_file_close+0x10a/0x1c0 [xe]
[22982.117074]  drm_file_free+0x22a/0x280 [drm]
[22982.117099]  drm_release_noglobal+0x22/0x70 [drm]
[22982.117119]  __fput+0xf1/0x2d0
[22982.117122]  task_work_run+0x59/0x90
[22982.117125]  do_exit+0x330/0xb40
[22982.117127]  do_group_exit+0x36/0xa0
[22982.117129]  get_signal+0xbd2/0xbe0
[22982.117131]  arch_do_signal_or_restart+0x3e/0x240
[22982.117134]  syscall_exit_to_user_mode+0x1e7/0x290
[22982.117137]  do_syscall_64+0xa1/0x180
[22982.117139]  ? lock_acquire+0xcb/0x2d0
[22982.117140]  ? __set_task_comm+0x28/0x1e0
[22982.117141]  ? find_held_lock+0x2b/0x80
[22982.117144]  ? __set_task_comm+0xe1/0x1e0
[22982.117145]  ? lock_release+0xca/0x290
[22982.117147]  ? __do_sys_prctl+0x245/0xab0
[22982.117149]  ? lockdep_hardirqs_on_prepare+0xde/0x190
[22982.117150]  ? syscall_exit_to_user_mode+0xb0/0x290
[22982.117152]  ? do_syscall_64+0xa1/0x180
[22982.117154]  ? __lock_acquire+0x417/0x2160
[22982.117155]  ? reacquire_held_locks+0xd1/0x1f0
[22982.117156]  ? do_user_addr_fault+0x30c/0x790
[22982.117158]  ? lock_acquire+0xcb/0x2d0
[22982.117160]  ? find_held_lock+0x2b/0x80
[22982.117162]  ? do_user_addr_fault+0x357/0x790
[22982.117163]  ? lock_release+0xca/0x290
[22982.117164]  ? do_user_addr_fault+0x361/0x790
[22982.117166]  ? trace_hardirqs_off+0x4b/0xc0
[22982.117168]  ? clear_bhb_loop+0x45/0xa0
[22982.117170]  ? clear_bhb_loop+0x45/0xa0
[22982.117172]  ? clear_bhb_loop+0x45/0xa0
[22982.117174]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[22982.117176] RIP: 0033:0x7f943d267169
[22982.117192] Code: Unable to access opcode bytes at 0x7f943d26713f.
[22982.117193] RSP: 002b:00007f9430bffc80 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca
[22982.117195] RAX: fffffffffffffe00 RBX: 0000000000000000 RCX: 00007f943d267169
[22982.117196] RDX: 0000000000000000 RSI: 0000000000000189 RDI: 00005622f89579d0
[22982.117197] RBP: 00007f9430bffcb0 R08: 0000000000000000 R09: 00000000ffffffff
[22982.117198] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
[22982.117199] R13: 0000000000000000 R14: 0000000000000000 R15: 00005622f89579d0
[22982.117202]  </TASK>

Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Cc: Christian König <christian.koenig@amd.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Cc: intel-xe@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v6.8+
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240628153848.4989-1-thomas.hellstrom@linux.intel.com
12 months agonilfs2: fix kernel bug on rename operation of broken directory
Ryusuke Konishi [Fri, 28 Jun 2024 16:51:07 +0000 (01:51 +0900)]
nilfs2: fix kernel bug on rename operation of broken directory

Syzbot reported that in rename directory operation on broken directory on
nilfs2, __block_write_begin_int() called to prepare block write may fail
BUG_ON check for access exceeding the folio/page size.

This is because nilfs_dotdot(), which gets parent directory reference
entry ("..") of the directory to be moved or renamed, does not check
consistency enough, and may return location exceeding folio/page size for
broken directories.

Fix this issue by checking required directory entries ("." and "..") in
the first chunk of the directory in nilfs_dotdot().

