Commit 28c86fda authored by Weston Andros Adamson's avatar Weston Andros Adamson Committed by Greg Kroah-Hartman

nfs: fix nonblocking calls to nfs_page_group_lock

commit bc8a309e upstream.

nfs_page_group_lock was calling wait_on_bit_lock even when told not to
block. Fix by first trying test_and_set_bit, followed by wait_on_bit_lock
if and only if blocking is allowed.  Return -EAGAIN if nonblocking and the
test_and_set of the bit was already locked.
Signed-off-by: default avatarWeston Andros Adamson <dros@primarydata.com>
Reviewed-by: default avatarPeng Tao <tao.peng@primarydata.com>
Signed-off-by: default avatarTrond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
parent 6483253a
...@@ -149,24 +149,29 @@ static int nfs_wait_bit_uninterruptible(void *word) ...@@ -149,24 +149,29 @@ static int nfs_wait_bit_uninterruptible(void *word)
* *
* this lock must be held if modifying the page group list * this lock must be held if modifying the page group list
* *
* returns result from wait_on_bit_lock: 0 on success, < 0 on error * return 0 on success, < 0 on error: -EDELAY if nonblocking or the
* result from wait_on_bit_lock
*
* NOTE: calling with nonblock=false should always have set the
* lock bit (see fs/buffer.c and other uses of wait_on_bit_lock
* with TASK_UNINTERRUPTIBLE), so there is no need to check the result.
*/ */
int int
nfs_page_group_lock(struct nfs_page *req, bool nonblock) nfs_page_group_lock(struct nfs_page *req, bool nonblock)
{ {
struct nfs_page *head = req->wb_head; struct nfs_page *head = req->wb_head;
int ret;
WARN_ON_ONCE(head != head->wb_head); WARN_ON_ONCE(head != head->wb_head);
do { if (!test_and_set_bit(PG_HEADLOCK, &head->wb_flags))
ret = wait_on_bit_lock(&head->wb_flags, PG_HEADLOCK, return 0;
if (!nonblock)
return wait_on_bit_lock(&head->wb_flags, PG_HEADLOCK,
nfs_wait_bit_uninterruptible, nfs_wait_bit_uninterruptible,
TASK_UNINTERRUPTIBLE); TASK_UNINTERRUPTIBLE);
} while (!nonblock && ret != 0);
WARN_ON_ONCE(ret > 0); return -EAGAIN;
return ret;
} }
/* /*
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment