| Age | Commit message (Collapse) | Author |
|
Patch series "fat: allocate multiple clusters in fallocate", v2.
This series adds support for allocating multiple clusters in a single
fat_add_clusters() call in fat_fallocate().
The multi-cluster TODO in inode.c is not implemented, as doing so would
require speculative preallocation or an iomap conversion.
This patch (of 3):
fat_alloc_clusters() currently BUG()s when asked to allocate more than
MAX_BUF_PER_PAGE / 2 clusters. This limit comes from the buffer_head
array used to defer syncing and mirroring.
Remove the allocation size limit by doing buffer_head batching like
fat_free_clusters() does.
Link: https://lore.kernel.org/20261012171106.146243-1-matheus.aalmeida@inf.ufrgs.br
Link: https://lore.kernel.org/20261012171106.146243-2-matheus.aalmeida@inf.ufrgs.br
Signed-off-by: Matheus Alves de Almeida <matheus.aalmeida@inf.ufrgs.br>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
|
|
vfat_create(), vfat_mkdir(), msdos_create() and msdos_mkdir() add the
on-disk entry before calling fat_build_inode() to set up the VFS inode.
If fat_build_inode() fails, the error is returned but the entry stays on
disk and the file or directory shows up after remount.
Remove the entry in that case. For mkdir, also free the directory cluster
and decrement the parent's link count.
[akpm@linux-foundation.org" export _fat_msg() to modules]
Link: https://lore.kernel.org/tencent_4EDD66523CED299DB0C99498B4D8C939A009@qq.com
Signed-off-by: zhoumin <teczm@foxmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
|
|
Use isdigit() to replace the open-coded implementation.
Link: https://lore.kernel.org/20261013072108.4407-3-blum@kernel.org
Signed-off-by: Thorsten Blum <blum@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
|
|
ocfs2_filecheck_create_sysfs() initializes the filecheck allocation before
calling kobject_init_and_add(), but assigns entry->fs_fcheck only after
that call returns. The default attribute group can already be visible
while registration is in progress. A concurrent read can therefore enter
ocfs2_filecheck_attr_show() with a NULL fs_fcheck and try to lock its
fc_lock.
Assign the initialized owner before adding the kobject, so every activated
attribute has valid state. Clear the pointer when addition fails,
preserving the existing teardown guard and error cleanup.
A controlled probe-instrumented test kernel with the original OCFS2
registration ordering produced this Oops during a concurrent sysfs read:
KASAN: null-ptr-deref in range [0x0000000000000028-0x000000000000002f]
RIP: 0010:kasan_byte_accessible+0x15/0x30
Call Trace:
<TASK>
__kasan_check_byte+0x13/0x50
lock_acquire+0x113/0x300
_raw_spin_lock+0x31/0x40
ocfs2_filecheck_attr_show+0x138/0x5f0
ocfs2_filecheck_show+0x5d/0x90
sysfs_kf_seq_show+0x1d3/0x360
seq_read_iter+0x480/0x1110
kernfs_fop_read_iter+0x419/0x5a0
vfs_read+0x7d2/0xc20
ksys_read+0x111/0x200
do_syscall_64+0x114/0x620
entry_SYSCALL_64_after_hwframe+0x77/0x7f
</TASK>
Modules linked in:
Link: https://lore.kernel.org/pm-ocfs2-objects-candidate-0086-v4-1b3a3c5fd05c1fb41df9@gmail.com
Fixes: 5f483c4abb50 ("ocfs2: add kobject for online file check")
Signed-off-by: Cen Zhang <zzzccc427@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Heming Zhao <heming.zhao@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Assisted-by: LLM
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Gang He <ghe@suse.com>
Cc: Jia-Ju Bai <baijiaju1990@gmail.com>
Cc: <jjzuming@gmail.com>
Cc: <stable@vger.kernel.org>
|
|
o2net_prep_nsw() publishes its stack waiter in nn_status_idr and
nn_status_list under nn_lock, then initializes the wait queue and
completion status after releasing the lock. A concurrent disconnect can
find that waiter through o2net_complete_nodes_nsw() and call wake_up()
before the wait queue has a valid lock and list head. The sender can also
overwrite a completion status with its initial value.
Initialize the wait queue and status before publishing the waiter.
nn_lock then orders publication against completion, and no reader can
observe a partially initialized object. Leave the IDR allocation and its
failure handling unchanged.
A controlled case-modified test kernel used a synthetic connection to
enter the real send path and overlap publication with disconnect. The
candidate completion and wake-up code were unchanged. Its Oops includes:
KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
RIP: 0010:__wake_up_common+0xa0/0x1f0
Call Trace:
<TASK>
__wake_up+0x36/0x60
o2net_complete_nsw_locked+0x222/0x370
o2net_set_nn_state+0x917/0xea0
o2net_disconnect_node+0xd0/0x190
o2net_validate_control_write+0x454/0x500
full_proxy_write+0x11f/0x180
vfs_write+0x25a/0x1010
ksys_write+0x111/0x200
do_syscall_64+0x114/0x620
entry_SYSCALL_64_after_hwframe+0x77/0x7f
</TASK>
Modules linked in:
Link: https://lore.kernel.org/pm-ocfs2-objects-candidate-0066-v1-437a0422801b292a2879@gmail.com
Fixes: 98211489d414 ("[PATCH] OCFS2: The Second Oracle Cluster Filesystem")
Signed-off-by: Cen Zhang <zzzccc427@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Assisted-by: LLM
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Kees Cook <kees@kernel.org>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Kurt Hackel <kurt.hackel@oracle.com>
Cc: Jia-Ju Bai <baijiaju1990@gmail.com>
Cc: <jjzuming@gmail.com>
|
|
Commit 3bc753c06dd0 ("kbuild: treat char as always unsigned") set
-funsigned-char globally, which changed the result of the naked 'char'
load in ocfs2_xattr_name_hash():
hash = (hash << OCFS2_HASH_SHIFT) ^
(hash >> (8*sizeof(hash) - OCFS2_HASH_SHIFT)) ^
*name++;
A name byte >= 0x80 used to sign-extend and now zero-extends, so the
hash no longer matches the xe_name_hash an older kernel stored. An
indexed xattr tree is searched by that hash alone, and both the bucket
binary search and the entry scan within it stop as soon as the wanted
hash falls below an entry's, so the entry is never reached: getxattr,
setxattr and removexattr return -ENODATA for a name that listxattr
still lists.
Search with the current unsigned hash and, on a miss, retry with the
legacy signed one, as ext4 does in commit f3bbac32475b2 ("ext4: deal
with legacy signed xattr name hash values"). New entries are always
stored under the unsigned hash. Skip the retry when the two hashes are
equal, so that a miss on an ASCII name does not walk the tree twice.
A miss is not empty handed: ocfs2_xattr_bucket_find() leaves xs->bucket
holding the bucket a new entry would go into. So after a double miss
drop it and search once more with the unsigned hash, otherwise a new
entry would be placed by its legacy hash and stored under its unsigned
one, breaking the ordering the search relies on.
Only indexed trees are affected; inline xattrs and non-indexed xattr
blocks compare names with memcmp and never look at the hash.
Also spell out the signedness instead of leaving the current hash to
-funsigned-char, as commit 854f0912f813 ("ext4: make xattr char
unsignedness in hash explicit") did, so that both variants stay correct
if this is backported to a kernel without the flag.
Link: https://lore.kernel.org/20261009083002.2621201-4-joseph.qi@linux.alibaba.com
Fixes: 3bc753c06dd0 ("kbuild: treat char as always unsigned")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Heming Zhao <heming.zhao@suse.com>
Cc: Anderson Ferneda <anderson.ferneda@braza.com.br>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: <stable@vger.kernel.org> # 6.2+
|
|
ocfs2_check_xattr_bucket_collision() decides whether splitting a full
bucket can make room for the entry being set by recomputing the hash of
the name:
u32 name_hash = ocfs2_xattr_name_hash(inode, name, strlen(name));
if (name_hash != le32_to_cpu(xh->xh_entries[0].xe_name_hash))
return 0;
For a new entry that is the hash it will be stored under, so comparing it
is correct. An existing entry keeps the hash it already has: an update
never rewrites xe_name_hash, since ocfs2_xa_add_entry() is the only
writer of that field for a real entry and ocfs2_xa_prepare_entry() calls
it only when loc->xl_entry is NULL.
An entry stored by a kernel that sign-extended the name bytes when
hashing can still be updated, because ocfs2_xattr_find_entry() searches
a non-indexed xattr block by memcmp on the name and never looks at
xe_name_hash. Growing it past the space left in the block converts the
block into a tree, ocfs2_cp_xattr_block_to_bucket() fills the bucket in
stored hash order, and the update runs out of room in the bucket too.
