Linus Torvalds [Mon, 20 May 2024 23:00:04 +0000 (16:00 -0700)]
Merge tag 'cocci-for-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/jlawall/linux
Pull coccinelle updates from Julia Lawall:
"One patch slightly improves the text in a comment.
The other patch (on minmax.cocci) removes a report about ? being used
in return statements that has been generating not very useful
suggestions to change idiomatic code"
* tag 'cocci-for-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/jlawall/linux:
Coccinelle: pm_runtime: Fix grammar in comment
coccinelle: misc: minmax: Suppress reports for err returns
Linus Torvalds [Mon, 20 May 2024 22:18:34 +0000 (15:18 -0700)]
Merge tag 'asm-generic-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic
Pull asm-generic cleanups from Arnd Bergmann:
"These are a few cross-architecture cleanup patches:
- separate out fbdev support from the asm/video.h contents that may
be used by either the old fbdev drivers or the newer drm display
code (Thomas Zimmermann)
- cleanups for the generic bitops code and asm-generic/bug.h
(Thorsten Blum)
- remove the orphaned include/asm-generic/page.h header that used to
be included by long-removed mmu-less architectures (me)"
* tag 'asm-generic-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic:
arch: Fix name collision with ACPI's video.o
bug: Improve comment
asm-generic: remove unused asm-generic/page.h
arch: Rename fbdev header and source files
arch: Remove struct fb_info from video helpers
arch: Select fbdev helpers with CONFIG_VIDEO
bitops: Change function return types from long to int
Linus Torvalds [Mon, 20 May 2024 22:11:53 +0000 (15:11 -0700)]
Merge tag 'soc-dt-late-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull more SoC devicetree updates from Arnd Bergmann:
"This is a follow-up to an earlier pull request for device tree
changes, as three platform maintainers sent their contents too late to
be included in the main set, but had not caused any further problems
since then:
- The Amlogic platform now containts support for two new SoC types,
the A4 and A5 chips for audio applications. Both come with a
reference board, and one more dts file gets addded for the
combination of the MNT Reform Laptop with the BPI-CM4 CPU module
- The ASpeed platform adds support for six addititional server
platforms that use ast2500 or ast2600 as their BMC, while another
one gets removed
- The RISC-V platforms from Microchip, Starfive and and T-HEAD get
additional features for existing hardware, plus the addition of the
Milk-V Mars based on the StarFive VisionFive v2 board"
* tag 'soc-dt-late-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (76 commits)
riscv: dts: microchip: add pac1934 power-monitor to icicle
riscv: dts: thead: Fix node ordering in TH1520 device tree
ARM: dts: aspeed: Add ASRock E3C256D4I BMC
dt-bindings: arm: aspeed: document ASRock E3C256D4I
dt-bindings: trivial-devices: add isil,isl69269
ARM: dts: aspeed: x4tf: Add dts for asus x4tf project
dt-bindings: arm: aspeed: add ASUS X4TF board
ARM: dts: aspeed: Remove Facebook Cloudripper dts
ARM: dts: aspeed: drop unused ref_voltage ADC property
ARM: dts: aspeed: harma: correct Mellanox multi-host property
ARM: dts: aspeed: yosemitev2: correct Mellanox multi-host property
ARM: dts: aspeed: yosemite4: correct Mellanox multi-host property
ARM: dts: aspeed: greatlakes: correct Mellanox multi-host property
ARM: dts: aspeed: Modify I2C bus configuration
ARM: dts: aspeed: Disable unused ADC channels for Asrock X570D4U BMC
ARM: dts: aspeed: Modify GPIO table for Asrock X570D4U BMC
ARM: dts: aspeed: yosemite4: set bus13 frequency to 100k
ARM: dts: Aspeed: Bonnell: Fix NVMe LED labels
ARM: dts: aspeed: yosemite4: Enable ipmb device for OCP debug card
ARM: dts: aspeed: ahe50dc: Update lm25066 regulator name
...
Linus Torvalds [Mon, 20 May 2024 21:56:50 +0000 (14:56 -0700)]
Merge tag 'vfio-v6.10-rc1' of https://github.com/awilliam/linux-vfio
Pull vfio updates from Alex Williamson:
- The vfio fsl-mc bus driver has become orphaned. We'll consider
removing it in future releases if a new maintainer isn't found (Alex
Williamson)
- Improved usage of opaque data in vfio-pci INTx handling, avoiding
lookups of the eventfd through the interrupt and irqfd runtime paths
(Alex Williamson)
- Resolve an error path memory leak introduced in vfio-pci interrupt
code (Ye Bin)
- Addition of interrupt support for vfio devices exposed on the CDX
bus, including a new MSI allocation helper and export of existing
helpers for MSI alloc and free (Nipun Gupta)
- A new vfio-pci variant driver supporting migration of Intel QAT VF
devices for the GEN4 PFs (Xin Zeng & Yahui Cao)
- Resolve a possibly circular locking dependency in vfio-pci by
avoiding copy_to_user() from a PCI bus walk callback (Alex
Williamson)
- Trivial docs update to remove a duplicate semicolon (Foryun Ma)
* tag 'vfio-v6.10-rc1' of https://github.com/awilliam/linux-vfio:
vfio/pci: Restore zero affected bus reset devices warning
vfio: remove an extra semicolon
vfio/pci: Collect hot-reset devices to local buffer
vfio/qat: Add vfio_pci driver for Intel QAT SR-IOV VF devices
vfio/cdx: add interrupt support
genirq/msi: Add MSI allocation helper and export MSI functions
vfio/pci: fix potential memory leak in vfio_intx_enable()
vfio/pci: Pass eventfd context object through irqfd
vfio/pci: Pass eventfd context to IRQ handler
MAINTAINERS: Orphan vfio fsl-mc bus driver
Linus Torvalds [Mon, 20 May 2024 21:49:39 +0000 (14:49 -0700)]
Merge tag 'linux_kselftest-next-6.10-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest
Pull kselftest fixes from Shuah Khan:
"Revert framework change to add D_GNU_SOURCE to KHDR_INCLUDES to
Makefile, lib.mk, and kselftest_harness.h and follow-on changes to
cgroup and sgx test as they are causing build failures and warnings"
* tag 'linux_kselftest-next-6.10-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
Revert "selftests/cgroup: Drop define _GNU_SOURCE"
Revert "selftests/sgx: Include KHDR_INCLUDES in Makefile"
Revert "selftests: Compile kselftest headers with -D_GNU_SOURCE"
Lang Yu [Fri, 26 Apr 2024 06:56:35 +0000 (14:56 +0800)]
drm/amdkfd: Let VRAM allocations go to GTT domain on small APUs
Small APUs(i.e., consumer, embedded products) usually have a small
carveout device memory which can't satisfy most compute workloads
memory allocation requirements.
We can't even run a Basic MNIST Example with a default 512MB carveout.
https://github.com/pytorch/examples/tree/main/mnist. Error Log:
"torch.cuda.OutOfMemoryError: HIP out of memory. Tried to allocate
84.00 MiB. GPU 0 has a total capacity of 512.00 MiB of which 0 bytes
is free. Of the allocated memory 103.83 MiB is allocated by PyTorch,
and 22.17 MiB is reserved by PyTorch but unallocated"
Though we can change BIOS settings to enlarge carveout size,
which is inflexible and may bring complaint. On the other hand,
the memory resource can't be effectively used between host and device.
