Tvheadend

Hi,

While running tvheadend 4.3-949~gebd339120, some "Alignment trap" kernel messages have started to appear in /var/log/messages. They look like this:

Alignment trap: tvh:service (1412) PC=0x7f6c0674 Instr=0xe8840012 Address=0xb5a77d8f FSR 0xa21
Alignment trap: tvh:save (1395) PC=0x7f6c0674 Instr=0xe8840012 Address=0xb5a87c5b FSR 0xa21
Alignment trap: tvh:notify (1393) PC=0x7f6c20d8 Instr=0xe8840014 Address=0xb5a87c5b FSR 0xa21
Alignment trap: tvh:mi-table (1837) PC=0x7f6c0674 Instr=0xe8840012 Address=0x8c3c5757 FSR 0xa21
Alignment trap: tvh:epgdata (1824) PC=0x7f6c0674 Instr=0xe8840012 Address=0x8cfccadf FSR 0xa21

Afterwards, tvheadend crashed with a SIGSEGV. I have a core dump and I have tried looking at it in gdb. Despite the fact that I have the tvheadend-dbg package installed, gdb would only show a few levels of backtrace and then report a corrupt stack. For instance:

(gdb) thread 5
[Switching to thread 5 (Thread 0xb6aaa270 (LWP 1390))]
#0  0xb6b5d044 in __GI___libc_free (mem=0xb5a72b88) at /usr/src/debug/glibc/2.26-r0/git/malloc/malloc.c:3099
3099        = atomic_forced_read (__free_hook);
(gdb) bt
#0  0xb6b5d044 in __GI___libc_free (mem=0xb5a72b88) at /usr/src/debug/glibc/2.26-r0/git/malloc/malloc.c:3099
#1  0x7f67a6ac in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

At the moment I do not know how to make the backtrace work properly. The full output is attached. It looks like most of the 35 threads were in some wait or sleep state. From the remaining threads, one of the more interesting ones seems to be thread 5, which was trying to free some memory.