linux/net/sunrpc
Trond Myklebust a6b31d18b0 SUNRPC: Fix a data corruption issue when retransmitting RPC calls
The following scenario can cause silent data corruption when doing
NFS writes. It has mainly been observed when doing database writes
using O_DIRECT.

1) The RPC client uses sendpage() to do zero-copy of the page data.
2) Due to networking issues, the reply from the server is delayed,
   and so the RPC client times out.

3) The client issues a second sendpage of the page data as part of
   an RPC call retransmission.

4) The reply to the first transmission arrives from the server
   _before_ the client hardware has emptied the TCP socket send
   buffer.
5) After processing the reply, the RPC state machine rules that
   the call to be done, and triggers the completion callbacks.
6) The application notices the RPC call is done, and reuses the
   pages to store something else (e.g. a new write).

7) The client NIC drains the TCP socket send buffer. Since the
   page data has now changed, it reads a corrupted version of the
   initial RPC call, and puts it on the wire.

This patch fixes the problem in the following manner:

The ordering guarantees of TCP ensure that when the server sends a
reply, then we know that the _first_ transmission has completed. Using
zero-copy in that situation is therefore safe.
If a time out occurs, we then send the retransmission using sendmsg()
(i.e. no zero-copy), We then know that the socket contains a full copy of
the data, and so it will retransmit a faithful reproduction even if the
RPC call completes, and the application reuses the O_DIRECT buffer in
the meantime.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@vger.kernel.org
2013-11-08 17:19:15 -05:00
..
auth_gss SUNRPC: Fix buffer overflow checking in gss_encode_v0_msg/gss_encode_v1_msg 2013-10-28 18:53:21 -04:00
xprtrdma Merge branch 'for-3.11' of git://linux-nfs.org/~bfields/linux 2013-07-17 13:43:55 -07:00
addr.c sunrpc: move address copy/cmp/convert routines and prototypes from clnt.h to addr.h 2013-02-05 09:41:14 -05:00
auth_generic.c sunrpc: Add missing kuids conversion for printing 2013-09-12 10:16:06 -04:00
auth_null.c SUNRPC refactor rpcauth_checkverf error returns 2013-09-03 15:25:09 -04:00
auth_unix.c SUNRPC refactor rpcauth_checkverf error returns 2013-09-03 15:25:09 -04:00
auth.c shrinker: convert remaining shrinkers to count/scan API 2013-09-10 18:56:32 -04:00
backchannel_rqst.c SUNRPC: remove two BUG_ON asserts 2012-11-04 14:43:41 -05:00
bc_svc.c SUNRPC: remove BUG_ON from bc_send 2012-11-04 14:43:41 -05:00
cache.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2013-07-14 11:42:26 -07:00
clnt.c SUNRPC: remove duplicated include from clnt.c 2013-10-30 11:13:34 -04:00
Kconfig Merge branch 'rpcsec_gss-from_cel' into linux-next 2013-04-23 15:40:40 -04:00
Makefile
netns.h SUNRPC: If the rpcbind channel is disconnected, fail the call to unregister 2013-08-07 17:07:18 -04:00
rpc_pipe.c SUNRPC: Add a helper to allow sharing of rpc_pipefs directory objects 2013-09-01 11:12:43 -04:00
rpcb_clnt.c SUNRPC: If the rpcbind channel is disconnected, fail the call to unregister 2013-08-07 17:07:18 -04:00
sched.c SUNRPC: Ensure rpc_task->tk_pid is available for tracepoints 2013-09-04 14:45:13 -04:00
socklib.c sunrpc: remove the second argument of k[un]map_atomic() 2012-03-20 21:48:28 +08:00
stats.c SUNRPC: Deprecate rpc_client->cl_protname 2013-08-30 09:19:34 -04:00
sunrpc_syms.c SUNRPC: register PipeFS file system after pernet sybsystem 2012-04-18 11:05:48 -04:00
sunrpc.h SUNRPC: subscribe RPC clients to pipefs notifications 2012-01-31 18:20:25 -05:00
svc_xprt.c svcrpc: fix rpc server shutdown races 2013-02-17 10:53:51 -05:00
svc.c drivers: avoid parsing names as kthread_run() format strings 2013-07-03 16:07:41 -07:00
svcauth_unix.c svcrpc: fix failures to handle -1 uid's 2013-07-08 17:27:23 -04:00
svcauth.c hlist: drop the node parameter from iterators 2013-02-27 19:10:24 -08:00
svcsock.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2013-08-03 21:36:46 -07:00
sysctl.c net: Convert uses of typedef ctl_table to struct ctl_table 2013-06-13 02:36:09 -07:00
timer.c net: cleanup unsigned to unsigned int 2012-04-15 12:44:40 -04:00
xdr.c SUNRPC: Fix memory corruption issue on 32-bit highmem systems 2013-08-28 15:43:43 -04:00
xprt.c SUNRPC: remove an unnecessary if statement 2013-10-28 18:16:56 -04:00
xprtsock.c SUNRPC: Fix a data corruption issue when retransmitting RPC calls 2013-11-08 17:19:15 -05:00