Skip to contentSkip to conversation

xfs_repair: avoid segfault if reporting progress early in repair

PATCH In review Eric Sandeen 4 participants 4 messages Oct 17, 2013 – Oct 18, 2013

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 xfs_repair: avoid segfault if reporting progress early in repair

For a very large filesystem, zeroing the log may take some time.

If we ask for progress reports frequently enough that one fires before we finish with log zeroing, we try to use a progress format which has not yet been set up, and segfault:

# mkfs.xfs -d size=60t,file,name=fsfile
# xfs_repair -m 9000 -o ag_stride=32 -t 1 fsfile 
Phase 1 - find and verify superblock...
        - reporting progress in intervals of 1 seconds
Phase 2 - using internal log
        - zero log...
Segmentation fault

(gdb) bt

#0  0x0000000000426962 in progress_rpt_thread (p=0x67ad20) at progress.c:234
#1  0x0000003b98a07851 in start_thread (arg=0x7f19d8e47700) at pthread_create.c:301
#2  0x0000003b982e767d in ?? ()
#3  0x0000000000000000 in ?? ()

(gdb) p msgp
$1 = (msg_block_t *) 0x67ad20
(gdb) p msgp->format
$2 = (progress_rpt_t *) 0x0
(gdb)

I suppose we could rig up progress reports for log zeroing, but that won't usually take terribly long; for now, be defensive and init the message->format to NULL, and just return early from the progress thread if we've not yet set up any message.

(Sure, global_msgs is global, and ->format is already NULL, but to me it's worth being explicit since we will test it).

---

repair/progress.c +5 -0
index ab320dc..45a412e 100644
@@ -124,6 +124,7 @@ init_progress_rpt (void)
 	 */
 
 	pthread_mutex_init(&global_msgs.mutex, NULL);
+	global_msgs.format = NULL;
 	global_msgs.count = glob_agcount;
 	global_msgs.interval = report_interval;
 	global_msgs.done   = prog_rpt_done;
@@ -169,6 +170,10 @@ progress_rpt_thread (void *p)
 	msg_block_t *msgp = (msg_block_t *)p;
 	__uint64_t percent;
 
+	/* It's possible to get here very early w/ no progress msg set */
+	if (!msgp->format)
+		return NULL;
+
 	if ((msgbuf = (char *)malloc(DURATION_BUF_SIZE)) == NULL)
 		do_error (_("progress_rpt: cannot malloc progress msg buffer\n"));
 

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#2 lore raw
In reply to Eric Sandeen
Quoted 49 lines · Eric Sandeen

For a very large filesystem, zeroing the log may take some time.

If we ask for progress reports frequently enough that one fires before we finish with log zeroing, we try to use a progress format which has not yet been set up, and segfault:

# mkfs.xfs -d size=60t,file,name=fsfile
# xfs_repair -m 9000 -o ag_stride=32 -t 1 fsfile 
Phase 1 - find and verify superblock...
        - reporting progress in intervals of 1 seconds
Phase 2 - using internal log
        - zero log...
Segmentation fault

(gdb) bt

#0  0x0000000000426962 in progress_rpt_thread (p=0x67ad20) at progress.c:234
#1  0x0000003b98a07851 in start_thread (arg=0x7f19d8e47700) at pthread_create.c:301
#2  0x0000003b982e767d in ?? ()
#3  0x0000000000000000 in ?? ()

(gdb) p msgp
$1 = (msg_block_t *) 0x67ad20
(gdb) p msgp->format
$2 = (progress_rpt_t *) 0x0
(gdb)

I suppose we could rig up progress reports for log zeroing, but that won't usually take terribly long; for now, be defensive and init the message->format to NULL, and just return early from the progress thread if we've not yet set up any message.

(Sure, global_msgs is global, and ->format is already NULL, but to me it's worth being explicit since we will test it).

---

repair/progress.c +1 -0
index ab320dc..45a412e 100644
@@ -124,6 +124,7 @@ init_progress_rpt (void)
 	 */
 
 	pthread_mutex_init(&global_msgs.mutex, NULL);
+	global_msgs.format = NULL;
 	global_msgs.count = glob_agcount;
 	global_msgs.interval = report_interval;
 	global_msgs.done   = prog_rpt_done;

repair/progress.c progress_rpt_thread (void *p) · line 178

Quoted 9 lines
repair/progress.c +4 -0
@@ -169,6 +170,10 @@ progress_rpt_thread (void *p)
 	msg_block_t *msgp = (msg_block_t *)p;
 	__uint64_t percent;
 
+	/* It's possible to get here very early w/ no progress msg set */
+	if (!msgp->format)
+		return NULL;
+
 	if ((msgbuf = (char *)malloc(DURATION_BUF_SIZE)) == NULL)
 		do_error (_("progress_rpt: cannot malloc progress msg buffer\n"));

Dammit:

CID 1107596: Data race condition (MISSING_LOCK)

/repair/progress.c: 127 ( missing_lock)
124 */
125
126 pthread_mutex_init(&global_msgs.mutex, NULL);

Quoted 2 lines
repair/progress.c +0 -0
 >> CID 1107596: Data race condition (MISSING_LOCK)
Quoted 1 line
Quoted 1 line

Accessing "global_msgs.format" without holding lock "msg_block_s.mutex". Elsewhere, "global_msgs.format" is accessed with "msg_block_s.mutex" held 2 out of 2 times.

   127    	global_msgs.format = NULL;
   128    	global_msgs.count = glob_agcount;
   129    	global_msgs.interval = report_interval;
   130    	global_msgs.done   = prog_rpt_done;
   131    	global_msgs.total  = &prog_rpt_total;

Probably best to just drop the new NULL assignment, since it's a global init'd to 0 anyway, to shut up coverity?

-Eric

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs