Zhang Yi [Mon, 5 Jan 2026 01:45:21 +0000 (09:45 +0800)]
ext4: simplify the mapping query logic in ext4_iomap_begin()
In the write path mapping check of ext4_iomap_begin(), the return value
'ret' should never greater than orig_mlen. If 'ret' equals 'orig_mlen',
it can be returned directly without checking IOMAP_ATOMIC.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Link: https://patch.msgid.link/20260105014522.1937690-7-yi.zhang@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Mon, 5 Jan 2026 01:45:19 +0000 (09:45 +0800)]
ext4: remove useless ext4_iomap_overwrite_ops
ext4_iomap_overwrite_ops was introduced in commit 8cd115bdda17 ("ext4:
Optimize ext4 DIO overwrites"), which can optimize pure overwrite
performance by dropping the IOMAP_WRITE flag to only query the mapped
mapping information. This avoids starting a new journal handle, thereby
improving speed. Later, commit 9faac62d4013 ("ext4: optimize file
overwrites") also optimized similar scenarios, but it performs the check
later, examining the mappings status only when the actual block mapping
is needed. Thus, it can handle the previous commit scenario. That means
in the case of an overwrite scenario, the condition
"offset + length <= i_size_read(inode)" in the write path must always be
true.
Therefore, it is acceptable to remove the ext4_iomap_overwrite_ops,
which will also clarify the write and read paths of ext4_iomap_begin.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Link: https://patch.msgid.link/20260105014522.1937690-5-yi.zhang@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Mon, 5 Jan 2026 01:45:18 +0000 (09:45 +0800)]
ext4: avoid starting handle when dio writing an unwritten extent
Since we have deferred the split of the unwritten extent until after I/O
completion, it is not necessary to initiate the journal handle when
submitting the I/O.
This can improve the write performance of concurrent DIO for multiple
files. The fio tests below show a ~25% performance improvement when
wirting to unwritten files on my VM with a mem disk.
Zhang Yi [Mon, 5 Jan 2026 01:45:17 +0000 (09:45 +0800)]
ext4: don't split extent before submitting I/O
Currently, when writing back dirty pages to the filesystem with the
dioread_nolock feature enabled and when doing DIO, if the area to be
written back is part of an unwritten extent, the
EXT4_GET_BLOCKS_IO_CREATE_EXT flag is set during block allocation before
submitting I/O. The function ext4_split_convert_extents() then attempts
to split this extent in advance. This approach is designed to prevents
extent splitting and conversion to the written type from failing due to
insufficient disk space at the time of I/O completion, which could
otherwise result in data loss.
However, we already have two mechanisms to ensure successful extent
conversion. The first is the EXT4_GET_BLOCKS_METADATA_NOFAIL flag, which
is a best effort, it permits the use of 2% of the reserved space or
4,096 blocks in the file system when splitting extents. This flag covers
most scenarios where extent splitting might fail. The second is the
EXT4_EXT_MAY_ZEROOUT flag, which is also set during extent splitting. If
the reserved space is insufficient and splitting fails, it does not
retry the allocation. Instead, it directly zeros out the extra part of
the extent, thereby avoiding splitting and directly converting the
entire extent to the written type.
These two mechanisms also exist when I/Os are completed because there is
a concurrency window between write-back and fallocate, which may still
require us to split extents upon I/O completion. There is no much
difference between splitting extents before submitting I/O. Therefore,
It seems possible to defer the splitting until I/O completion, it won't
increase the risk of I/O failure and data loss. On the contrary, if some
I/Os can be merged when I/O completion, it can also reduce unnecessary
splitting operations, thereby alleviating the pressure on reserved
space.
In addition, deferring extent splitting until I/O completion can
also simplify the IO submission process and avoid initiating unnecessary
journal handles when writing unwritten extents.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Link: https://patch.msgid.link/20260105014522.1937690-3-yi.zhang@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Mon, 5 Jan 2026 01:45:16 +0000 (09:45 +0800)]
ext4: use reserved metadata blocks when splitting extent on endio
When performing buffered writes, we may need to split and convert an
unwritten extent into a written one during the end I/O process. However,
we do not reserve space specifically for these metadata changes, we only
reserve 2% of space or 4096 blocks. To address this, we use
EXT4_GET_BLOCKS_PRE_IO to potentially split extents in advance and
EXT4_GET_BLOCKS_METADATA_NOFAIL to utilize reserved space if necessary.
These two approaches can reduce the likelihood of running out of space
and losing data. However, these methods are merely best efforts, we
could still run out of space, and there is not much difference between
converting an extent during the writeback process and the end I/O
process, it won't increase the risk of losing data if we postpone the
conversion.
Therefore, also use EXT4_GET_BLOCKS_METADATA_NOFAIL in
ext4_convert_unwritten_extents_endio() to prepare for the buffered I/O
iomap conversion, which may perform extent conversion during the end I/O
process.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Link: https://patch.msgid.link/20260105014522.1937690-2-yi.zhang@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zilin Guan [Thu, 25 Dec 2025 08:48:00 +0000 (08:48 +0000)]
ext4: fix memory leak in ext4_ext_shift_extents()
In ext4_ext_shift_extents(), if the extent is NULL in the while loop, the
function returns immediately without releasing the path obtained via
ext4_find_extent(), leading to a memory leak.
Fix this by jumping to the out label to ensure the path is properly
released.
Zhang Yi [Tue, 23 Dec 2025 01:19:27 +0000 (09:19 +0800)]
ext4: don't order data when zeroing unwritten or delayed block
When zeroing out a written partial block, it is necessary to order the
data to prevent exposing stale data on disk. However, if the buffer is
unwritten or delayed, it is not allocated as written, so ordering the
data is not required. This can prevent strange and unnecessary ordered
writes when appending data across a region within a block.
Assume we have a 2K unwritten file on a filesystem with 4K blocksize,
and buffered write from 3K to 4K. Before this patch,
__ext4_block_zero_page_range() would add the range [2k,3k) to the
ordered range, and then the JBD2 commit process would write back this
block. However, it does nothing since the block is not mapped as
written, this folio will be redirtied and written back agian through the
normal write back process.
Li Chen [Thu, 11 Dec 2025 11:51:42 +0000 (19:51 +0800)]
ext4: mark group extend fast-commit ineligible
Fast commits only log operations that have dedicated replay support.
EXT4_IOC_GROUP_EXTEND grows the filesystem to the end of the last
block group and updates the same on-disk metadata without going
through the fast commit tracking paths.
In practice these operations are rare and usually followed by further
updates, but mixing them into a fast commit makes the overall
semantics harder to reason about and risks replay gaps if new call
sites appear.
Teach ext4 to mark the filesystem fast-commit ineligible when
EXT4_IOC_GROUP_EXTEND grows the filesystem.
This forces those transactions to fall back to a full commit,
ensuring that the group extension changes are captured by the normal
journal rather than partially encoded in fast commit TLVs.
This change should not affect common workloads but makes online
resize via GROUP_EXTEND safer and easier to reason about under fast
commit.
Testing:
1. prepare:
dd if=/dev/zero of=/root/fc_resize.img bs=1M count=0 seek=256
mkfs.ext4 -O fast_commit -F /root/fc_resize.img
mkdir -p /mnt/fc_resize && mount -t ext4 -o loop /root/fc_resize.img /mnt/fc_resize
2. Extended the filesystem to the end of the last block group using a
helper that calls EXT4_IOC_GROUP_EXTEND on the mounted filesystem
and checked fc_info:
./group_extend_helper /mnt/fc_resize
cat /proc/fs/ext4/loop0/fc_info
shows the "Resize" ineligible reason increased.
3. Fsynced a file on the resized filesystem and confirmed that the fast
commit ineligible counter incremented for the resize transaction:
touch /mnt/fc_resize/file
/root/fsync_file /mnt/fc_resize/file
sync
cat /proc/fs/ext4/loop0/fc_info
Li Chen [Thu, 11 Dec 2025 11:51:41 +0000 (19:51 +0800)]
ext4: mark group add fast-commit ineligible
Fast commits only log operations that have dedicated replay support.
Online resize via EXT4_IOC_GROUP_ADD updates the superblock and group
descriptor metadata without going through the fast commit tracking
paths.
In practice these operations are rare and usually followed by further
updates, but mixing them into a fast commit makes the overall
semantics harder to reason about and risks replay gaps if new call
sites appear.
Teach ext4 to mark the filesystem fast-commit ineligible when
ext4_ioctl_group_add() adds new block groups.
This forces those transactions to fall back to a full commit,
ensuring that the filesystem geometry updates are captured by the
normal journal rather than partially encoded in fast commit TLVs.
This change should not affect common workloads but makes online
resize via GROUP_ADD safer and easier to reason about under fast
commit.
Testing:
1. prepare:
dd if=/dev/zero of=/root/fc_resize.img bs=1M count=0 seek=256
mkfs.ext4 -O fast_commit -F /root/fc_resize.img
mkdir -p /mnt/fc_resize && mount -t ext4 -o loop /root/fc_resize.img /mnt/fc_resize
2. Ran a helper that issues EXT4_IOC_GROUP_ADD on the mounted
filesystem and checked the resize ineligible reason:
./group_add_helper /mnt/fc_resize
cat /proc/fs/ext4/loop0/fc_info
shows "Resize": > 0.
3. Fsynced a file on the resized filesystem and verified that the fast
commit stats report at least one ineligible commit:
touch /mnt/fc_resize/file
/root/fsync_file /mnt/fc_resize/file
sync
cat /proc/fs/ext4/loop0/fc_info
shows fc stats ineligible > 0.
Li Chen [Thu, 11 Dec 2025 11:51:40 +0000 (19:51 +0800)]
ext4: mark move extents fast-commit ineligible
Fast commits only log operations that have dedicated replay support.
EXT4_IOC_MOVE_EXT swaps extents between regular files and may copy
data, rewriting the affected inodes' block mapping layout without
going through the fast commit tracking paths.
