kernel/linux.git/fs/nfs/dir.c, branch v6.19.11

fs/nfs: Fix readdir slow-start regression

2026-02-26T23:01:19+00:00

[ Upstream commit 42e7c876b182da65723700f6bc507a8aecb10d3b ] Commit 580f236737d1 ("NFS: Adjust the amount of readahead performed by NFS readdir") reduces the amount of readahead names caching done by the client. The downside of this approach is READDIR now may suffer from a slow-start issue, where initially it will fetch names that fit in a single page, then in 2, 4, 8 until the maximum supported transfer size (usually 1M). This patch tries to take a balanced approach between mitigating the slow-start issue still maintaining some efficiency gains. Fixes: 580f236737d1 ("NFS: Adjust the amount of readahead performed by NFS readdir") Signed-off-by: Sagi Grimberg Signed-off-by: Anna Schumaker Signed-off-by: Sasha Levin

Merge tag 'vfs-6.19-rc8.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs

2026-01-26T17:30:48+00:00

Pull vfs fixes from Christian Brauner: - Fix the the buggy conversion of fuse_reverse_inval_entry() introduced during the creation rework - Disallow nfs delegation requests for directories by setting simple_nosetlease() - Require an opt-in for getting readdir flag bits outside of S_DT_MASK set in d_type - Fix scheduling delayed writeback work by only scheduling when the dirty time expiry interval is non-zero and cancel the delayed work if the interval is set to zero - Use rounded_jiffies_interval for dirty time work - Check the return value of sb_set_blocksize() for romfs - Wait for batched folios to be stable in __iomap_get_folio() - Use private naming for fuse hash size - Fix the stale dentry cleanup to prevent a race that causes a UAF * tag 'vfs-6.19-rc8.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: vfs: document d_dispose_if_unused() fuse: shrink once after all buckets have been scanned fuse: clean up fuse_dentry_tree_work() fuse: add need_resched() before unlocking bucket fuse: make sure dentry is evicted if stale fuse: fix race when disposing stale dentries fuse: use private naming for fuse hash size writeback: use round_jiffies_relative for dirtytime_work iomap: wait for batched folios to be stable in __iomap_get_folio romfs: check sb_set_blocksize() return value docs: clarify that dirtytime_expire_seconds=0 disables writeback writeback: fix 100% CPU usage when dirtytime_expire_interval is 0 readdir: require opt-in for d_type flags vboxsf: don't allow delegations to be set on directories ceph: don't allow delegations to be set on directories gfs2: don't allow delegations to be set on directories 9p: don't allow delegations to be set on directories smb/client: properly disallow delegations on directories nfs: properly disallow delegation requests on directories fuse: fix conversion of fuse_reverse_inval_entry() to start_removing()

nfs: properly disallow delegation requests on directories

2026-01-12T09:54:46+00:00

Checking for S_ISREG() in nfs4_setlease() is incorrect, since that op is never called for directories. The right way to deny lease requests on directories is to set the ->setlease() operation to simple_nosetlease() in the directory file_operations. Fixes: e6d28ebc17eb ("filelock: push the S_ISREG check down to ->setlease handlers") Reported-by: Christoph Hellwig Closes: https://lore.kernel.org/linux-fsdevel/aV316LhsVSl0n9-E@infradead.org/ Signed-off-by: Jeff Layton Link: https://patch.msgid.link/20260107-setlease-6-19-v1-1-85f034abcc57@kernel.org Tested-by: Christoph Hellwig Signed-off-by: Christian Brauner

NFSv4: Fix nfs_clear_verifier_delegated() for delegated directories

2026-01-05T04:03:26+00:00

If the client returns a directory delegation, then look up all the child dentries, and clear their 'verifier delegated' bit, unless subject to a file delegation. Similarly, if a file delegation is being returned, check if there is a directory delegation before clearing a 'verifier delegated' bit. Reported-by: Christoph Hellwig Fixes: 156b09482933 ("NFS: Request a directory delegation on ACCESS, CREATE, and UNLINK") Signed-off-by: Trond Myklebust

NFS: Fix directory delegation verifier checks

2026-01-05T04:03:25+00:00

Doing this check in nfs_check_verifier() resulted in many, many more lookups on the wire when running Christoph's delegation benchmarking script. After some experimentation, I found that we can treat directory delegations exactly the same as having a delegated verifier when we reach nfs4_lookup_revalidate() for the best performance. Reported-by: Christoph Hellwig Fixes: 156b09482933 ("NFS: Request a directory delegation on ACCESS, CREATE, and UNLINK") Signed-off-by: Anna Schumaker Signed-off-by: Trond Myklebust

NFS: Shortcut lookup revalidations if we have a directory delegation

2025-11-23T21:01:47+00:00

Holding a directory delegation means we know that nobody else has modified the directory on the server, so we can take a few revalidation shortcuts. Signed-off-by: Anna Schumaker Signed-off-by: Trond Myklebust

NFS: Initialise verifiers for visible dentries in nfs_atomic_open()

2025-11-19T13:46:21+00:00

Ensure that the verifiers are initialised before calling d_splice_alias() in nfs_atomic_open(). Reported-by: Michael Stoler Fixes: 809fd143de88 ("NFSv4: Ensure nfs_atomic_open set the dentry verifier on ENOENT") Signed-off-by: Trond Myklebust

NFS: Initialise verifiers for visible dentries in readdir and lookup

2025-11-19T13:45:53+00:00

Ensure that the verifiers are initialised before calling d_splice_alias() in both nfs_prime_dcache() and nfs_lookup(). Reported-by: Michael Stoler Fixes: a1147b8281bd ("NFS: Fix up directory verifier races") Signed-off-by: Trond Myklebust

NFS: Avoid changing nlink when file removes and attribute updates race

2025-11-17T22:31:32+00:00

If a file removal races with another operation that updates its attributes, then skip the change to nlink, and just mark the attributes as being stale. Reported-by: Aiden Lambert Fixes: 59a707b0d42e ("NFS: Ensure we revalidate the inode correctly after remove or rename") Signed-off-by: Trond Myklebust

NFSv2/v3: Fix error handling in nfs_atomic_open_v23()

2025-11-10T19:30:45+00:00

When nfs_do_create() returns an EEXIST error, it means that a regular file could not be created. That could mean that a symlink needs to be resolved. If that's the case, a lookup needs to be kicked off. Reported-by: Stephen Abbene Link: https://bugzilla.kernel.org/show_bug.cgi?id=220710 Fixes: 7c6c5249f061 ("NFS: add atomic_open for NFSv3 to handle O_TRUNC correctly.") Signed-off-by: Trond Myklebust Reviewed-by: NeilBrown Signed-off-by: Anna Schumaker