Link: https://lkml.kernel.org/r/20240628165107.9006-1-konishi.ryusuke@gmail.com
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Reported-by: syzbot+d3abed1ad3d367fa2627@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=d3abed1ad3d367fa2627
Fixes: 2ba466d74ed7 ("nilfs2: directory entry operations")
Tested-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm/hugetlb_vmemmap: fix race with speculative PFN walkers
Yu Zhao [Thu, 27 Jun 2024 22:27:05 +0000 (16:27 -0600)]
mm/hugetlb_vmemmap: fix race with speculative PFN walkers

While investigating HVO for THPs [1], it turns out that speculative PFN
walkers like compaction can race with vmemmap modifications, e.g.,

  CPU 1 (vmemmap modifier)         CPU 2 (speculative PFN walker)
  -------------------------------  ------------------------------
  Allocates an LRU folio page1
                                   Sees page1
  Frees page1

  Allocates a hugeTLB folio page2
  (page1 being a tail of page2)

  Updates vmemmap mapping page1
                                   get_page_unless_zero(page1)

Even though page1->_refcount is zero after HVO, get_page_unless_zero() can
still try to modify this read-only field, resulting in a crash.

An independent report [2] confirmed this race.

There are two discussed approaches to fix this race:
1. Make RO vmemmap RW so that get_page_unless_zero() can fail without
   triggering a PF.
2. Use RCU to make sure get_page_unless_zero() either sees zero
   page->_refcount through the old vmemmap or non-zero page->_refcount
   through the new one.

The second approach is preferred here because:
1. It can prevent illegal modifications to struct page[] that has been
   HVO'ed;
2. It can be generalized, in a way similar to ZERO_PAGE(), to fix
   similar races in other places, e.g., arch_remove_memory() on x86
   [3], which frees vmemmap mapping offlined struct page[].

While adding synchronize_rcu(), the goal is to be surgical, rather than
optimized.  Specifically, calls to synchronize_rcu() on the error handling
paths can be coalesced, but it is not done for the sake of Simplicity:
noticeably, this fix removes ~50% more lines than it adds.

According to the hugetlb_optimize_vmemmap section in
Documentation/admin-guide/sysctl/vm.rst, enabling HVO makes allocating or
freeing hugeTLB pages "~2x slower than before".  Having synchronize_rcu()
on top makes those operations even worse, and this also affects the user
interface /proc/sys/vm/nr_overcommit_hugepages.

This is *very* hard to trigger:

1. Most hugeTLB use cases I know of are static, i.e., reserved at
   boot time, because allocating at runtime is not reliable at all.

2. On top of that, someone has to be very unlucky to get tripped
   over above, because the race window is so small -- I wasn't able to
   trigger it with a stress testing that does nothing but that (with
   THPs though).

[1] https://lore.kernel.org/20240229183436.4110845-4-yuzhao@google.com/
[2] https://lore.kernel.org/917FFC7F-0615-44DD-90EE-9F85F8EA9974@linux.dev/
[3] https://lore.kernel.org/be130a96-a27e-4240-ad78-776802f57cad@redhat.com/

Link: https://lkml.kernel.org/r/20240627222705.2974207-1-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Acked-by: Muchun Song <muchun.song@linux.dev>
Cc: David Hildenbrand <david@redhat.com>
Cc: Frank van der Linden <fvdl@google.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Peter Xu <peterx@redhat.com>
Cc: Yang Shi <yang@os.amperecomputing.com>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agocachestat: do not flush stats in recency check
Nhat Pham [Thu, 27 Jun 2024 20:17:37 +0000 (13:17 -0700)]
cachestat: do not flush stats in recency check

syzbot detects that cachestat() is flushing stats, which can sleep, in its
RCU read section (see [1]).  This is done in the workingset_test_recent()
step (which checks if the folio's eviction is recent).

Move the stat flushing step to before the RCU read section of cachestat,
and skip stat flushing during the recency check.

[1]: https://lore.kernel.org/cgroups/000000000000f71227061bdf97e0@google.com/

Link: https://lkml.kernel.org/r/20240627201737.3506959-1-nphamcs@gmail.com
Fixes: b00684722262 ("mm: workingset: move the stats flush into workingset_test_recent()")
Signed-off-by: Nhat Pham <nphamcs@gmail.com>
Reported-by: syzbot+b7f13b2d0cc156edf61a@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/cgroups/000000000000f71227061bdf97e0@google.com/
Debugged-by: Johannes Weiner <hannes@cmpxchg.org>
Suggested-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: David Hildenbrand <david@redhat.com>
Cc: "Huang, Ying" <ying.huang@intel.com>
Cc: Kairui Song <kasong@tencent.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Yosry Ahmed <yosryahmed@google.com>
Cc: <stable@vger.kernel.org> [6.8+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm/shmem: disable PMD-sized page cache if needed
Gavin Shan [Thu, 27 Jun 2024 00:39:52 +0000 (10:39 +1000)]
mm/shmem: disable PMD-sized page cache if needed

For shmem files, it's possible that PMD-sized page cache can't be
supported by xarray.  For example, 512MB page cache on ARM64 when the base
page size is 64KB can't be supported by xarray.  It leads to errors as the
following messages indicate when this sort of xarray entry is split.

WARNING: CPU: 34 PID: 7578 at lib/xarray.c:1025 xas_split_alloc+0xf8/0x128
Modules linked in: binfmt_misc nft_fib_inet nft_fib_ipv4 nft_fib_ipv6   \
nft_fib nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject        \
nft_ct nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4  \
ip_set rfkill nf_tables nfnetlink vfat fat virtio_balloon drm fuse xfs  \
libcrc32c crct10dif_ce ghash_ce sha2_ce sha256_arm64 sha1_ce virtio_net \
net_failover virtio_console virtio_blk failover dimlib virtio_mmio
CPU: 34 PID: 7578 Comm: test Kdump: loaded Tainted: G W 6.10.0-rc5-gavin+ #9
Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20240524-1.el9 05/24/2024
pstate: 83400005 (Nzcv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)
pc : xas_split_alloc+0xf8/0x128
lr : split_huge_page_to_list_to_order+0x1c4/0x720
sp : ffff8000882af5f0
x29: ffff8000882af5f0 x28: ffff8000882af650 x27: ffff8000882af768
x26: 0000000000000cc0 x25: 000000000000000d x24: ffff00010625b858
x23: ffff8000882af650 x22: ffffffdfc0900000 x21: 0000000000000000
x20: 0000000000000000 x19: ffffffdfc0900000 x18: 0000000000000000
x17: 0000000000000000 x16: 0000018000000000 x15: 52f8004000000000
x14: 0000e00000000000 x13: 0000000000002000 x12: 0000000000000020
x11: 52f8000000000000 x10: 52f8e1c0ffff6000 x9 : ffffbeb9619a681c
x8 : 0000000000000003 x7 : 0000000000000000 x6 : ffff00010b02ddb0
x5 : ffffbeb96395e378 x4 : 0000000000000000 x3 : 0000000000000cc0
x2 : 000000000000000d x1 : 000000000000000c x0 : 0000000000000000
Call trace:
 xas_split_alloc+0xf8/0x128
 split_huge_page_to_list_to_order+0x1c4/0x720
 truncate_inode_partial_folio+0xdc/0x160
 shmem_undo_range+0x2bc/0x6a8
 shmem_fallocate+0x134/0x430
 vfs_fallocate+0x124/0x2e8
 ksys_fallocate+0x4c/0xa0
 __arm64_sys_fallocate+0x24/0x38
 invoke_syscall.constprop.0+0x7c/0xd8
 do_el0_svc+0xb4/0xd0
 el0_svc+0x44/0x1d8
 el0t_64_sync_handler+0x134/0x150
 el0t_64_sync+0x17c/0x180

Fix it by disabling PMD-sized page cache when HPAGE_PMD_ORDER is larger
than MAX_PAGECACHE_ORDER.  As Matthew Wilcox pointed, the page cache in a
shmem file isn't represented by a multi-index entry and doesn't have this
limitation when the xarry entry is split until commit 6b24ca4a1a8d ("mm:
Use multi-index entries in the page cache").

Link: https://lkml.kernel.org/r/20240627003953.1262512-5-gshan@redhat.com
Fixes: 6b24ca4a1a8d ("mm: Use multi-index entries in the page cache")
Signed-off-by: Gavin Shan <gshan@redhat.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Darrick J. Wong <djwong@kernel.org>
Cc: Don Dutile <ddutile@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Zhenyu Zhang <zhenyzha@redhat.com>
Cc: <stable@vger.kernel.org> [5.17+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm/filemap: skip to create PMD-sized page cache if needed
Gavin Shan [Thu, 27 Jun 2024 00:39:51 +0000 (10:39 +1000)]
mm/filemap: skip to create PMD-sized page cache if needed

On ARM64, HPAGE_PMD_ORDER is 13 when the base page size is 64KB.  The
PMD-sized page cache can't be supported by xarray as the following error
messages indicate.