The solution is MI300A approach, i.e., let VRAM allocations go to GTT.
Then device and host can flexibly and effectively share memory resource.
Commit 2fd001cd3600 ("arch: Rename fbdev header and source files")
renames the video source files under arch/ such that they do not
refer to fbdev any longer. The new files named video.o conflict with
ACPI's video.ko module. Modprobing the ACPI module can then fail with
warnings about missing symbols, as shown below.
(i915_selftest:1107) igt_kmod-WARNING: i915: Unknown symbol acpi_video_unregister (err -2)
(i915_selftest:1107) igt_kmod-WARNING: i915: Unknown symbol acpi_video_register_backlight (err -2)
(i915_selftest:1107) igt_kmod-WARNING: i915: Unknown symbol __acpi_video_get_backlight_type (err -2)
(i915_selftest:1107) igt_kmod-WARNING: i915: Unknown symbol acpi_video_register (err -2)
Fix the issue by renaming the architecture's video.o to video-common.o.
Linus Torvalds [Mon, 20 May 2024 20:23:43 +0000 (13:23 -0700)]
Merge tag 'f2fs-for-6.10.rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs
Pull f2fs updates from Jaegeuk Kim:
"In this round, we've tried to address some performance issues on zoned
storage such as direct IO and write_hints. In addition, we've migrated
some IO paths using folio. Meanwhile, there are multiple bug fixes in
the compression paths, sanity check conditions, and error handlers.
Enhancements:
- allow direct io of pinned files for zoned storage
- assign the write hint per stream by default
- convert read paths and test_writeback to folio
- avoid allocating WARM_DATA segment for direct IO
Bug fixes:
- fix false alarm on invalid block address
- fix to add missing iput() in gc_data_segment()
- fix to release node block count in error path of
f2fs_new_node_page()
- compress:
- don't allow unaligned truncation on released compress inode
- cover {reserve,release}_compress_blocks() w/ cp_rwsem lock
- fix error path of inc_valid_block_count()
- fix to update i_compr_blocks correctly
- fix block migration when section is not aligned to pow2
- don't trigger OPU on pinfile for direct IO
- fix to do sanity check on i_xattr_nid in sanity_check_inode()
- write missing last sum blk of file pinning section
- clear writeback when compression failed
- fix to adjust appropirate defragment pg_end
As usual, there are several minor code clean-ups, and fixes to manage
missing corner cases in the error paths"
* tag 'f2fs-for-6.10.rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (50 commits)
f2fs: initialize last_block_in_bio variable
f2fs: Add inline to f2fs_build_fault_attr() stub
f2fs: fix some ambiguous comments
f2fs: fix to add missing iput() in gc_data_segment()
f2fs: allow dirty sections with zero valid block for checkpoint disabled
f2fs: compress: don't allow unaligned truncation on released compress inode
f2fs: fix to release node block count in error path of f2fs_new_node_page()
f2fs: compress: fix to cover {reserve,release}_compress_blocks() w/ cp_rwsem lock
f2fs: compress: fix error path of inc_valid_block_count()
f2fs: compress: fix typo in f2fs_reserve_compress_blocks()
f2fs: compress: fix to update i_compr_blocks correctly
f2fs: check validation of fault attrs in f2fs_build_fault_attr()
f2fs: fix to limit gc_pin_file_threshold
f2fs: remove unused GC_FAILURE_PIN
f2fs: use f2fs_{err,info}_ratelimited() for cleanup
f2fs: fix block migration when section is not aligned to pow2
f2fs: zone: fix to don't trigger OPU on pinfile for direct IO
f2fs: fix to do sanity check on i_xattr_nid in sanity_check_inode()
f2fs: fix to avoid allocating WARM_DATA segment for direct IO
f2fs: remove redundant parameter in is_next_segment_free()
...
Linus Torvalds [Mon, 20 May 2024 19:55:12 +0000 (12:55 -0700)]
Merge tag 'xfs-6.10-merge-6' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Pull xfs updates from Chandan Babu:
"Online repair feature continues to be expanded. Also, we now support
delayed allocation for realtime devices which have an extent size that
is equal to filesystem's block size.
New code:
- Introduce Parent Pointer extended attribute for inodes
- Bring back delalloc support for realtime devices which have an
extent size that is equal to filesystem's block size
- Improve performance of log incompat feature handling
Online Repair:
- Implement atomic file content exchanges i.e. exchange ranges of
bytes between two files atomically
- Create temporary files to repair file-based metadata. This uses
atomic file content exchange facility to swap file fork mappings
between the temporary file and the metadata inode
- Allow callers of directory/xattr code to set an explicit owner
number to be written into the header fields of any new blocks that
are created. This is required to avoid walking every block of the
new structure and modify their ownership during online repair
- Repair more data structures:
- Extended attributes
- Inode unlinked state
- Directories
- Symbolic links
- AGI's unlinked inode list
- Parent pointers
- Move Orphan files to lost and found directory
- Fixes for Inode repair functionality
- Introduce a new sub-AG FITRIM implementation to reduce the duration
for which the AGF lock is held
- Updates for the design documentation
- Use Parent Pointers to assist in checking directories, parent
pointers, extended attributes, and link counts
Fixes:
- Prevent userspace from reading invalid file data due to incorrect.
updation of file size when performing a non-atomic clone operation
- Minor fixes to online repair
- Fix confusing return values from xfs_bmapi_write()
- Fix an out of bounds access due to incorrect h_size during log
recovery
- Defer upgrading the extent counters in xfs_reflink_end_cow_extent()
until we know we are going to modify the extent mapping
- Remove racy access to if_bytes check in
xfs_reflink_end_cow_extent()
- Fix sparse warnings
Cleanups:
- Hold inode locks on all files involved in a rename until the
completion of the operation. This is in preparation for the parent
pointers patchset where parent pointers are applied in a separate
chained update from the actual directory update
- Compile out v4 support when disabled
- Cleanup xfs_extent_busy_clear()
- Remove unused flags and fields from struct xfs_da_args
- Remove definitions of unused functions
- Improve extended attribute validation
- Add higher level directory operations helpers to remove duplication
of code
- Cleanup quota (un)reservation interfaces"
* tag 'xfs-6.10-merge-6' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (221 commits)
xfs: simplify iext overflow checking and upgrade
xfs: remove a racy if_bytes check in xfs_reflink_end_cow_extent
xfs: upgrade the extent counters in xfs_reflink_end_cow_extent later
xfs: xfs_quota_unreserve_blkres can't fail
xfs: consolidate the xfs_quota_reserve_blkres definitions
xfs: clean up buffer allocation in xlog_do_recovery_pass
xfs: fix log recovery buffer allocation for the legacy h_size fixup
xfs: widen flags argument to the xfs_iflags_* helpers
xfs: minor cleanups of xfs_attr3_rmt_blocks
xfs: create a helper to compute the blockcount of a max sized remote value
xfs: turn XFS_ATTR3_RMT_BUF_SPACE into a function
xfs: use unsigned ints for non-negative quantities in xfs_attr_remote.c
xfs: do not allocate the entire delalloc extent in xfs_bmapi_write
xfs: fix xfs_bmap_add_extent_delay_real for partial conversions
xfs: remove the xfs_iext_peek_prev_extent call in xfs_bmapi_allocate
xfs: pass the actual offset and len to allocate to xfs_bmapi_allocate
xfs: don't open code XFS_FILBLKS_MIN in xfs_bmapi_write
xfs: lift a xfs_valid_startblock into xfs_bmapi_allocate
xfs: remove the unusued tmp_logflags variable in xfs_bmapi_allocate
xfs: fix error returns from xfs_bmapi_write
...
