Skip to contentSkip to conversation

CIFS: Fix read after write for files with read caching

PATCH In review Pavel Shilovsky 2 participants 2 messages Mar 5, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 CIFS: Fix read after write for files with read caching

When we have a READ lease for a file and have just issued a write operation to the server we need to purge the cache and set oplock/lease level to NONE to avoid reading stale data. Currently we do that only if a write operation succedeed thus not covering cases when a request was sent to the server but a negative error code was returned later for some other reasons (e.g. -EIOCBQUEUED or -EINTR). Fix this by turning off caching regardless of the error code being returned.

The patches fixes generic tests 075 and 112 from the xfs-tests.

---
 fs/cifs/file.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)
fs/cifs/file.c +7 -5
index 9b53f33..4c144c1 100644
@@ -3096,14 +3096,16 @@ cifs_strict_writev(struct kiocb *iocb, struct iov_iter *from)
 	 * these pages but not on the region from pos to ppos+len-1.
 	 */
 	written = cifs_user_writev(iocb, from);
-	if (written > 0 && CIFS_CACHE_READ(cinode)) {
+	if (CIFS_CACHE_READ(cinode)) {
 		/*
-		 * Windows 7 server can delay breaking level2 oplock if a write
-		 * request comes - break it on the client to prevent reading
-		 * an old data.
+		 * We have read level caching and we have just sent a write
+		 * request to the server thus making data in the cache stale.
+		 * Zap the cache and set oplock/lease level to NONE to avoid
+		 * reading stale data from the cache. All subsequent read
+		 * operations will read new data from the server.
 		 */
 		cifs_zap_mapping(inode);
-		cifs_dbg(FYI, "Set no oplock for inode=%p after a write operation\n",
+		cifs_dbg(FYI, "Set Oplock/Lease to NONE for inode=%p after write\n",
 			 inode);
 		cinode->oplock = 0;
 	}
Signature · 1 line
2.7.4
reviewer Reviewed-by on v1 1/1
#1 lore raw
Quoted 47 lines · Pavel Shilovsky

When we have a READ lease for a file and have just issued a write operation to the server we need to purge the cache and set oplock/lease level to NONE to avoid reading stale data. Currently we do that only if a write operation succedeed thus not covering cases when a request was sent to the server but a negative error code was returned later for some other reasons (e.g. -EIOCBQUEUED or -EINTR). Fix this by turning off caching regardless of the error code being returned.

The patches fixes generic tests 075 and 112 from the xfs-tests.

---
 fs/cifs/file.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)
fs/cifs/file.c +7 -6
index 9b53f33..4c144c1 100644
@@ -3096,14 +3096,16 @@ cifs_strict_writev(struct kiocb *iocb, struct iov_iter *from)
         * these pages but not on the region from pos to ppos+len-1.
         */
        written = cifs_user_writev(iocb, from);
-       if (written > 0 && CIFS_CACHE_READ(cinode)) {
+       if (CIFS_CACHE_READ(cinode)) {
                /*
-                * Windows 7 server can delay breaking level2 oplock if a write
-                * request comes - break it on the client to prevent reading
-                * an old data.
+                * We have read level caching and we have just sent a write
+                * request to the server thus making data in the cache stale.
+                * Zap the cache and set oplock/lease level to NONE to avoid
+                * reading stale data from the cache. All subsequent read
+                * operations will read new data from the server.
                 */
                cifs_zap_mapping(inode);
-               cifs_dbg(FYI, "Set no oplock for inode=%p after a write operation\n",
+               cifs_dbg(FYI, "Set Oplock/Lease to NONE for inode=%p after write\n",
                         inode);
                cinode->oplock = 0;
        }
--

2.7.4