------------[ cut here ]------------
WARNING: CPU: 35 PID: 7484 at lib/xarray.c:1025 xas_split_alloc+0xf8/0x128
Modules linked in: nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib  \
nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct    \
nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4    \
ip_set rfkill nf_tables nfnetlink vfat fat virtio_balloon drm      \
fuse xfs libcrc32c crct10dif_ce ghash_ce sha2_ce sha256_arm64      \
sha1_ce virtio_net net_failover virtio_console virtio_blk failover \
dimlib virtio_mmio
CPU: 35 PID: 7484 Comm: test Kdump: loaded Tainted: G W 6.10.0-rc5-gavin+ #9
Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20240524-1.el9 05/24/2024
pstate: 83400005 (Nzcv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)
pc : xas_split_alloc+0xf8/0x128
lr : split_huge_page_to_list_to_order+0x1c4/0x720
sp : ffff800087a4f6c0
x29: ffff800087a4f6c0 x28: ffff800087a4f720 x27: 000000001fffffff
x26: 0000000000000c40 x25: 000000000000000d x24: ffff00010625b858
x23: ffff800087a4f720 x22: ffffffdfc0780000 x21: 0000000000000000
x20: 0000000000000000 x19: ffffffdfc0780000 x18: 000000001ff40000
x17: 00000000ffffffff x16: 0000018000000000 x15: 51ec004000000000
x14: 0000e00000000000 x13: 0000000000002000 x12: 0000000000000020
x11: 51ec000000000000 x10: 51ece1c0ffff8000 x9 : ffffbeb961a44d28
x8 : 0000000000000003 x7 : ffffffdfc0456420 x6 : ffff0000e1aa6eb8
x5 : 20bf08b4fe778fca x4 : ffffffdfc0456420 x3 : 0000000000000c40
x2 : 000000000000000d x1 : 000000000000000c x0 : 0000000000000000
Call trace:
 xas_split_alloc+0xf8/0x128
 split_huge_page_to_list_to_order+0x1c4/0x720
 truncate_inode_partial_folio+0xdc/0x160
 truncate_inode_pages_range+0x1b4/0x4a8
 truncate_pagecache_range+0x84/0xa0
 xfs_flush_unmap_range+0x70/0x90 [xfs]
 xfs_file_fallocate+0xfc/0x4d8 [xfs]
 vfs_fallocate+0x124/0x2e8
 ksys_fallocate+0x4c/0xa0
 __arm64_sys_fallocate+0x24/0x38
 invoke_syscall.constprop.0+0x7c/0xd8
 do_el0_svc+0xb4/0xd0
 el0_svc+0x44/0x1d8
 el0t_64_sync_handler+0x134/0x150
 el0t_64_sync+0x17c/0x180

Fix it by skipping to allocate PMD-sized page cache when its size is
larger than MAX_PAGECACHE_ORDER.  For this specific case, we will fall to
regular path where the readahead window is determined by BDI's sysfs file
(read_ahead_kb).

Link: https://lkml.kernel.org/r/20240627003953.1262512-4-gshan@redhat.com
Fixes: 4687fdbb805a ("mm/filemap: Support VM_HUGEPAGE for file mappings")
Signed-off-by: Gavin Shan <gshan@redhat.com>
Suggested-by: David Hildenbrand <david@redhat.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Darrick J. Wong <djwong@kernel.org>
Cc: Don Dutile <ddutile@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Zhenyu Zhang <zhenyzha@redhat.com>
Cc: <stable@vger.kernel.org> [5.18+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm/readahead: limit page cache size in page_cache_ra_order()
Gavin Shan [Thu, 27 Jun 2024 00:39:50 +0000 (10:39 +1000)]
mm/readahead: limit page cache size in page_cache_ra_order()

In page_cache_ra_order(), the maximal order of the page cache to be
allocated shouldn't be larger than MAX_PAGECACHE_ORDER.  Otherwise, it's
possible the large page cache can't be supported by xarray when the
corresponding xarray entry is split.

