https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #131 from Eugene Grosbein ---
(In reply to Harald Schmalzbauer from comment #129)
And syslogd(8) problem has no connection with this mpd5 problem.
--
You are receiving this mail because:
You are on the CC list for the bug.
__
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #130 from Eugene Grosbein ---
(In reply to Harald Schmalzbauer from comment #129)
11.1 release branch releng/11.1 was created with r320486 from stable/11@r320475
https://svnweb.freebsd.org/base?view=revision&revision=320486
11
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #129 from Harald Schmalzbauer ---
(In reply to Eugene Grosbein from comment #128)
Could you please tell us the revision? Hope I really missed something and you
are right, but if I e.g. watch the log of lib/libc/stdio/fclose.c,
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #128 from Eugene Grosbein ---
(In reply to Harald Schmalzbauer from comment #127)
Additionally, there were specific merges to release branch releng/11.1 before
release was tagged. So, 11.1-RELEASE does have those fixes.
--
Yo
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Harald Schmalzbauer changed:
What|Removed |Added
CC||bugzilla.free...@omnilan.de
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #126 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #125)
It is not needed to fix things. It was produced to discover roots of the
problem only.
--
You are receiving this mail because:
You are on the CC l
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #125 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #124)
Hi Eugene,
Just one question, how about libthread patch? Is it has been merged as well?
Take a look on comment #53:
https://bugs.freebsd.org/bugzi
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Eugene Grosbein changed:
What|Removed |Added
Status|In Progress |Closed
Resolution|---
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Bug 186114 depends on bug 220151, which changed state.
Bug 220151 Summary: [libc] syslog() thread unsafety: mutex lock leak
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=220151
What|Removed |Added
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #123 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #116)
Hi Eugene,
The server has been running for 14 days without freezing. I think the issue has
been fixed :)
How would you like to proceed? Are you goi
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #122 from Eugene Grosbein ---
(In reply to Konstantin Belousov from comment #121)
mpd5 uses libpdel that uses funopen() extensively. For example, it uses
funopen() while loading HTTP headers from a stream (mpd5 web console fea
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #121 from Konstantin Belousov ---
(In reply to Eugene Grosbein from comment #120)
Comment about fseeko was that the function calls user method, when the
operating stream is from funopen(3).
--
You are receiving this mail becau
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #120 from Eugene Grosbein ---
(In reply to Konstantin Belousov from comment #119)
Well, comment #83 refers to some problem in "fseeko" that seems to be fixed by
your "stdio cancel-safe" patch among other things as this patch el
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #119 from Konstantin Belousov ---
(In reply to Eugene Grosbein from comment #118)
I already mentioned that the r320472 change only matters for rare case of
user-IO streams, created e.g. by funopen(3). As such, I do not think th
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #118 from Eugene Grosbein ---
(In reply to Konstantin Belousov from comment #63)
I see you already MFC'd "cancel-safe" patches to stable/11. Do you have any
plans to MFC them to stable/10, so upcoming 10.4-RELEASE be fixed too?
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #117 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #116)
I'm using for PPPoE only. I'll enable the logs and will let you know. But
anyway for now no freezes happened.
--
You are receiving this mail becaus
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #116 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #114)
I guess, this message is harmless and comes from the fact that mpd5 may destroy
ngXXX interface and corresponding bundle before some last packets fr
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #115 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #113)
"Accessing freed hook: name right" is another case and has no connection with
mpd5 process hanging. And that comes from ng_tee(4), not ng_iface(4).
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #114 from Cassiano Peixoto ---
(In reply to Cassiano Peixoto from comment #113)
Just to add more info about comment #112. I got some relevant messages in
mpd5.log:
Jul 10 01:22:32 B-ras mpd: Link: Packet from unexisting bundle
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #113 from Cassiano Peixoto ---
(In reply to Cassiano Peixoto from comment #112)
Eugene, patch has been applied but i still can see the kernel error messages:
Jul 10 01:22:30 B-ras kernel: KDB: stack backtrace:
Jul 10 01:22:30 B
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #112 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #111)
Ok, just applied and rebooted. I'll keep you posted.
--
You are receiving this mail because:
You are on the CC list for the bug.
__
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #111 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #110)
Yes, just like I've asked in the comment 107.
--
You are receiving this mail because:
You are on the CC list for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #110 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #109)
Hi,
Should i update with this last patch (PR 220076) and rebuild the kernel?
--
You are receiving this mail because:
You are on the CC list for t
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #109 from Eugene Grosbein ---
(In reply to Donald Baud from comment #108)
It does not really matter what triggered the panic as its reason is known:
ng_iface lacks protection of its private data from modifications while it is
b
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Donald Baud changed:
What|Removed |Added
CC||donaldb...@yahoo.com
--- Comment #10
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #107 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #106)
I've just updated the patch for ng_iface(4). Apply it to kernel sources instead
of its previous version, rebuild, install and boot new kernel.
