Red Hat Bugzilla – Bug 1465302
[abrt] [faf] gvfs: recompute_files(): /usr/libexec/gvfsd-network killed by 11
Last modified: 2017-11-14 02:45:43 EST
This bug has been created based on an anonymous crash report requested by the package maintainer.
Report URL: http://faf.lab.eng.brq.redhat.com/faf/reports/bthash/a9cbf0a62d67957ca7ba5c82d15087088da72ce2/
Note for QE, seems to be triggered by gnome-boxes or control-center tests.
Another faf that might be related is http://faf.lab.eng.brq.redhat.com/faf/reports/2791/
I think that it is a pretty late for 7.4 at this point. I am convinced that this bug isn't anything new, it is there since ever, you just probably changed your tests and it triggers this somehow currently. Thus I think it is not a blocker and it is ok to postpone this to 7.5.
I've already pushed upstream some patches for this crashes [1,2] and it seems it fixes the issue as per the retrace server. The patches from the first bug reports are already part of the rebased gvfs package, the patches from second are not... however the second is not a real fix, just workaround, because I have no idea how this could happen. I need definitely some reproducer to fix it properly...
Can you please provide full backtrace to be sure this is the second issue? Can you please try to figure out what is triggering this issue? May I download the mentioned tests somewhere?
And this i ok and expected, thanks Ondro.
*** Bug 1472371 has been marked as a duplicate of this bug. ***
Can you please provide me full backtrace of that crash? Can't you please figure out what is concretely causing those crashes and give me some reproducer?
*** Bug 1510937 has been marked as a duplicate of this bug. ***
Tentatively the following commit should workaround that issue (see Bug 1510937 for additional info):
But it would be nice to fix that properly, which requires info about reproducing... Tomas? Vladimir?
We can backport the patch anyway, so setting devel_ack+, but not sure it isn't too late at this point...
(In reply to Ondrej Holy from comment #8)
> Tentatively the following commit should workaround that issue (see Bug
> 1510937 for additional info):
> But it would be nice to fix that properly, which requires info about
> reproducing... Tomas? Vladimir?
> We can backport the patch anyway, so setting devel_ack+, but not sure it
> isn't too late at this point...
It's not late and I am quite unsure how to reproduce, but I can verify the fix in our automation. Could you give me a scratch build?
The ccpp file provided by Vladimir confirmed that this is caused by dereferencing error variable set to NULL and that the correct fix is the following commit from the same upstream bug report as the workaround:
It seems as per the FAF that there is only one report for RHEL 7.5, so a number of crashes have been probably reduced by the two already included fixes, so this bug is not probably anything urgent...
On host wlan-r2s33
under RHEL-7.5-20171110.n.0 Workstation x86_64
1. Reserve a test machine with a wireless and network controller.
2. Download the latest package of gvfs from Brew server.
3. Install the package and its dependencies.
4. Run automated tests from TCMS plan 22767
These were causing issues to 1.30.4-4.el7 and earlier.
Upgrading the component: gvfs
has fixed the bug.