The collision check then compares the unsigned hash against the legacy
hashes in the bucket and reports no collision.
ocfs2_xattr_set_entry_index_block() goes on to allocate a bucket that
ocfs2_divide_xattr_bucket() cannot fill: a bucket whose entries all share
one hash has no divide position, so all it does is append an empty bucket
with a sentinel hash one above the last entry's.
The re-search that follows depends on where the unsigned hash sorts.
Below the legacy ones, it comes back to the full bucket and the set fails
with -ENOSPC, having grown the tree for nothing. Above them, it lands on
the new empty bucket, which ocfs2_xattr_bucket_find() handles explicitly
and ocfs2_find_xe_in_bucket() scans zero times, so the set stores a second
copy of the entry under the unsigned hash and returns success. The
original stays in the full bucket, listxattr reports the name twice and
getxattr returns the new copy.
Pass the hash the entry is stored under instead: the stored one for an
existing entry, the unsigned one for a new entry. A tree whose entries
are all stored under the unsigned hash sees no change, since there the
two are the same value.
Link: https://lore.kernel.org/20261009083002.2621201-3-joseph.qi@linux.alibaba.com
Fixes: 3bc753c06dd0 ("kbuild: treat char as always unsigned")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Heming Zhao <heming.zhao@suse.com>
Cc: Anderson Ferneda <anderson.ferneda@braza.com.br>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: <stable@vger.kernel.org> # 6.2+
|
|
Patch series "ocfs2: deal with legacy signed name hash values", v2.
Commit 3bc753c06dd0 ("kbuild: treat char as always unsigned") turned on
-funsigned-char globally, which changed the result of two naked 'char'
loads in ocfs2 hashing code:
str2hashbuf() val = msg[i] + (val << 8);
ocfs2_xattr_name_hash() hash = (hash << 5) ^ (hash >> 27) ^ *name++;
A name byte >= 0x80 used to sign-extend and now zero-extends. Both
results are written to disk, the first into the dx leaf when a directory
entry is indexed and the second into xe_name_hash when an xattr is
stored, so anything written by a pre-6.2 kernel is looked up under a
different hash and is no longer found. readdir and listxattr still list
those names because neither of them hashes, which makes this look like
the filesystem losing entries rather than like a lookup bug. Anderson
Ferneda reported it that way for directories with non-ASCII names [1].
Patches 1 and 3 do what ext4 did in commit f3bbac32475b2 ("ext4: deal
with legacy signed xattr name hash values"): look up with the current
unsigned hash, retry with the legacy signed one on a miss, and always
store new entries under the unsigned hash. The retry is skipped when the
two hashes come out equal, so an ASCII name does not walk the index
twice. Both hash functions now spell out the signedness instead of
leaving it to -funsigned-char, as commit 854f0912f813 ("ext4: make xattr
char unsignedness in hash explicit") did, so a backport to a kernel
without the flag still computes both variants correctly.
Patch 2 is a separate fix, placed before patch 3 so that every commit in
the series is correct on its own. ocfs2_check_xattr_bucket_collision()
recomputes the hash of the name to decide whether splitting a full bucket
can make room for the entry being set. That is right for a new entry and
wrong for one that already exists, since an update leaves xe_name_hash
alone. Against a bucket holding legacy hashes the comparison does not
match, so the split goes ahead and cannot help: a bucket whose entries
all share one hash has no divide position. How that ends depends on
where the unsigned hash sorts. Below the legacy ones, the set fails with
-ENOSPC after growing the tree for nothing. Above them, it lands in the
empty bucket the split just appended and stores a second copy of the
entry there while the original stays behind, so listxattr reports the
name twice. Using the stored hash closes both.
Nothing retires a legacy hash at runtime. Deleting or updating an entry
does not rewrite it, and the only code that rehashes existing dirents is
ocfs2_expand_inline_dir() on the inline to extent conversion, so the
retry stays for as long as the legacy entries do. Rebuilding the index
offline is what removes it for good.
This patch (of 3):
Commit 3bc753c06dd0 ("kbuild: treat char as always unsigned") set
-funsigned-char globally, which changed the result of the naked 'char'
load in str2hashbuf():
val = msg[i] + (val << 8);
A name byte >= 0x80 used to sign-extend and now zero-extends, so
ocfs2_dx_dir_name_hash() computes a different hash pair for every name
containing one. The pair is written into the dx leaf when the entry is
created, so an index built by an older kernel no longer matches and
ocfs2_dx_dir_search() returns -ENOENT for a name that readdir still
lists.
Search with the current unsigned hash and, on a miss, retry with the
legacy signed one, as ext4 does in commit f3bbac32475b2 ("ext4: deal
with legacy signed xattr name hash values"). New entries are always
indexed under the unsigned hash. Skip the retry when the two hashes
are equal, so that a miss on an ASCII name does not walk the index
twice.
The retry costs a second walk of the dx tree, so a negative lookup of a
name containing a byte >= 0x80 does twice the work. That is bounded and
worth it, and directories whose dx root is still inline pay nothing
extra, since there the retry only rescans the root block that is already
loaded. Nothing retires it at runtime: a legacy entry keeps its legacy
hash, because neither deleting nor updating an entry rewrites the dx
hash, and the only code that rehashes existing dirents is
ocfs2_expand_inline_dir() on the inline to extent conversion.
Rebuilding the index offline is what removes the cost for good.
Also spell out the signedness in str2hashbuf() instead of leaving the
current hash to -funsigned-char, as commit 854f0912f813 ("ext4: make
xattr char unsignedness in hash explicit") did, so that both variants
stay correct if this is backported to a kernel without the flag.
Link: https://lore.kernel.org/20261009083002.2621201-1-joseph.qi@linux.alibaba.com
Link: https://lore.kernel.org/20261009083002.2621201-2-joseph.qi@linux.alibaba.com
Fixes: 3bc753c06dd0 ("kbuild: treat char as always unsigned")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reported-by: Anderson Ferneda <anderson.ferneda@braza.com.br>
Closes: https://lore.kernel.org/ocfs2-devel/CP5P284MB2780AC2C3C2AB2CF2B6CB7AFBE942@CP5P284MB2780.BRAP284.PROD.OUTLOOK.COM/
Reviewed-by: Heming Zhao <heming.zhao@suse.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: <stable@vger.kernel.org> # 6.2+
|
|
get_range() stops writing once the array is full, but get_options() still
advances its index by the whole range. So "1-100" parsed into four ints
reports 99 numbers, and a range like "0-2147483647" wraps the index
negative and writes outside the array.
Stop once a range fills the array, bail out in validation mode before the
count overflows, and saturate the range length in get_range(). Add KUnit
cases; without the fix "0-2147483647" panics the test kernel.
Only root can supply this input, so this is a robustness fix.
[akpm@linux-foundation.org: add additional test case from Leizhen]
Link: https://lore.kernel.org/20261005213945.359905-1-armaan.sandhu0504@gmail.com
Fixes: 22f2e2801799 ("[PATCH] get_options to allow a hypenated range for isolcpus")
Signed-off-by: Armaan Sandhu <armaan.sandhu0504@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reported-by: Leizhen Zhang <lzsx618@gmail.com>
Reviewed-by: Leizhen Zhang <lzsx618@gmail.com>
Cc: Andy Shevchenko <andriy.shevchenko@intel.com>
Cc: <stable@vger.kernel.org>
|
|
o2hb_shutdown_slot() returns immediately when its node is no longer in the
configfs registry. A slot can still be linked in o2hb_live_slots when
that lookup fails. Region teardown or startup-error cleanup then frees
the slot array without removing its embedded list member.
With the same peer live in another region, the next list update can follow
a predecessor into the freed array. Reusing the node number can also make
a later insertion encounter that stale member.
Always remove the slot under o2hb_live_lock, even when node lookup fails.
Preserve aggregate membership accounting and the last-region DOWN
callback: the existing callback interface explicitly permits a NULL node
for DOWN events. Only drop the node reference when lookup returned one.
This closes both normal release and startup-abort paths without changing
the publication or callback locking protocol.
Also require a configured node before admitting a new live slot in
o2hb_check_slot(). The early live-bitmap check is not held through
admission: the last live slot in another region can be removed in between.
Without a node, admission can then queue an UP event with a NULL argument
and hit the event helper's BUG_ON. Existing live slots can still leave
and generate NULL-node DOWN events.
