diff options
author | Jeff Layton <jlayton@redhat.com> | 2008-10-16 14:46:39 -0400 |
---|---|---|
committer | Steve French <sfrench@us.ibm.com> | 2008-10-16 14:46:39 -0400 |
commit | 469ee614aaa367d9cde01cbdd2027212f56c6cc6 (patch) | |
tree | b3adf11b97e14ece7c10c81ea9f156f16491ca87 /fs/cifs | |
parent | 2c1b861539c15491593625920058e06452cd3747 (diff) |
[CIFS] eliminate usage of kthread_stop for cifsd
When cifs_demultiplex_thread was converted to a kthread based kernel
thread, great pains were taken to make it so that kthread_stop would be
used to bring it down. This just added unnecessary complexity since we
needed to use a signal anyway to break out of kernel_recvmsg.
Also, cifs_demultiplex_thread does a bit of cleanup as it's exiting, and
we need to be certain that this gets done. It's possible for a kthread
to exit before its main function is ever run if kthread_stop is called
soon after its creation. While I'm not sure that this is a real problem
with cifsd now, it could be at some point in the future if cifs_mount is
ever changed to bring down the thread quickly.
The upshot here is that using kthread_stop to bring down the thread just
adds extra complexity with no real benefit. This patch changes the code
to use the original method to bring down the thread, but still leaves it
so that the thread is actually started with kthread_run.
This seems to fix the deadlock caused by the reproducer in this bug
report:
https://bugzilla.samba.org/show_bug.cgi?id=5720
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Diffstat (limited to 'fs/cifs')
-rw-r--r-- | fs/cifs/connect.c | 44 |
1 files changed, 15 insertions, 29 deletions
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 17ca8ce81bb7..1126f7ab4606 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c | |||
@@ -125,7 +125,7 @@ cifs_reconnect(struct TCP_Server_Info *server) | |||
125 | struct mid_q_entry *mid_entry; | 125 | struct mid_q_entry *mid_entry; |
126 | 126 | ||
127 | spin_lock(&GlobalMid_Lock); | 127 | spin_lock(&GlobalMid_Lock); |
128 | if (kthread_should_stop()) { | 128 | if (server->tcpStatus == CifsExiting) { |
129 | /* the demux thread will exit normally | 129 | /* the demux thread will exit normally |
130 | next time through the loop */ | 130 | next time through the loop */ |
131 | spin_unlock(&GlobalMid_Lock); | 131 | spin_unlock(&GlobalMid_Lock); |
@@ -185,7 +185,8 @@ cifs_reconnect(struct TCP_Server_Info *server) | |||
185 | spin_unlock(&GlobalMid_Lock); | 185 | spin_unlock(&GlobalMid_Lock); |
186 | up(&server->tcpSem); | 186 | up(&server->tcpSem); |
187 | 187 | ||
188 | while ((!kthread_should_stop()) && (server->tcpStatus != CifsGood)) { | 188 | while ((server->tcpStatus != CifsExiting) && |
189 | (server->tcpStatus != CifsGood)) { | ||
189 | try_to_freeze(); | 190 | try_to_freeze(); |
190 | if (server->protocolType == IPV6) { | 191 | if (server->protocolType == IPV6) { |
191 | rc = ipv6_connect(&server->addr.sockAddr6, | 192 | rc = ipv6_connect(&server->addr.sockAddr6, |
@@ -202,7 +203,7 @@ cifs_reconnect(struct TCP_Server_Info *server) | |||
202 | } else { | 203 | } else { |
203 | atomic_inc(&tcpSesReconnectCount); | 204 | atomic_inc(&tcpSesReconnectCount); |
