You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@trafficserver.apache.org by "Eric Balsa (JIRA)" <ji...@apache.org> on 2010/04/01 02:28:27 UTC
[jira] Created: (TS-286) Seg fault on startup on OSX inside
RecSetRawStatSum
Seg fault on startup on OSX inside RecSetRawStatSum
---------------------------------------------------
Key: TS-286
URL: https://issues.apache.org/jira/browse/TS-286
Project: Traffic Server
Issue Type: Bug
Reporter: Eric Balsa
Fix For: 2.1.0
OS X debug (--enable-debug) build:
Getting the following segfault about 5 seconds after startup with latest trunk:
#0 0x00007fff88b4cfe6 in __kill ()
#1 0x00007fff88bede32 in abort ()
#2 0x0000000100280639 in ink_mutex_acquire (m=0x100d56368) at ink_mutex.h:96
#3 0x0000000100280924 in raw_stat_clear_sum (rsb=0x100d56350, id=0) at RecProcess.cc:126
#4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
#5 0x00000001002588a3 in register_net_stats () at Net.cc:66
#6 0x0000000100258e04 in ink_net_init (version=16842752) at Net.cc:152
#7 0x0000000100075f9f in main (argc=1, argv=0x7fff5fbff948) at Main.cc:1896
(gdb) fr 4
#4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
592 raw_stat_clear_sum(rsb, id);
(gdb) p *rsb
$12 = {
ethr_stat_offset = 2920,
global = 0x100d563b0,
num_stats = 0,
max_stats = 16,
mutex = {
__sig = 0,
__opaque = '\0' <repeats 55 times>
}
}
(gdb) fr 2
#2 0x0000000100280639 in ink_mutex_acquire (m=0x101801448) at ink_mutex.h:96
96 abort();
(gdb) l
91
92 static inline int
93 ink_mutex_acquire(ink_mutex * m)
94 {
95 if (pthread_mutex_lock(m) != 0) {
96 abort();
97 }
98 return 0;
99 }
Looks like the mutex for rsb is NULL.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (TS-286) Seg fault on startup on OSX inside
RecSetRawStatSum
Posted by "Eric Balsa (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-286?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Eric Balsa updated TS-286:
--------------------------
Attachment: TS-286.patch
Bryan, this seems to fix it. Comments? I didn't delve too much into this code...
> Seg fault on startup on OSX inside RecSetRawStatSum
> ---------------------------------------------------
>
> Key: TS-286
> URL: https://issues.apache.org/jira/browse/TS-286
> Project: Traffic Server
> Issue Type: Bug
> Reporter: Eric Balsa
> Fix For: 2.1.0
>
> Attachments: TS-286.patch
>
>
> OS X debug (--enable-debug) build:
> Getting the following segfault about 5 seconds after startup with latest trunk:
> #0 0x00007fff88b4cfe6 in __kill ()
> #1 0x00007fff88bede32 in abort ()
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x100d56368) at ink_mutex.h:96
> #3 0x0000000100280924 in raw_stat_clear_sum (rsb=0x100d56350, id=0) at RecProcess.cc:126
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> #5 0x00000001002588a3 in register_net_stats () at Net.cc:66
> #6 0x0000000100258e04 in ink_net_init (version=16842752) at Net.cc:152
> #7 0x0000000100075f9f in main (argc=1, argv=0x7fff5fbff948) at Main.cc:1896
> (gdb) fr 4
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> 592 raw_stat_clear_sum(rsb, id);
> (gdb) p *rsb
> $12 = {
> ethr_stat_offset = 2920,
> global = 0x100d563b0,
> num_stats = 0,
> max_stats = 16,
> mutex = {
> __sig = 0,
> __opaque = '\0' <repeats 55 times>
> }
> }
> (gdb) fr 2
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x101801448) at ink_mutex.h:96
> 96 abort();
> (gdb) l
> 91
> 92 static inline int
> 93 ink_mutex_acquire(ink_mutex * m)
> 94 {
> 95 if (pthread_mutex_lock(m) != 0) {
> 96 abort();
> 97 }
> 98 return 0;
> 99 }
> Looks like the mutex for rsb is NULL.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (TS-286) Seg fault on startup on OSX inside
RecSetRawStatSum
Posted by "George Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-286?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12852407#action_12852407 ]
George Paul commented on TS-286:
--------------------------------
Looked at the code and the mutex needs to be initialized. The patch works for me also and is good to check in.