A controlled probe-instrumented test reproduced this while stopping one of
two regions after removing the peer from configfs. Its KASAN report
includes:
BUG: KASAN: slab-use-after-free in __list_del_entry_valid_or_report+0x1e0/0x210
Read of size 8 at addr ffff888106a4ac60 by task o2hb-pmbd0058_b/497
Call Trace:
<TASK>
dump_stack_lvl+0x93/0xd0
print_report+0xce/0x630
kasan_report+0xe0/0x110
__list_del_entry_valid_or_report+0x1e0/0x210
o2hb_do_disk_heartbeat+0x1591/0x2c50
o2hb_thread+0x1ed/0xe70
kthread+0x351/0x460
ret_from_fork+0x659/0x940
ret_from_fork_asm+0x1a/0x30
</TASK>
Allocated by task 491:
kasan_save_stack+0x33/0x60
kasan_save_track+0x14/0x30
__kasan_kmalloc+0xaa/0xb0
__kmalloc_noprof+0x2de/0x780
o2hb_region_dev_store+0x7c2/0x1b60
configfs_write_iter+0x2f9/0x4f0
vfs_write+0x639/0x1010
ksys_write+0x111/0x200
do_syscall_64+0x114/0x620
entry_SYSCALL_64_after_hwframe+0x77/0x7f
Freed by task 491:
kasan_save_stack+0x33/0x60
kasan_save_track+0x14/0x30
kasan_save_free_info+0x3b/0x60
__kasan_slab_free+0x5f/0x80
kfree+0x308/0x580
o2hb_unmap_slot_data+0x1db/0x330
o2hb_region_release+0x14b/0x4f0
config_item_cleanup+0x14b/0x230
config_item_put+0x79/0x90
configfs_rmdir+0x58a/0x910
vfs_rmdir+0x2e8/0x820
filename_rmdir+0x3b1/0x520
__x64_sys_rmdir+0x4b/0x70
do_syscall_64+0x114/0x620
entry_SYSCALL_64_after_hwframe+0x77/0x7f
Link: https://lore.kernel.org/pm-ocfs2-objects-candidate-0058-v2-9446113c2b16c817c538@gmail.com
Fixes: a7f6a5fb4bde ("[PATCH] OCFS2: The Second Oracle Cluster Filesystem")
Signed-off-by: Cen Zhang <zzzccc427@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Assisted-by: LLM
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Yufan Chen <ericterminal@gmail.com>
Cc: Kees Cook <kees@kernel.org>
Cc: Jia-Ju Bai <baijiaju1990@gmail.com>
Cc: <jjzuming@gmail.com>
Cc: <stable@vger.kernel.org>
|
|
ocfs2_calc_xattr_set_need() asks for the new value's extent tree twice
when it replaces an inline value smaller than a value root with one large
enough to be stored outside. The "stored outside" branch reserves the
tree and then falls through to meta_guess instead of going out:
if (!ocfs2_xattr_is_local(xe)) {
...
xv = (struct ocfs2_xattr_value_root *)
(base + name_offset + name_len);
value_size = OCFS2_XATTR_ROOT_SIZE;
} else
xv = &def_xv.xv;
if (old_clusters >= new_clusters) {
...
goto out;
} else {
meta_add += ocfs2_extend_meta_needed(&xv->xr_list);
...
if (value_size >= OCFS2_XATTR_ROOT_SIZE)
goto out;
}
On the fall-through xv is def_xv, so what gets added here is the two
blocks ocfs2_extend_meta_needed() wants for an empty value root -- and
both sides of meta_guess reserve that same tree again. The create side
has done so since commit 3ed2be719eb9 ("ocfs2: allow for more than one
data extent when creating xattr"), which ported this reservation to the
create case without noticing it was already on the way in, and the
existing-block side since commit 0cdc7dde00ec ("ocfs2: fix missing
metadata reservation for large xattrs"). Both were chasing the same
RESTART_META failure in ocfs2_xattr_extend_allocation(), for xattr create
and then for xattr update.
ocfs2_init_xattr_set_ctxt() hands the total straight to
ocfs2_reserve_new_metadata_blocks(), so the surplus is two blocks held for
the whole transaction. A setxattr that would otherwise fit can come back
-ENOSPC on a filesystem whose metadata allocator is nearly exhausted.
Move the reservation into the branch that stops there, leaving meta_guess
as the only place that reserves for a value tree the fall through is about
to create. The paths that do go out keep reserving against the value root
they already have, which for an external old value is the on-disk one and
so can legitimately ask for more than two blocks.
Both sides of meta_guess have to come out even on their own now that the
copy covering for them is gone. The existing-block side only started
reserving the value tree in the commit this one is tagged against. The
create side has reserved the tree since 3ed2be719eb9, but only the tree:
the block ocfs2_create_xattr_block() claims came out of the same two, and
on the fall-through it was the duplicate removed here that made up the
difference. So "ocfs2: reserve metadata for a new xattr block and its
value tree" has to land first -- without it the create side reserves 2
where 3 are needed and the value tree is back to RESTART_META. That one
fixes a shortfall of its own on the path that reaches meta_guess directly,
where there was never a duplicate to cover it, so it is worth taking
alone; this patch is not.
Link: https://lore.kernel.org/20261010113454.1420939-4-joseph.qi@linux.alibaba.com
Fixes: 0cdc7dde00ec ("ocfs2: fix missing metadata reservation for large xattrs")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Heming Zhao <heming.zhao@suse.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
|
|
ocfs2_calc_xattr_set_need() charges no metadata for the case where an
xattr living in an xattr block or bucket moves back into the inode:
if (old_in_xb) {
if (ocfs2_xattr_can_be_in_inode(inode, xi, xis)) {
clusters_add += new_clusters;
credits += ...;
if (!ocfs2_xattr_is_local(xe))
credits += ocfs2_calc_extend_credits(...);
goto out;
}
}
The credits for extending a value tree are there but the matching blocks
are not, so meta_add stays 0, ocfs2_init_xattr_set_ctxt() skips
ocfs2_reserve_new_metadata_blocks() and ctxt->meta_ac is left NULL.
This is the shape commit 0cdc7dde00ec ("ocfs2: fix missing metadata
reservation for large xattrs") fixed for the xattr block case, and it is
reached the same way: namevalue_size() charges a value larger than
OCFS2_XATTR_INLINE_SIZE as only OCFS2_XATTR_ROOT_SIZE, so
ocfs2_xattr_can_be_in_inode() accepts a new value that will be stored
outside. ocfs2_xa_prepare_entry() installs a value root from def_xv and
grows it, and once the cluster allocator is fragmented enough to need a
second extent ocfs2_add_clusters_in_btree() finds the root full with no
meta_ac at all and answers RESTART_META. ocfs2_xattr_extend_allocation()
turns that into -ENOSPC, or into a BUG_ON() without that commit, and
ocfs2_xa_cleanup_value_truncate() takes its !orig_clusters arm, dropping
the entry and leaking the clusters already handed out until fsck.ocfs2
reclaims them.
Reserve for the value tree here too. The entry moves back into the inode
rather than into a new xattr block, so unlike the create side of
meta_guess this needs no block of its own.
Link: https://lore.kernel.org/20261010113454.1420939-3-joseph.qi@linux.alibaba.com
Fixes: 78f30c314a74 ("ocfs2/xattr: Reserve meta/data at the beginning of ocfs2_xattr_set.")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Heming Zhao <heming.zhao@suse.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: <stable@vger.kernel.org>
|
|
Patch series "ocfs2: fix metadata reservation for xattr value trees".
ocfs2_calc_xattr_set_need() reserves too little metadata for a new xattr
value's extent tree on three paths. The reservation cannot be topped up
once the transaction has started, so ocfs2_add_clusters_in_btree() gives
up with RESTART_META when the value root fills and nothing is left.
ocfs2_xattr_extend_allocation() turns that into -ENOSPC, after
ocfs2_xa_cleanup_value_truncate() has dropped the entry and leaked the
clusters already handed out; before commit 0cdc7dde00ec ("ocfs2: fix
missing metadata reservation for large xattrs") it was a BUG_ON() instead.
It takes a fragmented cluster allocator to get there, since the first
extent always fits in the root.
Patch 1 fixes the create side of meta_guess: commit 3ed2be719eb9 ("ocfs2:
allow for more than one data extent when creating xattr") replaced the
block reserved for the new xattr block with the two the value tree needs
instead of adding them together.
Patch 2 fixes the arm that moves an xattr back into the inode, which
charges credits for a value tree but no metadata for one.
Patch 3 drops a second copy of the same reservation on the path that falls
through to meta_guess. That one is cleanup rather than a fix and is not
for stable, and it must not be backported without patch 1: the duplicate
it removes is what currently covers the shortfall patch 1 fixes.
This patch (of 3):
The create side of meta_guess in ocfs2_calc_xattr_set_need() reserves
metadata for the new value's extent tree and nothing for the xattr block
that will hold it:
} else {
credits += OCFS2_XATTR_BLOCK_CREATE_CREDITS;
if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) {
struct ocfs2_extent_list *el = &def_xv.xv.xr_list;
meta_add += ocfs2_extend_meta_needed(el);
...