For example, HPAGE_PMD_ORDER is 13 on ARM64 when the base page size is
64KB.  The PMD-sized page cache can't be supported by xarray.

Link: https://lkml.kernel.org/r/20240627003953.1262512-3-gshan@redhat.com
Fixes: 793917d997df ("mm/readahead: Add large folio readahead")
Signed-off-by: Gavin Shan <gshan@redhat.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Darrick J. Wong <djwong@kernel.org>
Cc: Don Dutile <ddutile@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Zhenyu Zhang <zhenyzha@redhat.com>
Cc: <stable@vger.kernel.org> [5.18+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm/filemap: make MAX_PAGECACHE_ORDER acceptable to xarray
Gavin Shan [Thu, 27 Jun 2024 00:39:49 +0000 (10:39 +1000)]
mm/filemap: make MAX_PAGECACHE_ORDER acceptable to xarray

Patch series "mm/filemap: Limit page cache size to that supported by
xarray", v2.

Currently, xarray can't support arbitrary page cache size.  More details
can be found from the WARN_ON() statement in xas_split_alloc().  In our
test whose code is attached below, we hit the WARN_ON() on ARM64 system
where the base page size is 64KB and huge page size is 512MB.  The issue
was reported long time ago and some discussions on it can be found here
[1].

[1] https://www.spinics.net/lists/linux-xfs/msg75404.html

In order to fix the issue, we need to adjust MAX_PAGECACHE_ORDER to one
supported by xarray and avoid PMD-sized page cache if needed.  The code
changes are suggested by David Hildenbrand.

PATCH[1] adjusts MAX_PAGECACHE_ORDER to that supported by xarray
PATCH[2-3] avoids PMD-sized page cache in the synchronous readahead path
PATCH[4] avoids PMD-sized page cache for shmem files if needed

Test program
============
# cat test.c
#define _GNU_SOURCE
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <string.h>
#include <fcntl.h>
#include <errno.h>
#include <sys/syscall.h>
#include <sys/mman.h>

#define TEST_XFS_FILENAME "/tmp/data"
#define TEST_SHMEM_FILENAME "/dev/shm/data"
#define TEST_MEM_SIZE 0x20000000

int main(int argc, char **argv)
{
const char *filename;
int fd = 0;
void *buf = (void *)-1, *p;
int pgsize = getpagesize();
int ret;

if (pgsize != 0x10000) {
fprintf(stderr, "64KB base page size is required\n");
return -EPERM;
}

system("echo force > /sys/kernel/mm/transparent_hugepage/shmem_enabled");
system("rm -fr /tmp/data");
system("rm -fr /dev/shm/data");
system("echo 1 > /proc/sys/vm/drop_caches");

/* Open xfs or shmem file */
filename = TEST_XFS_FILENAME;
if (argc > 1 && !strcmp(argv[1], "shmem"))
filename = TEST_SHMEM_FILENAME;

fd = open(filename, O_CREAT | O_RDWR | O_TRUNC);
if (fd < 0) {
fprintf(stderr, "Unable to open <%s>\n", filename);
return -EIO;
}

/* Extend file size */
ret = ftruncate(fd, TEST_MEM_SIZE);
if (ret) {
fprintf(stderr, "Error %d to ftruncate()\n", ret);
goto cleanup;
}

/* Create VMA */
buf = mmap(NULL, TEST_MEM_SIZE,
   PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0);
if (buf == (void *)-1) {
fprintf(stderr, "Unable to mmap <%s>\n", filename);
goto cleanup;
}

fprintf(stdout, "mapped buffer at 0x%p\n", buf);
ret = madvise(buf, TEST_MEM_SIZE, MADV_HUGEPAGE);
        if (ret) {
fprintf(stderr, "Unable to madvise(MADV_HUGEPAGE)\n");
goto cleanup;
}

/* Populate VMA */
ret = madvise(buf, TEST_MEM_SIZE, MADV_POPULATE_WRITE);
if (ret) {
fprintf(stderr, "Error %d to madvise(MADV_POPULATE_WRITE)\n", ret);
goto cleanup;
}

