[vdr] dxr3, segfault at start/end
prakashp at arcor.de
Sun Apr 17 13:27:04 CEST 2005
I didn't have this problem with vdr 1.3.6 and dxr3 0.2.2, but now with vdr
1.3.23 and dxr3-0.2.3-cvs20050323. I remember I had a similar one with vdr and
dxr plugin with versions in between, but I have removed those versions then...
1) If I start vdr with -Pdxr3 (and using linux-threads, using NPTL failure is
more probable) starting will usually die with "Ungültiger Maschinenbefehl", ie
unkown maschine instruction or alike. Sometime vdr manages to start though,
interesting enough, first try is usually succesful. (If I abort with ctrl-c
and try again I will have a hard time starting it again.)
I managed to get it started now more reliable: in dxr3memcpy.c I commented out
the benchmarking and simply set dxr3_memcpy = linux_kernel_memcpy;
as this is fastest anyway on that system. Now I have no problem starting, even
Interesting enough: If I start with unmodified dxr3, but using strace, vdr
starts up normally, as well.
So I think there is an issue in threading.
2) vdr won't correctly end. I will *always* get
*** glibc detected *** free(): invalid pointer: 0xb7baabc0 ***
Any idea, how to find out, who is the trouble maker? I just use dxr3 plugin
and vdr. Will gdb help? (Though I have nearly no experience using gdb...)
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 189 bytes
Desc: OpenPGP digital signature
Url : http://www.linuxtv.org/pipermail/vdr/attachments/20050417/314c4858/signature.pgp
More information about the vdr