-George
> Seg fault on startup on OSX inside RecSetRawStatSum
> ---------------------------------------------------
>
> Key: TS-286
> URL: https://issues.apache.org/jira/browse/TS-286
> Project: Traffic Server
> Issue Type: Bug
> Reporter: Eric Balsa
> Fix For: 2.1.0
>
> Attachments: TS-286.patch
>
>
> OS X debug (--enable-debug) build:
> Getting the following segfault about 5 seconds after startup with latest trunk:
> #0 0x00007fff88b4cfe6 in __kill ()
> #1 0x00007fff88bede32 in abort ()
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x100d56368) at ink_mutex.h:96
> #3 0x0000000100280924 in raw_stat_clear_sum (rsb=0x100d56350, id=0) at RecProcess.cc:126
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> #5 0x00000001002588a3 in register_net_stats () at Net.cc:66
> #6 0x0000000100258e04 in ink_net_init (version=16842752) at Net.cc:152
> #7 0x0000000100075f9f in main (argc=1, argv=0x7fff5fbff948) at Main.cc:1896
> (gdb) fr 4
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> 592 raw_stat_clear_sum(rsb, id);
> (gdb) p *rsb
> $12 = {
> ethr_stat_offset = 2920,
> global = 0x100d563b0,
> num_stats = 0,
> max_stats = 16,
> mutex = {
> __sig = 0,
> __opaque = '\0' <repeats 55 times>
> }
> }
> (gdb) fr 2
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x101801448) at ink_mutex.h:96
> 96 abort();
> (gdb) l
> 91
> 92 static inline int
> 93 ink_mutex_acquire(ink_mutex * m)
> 94 {
> 95 if (pthread_mutex_lock(m) != 0) {
> 96 abort();
> 97 }
> 98 return 0;
> 99 }
> Looks like the mutex for rsb is NULL.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (TS-286) Seg fault on startup on OSX inside
RecSetRawStatSum
Posted by "Eric Balsa (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-286?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12852227#action_12852227 ]
Eric Balsa commented on TS-286:
-------------------------------
lol, ya, forgot to add this fixes it for me.
> Seg fault on startup on OSX inside RecSetRawStatSum
> ---------------------------------------------------
>
> Key: TS-286
> URL: https://issues.apache.org/jira/browse/TS-286
> Project: Traffic Server
> Issue Type: Bug
> Reporter: Eric Balsa
> Fix For: 2.1.0
>
> Attachments: TS-286.patch
>
>
> OS X debug (--enable-debug) build:
> Getting the following segfault about 5 seconds after startup with latest trunk:
> #0 0x00007fff88b4cfe6 in __kill ()
> #1 0x00007fff88bede32 in abort ()
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x100d56368) at ink_mutex.h:96
> #3 0x0000000100280924 in raw_stat_clear_sum (rsb=0x100d56350, id=0) at RecProcess.cc:126
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> #5 0x00000001002588a3 in register_net_stats () at Net.cc:66
> #6 0x0000000100258e04 in ink_net_init (version=16842752) at Net.cc:152
> #7 0x0000000100075f9f in main (argc=1, argv=0x7fff5fbff948) at Main.cc:1896
> (gdb) fr 4
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> 592 raw_stat_clear_sum(rsb, id);
> (gdb) p *rsb
> $12 = {
> ethr_stat_offset = 2920,
> global = 0x100d563b0,
> num_stats = 0,
> max_stats = 16,
> mutex = {
> __sig = 0,
> __opaque = '\0' <repeats 55 times>
> }
> }
> (gdb) fr 2
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x101801448) at ink_mutex.h:96
> 96 abort();
> (gdb) l
> 91
> 92 static inline int
> 93 ink_mutex_acquire(ink_mutex * m)
> 94 {
> 95 if (pthread_mutex_lock(m) != 0) {
> 96 abort();
> 97 }
> 98 return 0;
> 99 }
> Looks like the mutex for rsb is NULL.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (TS-286) Seg fault on startup on OSX inside
RecSetRawStatSum
Posted by "Bryan Call (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/TS-286?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12852220#action_12852220 ]
Bryan Call commented on TS-286:
-------------------------------
Did you test it? Makes sense that it might fail if the mutex was not initialized. I guess Linux can handle this...
> Seg fault on startup on OSX inside RecSetRawStatSum
> ---------------------------------------------------
>
> Key: TS-286
> URL: https://issues.apache.org/jira/browse/TS-286
> Project: Traffic Server
> Issue Type: Bug
> Reporter: Eric Balsa
> Fix For: 2.1.0
>
> Attachments: TS-286.patch
>
>
> OS X debug (--enable-debug) build:
> Getting the following segfault about 5 seconds after startup with latest trunk:
> #0 0x00007fff88b4cfe6 in __kill ()
> #1 0x00007fff88bede32 in abort ()
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x100d56368) at ink_mutex.h:96
> #3 0x0000000100280924 in raw_stat_clear_sum (rsb=0x100d56350, id=0) at RecProcess.cc:126
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> #5 0x00000001002588a3 in register_net_stats () at Net.cc:66
> #6 0x0000000100258e04 in ink_net_init (version=16842752) at Net.cc:152
> #7 0x0000000100075f9f in main (argc=1, argv=0x7fff5fbff948) at Main.cc:1896
> (gdb) fr 4
> #4 0x0000000100280a0c in RecSetRawStatSum (rsb=0x100d56350, id=0, data=0) at RecProcess.cc:592
> 592 raw_stat_clear_sum(rsb, id);
> (gdb) p *rsb
> $12 = {
> ethr_stat_offset = 2920,
> global = 0x100d563b0,
> num_stats = 0,
> max_stats = 16,
> mutex = {
> __sig = 0,
> __opaque = '\0' <repeats 55 times>
> }
> }
> (gdb) fr 2
> #2 0x0000000100280639 in ink_mutex_acquire (m=0x101801448) at ink_mutex.h:96
> 96 abort();
> (gdb) l
> 91
> 92 static inline int
> 93 ink_mutex_acquire(ink_mutex * m)
> 94 {
> 95 if (pthread_mutex_lock(m) != 0) {
> 96 abort();
> 97 }
> 98 return 0;
> 99 }
> Looks like the mutex for rsb is NULL.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.