Mike Snitzer [Mon, 20 May 2024 17:34:06 +0000 (13:34 -0400)]
dm: always manage discard support in terms of max_hw_discard_sectors
Commit 4f563a64732d ("block: add a max_user_discard_sectors queue
limit") changed block core to set max_discard_sectors to:
min(lim->max_hw_discard_sectors, lim->max_user_discard_sectors)
Since commit 1c0e720228ad ("dm: use queue_limits_set") it was reported
dm-thinp was failing in a few fstests (generic/347 and generic/405)
with the first WARN_ON_ONCE in dm_cell_key_has_valid_range() being
reported, e.g.:
WARNING: CPU: 1 PID: 30 at drivers/md/dm-bio-prison-v1.c:128 dm_cell_key_has_valid_range+0x3d/0x50
blk_set_stacking_limits() sets max_user_discard_sectors to UINT_MAX,
so given how block core now sets max_discard_sectors (detailed above)
it follows that blk_stack_limits() stacks up the underlying device's
max_hw_discard_sectors and max_discard_sectors is set to match it. If
max_hw_discard_sectors exceeds dm's BIO_PRISON_MAX_RANGE, then
dm_cell_key_has_valid_range() will trigger the warning with:
WARN_ON_ONCE(key->block_end - key->block_begin > BIO_PRISON_MAX_RANGE)
Aside from this warning, the discard will fail. Fix this and other DM
issues by governing discard support in terms of max_hw_discard_sectors
instead of max_discard_sectors.
Reported-by: Theodore Ts'o <[email protected]> Fixes: 1c0e720228ad ("dm: use queue_limits_set") Signed-off-by: Mike Snitzer <[email protected]>
Linus Torvalds [Mon, 20 May 2024 19:49:25 +0000 (12:49 -0700)]
Merge tag 'fs_for_v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Pull isofs, udf, quota, ext2, and reiserfs updates from Jan Kara:
- convert isofs to the new mount API
- cleanup isofs Makefile
- udf conversion to folios
- some other small udf cleanups and fixes
- ext2 cleanups
- removal of reiserfs .writepage method
- update reiserfs README file
* tag 'fs_for_v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
isofs: Use *-y instead of *-objs in Makefile
ext2: Remove LEGACY_DIRECT_IO dependency
isofs: Remove calls to set/clear the error flag
ext2: Remove call to folio_set_error()
udf: Use a folio in udf_write_end()
udf: Convert udf_page_mkwrite() to use a folio
udf: Convert udf_symlink_getattr() to use a folio
udf: Convert udf_adinicb_readpage() to udf_adinicb_read_folio()
udf: Convert udf_expand_file_adinicb() to use a folio
udf: Convert udf_write_begin() to use a folio
udf: Convert udf_symlink_filler() to use a folio
reiserfs: Trim some README bits
quota: fix to propagate error of mark_dquot_dirty() to caller
reiserfs: Convert to writepages
udf: udftime: prevent overflow in udf_disk_stamp_to_time()
ext2: set FMODE_CAN_ODIRECT instead of a dummy direct_IO method
udf: replace deprecated strncpy/strcpy with strscpy
udf: Remove second semicolon
isofs: convert isofs to use the new mount API
fs: quota: use group allocation of per-cpu counters API
Mikulas Patocka [Mon, 20 May 2024 14:48:31 +0000 (16:48 +0200)]
dm-integrity: set discard_granularity to logical block size
dm-integrity could set discard_granularity lower than the logical block
size. This could result in failures when sending discard requests to
dm-integrity.
These kinds of patches are only making the code worse.
Compilers don't care about the unnecessary check, but removing it makes
the code less obvious to a human. The declaration of 'len' is more than
80 lines earlier, so a human won't easily see that 'len' is of an
unsigned type, so to a human the range check that checks against zero is
much more explicit and obvious.
Any tool that complains about a range check like this just because the
variable is unsigned is actively detrimental, and should be ignored.
Maíra Canal [Sun, 12 May 2024 22:23:29 +0000 (19:23 -0300)]
drm/v3d: Deprecate the use of the Performance Counters enum
The Performance Counters enum used to identify the index of each
performance counter and provide the total number of performance
counters (V3D_PERFCNT_NUM). But, this enum is only valid for V3D 4.2,
not for V3D 7.1.
As we implemented a new flexible structure to retrieve performance
counters information, we can deprecate this enum.
Maíra Canal [Sun, 12 May 2024 22:23:28 +0000 (19:23 -0300)]
drm/v3d: Use V3D_MAX_COUNTERS instead of V3D_PERFCNT_NUM
V3D_PERFCNT_NUM represents the maximum number of performance counters
for V3D 4.2, but not for V3D 7.1. This means that, if we use
V3D_PERFCNT_NUM, we might go out-of-bounds on V3D 7.1.
Therefore, use the number of performance counters on V3D 7.1 as the
maximum number of counters. This will allow us to create arrays on the
stack with reasonable size. Note that userspace must use the value
provided by DRM_V3D_PARAM_MAX_PERF_COUNTERS.
Maíra Canal [Sun, 12 May 2024 22:23:27 +0000 (19:23 -0300)]
drm/v3d: Create new IOCTL to expose performance counters information
Userspace usually needs some information about the performance counters
available. Although we could replicate this information in the kernel
and user-space, let's use the kernel as the "single source of truth" to
avoid issues in the future (e.g. list of performance counters is updated
in user-space, but not in the kernel, generating invalid requests).
Therefore, create a new IOCTL to expose the performance counters
information, that is name, category, and description.
Maíra Canal [Sun, 12 May 2024 22:23:26 +0000 (19:23 -0300)]
drm/v3d: Create a new V3D parameter for the maximum number of perfcnt
The maximum number of performance counters can change from version to
version and it's important for userspace to know this value, as it needs
to use the counters for performance queries. Therefore, expose the
maximum number of performance counters to userspace as a parameter.
Maíra Canal [Sun, 12 May 2024 22:23:25 +0000 (19:23 -0300)]
drm/v3d: Different V3D versions can have different number of perfcnt
Currently, even though V3D 7.1 has 93 performance counters, it is not
possible to create counters bigger than 87, as
`v3d_perfmon_create_ioctl()` understands that counters bigger than 87
are invalid.
Therefore, create a device variable to expose the maximum
number of counters for a given V3D version and make
`v3d_perfmon_create_ioctl()` check this variable.
This commit fixes CTS failures in the performance queries tests
`dEQP-VK.query_pool.performance_query.*` [1]
Maíra Canal [Sun, 12 May 2024 22:23:24 +0000 (19:23 -0300)]
drm/v3d: Add Performance Counters descriptions for V3D 4.2 and 7.1
Add name, category and description for each one of the 93 performance
counters available on V3D.