In practice these operations are rare and usually followed by further
updates, but mixing them into a fast commit makes the overall
semantics harder to reason about and risks replay gaps if new call
sites appear.
Teach ext4 to mark the filesystem fast-commit ineligible for the
journal transactions used by move_extent_per_page() when
EXT4_IOC_MOVE_EXT runs.
This forces those transactions to fall back to a full commit,
ensuring that these multi-inode extent swaps are captured by the
normal journal rather than partially encoded in fast commit TLVs.
This change should not affect common workloads but makes online
defragmentation safer and easier to reason about under fast commit.
Testing:
1. prepare:
dd if=/dev/zero of=/root/fc_move.img bs=1M count=0 seek=256
mkfs.ext4 -O fast_commit -F /root/fc_move.img
mkdir -p /mnt/fc_move && mount -t ext4 -o loop \
/root/fc_move.img /mnt/fc_move
2. Created two files, ran EXT4_IOC_MOVE_EXT via e4defrag, and checked
the ineligible reason statistics:
fallocate -l 64M /mnt/fc_move/file1
cp /mnt/fc_move/file1 /mnt/fc_move/file2
e4defrag /mnt/fc_move/file1
cat /proc/fs/ext4/loop0/fc_info
shows "Move extents": > 0 and fc stats ineligible > 0.
Li Chen [Thu, 11 Dec 2025 11:51:39 +0000 (19:51 +0800)]
ext4: mark fs-verity enable fast-commit ineligible
Fast commits only log operations that have dedicated replay support.
Enabling fs-verity builds a Merkle tree and updates inode and orphan
state in ways that are not described by the fast commit replay tags.
In practice these operations are rare and usually followed by further
updates, but mixing them into a fast commit makes the overall
semantics harder to reason about and risks replay gaps if new call
sites appear.
Teach ext4 to mark the filesystem fast-commit ineligible when
ext4_end_enable_verity() starts its journal transaction.
This forces that transaction to fall back to a full commit, ensuring
that the fs-verity enable changes are captured by the normal journal
rather than partially encoded in fast commit TLVs.
This change should not affect common workloads but makes fs-verity
enable safer and easier to reason about under fast commit.
Testing:
1. prepare:
dd if=/dev/zero of=/root/fc_verity.img bs=1M count=0 seek=128
mkfs.ext4 -O fast_commit,verity -F /root/fc_verity.img
mkdir -p /mnt/fc_verity && mount -t ext4 -o loop /root/fc_verity.img /mnt/fc_verity
2. Enabled fs-verity on a file and verified reason accounting:
echo "data" > /mnt/fc_verity/verityfile
/root/enable_verity /mnt/fc_verity/verityfile
sync
tail -n 1 /proc/fs/ext4/loop0/fc_info
"fs-verity enable": 1
3. Enabled fs-verity on a second file, fsynced it, and checked that the
ineligible commit counter is updated too:
echo "data2" > /mnt/fc_verity/verityfile2
/root/enable_verity /mnt/fc_verity/verityfile2
/root/fsync_file /mnt/fc_verity/verityfile2
sync
/proc/fs/ext4/loop0/fc_info shows "fs-verity enable" incremented and
fc stats ineligible increased accordingly.
Li Chen [Thu, 11 Dec 2025 11:51:38 +0000 (19:51 +0800)]
ext4: mark inode format migration fast-commit ineligible
Fast commits only log operations that have dedicated replay support.
Inode format migration (indirect<->extent layout changes via
EXT4_IOC_MIGRATE or toggling EXT4_EXTENTS_FL) rewrites the block mapping
representation without going through the fast commit tracking paths.
In practice these migrations are rare and usually followed by further
updates, but mixing them into a fast commit makes the overall semantics
harder to reason about and risks replay gaps if new call sites appear.
Teach ext4 to mark the filesystem fast-commit ineligible when
ext4_ext_migrate() or ext4_ind_migrate() start their journal transactions.
This forces those transactions to fall back to a full commit, ensuring
that the entire inode layout change is captured by the normal journal
rather than partially encoded in fast commit TLVs. This change should
not affect common workloads but makes format migrations safer and easier
to reason about under fast commit.
Testing:
1. prepare:
dd if=/dev/zero of=/root/fc.img bs=1M count=0 seek=128
mkfs.ext4 -O fast_commit -F /root/fc.img
mkdir -p /mnt/fc && mount -t ext4 -o loop /root/fc.img /mnt/fc
2. Created a test file and toggled the extents flag to exercise both
ext4_ind_migrate() and ext4_ext_migrate():
touch /mnt/fc/migtest
chattr -e /mnt/fc/migtest
chattr +e /mnt/fc/migtest
3. Verified fast-commit ineligible statistics:
tail -n 1 /proc/fs/ext4/loop0/fc_info
"Inode format migration": 2
Baolin Liu [Thu, 11 Dec 2025 03:02:56 +0000 (11:02 +0800)]
ext4: add sysfs attribute err_report_sec to control s_err_report timer
Add a new sysfs attribute "err_report_sec" to control the s_err_report
timer in ext4_sb_info. Writing '0' disables the timer, while writing
a non-zero value enables the timer and sets the timeout in seconds.
Baokun Li [Tue, 9 Dec 2025 13:31:16 +0000 (21:31 +0800)]
ext4: move ext4_percpu_param_init() before ext4_mb_init()
When running `kvm-xfstests -c ext4/1k -C 1 generic/383` with the
`DOUBLE_CHECK` macro defined, the following panic is triggered:
==================================================================
EXT4-fs error (device vdc): ext4_validate_block_bitmap:423:
comm mount: bg 0: bad block bitmap checksum
BUG: unable to handle page fault for address: ff110000fa2cc000
PGD 3e01067 P4D 3e02067 PUD 0
Oops: Oops: 0000 [#1] SMP NOPTI
CPU: 0 UID: 0 PID: 2386 Comm: mount Tainted: G W 6.18.0-gba65a4e7120a-dirty #1152 PREEMPT(none)
RIP: 0010:percpu_counter_add_batch+0x13/0xa0
Call Trace:
<TASK>
ext4_mark_group_bitmap_corrupted+0xcb/0xe0
ext4_validate_block_bitmap+0x2a1/0x2f0
ext4_read_block_bitmap+0x33/0x50
mb_group_bb_bitmap_alloc+0x33/0x80
ext4_mb_add_groupinfo+0x190/0x250
ext4_mb_init_backend+0x87/0x290
ext4_mb_init+0x456/0x640
__ext4_fill_super+0x1072/0x1680
ext4_fill_super+0xd3/0x280
get_tree_bdev_flags+0x132/0x1d0
vfs_get_tree+0x29/0xd0
vfs_cmd_create+0x59/0xe0
__do_sys_fsconfig+0x4f6/0x6b0
do_syscall_64+0x50/0x1f0
entry_SYSCALL_64_after_hwframe+0x76/0x7e
==================================================================
This issue can be reproduced using the following commands:
mkfs.ext4 -F -q -b 1024 /dev/sda 5G
tune2fs -O quota,project /dev/sda
mount /dev/sda /tmp/test
With DOUBLE_CHECK defined, mb_group_bb_bitmap_alloc() reads
and validates the block bitmap. When the validation fails,
ext4_mark_group_bitmap_corrupted() attempts to update
sbi->s_freeclusters_counter. However, this percpu_counter has not been
initialized yet at this point, which leads to the panic described above.
Fix this by moving the execution of ext4_percpu_param_init() to occur
before ext4_mb_init(), ensuring the per-CPU counters are initialized
before they are used.
Zhang Yi [Sat, 29 Nov 2025 10:32:45 +0000 (18:32 +0800)]
ext4: replace ext4_es_insert_extent() when caching on-disk extents
In ext4, the remaining places for inserting extents into the extent
status tree within ext4_ext_determine_insert_hole() and
ext4_map_query_blocks() directly cache on-disk extents. We can use
ext4_es_cache_extent() instead of ext4_es_insert_extent() in these
cases. This will help reduce unnecessary increases in extent sequence
numbers and cache invalidations after supporting IOMAP in the future.
Suggested-by: Jan Kara <jack@suse.cz> Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Baokun Li <libaokun1@huawei.com>
Message-ID: <20251129103247.686136-14-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:44 +0000 (18:32 +0800)]
ext4: adjust the debug info in ext4_es_cache_extent()
Print a trace point after successfully inserting an extent in the
ext4_es_cache_extent() function. Additionally, similar to other extent
cache operation functions, call ext4_print_pending_tree() to display the
extent debug information of the inode when in ES_DEBUG mode.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Baokun Li <libaokun1@huawei.com>
Message-ID: <20251129103247.686136-13-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:43 +0000 (18:32 +0800)]
ext4: make ext4_es_cache_extent() support overwrite existing extents
Currently, ext4_es_cache_extent() is used to load extents into the
extent status tree when reading on-disk extent blocks. But it inserts
information into the extent status tree if and only if there isn't
information about the specified range already. So it only used for the
initial loading and does not support overwrit extents.
However, there are many other places in ext4 where on-disk extents are
inserted into the extent status tree, such as in ext4_map_query_blocks().
Currently, they call ext4_es_insert_extent() to perform the insertion,
but they don't modify the extents, so ext4_es_cache_extent() would be a
more appropriate choice. However, when ext4_map_query_blocks() inserts
an extent, it may overwrite a short existing extent of the same type.
Therefore, to prepare for the replacements, we need to extend
ext4_es_cache_extent() to allow it to overwrite existing extents with
the same status. So it checks the found extents before removing and
inserting. (There is one exception, a hole in the on-disk extent but a
delayed extent in the extent status tree is allowed.)
