kernel.git/fs/nfs/dir.c, branch linux-4.12.y

NFS: only invalidate dentrys that are clearly invalid.

2017-07-27T22:10:20Z

commit cc89684c9a265828ce061037f1f79f4a68ccd3f7 upstream. Since commit bafc9b754f75 ("vfs: More precise tests in d_invalidate") in v3.18, a return of '0' from ->d_revalidate() will cause the dentry to be invalidated even if it has filesystems mounted on or it or on a descendant. The mounted filesystem is unmounted. This means we need to be careful not to return 0 unless the directory referred to truly is invalid. So -ESTALE or -ENOENT should invalidate the directory. Other errors such a -EPERM or -ERESTARTSYS should be returned from ->d_revalidate() so they are propagated to the caller. A particular problem can be demonstrated by: 1/ mount an NFS filesystem using NFSv3 on /mnt 2/ mount any other filesystem on /mnt/foo 3/ ls /mnt/foo 4/ turn off network, or otherwise make the server unable to respond 5/ ls /mnt/foo & 6/ cat /proc/$!/stack # note that nfs_lookup_revalidate is in the call stack 7/ kill -9 $! # this results in -ERESTARTSYS being returned 8/ observe that /mnt/foo has been unmounted. This patch changes nfs_lookup_revalidate() to only treat -ESTALE from nfs_lookup_verify_inode() and -ESTALE or -ENOENT from ->lookup() as indicating an invalid inode. Other errors are returned. Also nfs_check_inode_attributes() is changed to return -ESTALE rather than -EIO. This is consistent with the error returned in similar circumstances from nfs_update_inode(). As this bug allows any user to unmount a filesystem mounted on an NFS filesystem, this fix is suitable for stable kernels. Fixes: bafc9b754f75 ("vfs: More precise tests in d_invalidate") Signed-off-by: NeilBrown Signed-off-by: Anna Schumaker Signed-off-by: Greg Kroah-Hartman

Revert "NFS: nfs_rename() handle -ERESTARTSYS dentry left behind"

2017-06-28T01:58:14Z

This reverts commit 920b4530fb80430ff30ef83efe21ba1fa5623731 which could call d_move() without holding the directory's i_mutex, and reverts commit d4ea7e3c5c0e341c15b073016dbf3ab6c65f12f3 "NFS: Fix old dentry rehash after move", which was a follow-up fix. Signed-off-by: Benjamin Coddington Fixes: 920b4530fb80 ("NFS: nfs_rename() handle -ERESTARTSYS dentry left behind") Cc: stable@vger.kernel.org # v4.10+ Reviewed-by: Jeff Layton Signed-off-by: Trond Myklebust

nfs: use kmap/kunmap directly

2017-05-05T17:01:33Z

This patch removes useless nfs_readdir_get_array() and nfs_readdir_release_array() as suggested by Trond Myklebust nfs_readdir() calls nfs_revalidate_mapping() before readdir_search_pagecache() , nfs_do_filldir(), uncached_readdir() so mapping should be correct. While kmap() can't fail, all subsequent error checks were removed as well as unused labels. Signed-off-by: Fabian Frederick Signed-off-by: Trond Myklebust

NFS: switch back to to ->iterate()

2017-04-20T17:33:09Z

NFS has some optimizations for readdir to choose between using READDIR or READDIRPLUS based on workload, and which NFS operation to use is determined by subsequent interactions with lookup, d_revalidate, and getattr. Concurrent use of nfs_readdir() via ->iterate_shared() can cause those optimizations to repeatedly invalidate the pagecache used to store directory entries during readdir(), which causes some very bad performance for directories with many entries (more than about 10000). There's a couple ways to fix this in NFS, but no fix would be as simple as going back to ->iterate() to serialize nfs_readdir(), and neither fix I tested performed as well as going back to ->iterate(). The first required taking the directory's i_lock for each entry, with the result of terrible contention. The second way adds another flag to the nfs_inode, and so keeps the optimizations working for large directories. The difference from using ->iterate() here is that much more memory is consumed for a given workload without any performance gain. The workings of nfs_readdir() are such that concurrent users are serialized within read_cache_page() waiting to retrieve pages of entries from the server. By serializing this work in iterate_dir() instead, contention for cache pages is reduced. Waiting processes can have an uncontended pass at the entirety of the directory's pagecache once previous processes have completed filling it. v2 - Keep the bits needed for parallel lookup Signed-off-by: Benjamin Coddington Signed-off-by: Trond Myklebust

NFS: Fix old dentry rehash after move

2017-03-27T17:30:49Z

Now that nfs_rename()'s d_move has moved within the RPC task's rpc_call_done callback, rehashing new_dentry will actually rehash the old dentry's name in nfs_rename(). d_move() is going to rehash the new dentry for us anyway, so doing it again here is unnecessary. Reported-by: Chuck Lever Fixes: 920b4530fb80 ("NFS: nfs_rename() handle -ERESTARTSYS dentry left behind") Signed-off-by: Benjamin Coddington Tested-by: Chuck Lever Signed-off-by: Anna Schumaker

NFS: nfs_rename() handle -ERESTARTSYS dentry left behind

2017-02-08T22:02:44Z

An interrupted rename will leave the old dentry behind if the rename succeeds. Fix this by moving the final local work of the rename to rpc_call_done so that the results of the RENAME can always be handled, even if the original process has already returned with -ERESTARTSYS. Signed-off-by: Benjamin Coddington Signed-off-by: Anna Schumaker

NFS: Fix and clean up the access cache validity checking

2016-12-19T22:29:39Z

The access cache needs to check whether or not the mode bits, ownership, or ACL has changed or the cache has timed out. Signed-off-by: Trond Myklebust

NFS: Only look at the change attribute cache state in nfs_weak_revalidate()

2016-12-19T22:29:37Z

Just like in nfs_check_verifier(), we want to use nfs_mapping_need_revalidate_inode() to check our knowledge of the change attribute is up to date. Signed-off-by: Trond Myklebust

nfs: add support for the umask attribute

2016-12-10T04:47:10Z

Clients can set the umask attribute when creating files to cause the server to apply it always except when inheriting permissions from the parent directory. That way, the new files will end up with the same permissions as files created locally. See https://tools.ietf.org/html/draft-ietf-nfsv4-umask-02 for more details. Signed-off-by: Andreas Gruenbacher Signed-off-by: J. Bruce Fields Signed-off-by: Trond Myklebust

NFS: Only look at the change attribute cache state in nfs_check_verifier

2016-12-04T23:34:34Z

When looking at whether or not our dcache is valid, we really don't care about the general state of the directory attribute cache. Instead, we we only care about the state of the change attribute. This fixes a performance issue when the client is responsible for changing the directory contents; a number of NFSv4 operations will atomically update the directory change attribute, but may not return all the other attributes. Signed-off-by: Trond Myklebust