Note that V3D 4.2 has 87 performance counters, while V3D 7.1 has 93.
Therefore, there are two performance counters arrays. The index of the
performance counter for each V3D version is represented by its position
on the array.
Linus Torvalds [Mon, 20 May 2024 19:31:43 +0000 (12:31 -0700)]
Merge tag 'fsnotify_for_v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Pull fsnotify updates from Jan Kara:
- reduce overhead of fsnotify infrastructure when no permission events
are in use
- a few small cleanups
* tag 'fsnotify_for_v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
fsnotify: fix UAF from FS_ERROR event on a shutting down filesystem
fsnotify: optimize the case of no permission event watchers
fsnotify: use an enum for group priority constants
fsnotify: move s_fsnotify_connectors into fsnotify_sb_info
fsnotify: lazy attach fsnotify_sb_info state to sb
fsnotify: create helper fsnotify_update_sb_watchers()
fsnotify: pass object pointer and type to fsnotify mark helpers
fanotify: merge two checks regarding add of ignore mark
fsnotify: create a wrapper fsnotify_find_inode_mark()
fsnotify: create helpers to get sb and connp from object
fsnotify: rename fsnotify_{get,put}_sb_connectors()
fsnotify: Avoid -Wflex-array-member-not-at-end warning
fanotify: remove unneeded sub-zero check for unsigned value
Matti Vaittinen [Mon, 20 May 2024 12:36:55 +0000 (15:36 +0300)]
regulator: tps6287x: Force writing VSEL bit
The data-sheet for TPS6287x-Q1
https://www.ti.com/lit/ds/symlink/tps62873-q1.pdf
states at chapter 9.3.6.1 Output Voltage Range:
"Note that every change to the VRANGE[1:0] bits must be followed by a
write to the VSET register, even if the value of the VSET[7:0] bits does
not change."
The current implementation of the driver uses the
regulator_set_voltage_sel_pickable_regmap() helper which further uses
regmap_update_bits() to write the VSET-register. The
regmap_update_bits() will not access the hardware if the new register
value is same as old. It is worth noting that this is true also when the
register is marked volatile, which I can't say is wrong because
'read-mnodify-write'-cycle with a volatile register is in any case
something user should carefully consider.
The 'range_applied_by_vsel'-flag in regulator desc was added to force
the vsel register upodates by using regmap_write_bits(). This variant
will always unconditionally write the bits to the hardware.
It is worth noting that the vsel is now forced to be written to the
hardware, whether the range was changed or not. This may cause a
performance drop if users are wrtiting same voltage value repeteadly.
It would be possible to read the range register to determine if it was
changed, but this would be a performance issue for users who don't use
reg cache for vsel.
Always write the VSET register to the hardware regardless the cache.
Some PMICs treat the vsel_reg same as apply-bit. Eg, when voltage range
is changed, the new voltage setting is not taking effect until the vsel
register is written.
Add a flag 'range_applied_by_vsel' to the regulator desc to indicate this
behaviour and to force the vsel value to be written to hardware if range
was changed, even if the old selector was same as the new one.
Linus Torvalds [Mon, 20 May 2024 17:23:39 +0000 (10:23 -0700)]
Merge tag 'dma-mapping-6.10-2024-05-20' of git://git.infradead.org/users/hch/dma-mapping
Pull dma-mapping updates from Christoph Hellwig:
- optimize DMA sync calls when they are no-ops (Alexander Lobakin)
- fix swiotlb padding for untrusted devices (Michael Kelley)
- add documentation for swiotb (Michael Kelley)
* tag 'dma-mapping-6.10-2024-05-20' of git://git.infradead.org/users/hch/dma-mapping:
dma: fix DMA sync for drivers not calling dma_set_mask*()
xsk: use generic DMA sync shortcut instead of a custom one
page_pool: check for DMA sync shortcut earlier
page_pool: don't use driver-set flags field directly
page_pool: make sure frag API fields don't span between cachelines
iommu/dma: avoid expensive indirect calls for sync operations
dma: avoid redundant calls for sync operations
dma: compile-out DMA sync op calls when not used
iommu/dma: fix zeroing of bounce buffer padding used by untrusted devices
swiotlb: remove alloc_size argument to swiotlb_tbl_map_single()
Documentation/core-api: add swiotlb documentation
Linus Torvalds [Mon, 20 May 2024 16:23:36 +0000 (09:23 -0700)]
Merge tag 'dmi-for-v6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging
Pull dmi updates from Jean Delvare:
"Bug fixes:
- KCFI violation in dmi-id
- stop decoding on broken (short) DMI table entry
New features:
- print info about populated memory slots at boot"
* tag 'dmi-for-v6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
firmware: dmi: Add info message for number of populated and total memory slots
firmware: dmi: Stop decoding on broken entry
firmware: dmi-id: add a release callback function
Linus Torvalds [Mon, 20 May 2024 16:07:27 +0000 (09:07 -0700)]
Merge tag 'linux-watchdog-6.10-rc1' of git://www.linux-watchdog.org/linux-watchdog
Pull watchdog updates from Wim Van Sebroeck:
- Add Lenovo SE10 platform Watchdog Driver
- Other small fixes and improvements
* tag 'linux-watchdog-6.10-rc1' of git://www.linux-watchdog.org/linux-watchdog:
watchdog: LENOVO_SE10_WDT should depend on X86 && DMI
watchdog: sa1100: Fix PTR_ERR_OR_ZERO() vs NULL check in sa1100dog_probe()
watchdog: rti_wdt: Set min_hw_heartbeat_ms to accommodate a safety margin
watchdog: add HAS_IOPORT dependencies
watchdog/wdt-main: Use cpumask_of() to avoid cpumask var on stack
watchdog: bd9576: Drop "always-running" property
watchdog: mtx-1: drop driver owner assignment
watchdog: cpu5wdt.c: Fix use-after-free bug caused by cpu5wdt_trigger
watchdog: lenovo_se10_wdt: Watchdog driver for Lenovo SE10 platform
Linus Torvalds [Mon, 20 May 2024 15:55:18 +0000 (08:55 -0700)]
Merge tag 'i2c-for-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux
Pull i2c updates from Wolfram Sang:
"i2c core removes an argument from the i2c_mux_add_adapter() call to
further deprecate class based I2C device instantiation. All users are
converted, too.
Other that that, Andi collected a number if I2C host driver patches.