In addition, since cached extents can be more lenient than the extents
they modify and do not involve modifying reserved blocks, it is not
necessary to ensure that the insertion operation succeeds as strictly as
in the ext4_es_insert_extent() function.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Baokun Li <libaokun1@huawei.com>
Message-ID: <20251129103247.686136-12-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:42 +0000 (18:32 +0800)]
ext4: make __es_remove_extent() check extent status
Currently, __es_remove_extent() unconditionally removes extent status
entries within the specified range. In order to prepare for extending
the ext4_es_cache_extent() function to cache on-disk extents, which may
overwrite some existing short-length extents with the same status, allow
__es_remove_extent() to check the specified extent type before removing
it, and return error and pass out the conflicting extent if the status
does not match.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Baokun Li <libaokun1@huawei.com>
Message-ID: <20251129103247.686136-11-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:41 +0000 (18:32 +0800)]
ext4: cleanup useless out label in __es_remove_extent()
The out label in __es_remove_extent() is just return err value, we can
return it directly if something bad happens. Therefore, remove the
useless out label and rename out_get_reserved to out.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Baokun Li <libaokun1@huawei.com>
Message-ID: <20251129103247.686136-10-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:40 +0000 (18:32 +0800)]
ext4: cleanup zeroout in ext4_split_extent_at()
zero_ex is a temporary variable used only for writing zeros and
inserting extent status entry, it will not be directly inserted into the
tree. Therefore, it can be assigned values from the target extent in
various scenarios, eliminating the need to explicitly assign values to
each variable individually.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Baokun Li <libaokun1@huawei.com>
Message-ID: <20251129103247.686136-9-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:39 +0000 (18:32 +0800)]
ext4: drop extent cache when splitting extent fails
When the split extent fails, we might leave some extents still being
processed and return an error directly, which will result in stale
extent entries remaining in the extent status tree. So drop all of the
remaining potentially stale extents if the splitting fails.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Baokun Li <libaokun1@huawei.com> Cc: stable@kernel.org Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Message-ID: <20251129103247.686136-8-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:38 +0000 (18:32 +0800)]
ext4: drop extent cache after doing PARTIAL_VALID1 zeroout
When splitting an unwritten extent in the middle and converting it to
initialized in ext4_split_extent() with the EXT4_EXT_MAY_ZEROOUT and
EXT4_EXT_DATA_VALID2 flags set, it could leave a stale unwritten extent.
Assume we have an unwritten file and buffered write in the middle of it
without dioread_nolock enabled, it will allocate blocks as written
extent.
0 A B N
[UUUUUUUUUUUU] on-disk extent U: unwritten extent
[UUUUUUUUUUUU] extent status tree
[--DDDDDDDD--] D: valid data
|<- ->| ----> this range needs to be initialized
ext4_split_extent() first try to split this extent at B with
EXT4_EXT_DATA_PARTIAL_VALID1 and EXT4_EXT_MAY_ZEROOUT flag set, but
ext4_split_extent_at() failed to split this extent due to temporary lack
of space. It zeroout B to N and leave the entire extent as unwritten.
0 A B N
[UUUUUUUUUUUU] on-disk extent
[UUUUUUUUUUUU] extent status tree
[--DDDDDDDDZZ] Z: zeroed data
ext4_split_extent() then try to split this extent at A with
EXT4_EXT_DATA_VALID2 flag set. This time, it split successfully and
leave an written extent from A to N.
0 A B N
[UUWWWWWWWWWW] on-disk extent W: written extent
[UUUUUUUUUUUU] extent status tree
[--DDDDDDDDZZ]
Finally ext4_map_create_blocks() only insert extent A to B to the extent
status tree, and leave an stale unwritten extent in the status tree.
0 A B N
[UUWWWWWWWWWW] on-disk extent W: written extent
[UUWWWWWWWWUU] extent status tree
[--DDDDDDDDZZ]
Fix this issue by always cached extent status entry after zeroing out
the second part.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Baokun Li <libaokun1@huawei.com> Cc: stable@kernel.org Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Message-ID: <20251129103247.686136-7-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:37 +0000 (18:32 +0800)]
ext4: don't cache extent during splitting extent
Caching extents during the splitting process is risky, as it may result
in stale extents remaining in the status tree. Moreover, in most cases,
the corresponding extent block entries are likely already cached before
the split happens, making caching here not particularly useful.
Assume we have an unwritten extent, and then DIO writes the first half.
[UUUUUUUUUUUUUUUU] on-disk extent U: unwritten extent
[UUUUUUUUUUUUUUUU] extent status tree
|<- ->| ----> dio write this range
First, when ext4_split_extent_at() splits this extent, it truncates the
existing extent and then inserts a new one. During this process, this
extent status entry may be shrunk, and calls to ext4_find_extent() and
ext4_cache_extents() may occur, which could potentially insert the
truncated range as a hole into the extent status tree. After the split
is completed, this hole is not replaced with the correct status.
[UUUUUUU|UUUUUUUU] on-disk extent U: unwritten extent
[UUUUUUU|HHHHHHHH] extent status tree H: hole
Then, the outer calling functions will not correct this remaining hole
extent either. Finally, if we perform a delayed buffer write on this
latter part, it will re-insert the delayed extent and cause an error in
space accounting.
In adition, if the unwritten extent cache is not shrunk during the
splitting, ext4_cache_extents() also conflicts with existing extents
when caching extents. In the future, we will add checks when caching
extents, which will trigger a warning. Therefore, Do not cache extents
that are being split.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Baokun Li <libaokun1@huawei.com> Cc: stable@kernel.org
Message-ID: <20251129103247.686136-6-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:36 +0000 (18:32 +0800)]
ext4: correct the mapping status if the extent has been zeroed
Before submitting I/O and allocating blocks with the
EXT4_GET_BLOCKS_PRE_IO flag set, ext4_split_convert_extents() may
convert the target extent range to initialized due to ENOSPC, ENOMEM, or
EQUOTA errors. However, it still marks the mapping as incorrectly
unwritten. Although this may not seem to cause any practical problems,
it will result in an unnecessary extent conversion operation after I/O
completion. Therefore, it's better to correct the returned mapping
status.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Baokun Li <libaokun1@huawei.com>
Message-ID: <20251129103247.686136-5-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:35 +0000 (18:32 +0800)]
ext4: don't set EXT4_GET_BLOCKS_CONVERT when splitting before submitting I/O
When allocating blocks during within-EOF DIO and writeback with
dioread_nolock enabled, EXT4_GET_BLOCKS_PRE_IO was set to split an
existing large unwritten extent. However, EXT4_GET_BLOCKS_CONVERT was
set when calling ext4_split_convert_extents(), which may potentially
result in stale data issues.
Assume we have an unwritten extent, and then DIO writes the second half.
[UUUUUUUUUUUUUUUU] on-disk extent U: unwritten extent
[UUUUUUUUUUUUUUUU] extent status tree
|<- ->| ----> dio write this range
First, ext4_iomap_alloc() call ext4_map_blocks() with
EXT4_GET_BLOCKS_PRE_IO, EXT4_GET_BLOCKS_UNWRIT_EXT and
EXT4_GET_BLOCKS_CREATE flags set. ext4_map_blocks() find this extent and
call ext4_split_convert_extents() with EXT4_GET_BLOCKS_CONVERT and the
above flags set.
Then, ext4_split_convert_extents() calls ext4_split_extent() with
EXT4_EXT_MAY_ZEROOUT, EXT4_EXT_MARK_UNWRIT2 and EXT4_EXT_DATA_VALID2
flags set, and it calls ext4_split_extent_at() to split the second half
with EXT4_EXT_DATA_VALID2, EXT4_EXT_MARK_UNWRIT1, EXT4_EXT_MAY_ZEROOUT
and EXT4_EXT_MARK_UNWRIT2 flags set. However, ext4_split_extent_at()
failed to insert extent since a temporary lack -ENOSPC. It zeroes out
the first half but convert the entire on-disk extent to written since
the EXT4_EXT_DATA_VALID2 flag set, but left the second half as unwritten
in the extent status tree.
[0000000000SSSSSS] data S: stale data, 0: zeroed
[WWWWWWWWWWWWWWWW] on-disk extent W: written extent
[WWWWWWWWWWUUUUUU] extent status tree
Finally, if the DIO failed to write data to the disk, the stale data in
the second half will be exposed once the cached extent entry is gone.
Fix this issue by not passing EXT4_GET_BLOCKS_CONVERT when splitting
an unwritten extent before submitting I/O, and make
ext4_split_convert_extents() to zero out the entire extent range
to zero for this case, and also mark the extent in the extent status
tree for consistency.
Fixes: b8a8684502a0 ("ext4: Introduce FALLOC_FL_ZERO_RANGE flag for fallocate") Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Baokun Li <libaokun1@huawei.com> Cc: stable@kernel.org
Message-ID: <20251129103247.686136-4-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:34 +0000 (18:32 +0800)]
ext4: don't zero the entire extent if EXT4_EXT_DATA_PARTIAL_VALID1
When allocating initialized blocks from a large unwritten extent, or
when splitting an unwritten extent during end I/O and converting it to
initialized, there is currently a potential issue of stale data if the
extent needs to be split in the middle.
0 A B N
[UUUUUUUUUUUU] U: unwritten extent
[--DDDDDDDD--] D: valid data
|<- ->| ----> this range needs to be initialized
ext4_split_extent() first try to split this extent at B with
EXT4_EXT_DATA_ENTIRE_VALID1 and EXT4_EXT_MAY_ZEROOUT flag set, but
ext4_split_extent_at() failed to split this extent due to temporary lack
of space. It zeroout B to N and mark the entire extent from 0 to N
as written.
0 A B N
[WWWWWWWWWWWW] W: written extent
[SSDDDDDDDDZZ] Z: zeroed, S: stale data
ext4_split_extent() then try to split this extent at A with
EXT4_EXT_DATA_VALID2 flag set. This time, it split successfully and left
a stale written extent from 0 to A.
