diff options
| author | Trond Myklebust <Trond.Myklebust@netapp.com> | 2012-05-09 22:30:35 +0400 |
|---|---|---|
| committer | Trond Myklebust <Trond.Myklebust@netapp.com> | 2012-05-09 23:16:07 +0400 |
| commit | 3aff4ebb95b20ad8db2c1447e8c52097d89af5a7 (patch) | |
| tree | ad56aceafbcbd0650e71e029ffa138eedffd33ba /fs/nfs/pagelist.c | |
| parent | dc327ed4cd320be689596365372a3683208c3ba0 (diff) | |
| download | linux-3aff4ebb95b20ad8db2c1447e8c52097d89af5a7.tar.xz | |
NFS: Prevent a deadlock in the new writeback code
We have to unlock the nfs_page before we call nfs_end_page_writeback
to avoid races with functions that expect the page to be unlocked
when PG_locked and PG_writeback are not set.
The problem is that nfs_unlock_request also releases the nfs_page,
causing a deadlock if the release of the nfs_open_context
triggers an iput() while the PG_writeback flag is still set...
The solution is to separate the unlocking and release of the nfs_page,
so that we can do the former before nfs_end_page_writeback and the
latter after.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Fred Isaman <iisaman@netapp.com>
Diffstat (limited to 'fs/nfs/pagelist.c')
| -rw-r--r-- | fs/nfs/pagelist.c | 12 |
1 files changed, 10 insertions, 2 deletions
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c index 33a21ca9b84b..69146f386989 100644 --- a/fs/nfs/pagelist.c +++ b/fs/nfs/pagelist.c @@ -128,10 +128,10 @@ nfs_create_request(struct nfs_open_context *ctx, struct inode *inode, } /** - * nfs_unlock_request - Unlock request and wake up sleepers. + * nfs_unlock_request_dont_release - Unlock request and wake up sleepers. * @req: */ -void nfs_unlock_request(struct nfs_page *req) +void nfs_unlock_request_dont_release(struct nfs_page *req) { if (!NFS_WBACK_BUSY(req)) { printk(KERN_ERR "NFS: Invalid unlock attempted\n"); @@ -141,6 +141,14 @@ void nfs_unlock_request(struct nfs_page *req) clear_bit(PG_BUSY, &req->wb_flags); smp_mb__after_clear_bit(); wake_up_bit(&req->wb_flags, PG_BUSY); +} + +/** + * nfs_unlock_request - Unlock request and release the nfs_page + */ +void nfs_unlock_request(struct nfs_page *req) +{ + nfs_unlock_request_dont_release(req); nfs_release_request(req); } |