Those merges have their own description"
* tag 'i2c-for-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux: (72 commits)
power: supply: sbs-manager: Remove class argument from i2c_mux_add_adapter()
i2c: mux: Remove class argument from i2c_mux_add_adapter()
i2c: synquacer: Fix an error handling path in synquacer_i2c_probe()
i2c: acpi: Unbind mux adapters before delete
i2c: designware: Replace MODULE_ALIAS() with MODULE_DEVICE_TABLE()
i2c: pxa: use 'time_left' variable with wait_event_timeout()
i2c: s3c2410: use 'time_left' variable with wait_event_timeout()
i2c: rk3x: use 'time_left' variable with wait_event_timeout()
i2c: qcom-geni: use 'time_left' variable with wait_for_completion_timeout()
i2c: jz4780: use 'time_left' variable with wait_for_completion_timeout()
i2c: synquacer: use 'time_left' variable with wait_for_completion_timeout()
i2c: stm32f7: use 'time_left' variable with wait_for_completion_timeout()
i2c: stm32f4: use 'time_left' variable with wait_for_completion_timeout()
i2c: st: use 'time_left' variable with wait_for_completion_timeout()
i2c: omap: use 'time_left' variable with wait_for_completion_timeout()
i2c: imx-lpi2c: use 'time_left' variable with wait_for_completion_timeout()
i2c: hix5hd2: use 'time_left' variable with wait_for_completion_timeout()
i2c: exynos5: use 'time_left' variable with wait_for_completion_timeout()
i2c: digicolor: use 'time_left' variable with wait_for_completion_timeout()
i2c: amd-mp2-plat: use 'time_left' variable with wait_for_completion_timeout()
...
Linus Torvalds [Mon, 20 May 2024 15:47:54 +0000 (08:47 -0700)]
Merge tag 'v6.10-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Pull crypto fixes from Herbert Xu:
"Fix a bug in the new ecc P521 code as well as a buggy fix in qat"
* tag 'v6.10-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
crypto: ecc - Prevent ecc_digits_from_bytes from reading too many bytes
crypto: qat - Fix ADF_DEV_RESET_SYNC memory leak
Dan Aloni [Mon, 6 May 2024 09:37:59 +0000 (12:37 +0300)]
rpcrdma: fix handling for RDMA_CM_EVENT_DEVICE_REMOVAL
Under the scenario of IB device bonding, when bringing down one of the
ports, or all ports, we saw xprtrdma entering a non-recoverable state
where it is not even possible to complete the disconnect and shut it
down the mount, requiring a reboot. Following debug, we saw that
transport connect never ended after receiving the
RDMA_CM_EVENT_DEVICE_REMOVAL callback.
The DEVICE_REMOVAL callback is irrespective of whether the CM_ID is
connected, and ESTABLISHED may not have happened. So need to work with
each of these states accordingly.
Anna Schumaker [Thu, 25 Apr 2024 20:34:40 +0000 (16:34 -0400)]
NFS: Don't enable NFS v2 by default
This came up during one of the Bake-a-thon discussions. NFS v2 support
was dropped from nfs-utils/mount.nfs in December 2021. Let's turn it
off by default in the kernel too, since this means there isn't a way
to mount and test it.
Anna Schumaker [Thu, 25 Apr 2024 20:24:29 +0000 (16:24 -0400)]
NFS: Fix READ_PLUS when server doesn't support OP_READ_PLUS
Olga showed me a case where the client was sending multiple READ_PLUS
calls to the server in parallel, and the server replied
NFS4ERR_OPNOTSUPP to each. The client would fall back to READ for the
first reply, but fail to retry the other calls.
I fix this by removing the test for NFS_CAP_READ_PLUS in
nfs4_read_plus_not_supported(). This allows us to reschedule any
READ_PLUS call that has a NFS4ERR_OPNOTSUPP return value, even after the
capability has been cleared.
Dan Aloni [Thu, 25 Apr 2024 10:49:38 +0000 (13:49 +0300)]
sunrpc: fix NFSACL RPC retry on soft mount
It used to be quite awhile ago since 1b63a75180c6 ('SUNRPC: Refactor
rpc_clone_client()'), in 2012, that `cl_timeout` was copied in so that
all mount parameters propagate to NFSACL clients. However since that
change, if mount options as follows are given:
These values lead to NFSACL operations not being retried under the
condition of transient network outages with soft mount. Instead, getacl
call fails after 60 seconds with EIO.
The simple fix is to pass the existing client's `cl_timeout` as the new
client timeout.
In the case where we have received a successful reply to an RPC request,
but while processing the reply the client in rpc_decode_header() finds
an expired context, the code ends up propagating the error to the caller
instead of getting a new context and retrying the request.
To give more details, in rpc_decode_header() we call rpcauth_checkverf()
will call into the gss and internally will at some point call
gss_validate() which has a check if the current’s context lifetime
expired, and it would fail. The reason for the failure gets ‘scrubbed’
and translated to EACCES so when we get back to rpc_decode_header() we
just go to “out_verifier” which for that error would get converted to
“out_garbage” (ie it’s treated as garballed reply) and the next
action is call_encode. Which (1) doesn’t reencode or re-send (not to
mention no upcall happens because context expires as that reason just
not known) and it again fails in the same decoding process. After
re-trying it 3 times the error is propagated back to the caller
(ie nfs4_write_done_cb() in the case a failing write).
To fix this, instead we need to look to the case where the server
decides that context has expired and replies with an RPC auth error.
In that case, the rpc_decode_header() goes to "out_msg_denied" in that
we return EKEYREJECTED which in call_decode() is sent to “call_reserve”
which triggers an upcalls and a re-try of the operation.
The proposed fix is in case of a failed rpc_decode_header() to check
if credentials were set to be invalid and use that as a proxy for
deciding that context has expired and then treat is same way as
receiving an auth error.
Martin Kaiser [Sun, 14 Apr 2024 17:01:09 +0000 (19:01 +0200)]
nfs: keep server info for remounts
With newer kernels that use fs_context for nfs mounts, remounts fail with
-EINVAL.
$ mount -t nfs -o nolock 10.0.0.1:/tmp/test /mnt/test/
$ mount -t nfs -o remount /mnt/test/
mount: mounting 10.0.0.1:/tmp/test on /mnt/test failed: Invalid argument
For remounts, the nfs server address and port are populated by
nfs_init_fs_context and later overwritten with 0x00 bytes by
nfs23_parse_monolithic. The remount then fails as the server address is
invalid.
Fix this by not overwriting nfs server info in nfs23_parse_monolithic if
we're doing a remount.
Fixes: f2aedb713c28 ("NFS: Add fs_context support.") Signed-off-by: Martin Kaiser <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
Dan Carpenter reports smatch warning for nfs4_try_migration() when a memory
allocation failure results in a zero return value. In this case, a
transient allocation failure error will likely be retried the next time the
server responds with NFS4ERR_MOVED.
We can fixup the smatch warning with a small refactor: attempt all three
allocations before testing and returning on a failure.
NFS: make sure lock/nolock overriding local_lock mount option
Currently, mount option lock/nolock and local_lock option
may override NFS_MOUNT_LOCAL_FLOCK NFS_MOUNT_LOCAL_FCNTL flags
when passing in different order:
mount -o vers=3,local_lock=all,lock:
local_lock=none
mount -o vers=3,lock,local_lock=all:
local_lock=all
This patch will let lock/nolock override local_lock option
as nfs(5) suggested.
NeilBrown [Mon, 25 Mar 2024 05:36:05 +0000 (16:36 +1100)]
NFS: add atomic_open for NFSv3 to handle O_TRUNC correctly.
With two clients, each with NFSv3 mounts of the same directory, the sequence:
client1 client2
ls -l afile
echo hello there > afile
echo HELLO > afile
cat afile
will show
HELLO
there
because the O_TRUNC requested in the final 'echo' doesn't take effect.
This is because the "Negative dentry, just create a file" section in
lookup_open() assumes that the file *does* get created since the dentry
was negative, so it sets FMODE_CREATED, and this causes do_open() to
clear O_TRUNC and so the file doesn't get truncated.