0 A B N
[WW|WWWWWWWWWW]
[SS|DDDDDDDDZZ]
Fix this by pass EXT4_EXT_DATA_PARTIAL_VALID1 to ext4_split_extent_at()
when splitting at B, don't convert the entire extent to written and left
it as unwritten after zeroing out B to N. The remaining work is just
like the standard two-part split. ext4_split_extent() will pass the
EXT4_EXT_DATA_VALID2 flag when it calls ext4_split_extent_at() for the
second time, allowing it to properly handle the split. If the split is
successful, it will keep extent from 0 to A as unwritten.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Baokun Li <libaokun1@huawei.com> Cc: stable@kernel.org
Message-ID: <20251129103247.686136-3-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sat, 29 Nov 2025 10:32:33 +0000 (18:32 +0800)]
ext4: subdivide EXT4_EXT_DATA_VALID1
When splitting an extent, if the EXT4_GET_BLOCKS_CONVERT flag is set and
it is necessary to split the target extent in the middle,
ext4_split_extent() first handles splitting the latter half of the
extent and passes the EXT4_EXT_DATA_VALID1 flag. This flag implies that
all blocks before the split point contain valid data; however, this
assumption is incorrect.
Therefore, subdivid EXT4_EXT_DATA_VALID1 into
EXT4_EXT_DATA_ENTIRE_VALID1 and EXT4_EXT_DATA_PARTIAL_VALID1, which
indicate that the first half of the extent is either entirely valid or
only partially valid, respectively. These two flags cannot be set
simultaneously.
This patch does not use EXT4_EXT_DATA_PARTIAL_VALID1, it only replaces
EXT4_EXT_DATA_VALID1 with EXT4_EXT_DATA_ENTIRE_VALID1 at the location
where it is set, no logical changes.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com> Reviewed-by: Baokun Li <libaokun1@huawei.com> Cc: stable@kernel.org
Message-ID: <20251129103247.686136-2-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Julian Sun [Mon, 8 Dec 2025 12:37:13 +0000 (20:37 +0800)]
ext4: add missing down_write_data_sem in mext_move_extent().
Commit 962e8a01eab9 ("ext4: introduce mext_move_extent()") attempts to
call ext4_swap_extents() on the failure path to recover the swapped
extents, but fails to acquire locks for the two inode->i_data_sem,
triggering the BUG_ON statement in ext4_swap_extents().
This issue can be fixed by calling ext4_double_down_write_data_sem()
before ext4_swap_extents().
Signed-off-by: Julian Sun <sunjunchao@bytedance.com> Reported-by: syzbot+4ea6bd8737669b423aae@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/69368649.a70a0220.38f243.0093.GAE@google.com/ Fixes: 962e8a01eab9 ("ext4: introduce mext_move_extent()") Reviewed-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Zhang Yi <yi.zhang@huawei.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu> Link: https://patch.msgid.link/20251208123713.1971068-1-sunjunchao@bytedance.com Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Arnd Bergmann [Thu, 4 Dec 2025 10:19:10 +0000 (11:19 +0100)]
ext4: fix ext4_tune_sb_params padding
The padding at the end of struct ext4_tune_sb_params is architecture
specific and in particular is different between x86-32 and x86-64,
since the __u64 member only enforces struct alignment on the latter.
This shows up as a new warning when test-building the headers with
-Wpadded:
include/linux/ext4.h:144:1: error: padding struct size to alignment boundary with 4 bytes [-Werror=padded]
All members inside the structure are naturally aligned, so the only
difference here is the amount of padding at the end. Make the padding
explicit, to have a consistent sizeof(struct ext4_tune_sb_params) of
232 on all architectures and avoid adding compat ioctl handling for
EXT4_IOC_GET_TUNE_SB_PARAM/EXT4_IOC_SET_TUNE_SB_PARAM.
This is an ABI break on x86-32 but hopefully this can go into 6.18.y early
enough as a fixup so no actual users will be affected. Alternatively, the
kernel could handle the ioctl commands for both sizes (232 and 228 bytes)
on all architectures.
Fixes: 04a91570ac67 ("ext4: implemet new ioctls to set and get superblock parameters") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://patch.msgid.link/20251204101914.1037148-1-arnd@kernel.org Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org
Linus Torvalds [Sun, 4 Jan 2026 15:21:18 +0000 (07:21 -0800)]
Merge tag 'core_urgent_for_v6.19_rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull core entry fix from Borislav Petkov:
- Make sure clang inlines trivial local_irq_* helpers
* tag 'core_urgent_for_v6.19_rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
entry: Always inline local_irq_{enable,disable}_exit_to_user()
Linus Torvalds [Fri, 2 Jan 2026 22:24:09 +0000 (14:24 -0800)]
Merge tag 'perf-tools-fixes-for-v6.19-2026-01-02' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools
Pull perf tool fixes and from Namhyung Kim:
- skip building BPF skeletons if libopenssl is missing
- a couple of test updates
- handle error cases of filename__read_build_id()
- support NVIDIA Olympus for ARM SPE profiling
- update tool headers to sync with the kernel
* tag 'perf-tools-fixes-for-v6.19-2026-01-02' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools:
tools build: Fix the common set of features test wrt libopenssl
tools headers: Sync syscall table with kernel sources
tools headers: Sync linux/socket.h with kernel sources
tools headers: Sync linux/gfp_types.h with kernel sources
tools headers: Sync arm64 headers with kernel sources
tools headers: Sync x86 headers with kernel sources
tools headers: Sync UAPI sound/asound.h with kernel sources
tools headers: Sync UAPI linux/mount.h with kernel sources
tools headers: Sync UAPI linux/fs.h with kernel sources
tools headers: Sync UAPI linux/fcntl.h with kernel sources
tools headers: Sync UAPI KVM headers with kernel sources
tools headers: Sync UAPI drm/drm.h with kernel sources
perf arm-spe: Add NVIDIA Olympus to neoverse list
tools headers arm64: Add NVIDIA Olympus part
perf tests top: Make the test exclusive
perf tests kvm: Avoid leaving perf.data.guest file around
perf symbol: Fix ENOENT case for filename__read_build_id
perf tools: Disable BPF skeleton if no libopenssl found
tools/build: Add a feature test for libopenssl
Linus Torvalds [Fri, 2 Jan 2026 20:28:24 +0000 (12:28 -0800)]
Merge tag 'libcrypto-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux
Pull crypto library fix from Eric Biggers:
"Fix the kunit_run_irq_test() function (which I recently added for the
CRC and crypto tests) to be less timing-dependent.
This fixes flakiness in the polyval kunit test suite"
* tag 'libcrypto-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux:
kunit: Enforce task execution in {soft,hard}irq contexts
Linus Torvalds [Fri, 2 Jan 2026 20:25:47 +0000 (12:25 -0800)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma fixes from Jason Gunthorpe:
- Fix several syzkaller found bugs:
- Poor parsing of the RDMA_NL_LS_OP_IP_RESOLVE netlink
- GID entry refcount leaking when CM destruction races with
multicast establishment
- Missing refcount put in ib_del_sub_device_and_put()
- Fixup recently introduced uABI padding for 32 bit consistency
- Avoid user triggered math overflow in MANA and AFA
- Reading invalid netdev data during an event
- kdoc fixes
- Fix never-working gid copying in ib_get_gids_from_rdma_hdr
- Typo in bnxt when validating the BAR
- bnxt mis-parsed IB_SEND_IP_CSUM so it didn't work always
- bnxt out of bounds access in bnxt related to the counters on new
devices
- Allocate the bnxt PDE table with the right sizing
- Use dma_free_coherent() correctly in bnxt
- Allow rxe to be unloadable when CONFIG_PROVE_LOCKING by adjusting the
tracking of the global sockets it uses
- Missing unlocking on error path in rxe
- Compute the right number of pages in a MR in rtrs
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
RDMA/bnxt_re: fix dma_free_coherent() pointer
RDMA/rtrs: Fix clt_path::max_pages_per_mr calculation
IB/rxe: Fix missing umem_odp->umem_mutex unlock on error path
RDMA/bnxt_re: Fix to use correct page size for PDE table
RDMA/bnxt_re: Fix OOB write in bnxt_re_copy_err_stats()
RDMA/bnxt_re: Fix IB_SEND_IP_CSUM handling in post_send
RDMA/core: always drop device refcount in ib_del_sub_device_and_put()
RDMA/rxe: let rxe_reclassify_recv_socket() call sk_owner_put()
RDMA/bnxt_re: Fix incorrect BAR check in bnxt_qplib_map_creq_db()
RDMA/core: Fix logic error in ib_get_gids_from_rdma_hdr()
RDMA/efa: Remove possible negative shift
RTRS/rtrs: clean up rtrs headers kernel-doc
RDMA/irdma: avoid invalid read in irdma_net_event
RDMA/mana_ib: check cqe length for kernel CQs
RDMA/irdma: Fix irdma_alloc_ucontext_resp padding
RDMA/ucma: Fix rdma_ucm_query_ib_service_resp struct padding
RDMA/cm: Fix leaking the multicast GID table reference
RDMA/core: Check for the presence of LS_NLA_TYPE_DGID correctly
Linus Torvalds [Fri, 2 Jan 2026 20:21:34 +0000 (12:21 -0800)]
Merge tag 'linux_kselftest-fixes-6.19-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest
Pull kselftest fixes from Shuah Khan:
- Fix for build failures in tests that use an empty FIXTURE() seen in
Android's build environment, which uses -D_FORTIFY_SOURCE=3, a build
failure occurs in tests that use an empty FIXTURE()
- Fix func_traceonoff_triggers.tc sometimes failures on Kunpeng-920
board resulting from including transient trace file name in checksum
compare
- Fix to remove available_events requirement from toplevel-enable for
instance as it isn't a valid requirement for this test
* tag 'linux_kselftest-fixes-6.19-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
kselftest/harness: Use helper to avoid zero-size memset warning
selftests/ftrace: Test toplevel-enable for instance
selftests/ftrace: traceonoff_triggers: strip off names
Linus Torvalds [Fri, 2 Jan 2026 20:15:59 +0000 (12:15 -0800)]
Merge tag 'block-6.19-20260102' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux
Pull block fixes from Jens Axboe:
- Scan partition tables asynchronously for ublk, similarly to how nvme
does it. This avoids potential deadlocks, which is why nvme does it
that way too. Includes a set of selftests as well.
- MD pull request via Yu:
- Fix null-pointer dereference in raid5 sysfs group_thread_cnt
store (Tuo Li)
- Fix possible mempool corruption during raid1 raid_disks update
via sysfs (FengWei Shih)
- Fix logical_block_size configuration being overwritten during
super_1_validate() (Li Nan)
- Fix forward incompatibility with configurable logical block size:
arrays assembled on new kernels could not be assembled on older
kernels (v6.18 and before) due to non-zero reserved pad rejection
(Li Nan)
- Fix static checker warning about iterator not incremented (Li Nan)
- Skip CPU offlining notifications on unmapped hardware queues
- bfq-iosched block stats fix
- Fix outdated comment in bfq-iosched
* tag 'block-6.19-20260102' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
block, bfq: update outdated comment
blk-mq: skip CPU offline notify on unmapped hctx
selftests/ublk: fix Makefile to rebuild on header changes
selftests/ublk: add test for async partition scan
ublk: scan partition in async way
block,bfq: fix aux stat accumulation destination
md: Fix forward incompatibility from configurable logical block size
md: Fix logical_block_size configuration being overwritten
md: suspend array while updating raid_disks via sysfs
md/raid5: fix possible null-pointer dereferences in raid5_store_group_thread_cnt()
md: Fix static checker warning in analyze_sbs
Linus Torvalds [Fri, 2 Jan 2026 20:07:55 +0000 (12:07 -0800)]
Merge tag 'io_uring-6.19-20260102' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux
Pull io_uring fixes from Jens Axboe:
- Removed dead argument length for io_uring_validate_mmap_request()
- Use GFP_NOWAIT for overflow CQEs on legacy ring setups rather than
GFP_ATOMIC, which makes it play nicer with memcg limits
- Fix a potential circular locking issue with tctx node removal and
exec based cancelations
* tag 'io_uring-6.19-20260102' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
io_uring/memmap: drop unused sz param in io_uring_validate_mmap_request()
io_uring/tctx: add separate lock for list of tctx's in ctx
io_uring: use GFP_NOWAIT for overflow CQEs on legacy rings
Linus Torvalds [Fri, 2 Jan 2026 20:04:51 +0000 (12:04 -0800)]
Merge tag 'x86-urgent-2026-01-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fix from Ingo Molnar:
"Fix the AMD microcode Entrysign signature checking code to include
more models"
* tag 'x86-urgent-2026-01-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/microcode/AMD: Fix Entrysign revision check for Zen5/Strix Halo
Linus Torvalds [Fri, 2 Jan 2026 19:33:33 +0000 (11:33 -0800)]
Merge tag 'loongarch-fixes-6.19-1' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson
Pull LoongArch fixes from Huacai Chen:
"Complete CPUCFG registers definition, set correct protection_map[] for
VM_NONE/VM_SHARED, fix some bugs in the orc stack unwinder, ftrace and
BPF JIT"
* tag 'loongarch-fixes-6.19-1' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson:
samples/ftrace: Adjust LoongArch register restore order in direct calls
LoongArch: BPF: Enhance the bpf_arch_text_poke() function
LoongArch: BPF: Enable trampoline-based tracing for module functions
LoongArch: BPF: Adjust the jump offset of tail calls
LoongArch: BPF: Save return address register ra to t0 before trampoline
LoongArch: BPF: Zero-extend bpf_tail_call() index
LoongArch: BPF: Sign extend kfunc call arguments
LoongArch: Refactor register restoration in ftrace_common_return
LoongArch: Enable exception fixup for specific ADE subcode
LoongArch: Remove unnecessary checks for ORC unwinder
LoongArch: Remove is_entry_func() and kernel_entry_end
LoongArch: Use UNWIND_HINT_END_OF_STACK for entry points
LoongArch: Set correct protection_map[] for VM_NONE/VM_SHARED
LoongArch: Complete CPUCFG registers definition
Linus Torvalds [Fri, 2 Jan 2026 17:53:45 +0000 (09:53 -0800)]
Merge tag 'drm-fixes-2026-01-02' of https://gitlab.freedesktop.org/drm/kernel
Pull drm fixes from Dave Airlie:
"Happy New Year, jetlagged fixes from me, still pretty quiet, xe is
most of this, with i915/nouveau/imagination fixes and some shmem
cleanups.
shmem:
- docs and MODULE_LICENSE fix
xe:
- Ensure svm device memory is idle before migration completes
- Fix a SVM debug printout
- Use READ_ONCE() / WRITE_ONCE() for g2h_fence
i915:
- Fix eb_lookup_vmas() failure path
nouveau:
- fix prepare_fb warnings
imagination:
- prevent export of protected objects"
* tag 'drm-fixes-2026-01-02' of https://gitlab.freedesktop.org/drm/kernel:
drm/i915/gem: Zero-initialize the eb.vma array in i915_gem_do_execbuffer
drm/xe/guc: READ/WRITE_ONCE g2h_fence->done
drm/pagemap, drm/xe: Ensure that the devmem allocation is idle before use
drm/xe/svm: Fix a debug printout
drm/gem-shmem: Fix the MODULE_LICENSE() string
drm/gem-shmem: Fix typos in documentation
drm/nouveau/dispnv50: Don't call drm_atomic_get_crtc_state() in prepare_fb
drm/imagination: Disallow exporting of PM/FW protected objects
Linus Torvalds [Fri, 2 Jan 2026 17:24:43 +0000 (09:24 -0800)]
Merge tag 'v6.19-rc3-smb3-server-fixes' of git://git.samba.org/ksmbd
Pull smb server fixes from Steve French:
- Fix memory leak
- Fix two refcount leaks
- Fix error path in create_smb2_pipe
* tag 'v6.19-rc3-smb3-server-fixes' of git://git.samba.org/ksmbd:
smb/server: fix refcount leak in smb2_open()
smb/server: fix refcount leak in parse_durable_handle_context()
smb/server: call ksmbd_session_rpc_close() on error path in create_smb2_pipe()
ksmbd: Fix memory leak in get_file_all_info()
Linus Torvalds [Fri, 2 Jan 2026 17:14:13 +0000 (09:14 -0800)]
Merge tag 'v6.19-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client fixes from Steve French:
- Fix array out of bounds error in copy_file_range
- Add tracepoint to help debug ioctl failures
* tag 'v6.19-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
smb: client: fix UBSAN array-index-out-of-bounds in smb2_copychunk_range
smb3 client: add missing tracepoint for unsupported ioctls
Julia Lawall [Wed, 31 Dec 2025 17:22:07 +0000 (18:22 +0100)]
block, bfq: update outdated comment
The function bfq_bfqq_may_idle() was renamed as bfq_better_to_idle()
in commit 277a4a9b56cd ("block, bfq: give a better name to
bfq_bfqq_may_idle"). Update the comment accordingly.
Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 31 Dec 2025 15:12:46 +0000 (08:12 -0700)]
io_uring/tctx: add separate lock for list of tctx's in ctx
ctx->tcxt_list holds the tasks using this ring, and it's currently
protected by the normal ctx->uring_lock. However, this can cause a
circular locking issue, as reported by syzbot, where cancelations off
exec end up needing to remove an entry from this list:
======================================================
WARNING: possible circular locking dependency detected
syzkaller #0 Tainted: G L
------------------------------------------------------
syz.0.9999/12287 is trying to acquire lock: ffff88805851c0a8 (&ctx->uring_lock){+.+.}-{4:4}, at: io_uring_del_tctx_node+0xf0/0x2c0 io_uring/tctx.c:179
but task is already holding lock: ffff88802db5a2e0 (&sig->cred_guard_mutex){+.+.}-{4:4}, at: prepare_bprm_creds fs/exec.c:1360 [inline] ffff88802db5a2e0 (&sig->cred_guard_mutex){+.+.}-{4:4}, at: bprm_execve+0xb9/0x1400 fs/exec.c:1733
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
Add a separate lock just for the tctx_list, tctx_lock. This can nest
under ->uring_lock, where necessary, and be used separately for list
manipulation. For the cancelation off exec side, this removes the
need to grab ->uring_lock, hence fixing the circular locking
dependency.
Dave Airlie [Thu, 1 Jan 2026 06:51:30 +0000 (16:51 +1000)]
Merge tag 'drm-misc-fixes-2025-12-29' of https://gitlab.freedesktop.org/drm/misc/kernel into drm-fixes
drm-misc-fixes for v6.19-rc4:
- Documentation fixes and MODULE_LICENSE fix for shmem helper.
- Fix warnings in nouveau prepare_fb().
- Prevent export of protected objects in imagination driver.
Shuah Khan [Wed, 31 Dec 2025 23:46:26 +0000 (16:46 -0700)]
wifi: mt76: Remove blank line after mt792x firmware version dmesg
An extra blank line gets printed after printing firmware version
because the build date is null terminated. Remove the "\n" from
dev_info() calls to print firmware version and build date to fix
the problem.
Wake Liu [Wed, 24 Dec 2025 08:41:20 +0000 (16:41 +0800)]
kselftest/harness: Use helper to avoid zero-size memset warning
When building kselftests with a toolchain that enables source
fortification (e.g., Android's build environment, which uses
-D_FORTIFY_SOURCE=3), a build failure occurs in tests that use an
empty FIXTURE().
The root cause is that an empty fixture struct results in
`sizeof(self_private)` evaluating to 0. The compiler's fortification
checks then detect the `memset()` call with a compile-time constant size
of 0, issuing a `-Wuser-defined-warnings` which is promoted to an error
by `-Werror`.
