[Archlinux] hplip 3.21.2-1 segfaults

Bug #1916874 reported by Mel34
10
This bug affects 1 person
Affects Status Importance Assigned to Milestone
HPLIP
New
Undecided
Unassigned

Bug Description

Details
Description:
hplip 3.21.2-1 in testing is segfaulting, effectively making printing impossible.
Downgrade to 3.20.11-2 from extra resolves it.
I own a HP 1102 printer.

Additional info:
* package version(s)3.21.2-1
* config and/or log files etc.
```
~ ▶ LANG=c journalctl -g cups
-- Journal begins at Wed 2021-02-24 10:57:27 CET, ends at Wed 2021-02-24 15:29:37 CET. --
Feb 24 10:57:29 loonix systemd[1]: Listening on CUPS Scheduler.
Feb 24 11:00:39 loonix systemd[1]: Starting CUPS Scheduler...
Feb 24 11:00:39 loonix systemd[1]: Started CUPS Scheduler.
Feb 24 14:48:37 loonix hpcups[6439]: prnt/hpcups/HPCupsFilter.cpp 489: m_Job initialization failed with error = 48
Feb 24 14:48:37 loonix kernel: hpcups[6439]: segfault at 560c44654 ip 00007f4003193a51 sp 00007ffd005b8810 error 4 cpu 1 in libc-2.33.so[7f4003132000+14c000]
Feb 24 14:48:37 loonix systemd-coredump[6443]: Process 6439 (hpcups) of user 209 dumped core.

Stack trace of thread 6439:
#0 0x00007f4003193a51 unlink_chunk.constprop.0 (libc.so.6 + 0x87a51)
#1 0x00007f4003193ec1 _int_free (libc.so.6 + 0x87ec1)
#2 0x00007f4003197ca8 __libc_free (libc.so.6 + 0x8bca8)
#3 0x00007f400286293f n/a (libtasn1.so.6 + 0xc93f)
#4 0x00007f4002ef0d1a n/a (libgnutls.so.30 + 0x60d1a)
#5 0x00007f40036b823b _dl_fini (ld-linux-x86-64.so.2 + 0x1023b)
#6 0x00007f400314b697 __run_exit_handlers (libc.so.6 + 0x3f697)
#7 0x00007f400314b83e exit (libc.so.6 + 0x3f83e)
#8 0x00007f4003133b2c __libc_start_main (libc.so.6 + 0x27b2c)
#9 0x0000560c444039ce n/a (hpcups + 0x89ce)
Feb 24 14:49:12 loonix hpcups[6489]: prnt/hpcups/HPCupsFilter.cpp 489: m_Job initialization failed with error = 48
Feb 24 14:49:12 loonix kernel: hpcups[6489]: segfault at 55c465e0e ip 00007fb6c5928a51 sp 00007fff40f54510 error 4 cpu 3 in libc-2.33.so[7fb6c58c7000+14c000]
Feb 24 14:49:12 loonix systemd-coredump[6493]: Process 6489 (hpcups) of user 209 dumped core.

Stack trace of thread 6489:
#0 0x00007fb6c5928a51 unlink_chunk.constprop.0 (libc.so.6 + 0x87a51)
#1 0x00007fb6c5928ec1 _int_free (libc.so.6 + 0x87ec1)
#2 0x00007fb6c592cca8 __libc_free (libc.so.6 + 0x8bca8)
#3 0x00007fb6c4ff793f n/a (libtasn1.so.6 + 0xc93f)
#4 0x00007fb6c5685d1a n/a (libgnutls.so.30 + 0x60d1a)
#5 0x00007fb6c5e4d23b _dl_fini (ld-linux-x86-64.so.2 + 0x1023b)
#6 0x00007fb6c58e0697 __run_exit_handlers (libc.so.6 + 0x3f697)
#7 0x00007fb6c58e083e exit (libc.so.6 + 0x3f83e)
#8 0x00007fb6c58c8b2c __libc_start_main (libc.so.6 + 0x27b2c)
#9 0x000055c4651b09ce n/a (hpcups + 0x89ce)
Feb 24 14:50:09 loonix hpcups[6599]: prnt/hpcups/HPCupsFilter.cpp 489: m_Job initialization failed with error = 48
Feb 24 14:50:09 loonix kernel: hpcups[6599]: segfault at 5625794cf ip 00007f6dc513aa51 sp 00007ffdb44c0520 error 4 cpu 2 in libc-2.33.so[7f6dc50d9000+14c000]
Feb 24 14:50:09 loonix systemd-coredump[6603]: Process 6599 (hpcups) of user 209 dumped core.

Stack trace of thread 6599:
#0 0x00007f6dc513aa51 unlink_chunk.constprop.0 (libc.so.6 + 0x87a51)
#1 0x00007f6dc513aec1 _int_free (libc.so.6 + 0x87ec1)
#2 0x00007f6dc513eca8 __libc_free (libc.so.6 + 0x8bca8)
#3 0x00007f6dc480993f n/a (libtasn1.so.6 + 0xc93f)
#4 0x00007f6dc4e97d1a n/a (libgnutls.so.30 + 0x60d1a)
#5 0x00007f6dc565f23b _dl_fini (ld-linux-x86-64.so.2 + 0x1023b)
#6 0x00007f6dc50f2697 __run_exit_handlers (libc.so.6 + 0x3f697)
#7 0x00007f6dc50f283e exit (libc.so.6 + 0x3f83e)
```

See also comment 1 here:
https://bugs.archlinux.org/task/69761

Revision history for this message
Mel34 (mel34) wrote :

Yo? Anybody? I literally cannot print with the latest release of hplip and no traction gained here whatsover?

Revision history for this message
zdohnal (zdohnal) wrote :

Hi Mel,

I'm Zdenek, RHEL/Fedora/CentOS HPLIP maintainer - I saw your bug here in emails from HPLIP launchpad (I check the upstream bugs for relevant bugs).

The segfault is a separate issue, which I cannot tell more about from the stacktraces - they're missing information where specifically (which line/variables used...) in hpcups segfault happened:

#8 0x00007fb6c58c8b2c __libc_start_main (libc.so.6 + 0x27b2c)
#9 0x000055c4651b09ce n/a (hpcups + 0x89ce)
-------------------------------------------

usually helps to install debuginfo packages and then investigate the coredump in coredumpctl.

But regarding the other error:

Feb 24 14:50:09 loonix hpcups[6599]: prnt/hpcups/HPCupsFilter.cpp 489: m_Job initialization failed with error = 48

that code refers to missing/old plugin according prnt/hpcups/CommonDefinitions.h (0x30 is 48 in decimal):

PLUGIN_LIBRARY_MISSING = 0x30,

so you can try to install/reinstall the plugin for now by 'hp-plugin'.

Revision history for this message
Mel34 (mel34) wrote :

Hi Zdenek, thank you for your reply.

AFAIK Archlinux strips debug symbols from their packages, so the stack trace sadly doesn't provide useful debug info.

Your reply pointed me in the right direction. I was able to get the printer to work with the updated "hp-plugin" which is "hplip-plugin" in AUR.

There are no more segfaults after updating the firmware files.

As such, this ticket can be closed.

Thank you again for your kind assistance.

Revision history for this message
zdohnal (zdohnal) wrote :

(In reply to Mel34, comment #3)
> As such, this ticket can be closed.

IMO hpcups should give you human-readable message about the issue and not RHEL/Fedora/CentOS maintainer, so the ticket can cover the improvement in that way.

Plus a segfault isn't the correct ending of process, regardless it hit an error before or not...

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.