Even mounting with -o lookupcache=none does not help as
nfs_neg_need_reval() always returns false if LOOKUP_CREATE is set.
This patch fixes the problem by providing an atomic_open inode operation
for NFSv3 (and v2). The code is largely the code from the branch in
lookup_open() when atomic_open is not provided. The significant change
is that the O_TRUNC flag is passed a new nfs_do_create() which add
'trunc' handling to nfs_create().
With this change we also optimise away an unnecessary LOOKUP before the
file is created.
Anna Schumaker [Wed, 20 Mar 2024 21:11:20 +0000 (17:11 -0400)]
pNFS/filelayout: Specify the layout segment range in LAYOUTGET
Move from only requesting full file layout segments to requesting layout
segments that match our I/O size. This means the server is still free to
return a full file layout if it wants, but partial layouts will no
longer cause an error.
Shenghao Ding [Sat, 18 May 2024 14:15:46 +0000 (22:15 +0800)]
ASoC: tas2781: Fix wrong loading calibrated data sequence
Calibrated data will be set to default after loading DSP config params,
which will cause speaker protection work abnormally. Reload calibrated
data after loading DSP config params. Remove declaration of unused API
which load calibrated data in wrong sequence, changed the copyright year
and correct file name in license
header.
The framework change to add D_GNU_SOURCE to KHDR_INCLUDES
to Makefile, lib.mk, and kselftest_harness.h is reverted
as it is causing build failures and warnings.
Revert this change as this change depends on the framework
change.
The framework change to add D_GNU_SOURCE to KHDR_INCLUDES
to Makefile, lib.mk, and kselftest_harness.h is reverted
as it is causing build failures and warnings.
Revert this change as this change depends on the framework
change.
Ryosuke Yasuoka [Sun, 19 May 2024 09:43:03 +0000 (18:43 +0900)]
nfc: nci: Fix uninit-value in nci_rx_work
syzbot reported the following uninit-value access issue [1]
nci_rx_work() parses received packet from ndev->rx_q. It should be
validated header size, payload size and total packet size before
processing the packet. If an invalid packet is detected, it should be
silently discarded.
Fixes: d24b03535e5e ("nfc: nci: Fix uninit-value in nci_dev_up and nci_ntf_packet") Reported-and-tested-by: [email protected] Closes: https://syzkaller.appspot.com/bug?extid=d7b4dc6cd50410152534 [1] Signed-off-by: Ryosuke Yasuoka <[email protected]> Reviewed-by: Krzysztof Kozlowski <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Taehee Yoo [Sat, 18 May 2024 13:20:52 +0000 (13:20 +0000)]
selftests: net: kill smcrouted in the cleanup logic in amt.sh
The amt.sh requires smcrouted for multicasting routing.
So, it starts smcrouted before forwarding tests.
It must be stopped after all tests, but it isn't.
To fix this issue, it kills smcrouted in the cleanup logic.
Fixes: c08e8baea78e ("selftests: add amt interface selftest script") Signed-off-by: Taehee Yoo <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Andrea Mayer [Fri, 17 May 2024 16:45:41 +0000 (18:45 +0200)]
ipv6: sr: fix missing sk_buff release in seg6_input_core
The seg6_input() function is responsible for adding the SRH into a
packet, delegating the operation to the seg6_input_core(). This function
uses the skb_cow_head() to ensure that there is sufficient headroom in
the sk_buff for accommodating the link-layer header.
In the event that the skb_cow_header() function fails, the
seg6_input_core() catches the error but it does not release the sk_buff,
which will result in a memory leak.
This issue was introduced in commit af3b5158b89d ("ipv6: sr: fix BUG due
to headroom too small after SRH push") and persists even after commit 7a3f5b0de364 ("netfilter: add netfilter hooks to SRv6 data plane"),
where the entire seg6_input() code was refactored to deal with netfilter
hooks.
The proposed patch addresses the identified memory leak by requiring the
seg6_input_core() function to release the sk_buff in the event that
skb_cow_head() fails.
Fixes: af3b5158b89d ("ipv6: sr: fix BUG due to headroom too small after SRH push") Signed-off-by: Andrea Mayer <[email protected]> Reviewed-by: Simon Horman <[email protected]> Reviewed-by: David Ahern <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Florian Fainelli [Thu, 16 May 2024 16:56:30 +0000 (09:56 -0700)]
net: Always descend into dsa/ folder with CONFIG_NET_DSA enabled
Stephen reported that he was unable to get the dsa_loop driver to get
probed, and the reason ended up being because he had CONFIG_FIXED_PHY=y
in his kernel configuration. As Masahiro explained it:
"obj-m += dsa/" means everything under dsa/ must be modular.
If there is a built-in object under dsa/ with CONFIG_NET_DSA=m,
you cannot do "obj-$(CONFIG_NET_DSA) += dsa/".
You need to change it back to "obj-y += dsa/".
This was the case here whereby CONFIG_NET_DSA=m, and so the
obj-$(CONFIG_FIXED_PHY) += dsa_loop_bdinfo.o rule is not executed and
the DSA loop mdio_board info structure is not registered with the
kernel, and eventually the device is simply not found.
To preserve the intention of the original commit of limiting the amount
of folder descending, conditionally descend into drivers/net/dsa when
CONFIG_NET_DSA is enabled.
Kent Overstreet [Mon, 20 May 2024 00:09:35 +0000 (20:09 -0400)]
bcachefs: bch2_checksum() returns 0 for unknown checksum type
This fixes missing guards on trying to calculate a checksum with an
invalid/unknown checksum type; moving the guards up to e.g. btree_io.c
might be "more correct", but doesn't buy us anything - an unknown
checksum type will always be flagged as at least a checksum error so we
aren't losing any safety doing it this way and it makes it less likely
to accidentally pop an assert we don't want.
Kent Overstreet [Sun, 19 May 2024 23:58:44 +0000 (19:58 -0400)]
bcachefs: Fix bch2_alloc_ciphers()
Don't put error pointers in bch_fs, that's gross.
This fixes (?) the check in bch2_checksum_type_valid() - depending on
our error paths, or depending on what our error paths are doing it at
least makes the code saner.
Kent Overstreet [Sun, 19 May 2024 21:28:44 +0000 (17:28 -0400)]
bcachefs: Fix shift overflows in replicas.c
We can't disallow unknown data_types in verify() - we have to preserve
them unchanged for backwards compat; that means we have to add a few
more guards.
Youling Tang [Fri, 10 May 2024 06:10:58 +0000 (14:10 +0800)]
bcachefs: set FMODE_CAN_ODIRECT instead of a dummy direct_IO method
Since commit a2ad63daa88b ("VFS: add FMODE_CAN_ODIRECT file flag") file
systems can just set the FMODE_CAN_ODIRECT flag at open time instead of
wiring up a dummy direct_IO method to indicate support for direct I/O.
Do that for bcachefs so that noop_direct_IO can eventually be removed.
Similar to commit b29434999371 ("xfs: set FMODE_CAN_ODIRECT instead of
a dummy direct_IO method").