/* Punch the file to enforce xarray split */
ret = fallocate(fd, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE,
         TEST_MEM_SIZE - pgsize, pgsize);
if (ret)
fprintf(stderr, "Error %d to fallocate()\n", ret);

cleanup:
if (buf != (void *)-1)
munmap(buf, TEST_MEM_SIZE);
if (fd > 0)
close(fd);

return 0;
}

# gcc test.c -o test
# cat /proc/1/smaps | grep KernelPageSize | head -n 1
KernelPageSize:       64 kB
# ./test shmem
   :
------------[ cut here ]------------
WARNING: CPU: 17 PID: 5253 at lib/xarray.c:1025 xas_split_alloc+0xf8/0x128
Modules linked in: nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib  \
nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct    \
nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4    \
ip_set nf_tables rfkill nfnetlink vfat fat virtio_balloon          \
drm fuse xfs libcrc32c crct10dif_ce ghash_ce sha2_ce sha256_arm64  \
virtio_net sha1_ce net_failover failover virtio_console virtio_blk \
dimlib virtio_mmio
CPU: 17 PID: 5253 Comm: test Kdump: loaded Tainted: G W 6.10.0-rc5-gavin+ #12
Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20240524-1.el9 05/24/2024
pstate: 83400005 (Nzcv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)
pc : xas_split_alloc+0xf8/0x128
lr : split_huge_page_to_list_to_order+0x1c4/0x720
sp : ffff80008a92f5b0
x29: ffff80008a92f5b0 x28: ffff80008a92f610 x27: ffff80008a92f728
x26: 0000000000000cc0 x25: 000000000000000d x24: ffff0000cf00c858
x23: ffff80008a92f610 x22: ffffffdfc0600000 x21: 0000000000000000
x20: 0000000000000000 x19: ffffffdfc0600000 x18: 0000000000000000
x17: 0000000000000000 x16: 0000018000000000 x15: 3374004000000000
x14: 0000e00000000000 x13: 0000000000002000 x12: 0000000000000020
x11: 3374000000000000 x10: 3374e1c0ffff6000 x9 : ffffb463a84c681c
x8 : 0000000000000003 x7 : 0000000000000000 x6 : ffff00011c976ce0
x5 : ffffb463aa47e378 x4 : 0000000000000000 x3 : 0000000000000cc0
x2 : 000000000000000d x1 : 000000000000000c x0 : 0000000000000000
Call trace:
 xas_split_alloc+0xf8/0x128
 split_huge_page_to_list_to_order+0x1c4/0x720
 truncate_inode_partial_folio+0xdc/0x160
 shmem_undo_range+0x2bc/0x6a8
 shmem_fallocate+0x134/0x430
 vfs_fallocate+0x124/0x2e8
 ksys_fallocate+0x4c/0xa0
 __arm64_sys_fallocate+0x24/0x38
 invoke_syscall.constprop.0+0x7c/0xd8
 do_el0_svc+0xb4/0xd0
 el0_svc+0x44/0x1d8
 el0t_64_sync_handler+0x134/0x150
 el0t_64_sync+0x17c/0x180

This patch (of 4):

The largest page cache order can be HPAGE_PMD_ORDER (13) on ARM64 with
64KB base page size.  The xarray entry with this order can't be split as
the following error messages indicate.