--
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #106 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #105)
Ok, i kept web console on.
I'm starting mpd5 using rc.d script:
# ps -ax | grep mpd
1278 - Ss 25:03.78 /usr/local/sbin/mpd5 -p /var/run/mpd5
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #105 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #104)
I do not think web console has any connection to this new problem, keep it on.
How do you start mpd5 - do you use its standard startup (rc.d) script
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #104 from Cassiano Peixoto ---
Hi,
It wasn't zombie, it looks like zombie because the kind of behavior. See
process top output below:
PID USERNAME THR PRI NICE SIZERES STATE C TIMEWCPU COMMAND
1295 root
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #103 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #101)
In fact, a zombie process is not a process anymore. It does not prevent your
from restarting new instance of mpd5. Did you look at ps output to chec
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Eugene Grosbein changed:
What|Removed |Added
Flags|maintainer-feedback?(mav@Fr |maintainer-feedback+
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #101 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #100)
Hi guys,
I had the issue today, but with a different behavior. It didn't show on CPU
state as uwrlck anymore. The process was running as usual,
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #100 from Konstantin Belousov ---
(In reply to Cassiano Peixoto from comment #99)
Uh, ok. So lets wait for the issue to re-appear on latest stable. There were
some fixes that might be relevant.
--
You are receiving this mail
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #99 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #97)
When it happend i was using 11.0-release-p0. But yesterday i updated to
11.1-beta3.
--
You are receiving this mail because:
You are on the CC lis
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #98 from Eugene Grosbein ---
(In reply to Konstantin Belousov from comment #97)
That's from releng/11.0:
https://svnweb.freebsd.org/base/releng/11.0/libexec/rtld-elf/rtld.c?annotate=304456#l4802
--
You are receiving this mai
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #97 from Konstantin Belousov ---
(In reply to Cassiano Peixoto from comment #92)
Indeed, everything I asked for, is in the comment #83. Now I cannot make a
sense from the reported line 4802 of rtld.c in the backtrace for the th
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #96 from Cassiano Peixoto ---
(In reply to Cassiano Peixoto from comment #94)
Let it go, something messed when applied the patch. I re-synced the src and
applied fine. So, patched applied and mpd5 recompiled. Let's watch.
Thank
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #95 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #94)
It seems your source tree is broken now. Try to clean is up and apply all
patches again:
cd /usr/src && svnlite cleanup --remove-unversioned && svnli
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #94 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #93)
Ok i reverted the patch and applied the new one, but i got an error when
compiling libc:
cc -O2 -pipe -I/usr/src/lib/libc/include -I/usr/src/includ
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #93 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #90)
Make sure you have latest revision of https://reviews.freebsd.org/D11246 patch
as Konstantin updated it several hours ago to include "fseeko" chunk an
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #92 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #91)
Did you read everything on my comment #83? Everything you asked on comment #52
is there. Or am i missed something?
--
You are receiving this mail
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #91 from Konstantin Belousov ---
(In reply to Cassiano Peixoto from comment #89)
See comment #52 for instructions.
--
You are receiving this mail because:
You are on the CC list for the bug.
___
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #90 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #87)
Eugene,
Sorry my delay but i was updating to 11.1-BETA3 r320450M.
On updated i've applied the following patches:
- libc/stdio from https://reviews.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #89 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #88)
Hi Konstantin,
rtld patch has been applied since them as i said on comment #67. Do you need
some additional debug info from core file?
--
You ar
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #88 from Konstantin Belousov ---
(In reply to Eugene Grosbein from comment #84)
I very much doubt that the rtld lock leak (?) can be caused by stdio bugs.
In other words, apply the debugging rtld patch I posted 06-16, and follo
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #87 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #85)
And there is no need to apply "syslog" patch to 11-STABLE as it has already
been merged.
--
You are receiving this mail because:
You are on the CC l
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #86 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #85)
Yes, that would be easiest thing. All these patches apply to 11-STABLE just
fine.
--
You are receiving this mail because:
You are on the CC list for
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #85 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #84)
Hi Eugene,
Don't worry, no problem.
I tried to apply Konstantin patch but i had some rejects:
# cat lib/libc/stdio/fgets.c.rej
@@ -53,17 +53,17 @@
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #84 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #83)
It seems you hit another problem in the libc/stdio. Konstantin produced another
patch for this problem:
https://reviews.freebsd.org/file/data/nthhi3og
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #83 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #81)
Eugene and Konstantin,
Bad news, it just stopped working. Eugene i hadn't enabled web server yet. So
it stucked the mpd5 process with patch applied li
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #82 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #81)
Humm i see. Maybe Konstantin Belousov could help us :)
--
You are receiving this mail because:
You are on the CC list for the bug.