Masahiro Yamada [Sun, 4 Feb 2024 12:20:03 +0000 (21:20 +0900)]
x86/kconfig: Select ARCH_WANT_FRAME_POINTERS again when UNWINDER_FRAME_POINTER=y
It took me some time to understand the purpose of the tricky code at
the end of arch/x86/Kconfig.debug.
Without it, the following would be shown:
WARNING: unmet direct dependencies detected for FRAME_POINTER
because
81d387190039 ("x86/kconfig: Consolidate unwinders into multiple choice selection")
removed 'select ARCH_WANT_FRAME_POINTERS'.
The correct and more straightforward approach should have been to move
it where 'select FRAME_POINTER' is located.
Several architectures properly handle the conditional selection of
ARCH_WANT_FRAME_POINTERS. For example, 'config UNWINDER_FRAME_POINTER'
in arch/arm/Kconfig.debug.
Allen Pais [Wed, 27 Mar 2024 16:03:10 +0000 (16:03 +0000)]
mailbox: Convert from tasklet to BH workqueue
The only generic interface to execute asynchronously in the BH context is
tasklet; however, it's marked deprecated and has some design flaws. To
replace tasklets, BH workqueue support was recently added. A BH workqueue
behaves similarly to regular workqueues except that the queued work items
are executed in the BH context.
Based on the work done by Tejun Heo <[email protected]>
Branch: https://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git for-6.10
mailbox: mtk-cmdq: Fix pm_runtime_get_sync() warning in mbox shutdown
The return value of pm_runtime_get_sync() in cmdq_mbox_shutdown()
will return 1 when pm runtime state is active, and we don't want to
get the warning message in this case.
So we change the return value < 0 for WARN_ON().
Fixes: 8afe816b0c99 ("mailbox: mtk-cmdq-mailbox: Implement Runtime PM with autosuspend") Signed-off-by: Jason-JH.Lin <[email protected]> Reviewed-by: AngeloGioacchino Del Regno <[email protected]> Signed-off-by: Jassi Brar <[email protected]>
Tanmay Shah [Fri, 3 May 2024 14:15:45 +0000 (07:15 -0700)]
mailbox: zynqmp: handle SGI for shared IPI
At least one IPI is used in TF-A for communication with PMC firmware.
If this IPI needs to be used by other agents such as RPU then, IPI
system interrupt can't be generated in mailbox driver. In such case
TF-A generates SGI to mailbox driver for IPI notification.
Andrew Davis [Wed, 10 Apr 2024 13:59:42 +0000 (08:59 -0500)]
mailbox: omap: Remove kernel FIFO message queuing
The kernel FIFO queue has a couple issues. The biggest issue is that
it causes extra latency in a path that can be used in real-time tasks,
such as communication with real-time remote processors.
The whole FIFO idea itself looks to be a leftover from before the
unified mailbox framework. The current mailbox framework expects
mbox_chan_received_data() to be called with data immediately as it
arrives. Remove the FIFO and pass the messages to the mailbox
framework directly as part of a threaded IRQ handler.
Andrew Davis [Wed, 10 Apr 2024 13:59:41 +0000 (08:59 -0500)]
mailbox: omap: Reverse FIFO busy check logic
It is much more clear to check if the hardware FIFO is full and return
EBUSY if true. This allows us to also remove one level of indention
from the core of this function. It also makes the similarities between
omap_mbox_chan_send_noirq() and omap_mbox_chan_send() more obvious.
Andrew Davis [Wed, 10 Apr 2024 13:59:40 +0000 (08:59 -0500)]
mailbox: omap: Remove mbox_chan_to_omap_mbox()
This function only checks if mbox_chan *chan is not NULL, but that cannot
be the case and if it was returning NULL which is not later checked
doesn't save us from this. The second check for chan->con_priv is
completely redundant as if it was NULL we would return NULL just the
same. Simply dereference con_priv directly and remove this function.
Andrew Davis [Wed, 10 Apr 2024 13:59:39 +0000 (08:59 -0500)]
mailbox: omap: Use mbox_controller channel list directly
The driver stores a list of omap_mbox structs so it can later use it to
lookup the mailbox names in of_xlate. This same information is already
available in the mbox_controller passed into of_xlate. Simply use that
data and remove the extra allocation and storage of the omap_mbox list.
Andrew Davis [Wed, 10 Apr 2024 13:59:38 +0000 (08:59 -0500)]
mailbox: omap: Use function local struct mbox_controller
The mbox_controller struct is only needed in the probe function. Make
it a local variable instead of storing a copy in omap_mbox_device
to simplify that struct.
Currently the driver loops through all mailbox child nodes twice, once
to read in data from each node, and again to make use of this data.
Instead read the data and make use of it in one pass. This removes
the need for several temporary data structures and reduces the
complexity of this main loop in probe.
Andrew Davis [Wed, 10 Apr 2024 13:59:35 +0000 (08:59 -0500)]
mailbox: omap: Remove device class
The driver currently creates a new device class "mbox". Then for each
mailbox adds a device to that class. This class provides no file
operations provided for any userspace users of this device class.
It may have been extended to be functional in our vendor tree at
some point, but that is not the case anymore, nor does it matter
for the upstream tree.
Remove this device class and related functions and variables.
This also allows us to switch to module_platform_driver() as
there is nothing left to do in module_init().
The type of message sent using omap-mailbox is always u32. The definition
of mbox_msg_t is uintptr_t which is wrong as that type changes based on
the architecture (32bit vs 64bit). This type should have been defined as
u32. Instead of making that change here, simply remove the header usage
and fix the last couple users of the same in this driver.
Andrew Davis [Wed, 10 Apr 2024 13:59:32 +0000 (08:59 -0500)]
mailbox: omap: Move omap_mbox_irq_t into driver
This is only used internal to the driver, move it out of the
public header and into the driver file. While we are here,
this is not used as a bitwise, so drop that and make it a
simple enum type.
If an input device declares too many capability bits then modalias
string for such device may become too long and not fit into uevent
buffer, resulting in failure of sending said uevent. This, in turn,
may prevent userspace from recognizing existence of such devices.
This is typically not a concern for real hardware devices as they have
limited number of keys, but happen with synthetic devices such as
ones created by xen-kbdfront driver, which creates devices as being
capable of delivering all possible keys, since it doesn't know what
keys the backend may produce.
To deal with such devices input core will attempt to trim key data,
in the hope that the rest of modalias string will fit in the given
buffer. When trimming key data it will indicate that it is not
complete by placing "+," sign, resulting in conversions like this:
This should allow existing udev rules continue to work with existing
devices, and will also allow writing more complex rules that would
recognize trimmed modalias and check input device characteristics by
other means (for example by parsing KEY= data in uevent or parsing
input device sysfs attributes).
Note that the driver core may try adding more uevent environment
variables once input core is done adding its own, so when forming
modalias we can not use the entire available buffer, so we reduce
it by somewhat an arbitrary amount (96 bytes).
Andy Shevchenko [Tue, 7 May 2024 20:01:31 +0000 (23:01 +0300)]
bitops: Move aligned_byte_mask() to wordpart.h
The bitops.h is for bit related operations. The aligned_byte_mask()
is about byte (or part of the machine word) operations, for which
we have a separate header, move the mentioned macro to wordpart.h
to consolidate similar operations.
Dave Chinner [Tue, 30 Apr 2024 05:28:25 +0000 (15:28 +1000)]
mm/page-owner: use gfp_nested_mask() instead of open coded masking
The page-owner tracking code records stack traces during page allocation.
To do this, it must do a memory allocation for the stack information from
inside an existing memory allocation context. This internal allocation
must obey the high level caller allocation constraints to avoid generating
false positive warnings that have nothing to do with the code they are
instrumenting/tracking (e.g. through lockdep reclaim state tracking)
We also don't want recording stack traces to deplete emergency memory
reserves - debug code is useless if it creates new issues that can't be
replicated when the debug code is disabled.
Switch the stack tracking allocation masking to use gfp_nested_mask() to
address these issues. gfp_nested_mask() naturally strips GFP_ZONEMASK,
too, which greatly simplifies this code.
Dave Chinner [Tue, 30 Apr 2024 05:28:24 +0000 (15:28 +1000)]
stackdepot: use gfp_nested_mask() instead of open coded masking
The stackdepot code is used by KASAN and lockdep for recoding stack
traces. Both of these track allocation context information, and so their
internal allocations must obey the caller allocation contexts to avoid
generating their own false positive warnings that have nothing to do with
the code they are instrumenting/tracking.
We also don't want recording stack traces to deplete emergency memory
reserves - debug code is useless if it creates new issues that can't be
replicated when the debug code is disabled.
Switch the stackdepot allocation masking to use gfp_nested_mask() to
address these issues. gfp_nested_mask() also strips GFP_ZONEMASK
naturally, so that greatly simplifies this code.
Tl;dr: Memory allocations that are done inside the public memory
allocation API need to obey the reclaim recursion constraints placed on
the allocation by the original caller, including the "don't track
recursion for this allocation" case defined by __GFP_NOLOCKDEP.
These nested allocations are generally in debug code that is tracking
something about the allocation (kmemleak, KASAN, etc) and so are
allocating private kernel objects that only that debug system will use.
Neither the page-owner code nor the stack depot code get this right. They
also also clear GFP_ZONEMASK as a separate operation, which is completely
redundant because the constraint filter applied immediately after
guarantees that GFP_ZONEMASK bits are cleared.
kmemleak gets this filtering right. It preserves the allocation
constraints for deadlock prevention and clears all other context flags
whilst also ensuring that the nested allocation will fail quickly,
silently and without depleting emergency kernel reserves if there is no
memory available.
This can be made much more robust, immune to whack-a-mole games and the
code greatly simplified by lifting gfp_kmemleak_mask() to
include/linux/gfp.h and using that everywhere. Also document it so that
there is no excuse for not knowing about it when writing new debug code
that nests allocations.
Tested with lockdep, KASAN + page_owner=on and kmemleak=on over multiple
fstests runs with XFS.
This patch (of 3):
Any "internal" nested allocation done from within an allocation context
needs to obey the high level allocation gfp_mask constraints. This is
necessary for debug code like KASAN, kmemleak, lockdep, etc that allocate
memory for saving stack traces and other information during memory
allocation. If they don't obey things like __GFP_NOLOCKDEP or
__GFP_NOWARN, they produce false positive failure detections.
kmemleak gets this right by using gfp_kmemleak_mask() to pass through the
relevant context flags to the nested allocation to ensure that the
allocation follows the constraints of the caller context.
KASAN recently was foudn to be missing __GFP_NOLOCKDEP due to stack depot
allocations, and even more recently the page owner tracking code was also
found to be missing __GFP_NOLOCKDEP support.
We also don't wan't want KASAN or lockdep to drive the system into OOM
kill territory by exhausting emergency reserves. This is something that
kmemleak also gets right by adding (__GFP_NORETRY | __GFP_NOMEMALLOC |
__GFP_NOWARN) to the allocation mask.
Hence it is clear that we need to define a common nested allocation filter
mask for these sorts of third party nested allocations used in debug code.
So to start this process, lift gfp_kmemleak_mask() to gfp.h and rename it
to gfp_nested_mask(), and convert the kmemleak callers to use it.
Ryusuke Konishi [Sat, 11 May 2024 00:29:42 +0000 (09:29 +0900)]
nilfs2: make block erasure safe in nilfs_finish_roll_forward()
The implementation of writing a zero-fill block in
nilfs_finish_roll_forward() is not safe. The buffer is being cleared
without acquiring a lock or setting the uptodate flag, so theoretically,
between the time the buffer's data is cleared and the time it is written
back to the block device using sync_dirty_buffer(), that zero data can be
undone by concurrent block device reads.
Since this buffer points to a location that has been read from disk once,
the uptodate flag will most likely remain, but since it was obtained with
__getblk(), that is not guaranteed. In other words, this is exceptional,
and this function itself is not normally called (only once when mounting
after a specific pattern of unclean shutdown), so it is highly unlikely
that this will actually cause a problem.
Anyway, eliminate this potential race issue by protecting the clearing of
buffer data with a buffer lock and setting the buffer's uptodate flag
within the protected section.
Tao Su [Thu, 9 May 2024 05:31:13 +0000 (13:31 +0800)]
selftests/harness: use 1024 in place of LINE_MAX
Android was seeing a compilation error because its C library does not
define LINE_MAX. Since LINE_MAX is only used to determine the size of
test_name[] and 1024 should be enough for the test name, use 1024 instead
of LINE_MAX.
Tao Su [Thu, 9 May 2024 05:31:12 +0000 (13:31 +0800)]
Revert "selftests/harness: remove use of LINE_MAX"
Patch series "Selftests: Fix compilation warnings due to missing
_GNU_SOURCE definition", v2.
Since kselftest_harness.h introduces asprintf()[1], many selftests have
compilation warnings or errors due to missing _GNU_SOURCE definitions.
The issue stems from a lack of a LINE_MAX definition in Android (see
commit 38c957f07038), which is the reason why asprintf() was introduced.
We tried adding _GNU_SOURCE definitions to more selftests to fix, but
asprintf() may continue to cause problems, and since it is quite late in
the 6.9 cycle, we would like to revert 809216233555 first to provide
testing for forks[2].
asprintf() is declared in stdio.h when defining _GNU_SOURCE, but stdio.h
is so common that many files don't define _GNU_SOURCE before including
stdio.h, and defining _GNU_SOURCE after including stdio.h will no longer
take effect, which causes warnings or even errors during compilation in
many selftests.
Revert 'commit 809216233555 ("selftests/harness: remove use of LINE_MAX")'
as that came in quite late in the 6.9 cycle.
Samuel Holland [Fri, 29 Mar 2024 07:18:29 +0000 (00:18 -0700)]
selftests/fpu: move FP code to a separate translation unit
This ensures no compiler-generated floating-point code can appear outside
kernel_fpu_{begin,end}() sections, and some architectures enforce this
separation.
Samuel Holland [Fri, 29 Mar 2024 07:18:28 +0000 (00:18 -0700)]
drm/amd/display: use ARCH_HAS_KERNEL_FPU_SUPPORT
Now that all previously-supported architectures select
ARCH_HAS_KERNEL_FPU_SUPPORT, this code can depend on that symbol instead
of the existing list of architectures. It can also take advantage of the
common kernel-mode FPU API and method of adjusting CFLAGS.
projects / linux.git / log