An initial attempt to guard the call with `if (sizeof(self_private) > 0)`
was insufficient. The compiler's static analysis is aggressive enough
to flag the `memset(..., 0)` pattern before evaluating the conditional,
thus still triggering the error.
To resolve this robustly, this change introduces a `static inline`
helper function, `__kselftest_memset_safe()`. This function wraps the
size check and the `memset()` call. By replacing the direct `memset()`
in the `__TEST_F_IMPL` macro with a call to this helper, we create an
abstraction boundary. This prevents the compiler's static analyzer from
"seeing" the problematic pattern at the macro expansion site, resolving
the build failure.
Build Context:
Compiler: Android (14488419, +pgo, +bolt, +lto, +mlgo, based on r584948) clang version 22.0.0 (https://android.googlesource.com/toolchain/llvm-project 2d65e4108033380e6fe8e08b1f1826cd2bfb0c99)
Relevant Options: -O2 -Wall -Werror -D_FORTIFY_SOURCE=3 -target i686-linux-android10000
Zheng Yejian [Tue, 9 May 2023 20:36:59 +0000 (04:36 +0800)]
selftests/ftrace: Test toplevel-enable for instance
'available_events' is actually not required by
'test.d/event/toplevel-enable.tc' and its Existence has been tested in
'test.d/00basic/basic4.tc'.
So the require of 'available_events' can be dropped and then we can add
'instance' flag to test 'test.d/event/toplevel-enable.tc' for instance.
Test result show as below:
# ./ftracetest test.d/event/toplevel-enable.tc
=== Ftrace unit tests ===
[1] event tracing - enable/disable with top level files [PASS]
[2] (instance) event tracing - enable/disable with top level files [PASS]
# of passed: 2
# of failed: 0
# of unresolved: 0
# of untested: 0
# of unsupported: 0
# of xfailed: 0
# of undefined(test bug): 0
Yipeng Zou [Fri, 18 Aug 2023 01:32:26 +0000 (09:32 +0800)]
selftests/ftrace: traceonoff_triggers: strip off names
The func_traceonoff_triggers.tc sometimes goes to fail
on my board, Kunpeng-920.
[root@localhost]# ./ftracetest ./test.d/ftrace/func_traceonoff_triggers.tc -l fail.log
=== Ftrace unit tests ===
[1] ftrace - test for function traceon/off triggers [FAIL]
[2] (instance) ftrace - test for function traceon/off triggers [UNSUPPORTED]
I look up the log, and it shows that the md5sum is different between csum1 and csum2.
Link: https://lore.kernel.org/r/20230818013226.2182299-1-zouyipeng@huawei.com Fixes: d87b29179aa0 ("selftests: ftrace: Use md5sum to take less time of checking logs") Suggested-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Yipeng Zou <zouyipeng@huawei.com> Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org> Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Linus Torvalds [Wed, 31 Dec 2025 18:38:48 +0000 (10:38 -0800)]
Merge tag 'vfio-v6.19-rc4' of https://github.com/awilliam/linux-vfio
Pull VFIO fixes from Alex Williamson:
- Restrict ROM access to dword to resolve a regression introduced with
qword access seen on some Intel NICs. Update VGA region access to the
same given lack of precedent for 64-bit users (Kevin Tian)
- Fix missing .get_region_info_caps callback in the xe-vfio-pci variant
driver due to integration through the DRM tree (Michal Wajdeczko)
- Add aligned 64-bit access macros to tools/include/linux/types.h,
allowing removal of uapi/linux/type.h includes from various vfio
selftest, resolving redefinition warnings for integration with KVM
selftests (David Matlack)
* tag 'vfio-v6.19-rc4' of https://github.com/awilliam/linux-vfio:
vfio/xe: Fix use-after-free in xe_vfio_pci_alloc_file()
vfio/pds: Fix memory leak in pds_vfio_dirty_enable()
vfio: selftests: Drop <uapi/linux/types.h> includes
tools include: Add definitions for __aligned_{l,b}e64
vfio/xe: Add default handler for .get_region_info_caps
vfio/pci: Disable qword access to the VGA region
vfio/pci: Disable qword access to the PCI ROM bar
Jens Axboe [Wed, 31 Dec 2025 13:55:07 +0000 (06:55 -0700)]
Merge tag 'md-6.19-20251231' of gitolite.kernel.org:pub/scm/linux/kernel/git/mdraid/linux into block-6.19
Pull MD fixes from Yu Kuai:
"- Fix null-pointer dereference in raid5 sysfs group_thread_cnt store
(Tuo Li)
- Fix possible mempool corruption during raid1 raid_disks update via
sysfs (FengWei Shih)
- Fix logical_block_size configuration being overwritten during
super_1_validate() (Li Nan)
- Fix forward incompatibility with configurable logical block size:
arrays assembled on new kernels could not be assembled on kernels
<=6.18 due to non-zero reserved pad rejection (Li Nan)
- Fix static checker warning about iterator not incremented (Li Nan)"
* tag 'md-6.19-20251231' of gitolite.kernel.org:pub/scm/linux/kernel/git/mdraid/linux:
md: Fix forward incompatibility from configurable logical block size
md: Fix logical_block_size configuration being overwritten
md: suspend array while updating raid_disks via sysfs
md/raid5: fix possible null-pointer dereferences in raid5_store_group_thread_cnt()
md: Fix static checker warning in analyze_sbs
drm/i915/gem: Zero-initialize the eb.vma array in i915_gem_do_execbuffer
Initialize the eb.vma array with values of 0 when the eb structure is
first set up. In particular, this sets the eb->vma[i].vma pointers to
NULL, simplifying cleanup and getting rid of the bug described below.
During the execution of eb_lookup_vmas(), the eb->vma array is
successively filled up with struct eb_vma objects. This process includes
calling eb_add_vma(), which might fail; however, even in the event of
failure, eb->vma[i].vma is set for the currently processed buffer.
If eb_add_vma() fails, eb_lookup_vmas() returns with an error, which
prompts a call to eb_release_vmas() to clean up the mess. Since
eb_lookup_vmas() might fail during processing any (possibly not first)
buffer, eb_release_vmas() checks whether a buffer's vma is NULL to know
at what point did the lookup function fail.
In eb_lookup_vmas(), eb->vma[i].vma is set to NULL if either the helper
function eb_lookup_vma() or eb_validate_vma() fails. eb->vma[i+1].vma is
set to NULL in case i915_gem_object_userptr_submit_init() fails; the
current one needs to be cleaned up by eb_release_vmas() at this point,
so the next one is set. If eb_add_vma() fails, neither the current nor
the next vma is set to NULL, which is a source of a NULL deref bug
described in the issue linked in the Closes tag.
When entering eb_lookup_vmas(), the vma pointers are set to the slab
poison value, instead of NULL. This doesn't matter for the actual
lookup, since it gets overwritten anyway, however the eb_release_vmas()
function only recognizes NULL as the stopping value, hence the pointers
are being set to NULL as they go in case of intermediate failure. This
patch changes the approach to filling them all with NULL at the start
instead, rather than handling that manually during failure.
Reported-by: Gangmin Kim <km.kim1503@gmail.com> Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/15062 Fixes: 544460c33821 ("drm/i915: Multi-BB execbuf") Cc: stable@vger.kernel.org # 5.16.x Signed-off-by: Krzysztof Niemiec <krzysztof.niemiec@intel.com> Reviewed-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com> Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com> Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> Signed-off-by: Andi Shyti <andi.shyti@kernel.org> Link: https://lore.kernel.org/r/20251216180900.54294-2-krzysztof.niemiec@intel.com
(cherry picked from commit 08889b706d4f0b8d2352b7ca29c2d8df4d0787cd) Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Chenghao Duan [Wed, 31 Dec 2025 07:19:25 +0000 (15:19 +0800)]
samples/ftrace: Adjust LoongArch register restore order in direct calls
Ensure that in the ftrace direct call logic, the CPU register state
(with ra = parent return address) is restored to the correct state after
the execution of the custom trampoline function and before returning to
the traced function. Additionally, guarantee the correctness of the jump
logic for jr t0 (traced function address).
Chenghao Duan [Wed, 31 Dec 2025 07:19:21 +0000 (15:19 +0800)]
LoongArch: BPF: Enable trampoline-based tracing for module functions
Remove the previous restrictions that blocked the tracing of kernel
module functions. Fix the issue that previously caused kernel lockups
when attempting to trace module functions.
Before entering the trampoline code, the return address register ra
shall store the address of the next assembly instruction after the
'bl trampoline' instruction, which is the traced function address, and
the register t0 shall store the parent function return address. Refine
the trampoline return logic to ensure that register data remains correct
when returning to both the traced function and the parent function.
Before this patch was applied, the module_attach test in selftests/bpf
encountered a deadlock issue. This was caused by an incorrect jump
address after the trampoline execution, which resulted in an infinite
loop within the module function.
Chenghao Duan [Wed, 31 Dec 2025 07:19:20 +0000 (15:19 +0800)]
LoongArch: BPF: Save return address register ra to t0 before trampoline
Modify the build_prologue() function to ensure the return address
register ra is saved to t0 before entering trampoline operations.
This change ensures the accurate return address handling when a BPF
program calls another BPF program, preventing errors in the BPF-to-BPF
call chain.
Hengqi Chen [Wed, 31 Dec 2025 07:19:20 +0000 (15:19 +0800)]
LoongArch: BPF: Zero-extend bpf_tail_call() index
The bpf_tail_call() index should be treated as a u32 value. Let's
zero-extend it to avoid calling wrong BPF progs. See similar fixes
for x86 [1]) and arm64 ([2]) for more details.
Hengqi Chen [Wed, 31 Dec 2025 07:19:20 +0000 (15:19 +0800)]
LoongArch: BPF: Sign extend kfunc call arguments
The kfunc calls are native calls so they should follow LoongArch calling
conventions. Sign extend its arguments properly to avoid kernel panic.
This is done by adding a new emit_abi_ext() helper. The emit_abi_ext()
helper performs extension in place meaning a value already store in the
target register (Note: this is different from the existing sign_extend()
helper and thus we can't reuse it).
Chenghao Duan [Wed, 31 Dec 2025 07:19:20 +0000 (15:19 +0800)]
LoongArch: Refactor register restoration in ftrace_common_return
Refactor the register restoration sequence in the ftrace_common_return
function to clearly distinguish between the logic of normal returns and
direct call returns in function tracing scenarios. The logic is as
follows:
1. In the case of a normal return, the execution flow returns to the
traced function, and ftrace must ensure that the register data is
consistent with the state when the function was entered.
ra = parent return address; t0 = traced function return address.
2. In the case of a direct call return, the execution flow jumps to the
custom trampoline function, and ftrace must ensure that the register
data is consistent with the state when ftrace was entered.
ra = traced function return address; t0 = parent return address.
Chenghao Duan [Wed, 31 Dec 2025 07:19:20 +0000 (15:19 +0800)]
LoongArch: Enable exception fixup for specific ADE subcode
This patch allows the LoongArch BPF JIT to handle recoverable memory
access errors generated by BPF_PROBE_MEM* instructions.
When a BPF program performs memory access operations, the instructions
it executes may trigger ADEM exceptions. The kernel’s built-in BPF
exception table mechanism (EX_TYPE_BPF) will generate corresponding
exception fixup entries in the JIT compilation phase; however, the
architecture-specific trap handling function needs to proactively call
the common fixup routine to achieve exception recovery.
Tiezhu Yang [Wed, 31 Dec 2025 07:19:19 +0000 (15:19 +0800)]
LoongArch: Remove unnecessary checks for ORC unwinder
According to the following function definitions, __kernel_text_address()
already checks __module_text_address(), so it should remove the check of
__module_text_address() in bt_address() at least.
int __kernel_text_address(unsigned long addr)
{
if (kernel_text_address(addr))
return 1;
...
return 0;
}
int kernel_text_address(unsigned long addr)
{
bool no_rcu;
int ret = 1;
...
if (is_module_text_address(addr))
goto out;
...
return ret;
}
Furthermore, there are two checks of __kernel_text_address(), one is in
bt_address() and the other is after calling bt_address(), it looks like
redundant.
Handle the exception address first and then use __kernel_text_address()
to validate the calculated address for exception or the normal address
in bt_address(), then it can remove the check of __kernel_text_address()
after calling bt_address().
Just remove unnecessary checks, no functional changes intended.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn> Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Tiezhu Yang [Wed, 31 Dec 2025 07:19:10 +0000 (15:19 +0800)]
LoongArch: Remove is_entry_func() and kernel_entry_end
For now, the related code of is_entry_func() is useless, so they can be
removed. Then the symbol kernel_entry_end is not used any more, so it can
be removed too.
Huacai Chen [Wed, 31 Dec 2025 07:19:10 +0000 (15:19 +0800)]
LoongArch: Complete CPUCFG registers definition
According to the "LoongArch Reference Manual Volume 1: Basic
Architecture", begin with LA664 CPU core there are more features
supported which are indicated in CPUCFG2 and CPUCFG3. This patch
completes the definitions of them so as to match the architecture
specification.
Signed-off-by: Bibo Mao <maobibo@loongson.cn> Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Linus Torvalds [Wed, 31 Dec 2025 01:56:26 +0000 (17:56 -0800)]
Merge tag 'nfsd-6.19-2' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux
Pull nfsd fixes from Chuck Lever:
"A set of NFSD fixes that arrived just a bit late for the 6.19 merge
window.
Regression fix:
- Avoid unnecessarily breaking a timestamp delegation
Stable fixes:
- Fix a crasher in nlm4svc_proc_test()
- Fix nfsd_file reference leak during write delegation
- Fix error flow in client_states_open()"
* tag 'nfsd-6.19-2' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux:
nfsd: Drop the client reference in client_states_open()
nfsd: use ATTR_DELEG in nfsd4_finalize_deleg_timestamps()
nfsd: fix nfsd_file reference leak in nfsd4_add_rdaccess_to_wrdeleg()
lockd: fix vfs_test_lock() calls
Alexandre Negrel [Tue, 30 Dec 2025 18:57:28 +0000 (19:57 +0100)]
io_uring: use GFP_NOWAIT for overflow CQEs on legacy rings
Allocate the overflowing CQE with GFP_NOWAIT instead of GFP_ATOMIC. This
changes causes allocations to fail earlier in out-of-memory situations,
rather than being deferred. Using GFP_ATOMIC allows a process to exceed
memory limits.
Linus Torvalds [Tue, 30 Dec 2025 16:45:58 +0000 (08:45 -0800)]
Merge tag 'net-6.19-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Paolo Abeni:
"Including fixes from Bluetooth and WiFi. Notably this includes the fix
for the iwlwifi issue you reported.
Current release - regressions:
- core: avoid prefetching NULL pointers
- wifi:
- iwlwifi: implement settime64 as stub for MVM/MLD PTP
- mac80211: fix list iteration in ieee80211_add_virtual_monitor()
- handshake: fix null-ptr-deref in handshake_complete()
- eth: mana: fix use-after-free in reset service rescan path
Previous releases - regressions:
- openvswitch: avoid needlessly taking the RTNL on vport destroy
- dsa: properly keep track of conduit reference
- ipv4:
- fix error route reference count leak with nexthop objects
- fib: restore ECMP balance from loopback
- mptcp: ensure context reset on disconnect()
- bluetooth: fix potential UaF in btusb
- nfc: fix deadlock between nfc_unregister_device and
rfkill_fop_write
- eth:
- gve: defer interrupt enabling until NAPI registration
- i40e: fix scheduling in set_rx_mode
- macb: relocate mog_init_rings() callback from macb_mac_link_up()
to macb_open()
- rtl8150: fix memory leak on usb_submit_urb() failure
- wifi: 8192cu: fix tid out of range in rtl92cu_tx_fill_desc()
Previous releases - always broken:
- ip6_gre: make ip6gre_header() robust
- ipv6: fix a BUG in rt6_get_pcpu_route() under PREEMPT_RT
- af_unix: don't post cmsg for SO_INQ unless explicitly asked for
- phy: mediatek: fix nvmem cell reference leak in
mt798x_phy_calibration
- wifi: mac80211: discard beacon frames to non-broadcast address
- eth:
- iavf: fix off-by-one issues in iavf_config_rss_reg()
- stmmac: fix the crash issue for zero copy XDP_TX action
- team: fix check for port enabled when priority changes"
* tag 'net-6.19-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (64 commits)
ipv6: fix a BUG in rt6_get_pcpu_route() under PREEMPT_RT
net: rose: fix invalid array index in rose_kill_by_device()
net: enetc: do not print error log if addr is 0
net: macb: Relocate mog_init_rings() callback from macb_mac_link_up() to macb_open()
selftests: fib_test: Add test case for ipv4 multi nexthops
net: fib: restore ECMP balance from loopback
selftests: fib_nexthops: Add test cases for error routes deletion
ipv4: Fix reference count leak when using error routes with nexthop objects
net: usb: sr9700: fix incorrect command used to write single register
ipv6: BUG() in pskb_expand_head() as part of calipso_skbuff_setattr()
usbnet: avoid a possible crash in dql_completed()
gve: defer interrupt enabling until NAPI registration
net: stmmac: fix the crash issue for zero copy XDP_TX action
octeontx2-pf: fix "UBSAN: shift-out-of-bounds error"
af_unix: don't post cmsg for SO_INQ unless explicitly asked for
net: mana: Fix use-after-free in reset service rescan path
net: avoid prefetching NULL pointers
net: bridge: Describe @tunnel_hash member in net_bridge_vlan_group struct
net: nfc: fix deadlock between nfc_unregister_device and rfkill_fop_write
net: usb: asix: validate PHY address before use
...
Cong Zhang [Tue, 30 Dec 2025 09:17:05 +0000 (17:17 +0800)]
blk-mq: skip CPU offline notify on unmapped hctx
If an hctx has no software ctx mapped, blk_mq_map_swqueue() never
allocates tags and leaves hctx->tags NULL. The CPU hotplug offline
notifier can still run for that hctx, return early since hctx cannot
hold any requests.
Signed-off-by: Cong Zhang <cong.zhang@oss.qualcomm.com> Fixes: bf0beec0607d ("blk-mq: drain I/O when all CPUs in a hctx are offline") Reviewed-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
smb: client: fix UBSAN array-index-out-of-bounds in smb2_copychunk_range
struct copychunk_ioctl_req::ChunkCount is annotated with
__counted_by_le() as the number of elements in Chunks[].
smb2_copychunk_range reuses ChunkCount to store the number of chunks
sent in the current iteration. If a later iteration populates more
chunks than a previous one, the stale smaller value trips UBSAN.
Set ChunkCount to chunk_count (allocated capacity) before populating
Chunks[].
Steve French [Mon, 29 Dec 2025 16:23:12 +0000 (10:23 -0600)]
smb3 client: add missing tracepoint for unsupported ioctls
In debugging a recent problem with an xfstest, noticed that we weren't
tracing cases where the ioctl was not supported. Add dynamic tracepoint:
"trace-cmd record -e smb3_unsupported_ioctl"
and then after running an app which calls unsupported ioctl,
"trace-cmd show"would display e.g.
xfs_io-7289 [012] ..... 1205.137765: smb3_unsupported_ioctl: xid=19 fid=0x4535bb84 ioctl cmd=0x801c581f
Acked-by: Bharath SM <bharathsm@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
Thomas Fourier [Tue, 30 Dec 2025 08:51:21 +0000 (09:51 +0100)]
RDMA/bnxt_re: fix dma_free_coherent() pointer
The dma_alloc_coherent() allocates a dma-mapped buffer, pbl->pg_arr[i].