204 | spin_lock(&GlobalMid_Lock); | 205 | spin_lock(&GlobalMid_Lock); |
205 | if (!kthread_should_stop()) | 206 | if (server->tcpStatus != CifsExiting) |
206 | server->tcpStatus = CifsGood; | 207 | server->tcpStatus = CifsGood; |
207 | server->sequence_number = 0; | 208 | server->sequence_number = 0; |
208 | spin_unlock(&GlobalMid_Lock); | 209 | spin_unlock(&GlobalMid_Lock); |
@@ -357,7 +358,7 @@ cifs_demultiplex_thread(struct TCP_Server_Info *server) | |||
357 | GFP_KERNEL); | 358 | GFP_KERNEL); |
358 | 359 | ||
359 | set_freezable(); | 360 | set_freezable(); |
360 | while (!kthread_should_stop()) { | 361 | while (server->tcpStatus != CifsExiting) { |
361 | if (try_to_freeze()) | 362 | if (try_to_freeze()) |
362 | continue; | 363 | continue; |
363 | if (bigbuf == NULL) { | 364 | if (bigbuf == NULL) { |
@@ -398,7 +399,7 @@ incomplete_rcv: | |||
398 | kernel_recvmsg(csocket, &smb_msg, | 399 | kernel_recvmsg(csocket, &smb_msg, |
399 | &iov, 1, pdu_length, 0 /* BB other flags? */); | 400 | &iov, 1, pdu_length, 0 /* BB other flags? */); |
400 | 401 | ||
401 | if (kthread_should_stop()) { | 402 | if (server->tcpStatus == CifsExiting) { |
402 | break; | 403 | break; |
403 | } else if (server->tcpStatus == CifsNeedReconnect) { | 404 | } else if (server->tcpStatus == CifsNeedReconnect) { |
404 | cFYI(1, ("Reconnect after server stopped responding")); | 405 | cFYI(1, ("Reconnect after server stopped responding")); |
@@ -523,7 +524,7 @@ incomplete_rcv: | |||
523 | total_read += length) { | 524 | total_read += length) { |
524 | length = kernel_recvmsg(csocket, &smb_msg, &iov, 1, | 525 | length = kernel_recvmsg(csocket, &smb_msg, &iov, 1, |
525 | pdu_length - total_read, 0); | 526 | pdu_length - total_read, 0); |
526 | if (kthread_should_stop() || | 527 | if ((server->tcpStatus == CifsExiting) || |
527 | (length == -EINTR)) { | 528 | (length == -EINTR)) { |
528 | /* then will exit */ | 529 | /* then will exit */ |
529 | reconnect = 2; | 530 | reconnect = 2; |
@@ -652,14 +653,6 @@ multi_t2_fnd: | |||
652 | spin_unlock(&GlobalMid_Lock); | 653 | spin_unlock(&GlobalMid_Lock); |
653 | wake_up_all(&server->response_q); | 654 | wake_up_all(&server->response_q); |
654 | 655 | ||
655 | /* don't exit until kthread_stop is called */ | ||
656 | set_current_state(TASK_UNINTERRUPTIBLE); | ||
657 | while (!kthread_should_stop()) { | ||
658 | schedule(); | ||
659 | set_current_state(TASK_UNINTERRUPTIBLE); | ||
660 | } | ||
661 | set_current_state(TASK_RUNNING); | ||
662 | |||
663 | /* check if we have blocked requests that need to free */ | 656 | /* check if we have blocked requests that need to free */ |
664 | /* Note that cifs_max_pending is normally 50, but | 657 | /* Note that cifs_max_pending is normally 50, but |
665 | can be set at module install time to as little as two */ | 658 | can be set at module install time to as little as two */ |
@@ -2234,14 +2227,7 @@ cifs_mount(struct super_block *sb, struct cifs_sb_info *cifs_sb, | |||
2234 | spin_lock(&GlobalMid_Lock); | 2227 | spin_lock(&GlobalMid_Lock); |
2235 | srvTcp->tcpStatus = CifsExiting; | 2228 | srvTcp->tcpStatus = CifsExiting; |
2236 | spin_unlock(&GlobalMid_Lock); | 2229 | spin_unlock(&GlobalMid_Lock); |
2237 | if (srvTcp->tsk) { | 2230 | force_sig(SIGKILL, srvTcp->tsk); |
2238 | /* If we could verify that kthread_stop would | ||
2239 | always wake up processes blocked in | ||
2240 | tcp in recv_mesg then we could remove the | ||
2241 | send_sig call */ | ||
2242 | force_sig(SIGKILL, srvTcp->tsk); | ||
2243 | kthread_stop(srvTcp->tsk); | ||
2244 | } | ||
2245 | } | 2231 | } |
2246 | /* If find_unc succeeded then rc == 0 so we can not end */ | 2232 | /* If find_unc succeeded then rc == 0 so we can not end */ |
2247 | if (tcon) /* up accidently freeing someone elses tcon struct */ | 2233 | if (tcon) /* up accidently freeing someone elses tcon struct */ |
@@ -2255,18 +2241,18 @@ cifs_mount(struct super_block *sb, struct cifs_sb_info *cifs_sb, | |||
2255 | /* if the socketUseCount is now zero */ | 2241 | /* if the socketUseCount is now zero */ |
2256 | if ((temp_rc == -ESHUTDOWN) && | 2242 | if ((temp_rc == -ESHUTDOWN) && |
2257 | (pSesInfo->server) && | 2243 | (pSesInfo->server) && |
2258 | (pSesInfo->server->tsk)) { | 2244 | (pSesInfo->server->tsk)) |
2259 | force_sig(SIGKILL, | 2245 | force_sig(SIGKILL, |
2260 | pSesInfo->server->tsk); | 2246 | pSesInfo->server->tsk); |
2261 | kthread_stop(pSesInfo->server->tsk); | ||
2262 | } | ||
2263 | } else { | 2247 | } else { |
2264 | cFYI(1, ("No session or bad tcon")); | 2248 | cFYI(1, ("No session or bad tcon")); |
2265 | if ((pSesInfo->server) && | 2249 | if ((pSesInfo->server) && |
2266 | (pSesInfo->server->tsk)) { | 2250 | (pSesInfo->server->tsk)) { |
2251 | spin_lock(&GlobalMid_Lock); | ||
2252 | srvTcp->tcpStatus = CifsExiting; | ||
2253 | spin_unlock(&GlobalMid_Lock); | ||
2267 | force_sig(SIGKILL, | 2254 | force_sig(SIGKILL, |
2268 | pSesInfo->server->tsk); | 2255 | pSesInfo->server->tsk); |
2269 | kthread_stop(pSesInfo->server->tsk); | ||
2270 | } | 2256 | } |
2271 | } | 2257 | } |
2272 | sesInfoFree(pSesInfo); | 2258 | sesInfoFree(pSesInfo); |
@@ -3577,10 +3563,8 @@ cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb) | |||
3577 | return 0; | 3563 | return 0; |
3578 | } else if (rc == -ESHUTDOWN) { | 3564 | } else if (rc == -ESHUTDOWN) { |
3579 | cFYI(1, ("Waking up socket by sending signal")); | 3565 | cFYI(1, ("Waking up socket by sending signal")); |
3580 | if (cifsd_task) { | 3566 | if (cifsd_task) |
3581 | force_sig(SIGKILL, cifsd_task); | 3567 | force_sig(SIGKILL, cifsd_task); |
3582 | kthread_stop(cifsd_task); | ||
3583 | } | ||
3584 | rc = 0; | 3568 | rc = 0; |
3585 | } /* else - we have an smb session | 3569 | } /* else - we have an smb session |
3586 | left on this socket do not kill cifsd */ | 3570 | left on this socket do not kill cifsd */ |
@@ -3710,7 +3694,9 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *pSesInfo, | |||
3710 | cERROR(1, ("Send error in SessSetup = %d", rc)); | 3694 | cERROR(1, ("Send error in SessSetup = %d", rc)); |
3711 | } else { | 3695 | } else { |
3712 | cFYI(1, ("CIFS Session Established successfully")); | 3696 | cFYI(1, ("CIFS Session Established successfully")); |
3697 | spin_lock(&GlobalMid_Lock); | ||
3713 | pSesInfo->status = CifsGood; | 3698 | pSesInfo->status = CifsGood; |
3699 | spin_unlock(&GlobalMid_Lock); | ||
3714 | } | 3700 | } |
3715 | 3701 | ||
3716 | ss_err_exit: | 3702 | ss_err_exit: |