summaryrefslogtreecommitdiff
path: root/fs
AgeCommit message (Collapse)Author
2016-05-12hfs: switch to ->iterate_shared()Al Viro
exact parallel of hfsplus analogue Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-12hfsplus: switch to ->iterate_shared()Al Viro
We need to protect the list of hfsplus_readdir_data against parallel insertions (in readdir) and removals (in release). Add a spinlock for that. Note that it has nothing to do with protection of hfsplus_readdir_data->key - we have an exclusion between hfsplus_readdir() and hfsplus_delete_cat() on directory lock and between several hfsplus_readdir() for the same struct file on ->f_pos_lock. The spinlock is strictly for list changes. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-12hostfs: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-12hpfs: switch to ->iterate_shared()Al Viro
NOTE: the only reason we can do that without ->i_rdir_offs races is that hpfs_lock() serializes everything in there anyway. It's not that hard to get rid of, but not as part of this series... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-12hpfs: handle allocation failures in hpfs_add_pos()Al Viro
pr_err() is nice, but we'd better propagate the error to caller and not proceed to violate the invariants (namely, "every file with f_pos tied to directory block should have its address visible in per-inode array"). Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-12gfs2: switch to ->iterate_shared()Al Viro
protected by glock and already used without locking the directory by gfs2_get_name() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-10f2fs: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-10afs: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-10befs: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-10befs: constify stuff a bitAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09isofs: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09get_acorn_filename(): deobfuscate a bitAl Viro
Lots of Idiotic Silly Parentheses is -> that way... What that condition checks is that there's exactly 32 bytes between the end of name and the end of entire drectory record. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09btrfs: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09logfs: no need to lock directory in lseekAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09switch ecryptfs to ->iterate_sharedAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09Merge branch 'for-linus' into work.lookupsAl Viro
2016-05-099p: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09fat: switch to ->iterate_shared()Al Viro
... and make that weird ioctl lock directory only shared. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09romfs, squashfs: switch to ->iterate_shared()Al Viro
don't need to lock directory in ->llseek(), either Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09more trivial ->iterate_shared conversionsAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09kernfs: no point locking directory around that generic_file_llseek()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09configfs_readdir(): make safe under shared lockAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-09nfs: per-name sillyunlink exclusionAl Viro
use d_alloc_parallel() for sillyunlink/lookup exclusion and explicit rwsem (nfs_rmdir() being a writer and nfs_call_unlink() - a reader) for rmdir/sillyunlink one. That ought to make lookup/readdir/!O_CREAT atomic_open really parallel on NFS. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-07get_rock_ridge_filename(): handle malformed NM entriesAl Viro
Payloads of NM entries are not supposed to contain NUL. When we run into such, only the part prior to the first NUL goes into the concatenation (i.e. the directory entry name being encoded by a bunch of NM entries). We do stop when the amount collected so far + the claimed amount in the current NM entry exceed 254. So far, so good, but what we return as the total length is the sum of *claimed* sizes, not the actual amount collected. And that can grow pretty large - not unlimited, since you'd need to put CE entries in between to be able to get more than the maximum that could be contained in one isofs directory entry / continuation chunk and we are stop once we'd encountered 32 CEs, but you can get about 8Kb easily. And that's what will be passed to readdir callback as the name length. 8Kb __copy_to_user() from a buffer allocated by __get_free_page() Cc: stable@vger.kernel.org # 0.98pl6+ (yes, really) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-04ecryptfs: fix handling of directory openingAl Viro
First of all, trying to open them r/w is idiocy; it's guaranteed to fail. Moreover, assigning ->f_pos and assuming that everything will work is blatantly broken - try that with e.g. tmpfs as underlying layer and watch the fireworks. There may be a non-trivial amount of state associated with current IO position, well beyond the numeric offset. Using the single struct file associated with underlying inode is really not a good idea; we ought to open one for each ecryptfs directory struct file. Additionally, file_operations both for directories and non-directories are full of pointless methods; non-directories should *not* have ->iterate(), directories should not have ->flush(), ->fasync() and ->splice_read(). Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02nfs: switch to ->iterate_shared()Al Viro
aside of the usual care about seeding dcache from readdir, we need to be careful about the pagecache evictions here. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02lookup_open(): lock the parent shared unless O_CREAT is givenAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02lookup_open(): put the dentry fed to ->lookup() or ->atomic_open() into ↵Al Viro
in-lookup hash Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02lookup_open(): expand the call of real_lookup()Al Viro
... and lose the duplicate IS_DEADDIR() - we'd already checked that. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02atomic_open(): reorder and clean up a bitAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02lookup_open(): lift the "fallback to !O_CREAT" logics from atomic_open()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02atomic_open(): be paranoid about may_open() return valueAl Viro
It should never return positives; however, with Linux S&M crowd involved, no bogosity is impossible. Results would be unpleasant... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02atomic_open(): delay open_to_namei_flags() until the method callAl Viro
nobody else needs that transformation. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02do_last(): take fput() on error after opening to out:Al Viro
make it conditional on *opened & FILE_OPENED; in addition to getting rid of exit_fput: thing, it simplifies atomic_open() cleanup on may_open() failure. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02do_last(): get rid of duplicate ELOOP checkAl Viro
may_open() will catch it Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02atomic_open(): massage the create_error logics a bitAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02atomic_open(): consolidate "overridden ENOENT" in open-yourself casesAl Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02atomic_open(): don't bother with EEXIST check - it's done in do_last()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02Merge branch 'for-linus' into work.lookupsAl Viro
2016-05-02lookup_open(): expand the call of vfs_create()Al Viro
Lift IS_DEADDIR handling up into the part common with atomic_open(), remove it from the latter. Collapse permission checks into the call of may_o_create(), getting it closer to atomic_open() case. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02path_openat(): take O_PATH handling out of do_last()Al Viro
do_last() and lookup_open() simpler that way and so does O_PATH itself. As it bloody well should: we find what the pathname resolves to, same way as in stat() et.al. and associate it with FMODE_PATH struct file. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02simple local filesystems: switch to ->iterate_shared()Al Viro
no changes needed (XFS isn't simple, but it has the same parallelism in the interesting parts exercised from CXFS). Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02dcache_{readdir,dir_lseek}() users: switch to ->iterate_sharedAl Viro
no need to lock directory in dcache_dir_lseek(), while we are at it - per-struct file exclusion is enough. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02cifs: switch to ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02fuse: switch to ->iterate_shared()Al Viro
Switch dcache pre-seeding on readdir to d_alloc_parallel(); nothing else is needed. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02switch all procfs directories ->iterate_shared()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02proc_sys_fill_cache(): switch to d_alloc_parallel()Al Viro
make it usable with directory locked shared Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02proc_fill_cache(): switch to d_alloc_parallel()Al Viro
... making it usable with directory locked shared Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02introduce a parallel variant of ->iterate()Al Viro
New method: ->iterate_shared(). Same arguments as in ->iterate(), called with the directory locked only shared. Once all filesystems switch, the old one will be gone. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-05-02give readdir(2)/getdents(2)/etc. uniform exclusion with lseek()Al Viro
same as read() on regular files has, and for the same reason. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
generated by cgit v1.2.3-58-ga151 (git 2.39.0) at 2024-12-27 12:56:55 +0000