The dma_free_coherent() should pass the same buffer to
dma_free_coherent() and not page-aligned.
Jiayuan Chen [Tue, 23 Dec 2025 05:14:12 +0000 (13:14 +0800)]
ipv6: fix a BUG in rt6_get_pcpu_route() under PREEMPT_RT
On PREEMPT_RT kernels, after rt6_get_pcpu_route() returns NULL, the
current task can be preempted. Another task running on the same CPU
may then execute rt6_make_pcpu_route() and successfully install a
pcpu_rt entry. When the first task resumes execution, its cmpxchg()
in rt6_make_pcpu_route() will fail because rt6i_pcpu is no longer
NULL, triggering the BUG_ON(prev). It's easy to reproduce it by adding
mdelay() after rt6_get_pcpu_route().
Using preempt_disable/enable is not appropriate here because
ip6_rt_pcpu_alloc() may sleep.
Fix this by handling the cmpxchg() failure gracefully on PREEMPT_RT:
free our allocation and return the existing pcpu_rt installed by
another task. The BUG_ON is replaced by WARN_ON_ONCE for non-PREEMPT_RT
kernels where such races should not occur.
Pwnverse [Mon, 22 Dec 2025 21:22:27 +0000 (21:22 +0000)]
net: rose: fix invalid array index in rose_kill_by_device()
rose_kill_by_device() collects sockets into a local array[] and then
iterates over them to disconnect sockets bound to a device being brought
down.
The loop mistakenly indexes array[cnt] instead of array[i]. For cnt <
ARRAY_SIZE(array), this reads an uninitialized entry; for cnt ==
ARRAY_SIZE(array), it is an out-of-bounds read. Either case can lead to
an invalid socket pointer dereference and also leaks references taken
via sock_hold().
Wei Fang [Mon, 22 Dec 2025 02:26:28 +0000 (10:26 +0800)]
net: enetc: do not print error log if addr is 0
A value of 0 for addr indicates that the IEB_LBCR register does not
need to be configured, as its default value is 0. However, the driver
will print an error log if addr is 0, so this issue needs to be fixed.
Fixes: 50bfd9c06f0f ("net: enetc: set external PHY address in IERB for i.MX94 ENETC") Signed-off-by: Wei Fang <wei.fang@nxp.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20251222022628.4016403-1-wei.fang@nxp.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Xiaolei Wang [Mon, 22 Dec 2025 01:56:24 +0000 (09:56 +0800)]
net: macb: Relocate mog_init_rings() callback from macb_mac_link_up() to macb_open()
In the non-RT kernel, local_bh_disable() merely disables preemption,
whereas it maps to an actual spin lock in the RT kernel. Consequently,
when attempting to refill RX buffers via netdev_alloc_skb() in
macb_mac_link_up(), a deadlock scenario arises as follows:
WARNING: possible circular locking dependency detected 6.18.0-08691-g2061f18ad76e #39 Not tainted
------------------------------------------------------
kworker/0:0/8 is trying to acquire lock: ffff00080369bbe0 (&bp->lock){+.+.}-{3:3}, at: macb_start_xmit+0x808/0xb7c
but task is already holding lock: ffff000803698e58 (&queue->tx_ptr_lock){+...}-{3:3}, at: macb_start_xmit
+0x148/0xb7c
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
Notably, invoking the mog_init_rings() callback upon link establishment
is unnecessary. Instead, we can exclusively call mog_init_rings() within
the ndo_open() callback. This adjustment resolves the deadlock issue.
Furthermore, since MACB_CAPS_MACB_IS_EMAC cases do not use mog_init_rings()
when opening the network interface via at91ether_open(), moving
mog_init_rings() to macb_open() also eliminates the MACB_CAPS_MACB_IS_EMAC
check.
Fixes: 633e98a711ac ("net: macb: use resolved link config in mac_link_up()") Cc: stable@vger.kernel.org Suggested-by: Kevin Hao <kexin.hao@windriver.com> Signed-off-by: Xiaolei Wang <xiaolei.wang@windriver.com> Link: https://patch.msgid.link/20251222015624.1994551-1-xiaolei.wang@windriver.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Vadim Fedorenko [Sun, 21 Dec 2025 19:26:39 +0000 (19:26 +0000)]
selftests: fib_test: Add test case for ipv4 multi nexthops
The test checks that with multi nexthops route the preferred route is the
one which matches source ip. In case when source ip is on dummy
interface, it checks that the routes are balanced.
Vadim Fedorenko [Sun, 21 Dec 2025 19:26:38 +0000 (19:26 +0000)]
net: fib: restore ECMP balance from loopback
Preference of nexthop with source address broke ECMP for packets with
source addresses which are not in the broadcast domain, but rather added
to loopback/dummy interfaces. Original behaviour was to balance over
nexthops while now it uses the latest nexthop from the group. To fix the
issue introduce next hop scoring system where next hops with source
address equal to requested will always have higher priority.
For the case with 198.51.100.1/32 assigned to dummy0 and routed using
192.0.2.0/24 and 203.0.113.0/24 networks:
2: dummy0: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000
link/ether d6:54:8a:ff:78:f5 brd ff:ff:ff:ff:ff:ff
inet 198.51.100.1/32 scope global dummy0
valid_lft forever preferred_lft forever
7: veth1@if6: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
link/ether 06:ed:98:87:6d:8a brd ff:ff:ff:ff:ff:ff link-netnsid 0
inet 192.0.2.2/24 scope global veth1
valid_lft forever preferred_lft forever
inet6 fe80::4ed:98ff:fe87:6d8a/64 scope link proto kernel_ll
valid_lft forever preferred_lft forever
9: veth3@if8: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
link/ether ae:75:23:38:a0:d2 brd ff:ff:ff:ff:ff:ff link-netnsid 0
inet 203.0.113.2/24 scope global veth3
valid_lft forever preferred_lft forever
inet6 fe80::ac75:23ff:fe38:a0d2/64 scope link proto kernel_ll
valid_lft forever preferred_lft forever
~ ip ro list:
default
nexthop via 192.0.2.1 dev veth1 weight 1
nexthop via 203.0.113.1 dev veth3 weight 1
192.0.2.0/24 dev veth1 proto kernel scope link src 192.0.2.2
203.0.113.0/24 dev veth3 proto kernel scope link src 203.0.113.2
before:
for i in {1..255} ; do ip ro get 10.0.0.$i; done | grep veth | awk ' {print $(NF-2)}' | sort | uniq -c:
255 veth3
after:
for i in {1..255} ; do ip ro get 10.0.0.$i; done | grep veth | awk ' {print $(NF-2)}' | sort | uniq -c:
122 veth1
133 veth3
Fixes: 32607a332cfe ("ipv4: prefer multipath nexthop that matches source address") Signed-off-by: Vadim Fedorenko <vadim.fedorenko@linux.dev> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://patch.msgid.link/20251221192639.3911901-1-vadim.fedorenko@linux.dev Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Ido Schimmel [Sun, 21 Dec 2025 14:48:28 +0000 (16:48 +0200)]
ipv4: Fix reference count leak when using error routes with nexthop objects
When a nexthop object is deleted, it is marked as dead and then
fib_table_flush() is called to flush all the routes that are using the
dead nexthop.
The current logic in fib_table_flush() is to only flush error routes
(e.g., blackhole) when it is called as part of network namespace
dismantle (i.e., with flush_all=true). Therefore, error routes are not
flushed when their nexthop object is deleted:
# ip link add name dummy1 up type dummy
# ip nexthop add id 1 dev dummy1
# ip route add 198.51.100.1/32 nhid 1
# ip route add blackhole 198.51.100.2/32 nhid 1
# ip nexthop del id 1
# ip route show
blackhole 198.51.100.2 nhid 1 dev dummy1
As such, they keep holding a reference on the nexthop object which in
turn holds a reference on the nexthop device, resulting in a reference
count leak:
# ip link del dev dummy1
[ 70.516258] unregister_netdevice: waiting for dummy1 to become free. Usage count = 2
Fix by flushing error routes when their nexthop is marked as dead.
IPv6 does not suffer from this problem.
Fixes: 493ced1ac47c ("ipv4: Allow routes to use nexthop objects") Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Closes: https://lore.kernel.org/netdev/d943f806-4da6-4970-ac28-b9373b0e63ac@I-love.SAKURA.ne.jp/ Reported-by: syzbot+881d65229ca4f9ae8c84@syzkaller.appspotmail.com Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/20251221144829.197694-1-idosch@nvidia.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
net: usb: sr9700: fix incorrect command used to write single register
This fixes the device failing to initialize with "error reading MAC
address" for me, probably because the incorrect write of NCR_RST to
SR_NCR is not actually resetting the device.
Fixes: c9b37458e95629b1d1171457afdcc1bf1eb7881d ("USB2NET : SR9700 : One chip USB 1.1 USB2NET SR9700Device Driver Support") Cc: stable@vger.kernel.org Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com> Link: https://patch.msgid.link/20251221082400.50688-1-enelsonmoore@gmail.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Li Zhijian [Fri, 26 Dec 2025 09:41:12 +0000 (17:41 +0800)]
IB/rxe: Fix missing umem_odp->umem_mutex unlock on error path
rxe_odp_map_range_and_lock() must release umem_odp->umem_mutex when an
error occurs, including cases where rxe_check_pagefault() fails.
Fixes: 2fae67ab63db ("RDMA/rxe: Add support for Send/Recv/Write/Read with ODP") Signed-off-by: Li Zhijian <lizhijian@fujitsu.com> Link: https://patch.msgid.link/20251226094112.3042583-1-lizhijian@fujitsu.com Reviewed-by: Zhu Yanjun <yanjun.zhu@linux.dev> Signed-off-by: Leon Romanovsky <leon@kernel.org>
projects / thirdparty / linux.git / log
