cifs: don't always drop malformed replies on the floor (try #3)
authorJeff Layton <jlayton@redhat.com>
Thu, 10 Feb 2011 13:03:50 +0000 (08:03 -0500)
committerSteve French <sfrench@us.ibm.com>
Fri, 11 Feb 2011 03:59:12 +0000 (03:59 +0000)
commit71823baff1978be892e7a36eddf6170e1cc6650d
tree2cc136ce754cc5c24ed8aefde5e486fc873ae139
parent195291e68c2ad59a046fc56d32bf59635b100e5c
cifs: don't always drop malformed replies on the floor (try #3)

Slight revision to this patch...use min_t() instead of conditional
assignment. Also, remove the FIXME comment and replace it with the
explanation that Steve gave earlier.

After receiving a packet, we currently check the header. If it's no
good, then we toss it out and continue the loop, leaving the caller
waiting on that response.

In cases where the packet has length inconsistencies, but the MID is
valid, this leads to unneeded delays. That's especially problematic now
that the client waits indefinitely for responses.

Instead, don't immediately discard the packet if checkSMB fails. Try to
find a matching mid_q_entry, mark it as having a malformed response and
issue the callback.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
fs/cifs/cifsglob.h
fs/cifs/connect.c
fs/cifs/transport.c