kernel/linux.git/fs/nfsd/filecache.c, branch linux-6.0.y

nfsd: fix net-namespace logic in __nfsd_file_cache_purge

2022-11-10T17:17:18+00:00

[ Upstream commit d3aefd2b29ff5ffdeb5c06a7d3191a027a18cdb8 ] If the namespace doesn't match the one in "net", then we'll continue, but that doesn't cause another rhashtable_walk_next call, so it will loop infinitely. Fixes: ce502f81ba88 ("NFSD: Convert the filecache to use rhashtable") Reported-by: Petr Vorel Link: https://lore.kernel.org/ltp/Y1%2FP8gDAcWC%2F+VR3@pevik/ Signed-off-by: Jeff Layton Signed-off-by: Chuck Lever Signed-off-by: Sasha Levin

nfsd: fix nfsd_file_unhash_and_dispose

2022-11-10T17:17:18+00:00

[ Upstream commit 8d0d254b15cc5b7d46d85fb7ab8ecede9575e672 ] nfsd_file_unhash_and_dispose() is called for two reasons: We're either shutting down and purging the filecache, or we've gotten a notification about a file delete, so we want to go ahead and unhash it so that it'll get cleaned up when we close. We're either walking the hashtable or doing a lookup in it and we don't take a reference in either case. What we want to do in both cases is to try and unhash the object and put it on the dispose list if that was successful. If it's no longer hashed, then we don't want to touch it, with the assumption being that something else is already cleaning up the sentinel reference. Instead of trying to selectively decrement the refcount in this function, just unhash it, and if that was successful, move it to the dispose list. Then, the disposal routine will just clean that up as usual. Also, just make this a void function, drop the WARN_ON_ONCE, and the comments about deadlocking since the nature of the purported deadlock is no longer clear. Signed-off-by: Jeff Layton Signed-off-by: Chuck Lever Stable-dep-of: d3aefd2b29ff ("nfsd: fix net-namespace logic in __nfsd_file_cache_purge") Signed-off-by: Sasha Levin

Merge tag 'nfsd-6.0' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux

2022-08-09T21:56:49+00:00

Pull nfsd updates from Chuck Lever: "Work on 'courteous server', which was introduced in 5.19, continues apace. This release introduces a more flexible limit on the number of NFSv4 clients that NFSD allows, now that NFSv4 clients can remain in courtesy state long after the lease expiration timeout. The client limit is adjusted based on the physical memory size of the server. The NFSD filecache is a cache of files held open by NFSv4 clients or recently touched by NFSv2 or NFSv3 clients. This cache had some significant scalability constraints that have been relieved in this release. Thanks to all who contributed to this work. A data corruption bug found during the most recent NFS bake-a-thon that involves NFSv3 and NFSv4 clients writing the same file has been addressed in this release. This release includes several improvements in CPU scalability for NFSv4 operations. In addition, Neil Brown provided patches that simplify locking during file lookup, creation, rename, and removal that enables subsequent work on making these operations more scalable. We expect to see that work materialize in the next release. There are also numerous single-patch fixes, clean-ups, and the usual improvements in observability" * tag 'nfsd-6.0' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux: (78 commits) lockd: detect and reject lock arguments that overflow NFSD: discard fh_locked flag and fh_lock/fh_unlock NFSD: use (un)lock_inode instead of fh_(un)lock for file operations NFSD: use explicit lock/unlock for directory ops NFSD: reduce locking in nfsd_lookup() NFSD: only call fh_unlock() once in nfsd_link() NFSD: always drop directory lock in nfsd_unlink() NFSD: change nfsd_create()/nfsd_symlink() to unlock directory before returning. NFSD: add posix ACLs to struct nfsd_attrs NFSD: add security label to struct nfsd_attrs NFSD: set attributes when creating symlinks NFSD: introduce struct nfsd_attrs NFSD: verify the opened dentry after setting a delegation NFSD: drop fh argument from alloc_init_deleg NFSD: Move copy offload callback arguments into a separate structure NFSD: Add nfsd4_send_cb_offload() NFSD: Remove kmalloc from nfsd4_do_async_copy() NFSD: Refactor nfsd4_do_copy() NFSD: Refactor nfsd4_cleanup_inter_ssc() (2/2) NFSD: Refactor nfsd4_cleanup_inter_ssc() (1/2) ...

NFSD: Ensure nf_inode is never dereferenced

2022-07-30T00:16:56+00:00

The documenting comment for struct nf_file states: /* * A representation of a file that has been opened by knfsd. These are hashed * in the hashtable by inode pointer value. Note that this object doesn't * hold a reference to the inode by itself, so the nf_inode pointer should * never be dereferenced, only used for comparison. */ Replace the two existing dereferences to make the comment always true. Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever

NFSD: NFSv4 CLOSE should release an nfsd_file immediately

2022-07-30T00:16:42+00:00

The last close of a file should enable other accessors to open and use that file immediately. Leaving the file open in the filecache prevents other users from accessing that file until the filecache garbage-collects the file -- sometimes that takes several seconds. Reported-by: Wang Yugui Link: https://bugzilla.linux-nfs.org/show_bug.cgi?387 Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever

NFSD: Move nfsd_file_trace_alloc() tracepoint

2022-07-30T00:16:07+00:00

Avoid recording the allocation of an nfsd_file item that is immediately released because a matching item was already inserted in the hash. Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever

NFSD: Separate tracepoints for acquire and create

2022-07-30T00:15:54+00:00

These tracepoints collect different information: the create case does not open a file, so there's no nf_file available. Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever

NFSD: Clean up unused code after rhashtable conversion

2022-07-30T00:15:38+00:00

Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever

NFSD: Convert the filecache to use rhashtable

2022-07-30T00:14:25+00:00

Enable the filecache hash table to start small, then grow with the workload. Smaller server deployments benefit because there should be lower memory utilization. Larger server deployments should see improved scaling with the number of open files. Suggested-by: Jeff Layton Suggested-by: Dave Chinner Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever

NFSD: Set up an rhashtable for the filecache

2022-07-30T00:12:02+00:00

Add code to initialize and tear down an rhashtable. The rhashtable is not used yet. Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever