Bug#1013277: opendkim crash

2022-10-13 Thread Jan Korbel
Hello. Some news:

- it is Proxmox virtualisation (KVM
- tested kernels: linux-image-cloud-amd64, linux-image-amd64,
  linux-image-686-pae
- dbgsym packages was installed
- no coredumps this month, but:
- configured opendkim autorestart function:

AutoRestart yes
AutoRestartRate 10/1m

- autorestarted 6x on server1 and 1x on server2 - log:

opendkim[23781]: terminated with signal 6, restarting

Old backtraces with DEBUGINFOD_URLS:

BT1:

Core was generated by `/usr/sbin/opendkim -u opendkim -P
/run/opendkim/opendkim.pid'. Program terminated with signal SIGQUIT,
Quit. --Type  for more, q to quit, c to continue without paging--c
#0  0xf7f7b559 in __kernel_vsyscall ()
[Current thread is 1 (Thread 0xf6bad900 (LWP 16352))]
(gdb) bt full
#0  0xf7f7b559 in __kernel_vsyscall ()
No symbol table info available.
#1  0xf7695333 in fstab_init (opt_rewind=0) at fstab.c:121
state = 0xf7783380 
buffer = 0x0
fp = 
#2  0x1388 in ?? ()
No symbol table info available.
#3  0x in ?? ()
No symbol table info available.

BT2:

Program terminated with signal SIGABRT, Aborted.
--Type  for more, q to quit, c to continue without paging--c
#0  0xf7f19559 in __kernel_vsyscall ()
[Current thread is 1 (Thread 0xf3bffb40 (LWP 25976))]
(gdb) bt full
#0  0xf7f19559 in __kernel_vsyscall ()
No symbol table info available.
#1  0xf756ee02 in mrand48 () at mrand48.c:24
result = 0
#2  0xf3f2c118 in ?? ()
No symbol table info available.
#3  0x in ?? ()
No symbol table info available.

BT3:

Program terminated with signal SIGABRT, Aborted.
--Type  for more, q to quit, c to continue without paging--c
#0  0xf7f0f559 in __kernel_vsyscall ()
[Current thread is 1 (Thread 0xf3bffb40 (LWP 23979))]
(gdb) bt full
#0  0xf7f0f559 in __kernel_vsyscall ()
No symbol table info available.
#1  0xf7564e02 in mrand48 () at mrand48.c:24
result = 0
#2  0xf3efcd40 in ?? ()
No symbol table info available.
#3  0x in ?? ()
No symbol table info available.

BT4:

Program terminated with signal SIGABRT, Aborted.
--Type  for more, q to quit, c to continue without paging--c
#0  0xf7f54559 in __kernel_vsyscall ()
[Current thread is 1 (Thread 0xf3bffb40 (LWP 32056))]
(gdb) bt full
#0  0xf7f54559 in __kernel_vsyscall ()
No symbol table info available.
#1  0xf75a9e02 in mrand48 () at mrand48.c:24
result = 0
#2  0x0004 in ?? ()
No symbol table info available.
#3  0xf784ce28 in ?? () from /usr/lib/i386-linux-gnu/libunbound.so.8
No symbol table info available.
#4  0x in ?? ()
No symbol table info available.

Regards,

JK

On Wed, 21 Sep 2022 16:32:06 +0200
Bernhard Übelacker  wrote:

> Hello Jan,
> 
> 
> Maybe you could backtraces with better/more information by having the
> dbgsym packages installed, or calling gdb with the DEBUGINFOD_URLS
> from [10].
> 
> Is this some kind of virtualisation?
> Does this also happen with a regular amd64- or i386-kernel?
> 
> Kind regards,
> Bernhard
> 
> [10]
> https://wiki.debian.org/HowToGetABacktrace#Automatically_loading_debugging_symbol_from_the_Internet



Bug#1013277: opendkim crash

2022-09-21 Thread Bernhard Übelacker

Hello Jan,


Am 20.09.22 um 10:24 schrieb Jan Korbel:

1.
(gdb) bt
#0  __kernel_vsyscall ()
 at 
/build/linux-q1ABpl/linux-5.10.136/arch/x86/entry/vdso/vdso32/system_call.S:72
#1  0xf7695333 in fstab_init (opt_rewind=0) at fstab.c:121





2.
(gdb) bt
#0  __kernel_vsyscall ()
 at 
/build/linux-q1ABpl/linux-5.10.136/arch/x86/entry/vdso/vdso32/system_call.S:72
#1  0xf756ee02 in mrand48 () at mrand48.c:24
#2  0xf3f2c118 in ?? ()



3.
(gdb) bt
#0  __kernel_vsyscall ()
 at 
/build/linux-q1ABpl/linux-5.10.136/arch/x86/entry/vdso/vdso32/system_call.S:72
#1  0xf7564e02 in mrand48 () at mrand48.c:24
#2  0xf3efcd40 in ?? ()



4.
(gdb) bt
#0  __kernel_vsyscall ()
 at 
/build/linux-q1ABpl/linux-5.10.136/arch/x86/entry/vdso/vdso32/system_call.S:72
#1  0xf75a9e02 in mrand48 () at mrand48.c:24
#2  0x0004 in ?? ()



Maybe you could backtraces with better/more information by having the
dbgsym packages installed, or calling gdb with the DEBUGINFOD_URLS from [10].

Is this some kind of virtualisation?
Does this also happen with a regular amd64- or i386-kernel?

Kind regards,
Bernhard

[10] 
https://wiki.debian.org/HowToGetABacktrace#Automatically_loading_debugging_symbol_from_the_Internet



Bug#1013277: opendkim crash

2022-06-20 Thread David Bürgin
Jan Korbel:
> Package: opendkim
> Version: 2.11.0~beta2-4
> 
> Debian stable (11.3) i386 with 5.10.0-15-cloud-amd64 kernel.
> 
> Hello.
> 
> We have some crashes of opendkim.
> 
> In unstable changelog i found:
> 
> opendkim (2.11.0~beta2-5) unstable; urgency=medium
>   * Add new patches:
> - cve-2020-12272.patch: Mitigation for CVE-2020-12272.
> ...
> - mlfi_close.patch: Fix segfault due to use-after-free in
> mlfi_close.
> 
> Is there a plan to patch / rebuild for stable?

Not at this time.

Have you verified that the version in testing would fix your problem?
Can you provide logs, a backtrace of the crash, etc. …