xfs
[Top] [All Lists]

[PATCH 2/6] xfs: avoid direct I/O write vs buffered I/O race

To: stable@xxxxxxxxxxxxxxx
Subject: [PATCH 2/6] xfs: avoid direct I/O write vs buffered I/O race
From: Ben Myers <bpm@xxxxxxx>
Date: Wed, 23 Nov 2011 15:38:09 -0600
Cc: xfs@xxxxxxxxxxx, Christoph Hellwig <hch@xxxxxxxxxxxxx>, Christoph Hellwig <hch@xxxxxx>, Alex Elder <aelder@xxxxxxx>
In-reply-to: <1322084293-23557-1-git-send-email-bpm@xxxxxxx>
References: <1322084293-23557-1-git-send-email-bpm@xxxxxxx>
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>

>From 1d443d87fbda3b1e01d3dba07008ad0064fd8c59

Currently a buffered reader or writer can add pages to the pagecache
while we are waiting for the iolock in xfs_file_dio_aio_write.  Prevent
this by re-checking mapping->nrpages after we got the iolock, and if
nessecary upgrade the lock to exclusive mode.  To simplify this a bit
only take the ilock inside of xfs_file_aio_write_checks.

Signed-off-by: Christoph Hellwig <hch@xxxxxx>
Reviewed-by: Dave Chinner <dchinner@xxxxxxxxxx>
Signed-off-by: Alex Elder <aelder@xxxxxxx>
---
 fs/xfs/xfs_file.c |   17 ++++++++++++++---
 1 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index 8fd4a07..b7e75c6 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -677,6 +677,7 @@ xfs_file_aio_write_checks(
        xfs_fsize_t             new_size;
        int                     error = 0;
 
+       xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
        error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
        if (error) {
                xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
@@ -768,14 +769,24 @@ xfs_file_dio_aio_write(
                *iolock = XFS_IOLOCK_EXCL;
        else
                *iolock = XFS_IOLOCK_SHARED;
-       xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
+       xfs_rw_ilock(ip, *iolock);
 
        ret = xfs_file_aio_write_checks(file, &pos, &count, iolock);
        if (ret)
                return ret;
 
+       /*
+        * Recheck if there are cached pages that need invalidate after we got
+        * the iolock to protect against other threads adding new pages while
+        * we were waiting for the iolock.
+        */
+       if (mapping->nrpages && *iolock == XFS_IOLOCK_SHARED) {
+               xfs_rw_iunlock(ip, *iolock);
+               *iolock = XFS_IOLOCK_EXCL;
+               xfs_rw_ilock(ip, *iolock);
+       }
+
        if (mapping->nrpages) {
-               WARN_ON(*iolock != XFS_IOLOCK_EXCL);
                ret = -xfs_flushinval_pages(ip, (pos & PAGE_CACHE_MASK), -1,
                                                        FI_REMAPF_LOCKED);
                if (ret)
@@ -820,7 +831,7 @@ xfs_file_buffered_aio_write(
        size_t                  count = ocount;
 
        *iolock = XFS_IOLOCK_EXCL;
-       xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
+       xfs_rw_ilock(ip, *iolock);
 
        ret = xfs_file_aio_write_checks(file, &pos, &count, iolock);
        if (ret)
-- 
1.7.8.rc0.46.g5ae0f

<Prev in Thread] Current Thread [Next in Thread>