diff options
| author | Trond Myklebust <Trond.Myklebust@netapp.com> | 2010-09-24 22:49:43 (GMT) |
|---|---|---|
| committer | Trond Myklebust <Trond.Myklebust@netapp.com> | 2010-10-23 19:27:30 (GMT) |
| commit | baf57a09e9d87b14be5e2788828169394a2525ab (patch) | |
| tree | 77b35fe3ef7868f3c642093a4a8c01ed6be584cd /fs | |
| parent | d1bacf9eb2fd0e7ef870acf84b9e3b157dcfa7dc (diff) | |
| download | linux-baf57a09e9d87b14be5e2788828169394a2525ab.tar.xz | |
NFS: Optimise the readdir searches
If we're going through the loop in nfs_readdir() more than once, we usually
do not want to restart searching from the beginning of the pages cache.
We only want to do that if the previous search failed...
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Diffstat (limited to 'fs')
| -rw-r--r-- | fs/nfs/dir.c | 6 |
1 files changed, 2 insertions, 4 deletions
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index f772439..33b0ce7 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -491,10 +491,6 @@ static inline int readdir_search_pagecache(nfs_readdir_descriptor_t *desc) { int res = -EAGAIN; - desc->page_index = 0; - - if (*desc->dir_cookie == 0) - desc->cache_entry_index = 0; while (1) { res = find_cache_page(desc); @@ -589,6 +585,7 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent, goto out_release; } + desc->page_index = 0; desc->page = page; status = nfs_do_filldir(desc, dirent, filldir); @@ -653,6 +650,7 @@ static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir) if (res == -ETOOSMALL && desc->plus) { clear_bit(NFS_INO_ADVISE_RDPLUS, &NFS_I(inode)->flags); nfs_zap_caches(inode); + desc->page_index = 0; desc->plus = 0; desc->eof = 0; continue; |