Commit 3ed2be719eb9 ("ocfs2: allow for more than one data extent when
creating xattr") meant to add that tree on top of the meta_add += 1 the
branch already had, but it moved the 1 into the else arm instead, so the
external-value case lost it. The 1 is not optional: it is what
ocfs2_create_xattr_block() takes from meta_ac through
ocfs2_claim_metadata(), and it is spent before the value tree exists.
ocfs2_extend_meta_needed() on the empty def_xv list returns 2, so the
branch reserves 2 where 3 are needed. The xattr block consumes one, and
when the cluster allocator is fragmented enough that the value needs more
than one extent, ocfs2_add_clusters_in_btree() comes back around with a
full value root and ocfs2_alloc_context_bits_left() of 1 against a
threshold of 2. It answers RESTART_META, which
ocfs2_xattr_extend_allocation() turns into a BUG_ON(). Commit
0cdc7dde00ec ("ocfs2: fix missing metadata reservation for large xattrs")
downgraded that to -ENOSPC, at the cost of leaking the clusters already
handed out: ocfs2_xa_cleanup_value_truncate() takes its !orig_clusters
arm, drops the entry and leaves the clusters for fsck.ocfs2 to reclaim.
Reserve for the block and the tree separately.
Link: https://lore.kernel.org/20261010113454.1420939-1-joseph.qi@linux.alibaba.com
Link: https://lore.kernel.org/20261010113454.1420939-2-joseph.qi@linux.alibaba.com
Fixes: 3ed2be719eb9 ("ocfs2: allow for more than one data extent when creating xattr")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Heming Zhao <heming.zhao@suse.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: <stable@vger.kernel.org>
|
|
The danger of doing the merge window while on vacation and traveling
with just the laptop: I only do full builds a couple of times a day
because they are so slow on my poor little laptop.
As a result, I missed the conflict between the newly added smb client
fileattr_set function (commit 273913b7c33a: "smb: client: implement
fileattr_set for compression flags") and the vfs layer having in the
meantime made the idmap argument 'const' (commit 025756ee7b47: "fs: port
mnt_idmap() and file_mnt_idmap() to const mnt_idmap")
Mea culpa.
Fixes: 32543b09681a ("Merge tag 'cifs-for-7.4' of https://git.manguebit.org/linux")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Maxime Chevallier says:
====================
selftests: drv-net: Introduce flow-control selftest
This test aims at validating the flow control implementation in drivers,
using the ethtool API.
The test can run in various conditions :
DUT <-----> Anything that has link up (not controllable)
DUT <-----> A directly connected device
It enumerates the local and remote device's capabilities, to test as
much as possibly testable. I've run it on 17 different boards connected
to other boards through various means : Usb to Eth dongles, direct
connection to the board, connection through a switch. Some of the ports
tested were user ports of DSA switches too.
Contrary to V2, on V3 we really only have a collection of actual
failures that show during tests, as one of the peer's I'm running these
on has a v6.18 kernel that pre-dates the realtek phylink port and show
actual bugs in the driver.
====================
Link: https://patch.msgid.link/20261015134350.629593-1-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Pause autonegotiation happens with the link partner using the same words
as the link negotiation, used for speed and duplex exchanges. However,
pause and link autonegotiation can be separately toggled :
ethtool -s eth0 autoneg on # Enable link negotiation
ethtool -A eth0 autoneg on # Enable Pause negotiation
Pause can't be negotiated if the link autoneg isn't enabled.
Pause autoneg and link autoneg settings must not interfere with one
another when user is configuring them :
- If Pause autoneg is on, it must stay on when link autoneg is disabled
(even though Pause won't actually be negotiated)
- If Pause autoneg is off, it must stay off when link autoneg is
enabled
Introduce a set of tests to verify that pause and link autoneg are
behaving correctly, in particular that the user intent is correctly
cached when toggling link autoneg.
Signed-off-by: Maxime Chevallier (Netdev Foundation) <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/20261015134350.629593-4-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Pause autonegotiation behaves differently than the regular link params
autonegotiation, in that the user intention may differ from the
negotiated pause parameters.
Introduce a test that validates we're correctly resolving pause
parameters according to 802.3.
The validation is done based on what the local interface reports in
terms of advertised and lp_advertised linkmodes.
If the remote can express what it sees in terms of its own advertised
and lp_advertised modes, validate the autoneg results on the peer as
well.
Signed-off-by: Maxime Chevallier (Netdev Foundation) <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/20261015134350.629593-3-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Ethernet flow control is a tricky thing to get right, especially when it
comes to correctly handling the negotiation of the parameters with a
link partner.
On one hand, the userspace API talks in terms of the device's ability to
send Pause frames when the device is overwhelmed by ingress traffic (TX
pause) and the ability to stop sending traffic upon receiving Pause
frames (RX pause).
On the other hand, 802.3 explains that devices can exchange their
abilities over link negotiation, but instead of echanging TX and RX
abilities, they exchange Pause and Asymmetric Pause capabilities :
RX unable TX unable => None
RX able TX unable => Pause + Asymmetric
RX unable TX able => Asymmetric
RX able TX able => Pause
The reported Pause and Asym abilities depend both on the MAC, that
eventually sends and processes these frames, and the PHY, that
advertises these to the partner.
Both MAC and PHYs can have their own limitations, and getting the
correct set of supported Pause/Asym parameters is non-trivial, unless
the MAC uses phylink, which deals with the complexity.
Introduce a set of Pause selftests that verify that the local
interface's supported pause parameters reported from ethtool (in terms
of Pause + Asym ) match the accepted parameters from "ethtool -A",
corresponding to ethtool's .set_pauseparams() ops, expressed in TX and
RX abilities.
Signed-off-by: Maxime Chevallier (Netdev Foundation) <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/20261015134350.629593-2-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
This reverts commit 587acb86d6465efdc03f6d8524461e6e725597b7.
While the i.MX95 MDIO MDC clock is improved by the commit
587acb86d646 ("enetc: Increase eMDIO MDC rate to 2.5 MHz"),
other SoCs use different base clock for their MDIO and the
MDC clock on those SoCs is set out of specification by the
aforementioned commit. The fix for the i.MX95 MDIO MDC has
to be done differently, and will be done so in a follow up
change. Until that is done, revert the commit to fix the
other SoCs.
Fixes: 587acb86d646 ("enetc: Increase eMDIO MDC rate to 2.5 MHz")
Signed-off-by: Marek Vasut <marex@nabladev.com>
Link: https://patch.msgid.link/20261016185817.54096-1-marex@nabladev.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm
Pull dlm updates from David Teigland:
"This contains several fixes for bugs found by static analysis, fuzzing
and code review. They mostly involve invalid input, e.g. from
userspace, configfs or remote nodes, and are unlikely to be seen in
normal usage"
* tag 'dlm-7.4' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm:
dlm: fix typos in comments
dlm: clean up a type in waiters_read()
dlm: validate node weights before building member array
dlm: wait for outstanding SRCU callbacks to complete in exit paths
dlm: fix variable key length lookup
dlm: fix NULL pointer dereference in dlm_dump_rsb_name()
dlm: validate lock modes in recovery messages
dlm: fix buffer overflow from negative len in dlm_search_rsb_tree
dlm: validate userspace lock resource name length
dlm: require CAP_SYS_ADMIN for dlm-monitor device
dlm: gate dlm_plock device on CAP_SYS_ADMIN
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs
Pull erofs updates from Gao Xiang:
"The most notable change is to support superblock sharing for
composefs, thanks to Giuseppe again.
Another one switches to iomap for file-backed mounts on unencoded
inodes, which enables efficient sub-folio I/Os. The others are all
random ones.
Summary:
- Support superblock sharing for file-backed mounts
- Use iomap for file-backed mounts on unencoded inodes
- Minor fixes and cleanups"
* tag 'erofs-for-7.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs:
erofs: enable iomap for file-backed mounts on unencoded inodes
erofs: use on-disk blkszbits for lz4 max pcluster bound
mailmap: update entry for Gao Xiang
MAINTAINERS: erofs: update my email address
erofs: reuse superblock for file-backed mounts
MAINTAINERS: erofs: add dev branch to git tree entry
erofs: fix xattrs handling when metadata compression is used
|
|
Pull xfs updates from Carlos Maiolino:
"The bulk of this are bug fixes, cleanups and code refactoring, lots of
them. Majority of those bug fixes are targeting scrub and online
repair code. There are no real new features here but lots of prep code
for future work. There are also some small performance optimizations
but nothing major"
* tag 'xfs-merge-7.4' of git://git.kernel.org:/pub/scm/fs/xfs/xfs-linux: (168 commits)
xfs: don't limit software atomic writes by the group alignment
xfs: add missing cursor stat checks after xfs_btree_delete
xfs: add missing cursor status check after xfs_inobt_get_rec
xfs: handle NULL open_zone for merged ioends in xfs_end_ioend_write_zoned
xfs: refactor inode forkoff to byte conversions
xfs: drop unnecessary sr parameter from xchk_rtgroup init functions
xfs: drop unnecessary sa parameter from xchk_ag init functions
xfs: drop unnecessary sr parameter from xrep_rtgroup init functions
xfs: drop unnecessary sa parameter from the xrep_ag init functions
xfs: improve dirent bounds checking in scrub and repair
xfs: refactor directory data block walking loops in scrub
xfs: update quota flags on secondary supers when metadir enabled
xfs: don't try to repair rtsummary if the summary file isn't computed
xfs: rescan the AGs if incore delalloc counter is negative
xfs: don't rescan ifree and icount when frozen
xfs: don't set a metadata directory's parent to sb_rootino
xfs: don't scan across metadata/regular directory trees for dirents
xfs: always report scrub errors in tracepoints
xfs: opportunistically skip btree block checks
xfs: clear set[ug]id mode bits when zapping attr fork
...
|
|
Sander Vanheule says:
====================
net: pse-pd: Plain I2C transport for Realtek Gen1
While for the Gen2 protocol both a plain I2C transport and an SMBus
transport were provided, Gen1 only got SMBus. The Unifi Switch Lite 16
PoE uses the Gen1 protocol to manage a single BCM59121, but does to
using plain I2C.
These patches simply add a new compatible for the plain I2C transport
and the associated match data. This is sufficient to let the existing
code talk to the management MCU.
====================
Link: https://patch.msgid.link/20261018161248.376205-1-sander@svanheule.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Provide the required data for the gen1 raw I2C binding, which like for
gen2 is simply the gen1 protocol data with the native_i2c flag set.
Signed-off-by: Sander Vanheule <sander@svanheule.net>
Reviewed-by: Kory Maincent <kory.maincent@bootlin.com>
Acked-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Jonas Jelonek <jonas@jonasjelonek.de>
Link: https://patch.msgid.link/20261018161248.376205-3-sander@svanheule.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Like the gen2 protocol, the gen1 protocol can also be used over both a
raw I2C transport and an SMBUs transport. Add a compatible for the
former and the compatible for the Ubiquiti Unifi Switch Lite 16 PoE
device that uses it.
Signed-off-by: Sander Vanheule <sander@svanheule.net>
Reviewed-by: Kory Maincent <kory.maincent@bootlin.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://patch.msgid.link/20261018161248.376205-2-sander@svanheule.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2
Pull gfs2 updates from Andreas Gruenbacher:
- Two journaled data inode fixes (fix journaled truncate range end
offset, fix the escaped block copy in gfs2_before_commit)
- When reading a new resource group from disk, validate its boundaries
- To eliminate lock contention, get rid of the per-filesystem
sd_async_glock_wait queue and wait for the individual glock locking
events instead
- Shut up sparse RCU warnings by marking ip->i_gl as __rcu. Accessing
it through gfs2_inode_glock() -> rcu_dereference_protected() where we
know that the pointer cannot change
- Some more minor fixes and cleanups
* tag 'gfs2-for-7.4' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
gfs2: remove unnecessary check in gfs2_freeze_lock_shared
gfs2: fix the escaped block copy in gfs2_before_commit()
gfs2: gfs2_create_inode excl cleanup
gfs2: Fix the return value in gfs2_get_link() kernel-doc
gfs2: fix NULL deref in gfs2_quota_lock() for meta inodes
gfs2: fix typos in comments
gfs2: Fix journaled truncate range end offset
gfs2: missing whitespace in error message
gfs2: Get rid of sd_async_glock_wait
gfs2: reject invalid inode sizes
gfs2: No quotas for meta inodes
gfs2: move brelse() after buffer head accesses
gfs2: Improve resource group validation and error handling
gfs2: Silence sparse signedness warning
gfs2: annotate i_gl with __rcu
gfs2: Introduce gfs2_inode_glock
gfs2: Add inode variables
gfs2: Get to super block from inode in tracepoints
gfs2: Avoid ip->i_gl dereferences in gfs2_inode_lookup and gfs2_create_inode
|
|
Add a YNL spec for nexthop objects and nexthop buckets in rtnetlink.
./tools/net/ynl/pyynl/cli.py \
--spec Documentation/netlink/specs/rt-nexthop.yaml \
--dump getnexthop
Exclude newnexthopbucket-ntf from the C code generation. ynl_gen_c.py
looks up a classic notification by the request of the same message
type, and no request uses RTM_NEWNEXTHOPBUCKET. The spec keeps the
notification, so the Python tools can still subscribe to it.
Signed-off-by: Demos Ra <demosthenes.ra@gmail.com>
Link: https://patch.msgid.link/20261019-rt-nexthop-spec-v1-1-9c586edf86ab@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
In tcp_bpf_sendmsg(), the value returned by sk_msg_memcopy_from_iter()
is checked for error but not propagated to 'out_err' label, so error
handling may not work as expected. Fix this by assigning error value
to 'err', thus making sure that error won't be silently ignored.
Reported-by: Sashiko <shodan.ai@mcst.ru>
Fixes: 5153a75ef34b ("tcp_bpf: Fix copied value in tcp_bpf_sendmsg")
Signed-off-by: Dmitry Antipov <dmantipov@yandex.ru>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@kernel.org>
Link: https://patch.msgid.link/20261019082214.102398-1-dmantipov@yandex.ru
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Pull ext2, isofs, udf and quota updates from Jan Kara:
- ext2 changes to support clang context tracking
- multiple isofs hardenings and cleanups
- several udf hardening fixes
- block use of fs-specific quota formats with incorrect filesystem
* tag 'fs_for_v7.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs: (26 commits)
udf: fix return value for non-recorded-allocated extents in udf_map_block()
isofs: shrink the name conversion buffer
isofs: pass the name buffer size to get_rock_ridge_filename()
isofs: report the Joliet name length in statfs
isofs: return long Joliet names whole
isofs: Simplify isofs_read_level3_size()
isofs: bound empty directory blocks in isofs_read_level3_size()
quota: fix shrinker scan budget accounting in dqcache_shrink_scan()
ext2: drain in-flight DIO before buffered write fallback
isofs: validate directory records in isofs_read_level3_size()
once_lite: Simplify condition handling and fix context analysis
isofs: convert the zisofs read path to use folios
isofs: fix Rock Ridge CE extent validation on multisession media
shmem: Refuse to enable tmpfs format for another filesystem
ocfs2: Refuse to enable ocfs2 format for another filesystem
ext2: enable context analysis support for ext2 filesystem
ext2: annotate s_rsv_window_root as requiring s_rsv_window_lock
ext2: annotate block-mapping helpers as requiring truncate_mutex
ext2: annotate ext2_init_block_alloc_info() as requiring truncate_mutex
ext2: mark s_mount_state as guarded by s_lock
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Pull fsnotify updates from Jan Kara:
"A couple of smaller changes for fsnotify"
* tag 'fsnotify_for_v7.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
fanotify: report names longer than NAME_MAX
fanotify: Re-add Matt Bobrowski as a reviewer
fsnotify: fix typo "anonynous" in comment
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/leitao/linux
Pull configfs updates from Breno Leitao:
- Attribute structures can now be const (Thomas Weißschuh)
- The first selftests for the configfs userspace interface, plus a
sample subsystem that accepts symlinks to test against. They include
a race of symlink(2) against rmdir of the target, which exercises the
use-after-free fixed in v7.3-rc3.
- Rust bindings, merged from Andreas Hindborg's signed tag: fix the
data offset calculation in subsystem callbacks (marked for stable),
add the missing Send and Sync bounds, and correct a safety comment.
* tag 'configfs-7.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/leitao/linux:
samples: configfs: constify the configfs_attribute structures
configfs: Allow the registration of const struct configfs_attribute
configfs: Constify configfs_bin_attribute
configfs: Treat attribute structures as const internally
configfs: Constify is_visible/is_visible_bin in configfs_group_operations
rust: configfs: remove mutability from some field initializers
selftests/configfs: race symlink against rmdir of the target
selftests/configfs: cover symlink and unlink
samples/configfs: add a subsystem that accepts symlinks
selftests/configfs: add tests for the userspace interface
rust: configfs: fix data offset calculation for subsystem callbacks
rust: configfs: require thread-safe callback data
rust: configfs: fix release safety documentation
|
|
Pull smb client updates from Paulo Alcantara:
- Implement ->fileattr_set so chattr(1) compression changes (chattr
+c/-c) go through the generic VFS fileattr path and its permission
and serialization checks instead of the raw FS_IOC_SETFLAGS fallback
and remove the now-unreachable file flags ioctl handlers
- Require stable pages when the connection is signed or SMB compression
is enabled, so a buffered write can no longer modify a folio that is
still in flight and corrupt its signature (surfaced as -EIO) or, with
compression, silently corrupt the data stored on the server
- Stop breaking a file's just-granted lease while probing it for a
Minshall+French symlink, by reusing an already open handle or opening
with the inode's lease key
- Fix several out-of-bounds reads and use-after-free bug in SMB1
- Fix an use-after-free bug in cifs_ses_add_channel()
- Fix a NULL pointer dereference in the demultiplex thread when an
encrypted read response matches no outstanding request and reclaim
the credits granted on such responses to avoid starving the
connection of credits
- Fail sends fast when the connection needs to reconnect
- Drop the redundant nls_table argument threaded through the session
setup, tree connect and path-based call paths, allocate the session
setup state on the stack and remove the internal CIFS version number
along with a batch of write-only fields and commented-out code
* tag 'cifs-for-7.4' of https://git.manguebit.org/linux: (61 commits)
smb: client: reclaim credits for decrypted reads with no matching mid
smb: client: do not pass a NULL buf to is_oplock_break()
smb: client: require stable pages for SMB compression
smb: client: allocate struct SMB2_sess_data on the stack
smb: client: allocate struct sess_data on the stack
smb: client: drop redundant nls_table arg from CIFSSMBRenameOpenFile()
smb: client: drop redundant nls_table arg from SMBLegacyOpen()
smb: client: drop redundant nls_table arg from CIFSPOSIXCreate()
smb: client: drop redundant nls_table arg from CIFSPOSIXDelFile()
smb: client: drop redundant nls_table arg from ->tree_connect()
smb: client: drop redundant nls_table arg from cifs_setup_session()
smb: client: drop redundant nls_table arg from ->sess_setup()
smb: client: drop unused nls_table arg from decode_ascii_ssetup()
smb: client: drop redundant nls_table arg from decode_unicode_ssetup()
smb: client: drop unused nls_table arg from ascii_ssetup_strings()
smb: client: drop redundant nls_table arg from unicode_ssetup_strings()
smb: client: drop unused nls_table arg from ascii_domain_string()
smb: client: drop redundant nls_table arg from unicode_domain_string()
smb: client: drop unused nls_table arg from ascii_oslm_strings()
smb: client: drop redundant nls_table arg from build_ntlmssp_auth_blob()
...
|
|
Several tc-testing cases assert an exact bytes/packets value from a single
injected frame, but count every frame that traverses an egress qdisc (or
match with "protocol all"). Every case creates a fresh veth pair (and, for
the gso_skb cases, a netdevsim device) and brings it up with IPv6 enabled;
a device whose link-local autoconfiguration runs then emits neighbour
solicitations, MLD reports and router solicitations for a few seconds.
Those are ordinary Ethernet frames (0x86DD) and are counted too, so the
exact assertions race with the autoconf burst and fail intermittently. The
$DEV0 peer of the veth pair lives in the init netns, so its frames follow
the init-netns IPv6 sysctls.
Conditions to recreate the bug:
Run the affected cases with IPv6 enabled (the tdc default) on net-next
(7.3.0-rc6+). 4e6b fails intermittently (4/6 here); the four netdevsim
"gso_skb" cases fail intermittently too (fb6c 1/25, 0235 4/15, 476f 4/8).
1922 did not trip in the sampled runs but shares the same setup and is
fixed for uniformity. Each failure carries a stray autoconf frame, e.g.
"bytes: 98 != 188".
The Fix:
Fix the affected cases:
- actions/ife.json 4e6b injects only an 0xED3E frame but matches with
"protocol all"; scope the filter to that ethertype.
- The four netdevsim "gso_skb" cases (fb6c, 1922, 476f, 0235) create $ETH
from within the case and bring it up immediately before installing the
qdiscs, so the autoconf burst overlaps the measurement. Set
"addrgenmode none" before the link is brought up, so the device
generates no link-local address and the counter holds only the injected
packet. IPv6 stays usable (a manually added address still works), so
cases that exercise IPv6 are unaffected. On a kernel without IPv6
(CONFIG_IPV6=n, or booted with ipv6.disable=1) the attribute is not
available, so the line is made non-fatal like every neighbouring setup
line.
With this change the affected cases pass 20/20, and the whole ife.json
(51/51) and infra/qdiscs.json (48/48) suites pass, as do the protocol-ipv6
matchall cases (filters/matchall.json 0596, 41df, 8bd2, 2a4a).
Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Victor Nogueira <victor@mojatatu.com>
Link: https://patch.msgid.link/QDISC-L9J8.v1.r0.20261014162217-1@mojatatu.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Bartosz Golaszewski says:
====================
net: stmmac: qcom-ethqos: add support for SCMI power domains
Add support for the firmware-managed variant of the DesignWare MAC on
the sa8255p platform. This series contains new DT bindings and driver
changes required to support the MAC in the STMMAC driver.
It also reorganizes the ethqos code quite a bit to make the introduction
of power domains into the driver a bit easier on the eye.
The DTS changes will go in separately.
====================
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-0-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Extend the driver to support a new model - sa8255p. Unlike the previously
supported variants, this one's power management is done in the firmware
over SCMI. This is modeled in linux using power domains so add a new
emac data variant and a separate setup callback.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-7-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Ahead of adding support for firmware-controlled EMAC variants, extend
the ethqos_emac_driver_data structure with a setup() callback, implement
it for the existing models and move all operations not required in SCMI
mode into it.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Reviewed-by: Mohd Ayaan Anwar <mohd.anwar@oss.qualcomm.com>
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-6-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Instead of needlessly copying the fields of ethqos_emac_driver_data into
struct qcom_ethqos, just use the address of the former as a reference.
It's .rodata after all. This is done in order to avoid having either two
calls to of_device_get_match_data() or having to extend the latter with
another field when adding support for SCMI.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Reviewed-by: Mohd Ayaan Anwar <mohd.anwar@oss.qualcomm.com>
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-5-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
mac_finish()
ethqos->phy_mode is read by serdes_powerup() and
qcom_ethqos_set_sgmii_loopback() to configure the SerDes and its
wrapper for the correct interface. It is set once at probe from the DT
phy-mode property and never updated.
phylink resolves the actual runtime interface and passes it to the
mac_finish() callback. On platforms that only ever operate at one speed
these two values are always the same, but when the interface changes
at runtime - for example switching between SGMII and 2500BASE-X - the
stored value becomes stale. The next serdes_powerup() call, on
a subsequent link cycle, then reinitializes the SerDes to the wrong mode.
Update ethqos->phy_mode from the interface argument in
ethqos_mac_finish_serdes() so that it always reflects the last
resolved interface.
With that, we now update the phy_mode to the runtime-resolved interface,
so any subsequent resume with a 1G SGMII link skips the loopback setup
and the DMA reset hangs. To remedy this, extend the guard to include
PHY_INTERFACE_MODE_SGMII.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Reviewed-by: Mohd Ayaan Anwar <mohd.anwar@oss.qualcomm.com>
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-4-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Call phy_set_mode_ext() before phy_power_on() in
qcom_ethqos_serdes_powerup(). This is harmless for existing users but on
SCMI systems this is required for the PHY driver to select the right
performance level - which translates to the link speed. This is done
ahead of adding support for the firmware-managed EMAC on Qualcomm sa8255p.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Reviewed-by: Mohd Ayaan Anwar <mohd.anwar@oss.qualcomm.com>
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-3-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Describe the firmware-managed variant of the QCom DesignWare MAC. As the
properties here differ a lot from the HLOS-managed variant, lets put it
in a separate file. Since we need to update the maximum number of power
domains, let's update existing bindings referencing the top-level
snps,dwmac.yaml and limit their maxItems for power-domains to 1.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Reviewed-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Romain Gantois <romain.gantois@bootlin.com> # For RZ/N1
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-2-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
dwmac4_core_init() derives the TIC counter value from the rate of the CSR
clock. If no such clock is available (for instance because clocks are
managed by firmware and not described in DT), the rate is 0 and
(0 / 1000000) - 1 underflows to 0xffffffff, which is then written
to the register and can break the LPI timers that count in microseconds.
Leave the register at its reset value when the CSR clock rate is
unknown. The RX watchdog conversion helpers already cope with a missing
clock by falling back to clk_ref_rate or returning 0, so nothing else
needs to change.
This is first hit with the Qualcomm sa8255p ethernet, whose DT has no
clocks.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/all/179052656185.2160803.1751830002645522416@kernel.org/
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20261014-qcom-sa8255p-emac-v16-1-0266c2ead980@oss.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/hfs
Pull hfs/hfsplus updates from Viacheslav Dubeyko:
"Fix {hfs,hfsplus}_btree_close() (Bruno Produit):
- B-tree nodes with a zero reference count remain in the node hash
until folio reclaim or tree teardown frees them. The folio release
callbacks remove nodes while holding hash_lock, but
hfs_btree_close() walks and frees the same hash without that lock.
Reclaim can therefore unhash and free a node after close has loaded
its pointer, causing a use-after-free or double-free. Detach each
node with hfs_bnode_unhash() while holding hash_lock before
inspecting and freeing it. Drop the lock before hfs_bnode_free() so
a large tree is not freed while holding a spinlock.
Fix hfs_inode_setattr() and hfs_brec_insert() (Davy Felipe):
- hfs_inode_setattr() rejects unsupported UID, GID and mode changes,
but returns the value stored in error when the filesystem is not
mounted with the quiet option. At this point error is zero, as
execution has already passed a successful setattr_prepare() call.
As a result, unsupported attribute changes are incorrectly reported
to userspace as successful even though the inode remains unchanged.
Return -EPERM for these operations while preserving the existing
behavior for quiet mounts.
- hfs_brec_insert() may fail while inserting a new extent record, but
__hfs_ext_write_extent() currently ignores its return value and
clears HFS_FLG_EXT_DIRTY and HFS_FLG_EXT_NEW as if the insertion
had succeeded. Use a reusable B-tree node range helper that takes
the find data and the expected record size. The helper validates
the bnode and tree pointers, entry offset and entry length, and
ensures that the write range fits within the node.
Fix failure in xfstests generic/339 (Viacheslav Dubeyko):
- The test left an HFS+ volume corrupted and, as a result,
fsck.hfsplus reported a long run of "Illegal name". The root cause
is that fs/hfsplus/unicode.c's decomposition table (Apple Technote
#1150) decomposes each source character entirely on its own and
never does anything further with the result. Real fsck_hfs has
carried a permanent "FixDecomps" fixup in CatalogCheck.c for
exactly what that leaves unhandled. The fix makes hfsplus store,
hash and compare names the same way current macOS does"
* tag 'hfs-v7.4-tag1' of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/hfs:
hfs: handle extent B-tree write errors
hfsplus: add KUnit coverage for canonical reordering and legacy fixups
hfsplus: canonicalize decomposed catalog names like fsck_hfs expects
hfsplus: add Unicode combining-class and legacy decomposition data
hfs/hfsplus: serialize B-tree close against folio release
hfs: return -EPERM for unsupported setattr changes
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/nilfs2
Pull nilfs2 updates from Viacheslav Dubeyko:
"Fix issues in nilfs_sufile_get_suinfo(), nilfs_scan_dsync_log(), and
nilfs_find_or_create_root() (Aldo Ariel Panzardo):
- If the requested segment number is greater than the total number of
segments, the function may process segment numbers outside the
filesystem.
- nilfs_scan_dsync_log() trusts the block counts in each on-disk file
information entry. If fi_ndatablk is greater than fi_nblocks, the
data block loop can consume excessive summary entries and the later
subtraction used to derive the number of node blocks underflows.
- nilfs_find_or_create_root() links a new checkpoint root into the
checkpoint tree and drops ns_cptree_lock before creating the root's
sysfs group. If nilfs_sysfs_create_snapshot_group() then fails, the
root is freed while it is still reachable from ns_cptree, so a
concurrent nilfs_lookup_root() can dereference freed memory.
Fix typos in comments, reported by scripts/checkpatch.pl using the
misspelling list in scripts/spelling.txt (Hemanth Selam)
Fix issues in nilfs_recover_dsync_blocks() and nilfs_readdir() (Jake
Labelle):
- During roll-forward recovery, nilfs_recover_dsync_blocks() obtains
the inode designated by fi_ino of a dsync log without checking its
type. Regular files, directories, and symlinks are the only inode
types with data blocks to recover; reject anything else before
touching its block mapping.
- nilfs_readdir() uses ctx->pos, which userspace can set to an
arbitrary value via llseek, as an offset into the directory folio
and parses whatever bytes sit there as a directory entry. Snap
unaligned positions onto a valid entry boundary by walking the
rec_len chain from the chunk start before parsing.
Fix issue in logic of clearing the dirty folios. When garbage
collection fails, nilfs2 rolls the DAT metadata file's page cache back
to a shadow copy taken before GC started. nilfs_clear_dirty_pages()
drops the dirty state of the folios in the DAT cache, then
nilfs_copy_back_pages() overwrites them with the saved contents and
warns if one is still dirty. Add a force flag to
nilfs_clear_dirty_pages() and nilfs_clear_folio_dirty() that skips the
busy buffer check, and set it in the two calls from
nilfs_mdt_restore_from_shadow_map() (Jiaming Zhang)
Remove dead code and comment code sections in nilfs_new_inode(),
nilfs_read_inode_common(), and nilfs_setup_super() (Joshua Crofts)
Multiple fixes (Ryusuke Konishi):
- While the recent out-of-bounds memory access issue in
nilfs_direct_propagate() has already been fixed, adds extra checks
to catch and gracefully handle similar potential bugs.
- Fix syzbot reported deadlock between nilfs_evict_inode() and
find_inode() by removing the synchronous log writer trigger from
nilfs_evict_inode().
- Fix an omitted clear of the dirty flag on backing device buffers,
which could cause an abnormal buffer state and potential block
overwriting data corruption.
- Improve stability by preventing errors caused by user interrupts
while waiting for log writes, which are triggered by synchronous
write flags following operations such as rename().
- Refactor the metadata file writeback callback function,
restructuring an inefficient legacy function organization rooted in
the era of aops->writepage() to align with aops->writepages().
- Fix a link count underflow during rmdir on a directory with a
corrupted link count (an issue recently reported by syzbot).
Eliminate blockdev_direct_IO() from nilfs2 entirely. Add iomap.h
with declaration of iomap operations and iomap.c with implementation
of nilfs_iomap_begin(). The nilfs_direct_IO() and the .direct_IO
callback on nilfs_aops are removed. The nilfs_file_read_iter()
dispatches O_DIRECT reads to iomap_dio_rw() using nilfs_iomap_ops.
Everything else still goes through generic_file_read_iter()
(Viacheslav Dubeyko)"
* tag 'nilfs2-v7.4-tag1' of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/nilfs2:
nilfs2: validate directory position after llseek in readdir
nilfs2: reject non-regular inodes in dsync recovery
nilfs2: fix link count underflow when removing corrupted directory
nilfs2: simplify nilfs_mdt_writeback()
nilfs2: ignore non-fatal signals during synchronous log write wait
nilfs2: clear dirty flag on bdev buffers on log write failure
nilfs2: fix deadlock between nilfs_evict_inode() and find_inode()
nilfs2: fix checkpoint root lifetime on sysfs errors
nilfs2: switch O_DIRECT to iomap based operations
nilfs2: add iomap operations for direct I/O
nilfs2: validate file block counts during recovery
nilfs2: force clear dirty state when restoring from the shadow map
nilfs2: validate segment number in nilfs_sufile_get_suinfo()
nilfs2: fix typos in comments
nilfs2: remove '#if 0' code blocks
nilfs2: strengthen consistency checks in nilfs_direct_propagate()
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux
Pull affs updates from David Sterba:
"AFFS has attracted the AI security scanners but there's probably
little real impact.
A few fixes, more will come in the future, but it's low priority"
* tag 'affs-for-7.4' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
affs: validate blocks before freeing them
affs: split hash-lock subclass and reject self-referencing dir entry
affs: don't instantiate dentry if affs_insert_hash failed
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Pablo Neira Ayuso says:
====================
Netfilter/IPVS fixes for net
The following patchset contains Netfilter/IPVS for net:
1) Fix IPv6 nexthdr offset truncation in fragment reassembly,
this has been uncovered by LLM, from Jeremy Jean.
2) Restrict checksums in nfqueue to known protocols and offset,
after this patch the plan is to re-enable nfqueue in userns,
this came via LLM report. from Florian Westhphal.
3) Fix races when handling overload flag in IPVS, this patch
is not perfect but Julian considers this improves the situation.
This was uncovered by LLM. From Julian Anastasov.
4) Error handling for IPVS sync_threshold sysctl, from Joel Granados.
5) Use skb_vlan_tag_get_id() to retrieve the VLAN id, discard tci
bits, when handling VLAN in the flowtable, from Shuangfeng He.
6) Initialize IPCB for inet/ingress reject, uncovered by LLM.
7) Restrict completion of binary search array to anonymous set.
8) Restrict DCCP option matching to avoid out-of-bound read,
from Sung Byeongchan.
9) Reset transport header offset in flowtable after tunnel
decapsulation to fix issue with qdisc xmit path,
from Lorenzo Bianconi.
10) Fix inverted matching in ctnetlink dump filtering,
from Liu Chao.
* tag 'nf-26-10-19' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
netfilter: nf_conntrack_netlink: fix inverted IPv6 tuple filter
netfilter: flowtable: reset transport header after nf_flow_encap_pop()
netfilter: xt_dccp: handle reserved options as single-byte
netfilter: nft_set_rbtree: narrow down range completion to anonymous sets
netfilter: nf_reject: initialize IPCB at inet ingress
netfilter: flowtable: use the vlan id, not the full tci, in the tuple key
ipvs: skip committing sync_threshold when proc_dointvec fails
ipvs: fix more races around the overload flag
netfilter: nft_payload: restrict checksum offsets to known values
netfilter: nf_conntrack: avoid truncating IPv6 nexthdr offset
====================
Link: https://patch.msgid.link/20261019150249.1879516-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux
Pull btrfs updates from David Sterba:
"User visible changes:
- remove tweakable sysfs way to enable features, used only for
extent-irefs which has been default since 3.12; the sysfs interface
makes it complicated due to locking and context, unlike ioctl
- change which backup superblock does mount option
'rescue=usebackuproot' choose, skip the last one due to rotation,
for more fine grained control add options selecting a specific
backup superblock by index (usebackuproot_N)
Notable fixes:
- syslog message fixup of error state description, duplicate 'E'
renamed to 'H' (emergency shutdown), added 'I 'when delayed iputs
are blocked
- use kvmalloc for b-tree buffer allocations, as we're using more
contig ranges because of large folios, the allocations started to
warn on fragmented memory
- handle unknown commands for the /dev/btrfs-control ioctls
- tune timing of transaction commit when the space is getting low
- in raid56, skip verifications if data checksums are not available
Performance improvements:
- reflink speedup, reorder locking and flushing to optimize for more
common cases where the reflinked ranges are accessed, avoid
unnecessary contention
- minor speedup in zstd decompression (+8MB/s)
- reduced contention of inode extent map lock, improves e.g. when
shrinker processes inodes without extent maps, measured time
reduction 10-20%
Core changes:
- move allocation errors out of some contexts (new
inode/dentry/rename) to avoid transaction aborts
- allocate system new chunks earlier, e.g. when consumed by
fallocate, and tune deleting unused block groups if the system
space is low
- tree checker: reject extents for special files
- many qgroup fixes regarding enabling/disabling, other state changes
and qgroup relation updates
- error handling fixes
- refactoring, cleanups:
- page to folio cleanups
- transaction state macros/enums simplification
- parameter simplifications"
* tag 'for-7.4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (92 commits)
btrfs: fix use-after-free on quota enable allocation failure
btrfs: prevent use-after-free in btrfs_transaction::switch_commits
btrfs: return -ENOTTY for unknown ioctls on the control device
btrfs: defrag: fix check to see if next extent can be defragged too
btrfs: defrag: fix key comparison with max defrag key in btrfs_defrag_leaves()
btrfs: use READ_ONCE() to read the chunk size from the data space_info
btrfs: qgroup: make btrfs_record_squota_delta() return void
btrfs: fix lost error return value in btrfs_alloc_logged_file_extent()
btrfs: qgroup: fix off-by-one max level check in qgroup_trace_new_subtree_blocks()
btrfs: abort transaction on qgroup failures in create_pending_snapshot()
btrfs: qgroup: merge error labels in btrfs_quota_enable()
btrfs: qgroup: fix leak of qgroups in rb tree after failure to enable quotas
btrfs: qgroup: abort transaction on failure to add qgroup relation
btrfs: qgroup: fix swapped blocks existence check when tracing after COW
btrfs: introduce more accurate usebackuproot options
btrfs: always use the second newest slot for rescue=usebackuproot
btrfs: raid56: do not verify the content if there is no data checksum
btrfs: qgroup: do not leak -ENOENT from btrfs_remove_qgroup()
btrfs: qgroup: use qgroup_mark_inconsistent() in quick_update_accounting()
btrfs: qgroup: do not treat "ret > 0" as error when deleting a qgroup
...
|
|
Commit 68822bdf76f1 ("net: generalize skb freeing deferral to
per-cpu lists") added alloc_cpu to struct sk_buff and placed it in the
headers group copied by __copy_skb_header().
This is wrong for newly allocated skbs. For example, skb_segment()
allocates a new skb and records the current CPU in alloc_cpu, but
__copy_skb_header() overwrites it with the value from the original skb.
If the original skb was allocated on another CPU, the segment may later
be treated as remotely allocated and unnecessarily deferred there when
it is freed.
Move alloc_cpu out of the headers group so it is preserved across
header copies. Copy it explicitly in skb_clone(), which allocates
the skb object directly instead of going through __alloc_skb().
This was discovered while investigating a throughput regression with
software IPsec. The load used 512 IPsec tunnels carrying traffic over
VXLAN, with 512 parallel iperf3 flows and an MTU of 4500.
Before the fix:
16.09% queued_spin_lock_slowpath
|
--15.90%--_raw_spin_lock_irqsave
|
--9.37%--__slab_free
kfree
skb_release_data
napi_consume_skb
mlx5e_poll_tx_cq
After the fix:
6.84% queued_spin_lock_slowpath
|
--6.57%--_raw_spin_lock_irqsave
|
--5.00%--__slab_free
kfree
skb_release_data
napi_consume_skb
mlx5e_poll_tx_cq
Fixes: 68822bdf76f1 ("net: generalize skb freeing deferral to per-cpu lists")
Signed-off-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Reviewed-by: Björn Töpel <bjorn@kernel.org>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20261019174152.2932829-1-cjubran@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Pull fsverity updates from Eric Biggers:
"Patches from Andrey Albershteyn to prepare for XFS fsverity support:
- Report verification errors through fserror, instead of only through
an unstructured log message and generic I/O error return
- Move the reporting of STATX_ATTR_VERITY into the VFS, so that each
fsverity-capable filesystem doesn't need to do it. This also fixes
a bug where btrfs didn't return STATX_ATTR_VERITY in
attributes_mask
- Make the VFS prevent enabling DAX on fsverity files, so that it
will be handled in a single place for ext4 and XFS
- Move ext4 and f2fs's code to read verity metadata from the
pagecache into shared code, so that XFS will be able to use it too
- Export fsverity_ensure_verity_info() for XFS scrub to use
- Pass all-zeroes hash to ->write_merkle_tree_block() so that XFS
will be able to implement an optimization for sparse files"
* tag 'fsverity-for-linus' of git://git.kernel.org/pub/scm/fs/fsverity/linux:
fsverity: hoist statx reporting of fs-verity flag
fsverity: don't allow setting DAX file attribute on fsverity files
fsverity: hoist pagecache_read from f2fs/ext4 to fsverity
fsverity: pass digest size and hash of the all-zeroes block to ->write
fsverity: expose ensure_fsverity_info()
fsverity: report validation errors through fserror to fsnotify
MAINTAINERS: List the fsverity branches
|
|
Pull fscrypt updates from Eric Biggers:
"A few minor cleanups for fs/crypto/"
* tag 'fscrypt-for-linus' of git://git.kernel.org/pub/scm/fs/fscrypt/linux:
MAINTAINERS: List the fscrypt branches
fscrypt: Move several items from fscrypt_private.h to .c files
fscrypt: Use aes_ecb_encrypt() for v1 key derivation
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux
Pull crypto library updates from Eric Biggers:
- Migrate the x86 and riscv optimized implementations of various AES
modes (ECB, CBC, CBC-CTS, CTR, XCTR, XTS) into lib/crypto/.
This makes the library APIs be properly accelerated on these
architectures, while still accelerating crypto_skcipher as well.
It removes a lot of redundant glue code, since the crypto API
boilerplate no longer needs to be duplicated for these architectures.
It also fixes the longstanding issue where optimizations for these
AES modes were disabled by default on these architectures.
- Optimize AES-XTS on SPARC64 using the SPARC crypto opcodes
- Move the SipHash code into lib/crypto/
- Update the ChaCha and Poly1305 functions to take size_t lengths,
resolving issues where lengths over INT_MAX were mishandled
- Make frame pointer unwinding work properly in some of the x86
assembly functions, by following the standard %rbp convention
- Add helper functions for zeroizing structs
- Make a few documentation improvements
* tag 'libcrypto-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux: (49 commits)
lib/crypto: md5: remove unused header
lib/crypto: mips/chacha: Handle lengths of 2 GiB or more
lib/crypto: arm/chacha: Handle lengths of 2 GiB or more in scalar code
lib/crypto: chacha: Use size_t length
lib/crypto: poly1305: Use size_t length
lib/crypto: x86/sha512: Follow the normal %rbp convention
lib/crypto: x86/sha256: Follow the normal %rbp convention
lib/crypto: x86/sha1: Follow the normal %rbp convention
lib/crypto: chacha20poly1305: zeroize state in __chacha20poly1305_decrypt
lib/crypto: sparc/aes-xts: Add optimization using the AES opcodes
lib: Move SipHash into lib/crypto/
lib/crypto: x86/aes-ctr: Remove some unreachable code
crypto: x86/aesni-intel - Add transitional selections
crypto: aes - Boost priority of encryption modes further
lib/crypto: x86/aes: Set RNDKEY in aes_cbc_encrypt_aesni()
lib/crypto: riscv/aes-xts: Migrate optimized code into library
lib/crypto: riscv/aes-ctr: Migrate optimized code into library
lib/crypto: riscv/aes-cbc: Migrate optimized code into library
lib/crypto: riscv/aes-ecb: Migrate optimized code into library
lib/crypto: riscv/aes: Pass key struct to assembly code
...
|