[mythtv-commits] Ticket #11318: Segfault in mythbackend (mainserver/autoexpire?)
MythTV
noreply at mythtv.org
Thu Jan 3 09:50:10 UTC 2013
#11318: Segfault in mythbackend (mainserver/autoexpire?)
--------------------------------+-----------------------------
Reporter: verycoldpenguin@… | Owner: stuartm
Type: Bug Report - Crash | Status: accepted
Priority: minor | Milestone: 0.26.1
Component: MythTV - General | Version: Master Head
Severity: medium | Resolution:
Keywords: | Ticket locked: 0
--------------------------------+-----------------------------
Comment (by verycoldpenguin@…):
Please do not use patch.txt (Patch 1).
Although it neatly stops the segfault, it completely renders useless the
two pieces of code :)
Hello Gregorio,
I am sorry, it looks like you might be right. I made the assumption that
if the protocol versions were different, then the master and slave
backends would not communicate. However, it appears that although the
slave may not connect to the master if the version is different, the
master can connect to the slave. Perhaps this should be listed as a
separate bug, or perhaps it is intentional, either way it is a pain in
this instance.
I have attached a patch2. This is an ugly patch, it has many debug/log
prints and areas commented out.
With this, it can be seen that the system correctly determines information
from the local database.
It also shows that often it does manage to communicate with the remote
backend (although again whilst this was running, there was a protocol mis-
match, 76 slave ->77 master).
A lot of the time the correct information was returned (the number of
lines being a multiple of 8).
However, often, the number of lines being returned (after the local
determination) is not a multiple of 8. This would suggest that the
information returned was truncated. In these instances, the patch2
supplied will ignore data that doesn't divide by 8. This masks the
symptoms of the problems, meaning that the segfault will not occur.
However, investigation should be made as to how to fix these truncations.
I will apply the patch from 11316 and see whether I see any more instances
of 11318. I think however that a refined version of patch2 would be
worthwhile incorporating anyway, in case communication problems occur from
elsewhere in the future.
--
Ticket URL: <http://code.mythtv.org/trac/ticket/11318#comment:5>
MythTV <http://code.mythtv.org/trac>
MythTV Media Center
More information about the mythtv-commits
mailing list