------------[ cut here ]------------
WARNING: CPU: 35 PID: 7484 at lib/xarray.c:1025 xas_split_alloc+0xf8/0x128
Modules linked in: nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib  \
nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct    \
nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4    \
ip_set rfkill nf_tables nfnetlink vfat fat virtio_balloon drm      \
fuse xfs libcrc32c crct10dif_ce ghash_ce sha2_ce sha256_arm64      \
sha1_ce virtio_net net_failover virtio_console virtio_blk failover \
dimlib virtio_mmio
CPU: 35 PID: 7484 Comm: test Kdump: loaded Tainted: G W 6.10.0-rc5-gavin+ #9
Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20240524-1.el9 05/24/2024
pstate: 83400005 (Nzcv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)
pc : xas_split_alloc+0xf8/0x128
lr : split_huge_page_to_list_to_order+0x1c4/0x720
sp : ffff800087a4f6c0
x29: ffff800087a4f6c0 x28: ffff800087a4f720 x27: 000000001fffffff
x26: 0000000000000c40 x25: 000000000000000d x24: ffff00010625b858
x23: ffff800087a4f720 x22: ffffffdfc0780000 x21: 0000000000000000
x20: 0000000000000000 x19: ffffffdfc0780000 x18: 000000001ff40000
x17: 00000000ffffffff x16: 0000018000000000 x15: 51ec004000000000
x14: 0000e00000000000 x13: 0000000000002000 x12: 0000000000000020
x11: 51ec000000000000 x10: 51ece1c0ffff8000 x9 : ffffbeb961a44d28
x8 : 0000000000000003 x7 : ffffffdfc0456420 x6 : ffff0000e1aa6eb8
x5 : 20bf08b4fe778fca x4 : ffffffdfc0456420 x3 : 0000000000000c40
x2 : 000000000000000d x1 : 000000000000000c x0 : 0000000000000000
Call trace:
 xas_split_alloc+0xf8/0x128
 split_huge_page_to_list_to_order+0x1c4/0x720
 truncate_inode_partial_folio+0xdc/0x160
 truncate_inode_pages_range+0x1b4/0x4a8
 truncate_pagecache_range+0x84/0xa0
 xfs_flush_unmap_range+0x70/0x90 [xfs]
 xfs_file_fallocate+0xfc/0x4d8 [xfs]
 vfs_fallocate+0x124/0x2e8
 ksys_fallocate+0x4c/0xa0
 __arm64_sys_fallocate+0x24/0x38
 invoke_syscall.constprop.0+0x7c/0xd8
 do_el0_svc+0xb4/0xd0
 el0_svc+0x44/0x1d8
 el0t_64_sync_handler+0x134/0x150
 el0t_64_sync+0x17c/0x180

Fix it by decreasing MAX_PAGECACHE_ORDER to the largest supported order
by xarray. For this specific case, MAX_PAGECACHE_ORDER is dropped from
13 to 11 when CONFIG_BASE_SMALL is disabled.

Link: https://lkml.kernel.org/r/20240627003953.1262512-1-gshan@redhat.com
Link: https://lkml.kernel.org/r/20240627003953.1262512-2-gshan@redhat.com
Fixes: 793917d997df ("mm/readahead: Add large folio readahead")
Signed-off-by: Gavin Shan <gshan@redhat.com>
Suggested-by: David Hildenbrand <david@redhat.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Darrick J. Wong <djwong@kernel.org>
Cc: Don Dutile <ddutile@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Zhenyu Zhang <zhenyzha@redhat.com>
Cc: <stable@vger.kernel.org> [5.18+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
12 months agomm/damon/core: merge regions aggressively when max_nr_regions is unmet
SeongJae Park [Mon, 24 Jun 2024 17:58:14 +0000 (10:58 -0700)]
mm/damon/core: merge regions aggressively when max_nr_regions is unmet

DAMON keeps the number of regions under max_nr_regions by skipping regions
split operations when doing so can make the number higher than the limit.
It works well for preventing violation of the limit.  But, if somehow the
violation happens, it cannot recovery well depending on the situation.  In
detail, if the real number of regions having different access pattern is
higher than the limit, the mechanism cannot reduce the number below the
limit.  In such a case, the system could suffer from high monitoring
overhead of DAMON.

The violation can actually happen.  For an example, the user could reduce
max_nr_regions while DAMON is running, to be lower than the current number
of regions.  Fix the problem by repeating the merge operations with
increasing aggressiveness in kdamond_merge_regions() for the case, until
the limit is met.

[sj@kernel.org: increase regions merge aggressiveness while respecting min_nr_regions]
Link: https://lkml.kernel.org/r/20240626164753.46270-1-sj@kernel.org
[sj@kernel.org: ensure max threshold attempt for max_nr_regions violation]
Link: https://lkml.kernel.org/r/20240627163153.75969-1-sj@kernel.org
Link: https://lkml.kernel.org/r/20240624175814.89611-1-sj@kernel.org
Fixes: b9a6ac4e4ede ("mm/damon: adaptively adjust regions")
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: <stable@vger.kernel.org> [5.15+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>