__
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #81 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #80)
I cannot commit kernel patches myself as I have no src commit bit. Any src
committed is needed to take a look at least, so I've filled my PRs.
--
Yo
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #80 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #79)
Yes, i did. All my tests were with web server disabled as you requested. Only
with console enabled. I'll run with web server enabled to try the patch.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #79 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #78)
I believe, the libthr patch was for debugging purposes only and is not needed
to fix the problem itself.
Did you test it with web server disabled? If
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #78 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #77)
Hi guys,
After 8 days working with no issues i think at last it has been fixed. Well
done Eugene :)
I could see libc patch has been committed to 11-S
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #77 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #76)
All patches are pretty simple and should just work for 10.x
--
You are receiving this mail because:
You are on the CC list for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #76 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #75)
Eugene, are you going to provide all these patches to FreeBSD 10-STABLE?
--
You are receiving this mail because:
You are on the CC list for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Kubilay Kocak changed:
What|Removed |Added
Depends on||220151
Referenced Bugs:
https://
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #75 from Eugene Grosbein ---
(In reply to Kubilay Kocak from comment #74)
Done: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=220151
Konstantin Belousov (author of the patch) is in the re@, so this team should be
aware of
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #74 from Kubilay Kocak ---
(In reply to Eugene Grosbein from comment #73)
I order to ensure that (base) change is tracked and merged correctly, I'd
recommend (as suggested in comment 69) creating a separate dependent issue so
t
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #73 from Eugene Grosbein ---
(In reply to Kubilay Kocak from comment #72)
This is the patch https://bugs.freebsd.org/bugzilla/attachment.cgi?id=183537
My tests show no issues with it. It really helps and I'd like to see it
inc
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Kubilay Kocak changed:
What|Removed |Added
See Also||https://bugs.freebsd.org/bu
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #72 from Kubilay Kocak ---
(In reply to Eugene Grosbein from comment #71)
Is there an issue for the syslog related changes/patch discussed with kib@ on
freebsd-current?
--
You are receiving this mail because:
You are on the C
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #71 from Eugene Grosbein ---
IFYI2: another reason for this task to break and new set of patches here:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=220078
--
You are receiving this mail because:
You are on the CC list for
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #70 from Eugene Grosbein ---
JFYI: these patches made my mpd server side pretty stable, so my stress test
can run long enough to hit another problem that produces repeatable kernel
panics at client side. I've digged that problem
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Kubilay Kocak changed:
What|Removed |Added
Assignee|e...@freebsd.org |eu...@freebsd.org
--- Comment #69
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Eugene Grosbein changed:
What|Removed |Added
CC||dgilb...@eicat.ca
--- Comment #6
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #67 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #66)
Konstantin,
Worked after copied _umtx.h. Thanks.
Konstantin and Eugene,
I applied the patches, to sum up what i've applied:
- libc patch
- libt
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #66 from Konstantin Belousov ---
(In reply to Cassiano Peixoto from comment #65)
Perhaps you need to manually copy patched sys/_umtx.h to /usr/include/sys. Do
not forget to revert this when cleaning your machine.
--
You are r
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #65 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #64)
Konstantin,
Libc patch compiled fine after using 'make WITHOUT_TESTS=yes'. But libthr patch
doesn't, see:
cc -O2 -pipe -DPTHREAD_KERNEL -I/usr
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #64 from Konstantin Belousov ---
(In reply to Konstantin Belousov from comment #63)
Sorry, the correct spelling is 'make WITHOUT_TESTS=yes', no -D is needed.
--
You are receiving this mail because:
You are on the CC list for t
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #63 from Konstantin Belousov ---
(In reply to Cassiano Peixoto from comment #62)
Use 'make -DWITHOUT_TESTS=yes' to compile both libc and libthr.
--
You are receiving this mail because:
You are on the CC list for the bug.
_
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #62 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #57)
Eugene, i got the following error compiling libc patch:
===> tests (all)
===> tests/tls_dso (all)
===> tests/c063 (all)
(cd /usr/src/lib/libc/tests/c0
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #61 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #59)
Eugene, i think it's because web option was still enabled in mpd.conf. Anyway i
disabled it now to avoid any issues.
--
You are receiving this mail b
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #60 from Cassiano Peixoto ---
(In reply to Dmitry Luhtionov from comment #56)
Hi Dmitry,
Yes, i can. Please just provide the link and i can download and compile.
Thanks.
--
You are receiving this mail because:
You are on the
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #59 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #49)
Note: Thread 5 shows "http_server_accept" in the trace that means that mpd web
interface is being used.
--
You are receiving this mail because:
You
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #58 from Eugene Grosbein ---
Created attachment 183538
--> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=183538&action=edit
patch for mpd/console locks by me
--
You are receiving this mail because:
You are on the CC li
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #57 from Eugene Grosbein ---
Created attachment 183537
--> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=183537&action=edit
patch for lib/syslog by kib
(In reply to Cassiano Peixoto from comment #51)
I'd advise you sto
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #56 from Dmitry Luhtionov ---
Although I may be wrong.
Can you compile mpd5 from sources from sf.net ?
I can add some helpful debug to than.
--
You are receiving this mail because:
You are on the CC list for the bug.
___
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #55 from Dmitry Luhtionov ---
Restart only.
--
You are receiving this mail because:
You are on the CC list for the bug.
___
freebsd-net@freebsd.org mailing list
https://lists.freebsd
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #54 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #53)
Should i restart or recompile mpd5 after patch applied?
--
You are receiving this mail because:
You are on the CC list for the bug.
_
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #53 from Konstantin Belousov ---
Created attachment 183536
--> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=183536&action=edit
debugging patch to track rtld bind lock write owner
--
You are receiving this mail because
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #52 from Konstantin Belousov ---
(In reply to Cassiano Peixoto from comment #51)
Please apply the following patch, recompile libthr.so _only_, and then
reproduce the issue again. I need the same information: backtraces from all
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #51 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #50)
Please, see below:
(gdb) frame 1
#1 0x000802479c81 in __thr_rwlock_wrlock (rwlock=0x802694500, tsp=) at /usr/src/lib/libthr/thread/thr_umtx.c
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #50 from Konstantin Belousov ---
(In reply to Cassiano Peixoto from comment #49)
>From the thread 6, in the frame 1, please do 'p *rwlock'
--
You are receiving this mail because:
You are on the CC list for the bug.
___
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #49 from Cassiano Peixoto ---
(In reply to Konstantin Belousov from comment #47)
I'm using FreeBSD 11.0-p1. But it happens on any FreeBSD version above 9x.
Here is the backtrace for all threads:
(gdb) thread apply all bt
Thre
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #48 from Dmitry Luhtionov ---
In comment #31
(root@B-ras)~# file /usr/local/sbin/mpd5
/usr/local/sbin/mpd5: ELF 64-bit LSB executable, x86-64, version 1 (FreeBSD),
dynamically linked, interpreter /libexec/ld-elf.so.1, for FreeB
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #47 from Konstantin Belousov ---
(In reply to Konstantin Belousov from comment #46)
And exactly what FreeBSD version is it ?
--
You are receiving this mail because:
You are on the CC list for the bug.
_
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
Konstantin Belousov changed:
What|Removed |Added
CC||k...@freebsd.org
--- Comment
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #45 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #43)
Hi Eugene,
I just got a stuck process this morning with uwrlck state. As i said before i'm
not using mpd web anymore.
Here is the bt result after it
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #44 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #43)
Hi Eugene,
Thank you very much for your support. I stopped using mpd5 web, i replaced my
script with mpd5 console to collect statistics. Let's see if
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #43 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #42)
I've repeated my stress test for mpd5 and without RADIUS it runs just fine but
with RADIUS I'm able to reliably reproduce this problem in a minute usi
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #42 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #41)
I've just found out. My coredumpsize limit has been configured to 0Kb :(
Now it worked fine. Thanks for your help. I'll keep you posted. Until now it'
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #41 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #40)
The signal is delivered but core generation is disabled for some reason. To
find it, please show output of commands:
sysctl kern.coredump
proccontrol
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #40 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #39)
# dmesg | grep mpd5
pid 53402 (mpd5), uid 0: exited on signal 3
pid 53441 (mpd5), uid 0: exited on signal 3
pid 53758 (mpd5), uid 0: exited on signal
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #39 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #38)
Please show output of the command: dmesg | grep mpd5
--
You are receiving this mail because:
You are on the CC list for the bug.
___
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #38 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #37)
Same behaviour:
# ps ax | grep mpd5
53396 1 R+ 0:00.00 grep mpd5
# /usr/local/sbin/mpd5 -p /var/run/mpd5.pid -b
# ps ax | grep mpd5
53402 -
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #37 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #36)
Do you use tcsh for your login shell?
Anyway, try starting mpd using direct command "/usr/local/sbin/mpd5 -p
/var/run/mpd5.pid -b" for a test instead
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #36 from Cassiano Peixoto ---
(In reply to Eugene Grosbein from comment #35)
it's not running anymore, look:
# ps ax | grep mpd5
52264 1 S+ 0:00.00 grep mpd5
# /usr/local/etc/rc.d/mpd5 forcestart
Starting mpd5.
# ps ax
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=186114
--- Comment #35 from Eugene Grosbein ---
(In reply to Cassiano Peixoto from comment #34)
Does "killall" really terminate the process or does it continue to run?
Perhaps, the signal is blocked for some reason. There are several ways to star
1 - 100 of 119 matches
Mail list logo
