Menu
▾
▴
#2 asd segfaults and leaves dsp locked
I wanted to try the asd daemon so I compiled the asd-0.2
, run the asd daemon and run the esdplay program to
play some wav file through it. It played a part of it
and then crashed leaving the dsp device locked. So I
had to reboot the machine to unlock it. I'm using alsa
sound drivers so maybe it's another source of problems.
Here is the console output of the asd:
Message: INET: Success created server socket on port 4566.
Message: UNIX: Successfully created server socket
'/home/mraz/.asd-socket'.
Message: INET: Success created server socket on port 16001.
Message: UNIX: Successfully created server socket
'/tmp/.esd/socket'.
Message: INET: Success created server socket on port 4568.
Message: ORIGIN: Registered 0.0.
Message: ORIGIN: Registered 1.0.
Message: Queue length/hold set to 4/1.
Message: Queue length/hold set to 4/1.
Message: Queue length/hold set to 4/1.
Message: Server startup complete.
Message: DSP: Fragment sizes for OSS/ASD: 4096b/8192b
Message: DSP: SNDCTL_DSP_GET_OSPACE says fragment size:
4096; fragments: 2
Message: Source 'capt0' successfully opened.
Message: Sink 'play0' successfully opened.
Message: UNIX: Connection recieved.
Message: ESOUND: New client 'esound0' registered.
Message: ESOUND: Client wants to play a stream.
Message: ESOUND: Sample format is PCM, 2 channels, 16
bit, signed, LE, 44100 Hz.Message: ORIGIN: Registered 2.0.
Message: Queue length/hold set to 4/1.
Segmentation fault
Logged In: YES
user_id=4197
Additional comments:
when I've switched from ALSA to kernel driver for my card (I
have on board via82c686a chipset) it segfaults only
sometimes - especially when I switch the sample which I'm
playing using the esdplay utility.
But it has other problems. When I play some short sound
repeatedly, it plays it sometimes (each second try) wrong.
Or it doesn't play anything at all.
The short sample is 44100 Hz mono 16bit and the longer one
is 44100 Hz stereo 16bit.
Logged In: YES
user_id=47996
i'd appreciate it, when you'd run asd with gdb und give me a
back trace from the segfault:
gdb asd
[...]
run
[...]
bt
thanks in advance!
Logged In: YES
user_id=4197
Here is the backtrace from the crash with the kernel driver
(not ALSA):
#0 0x40038053 in g_mem_chunk_alloc () from
/usr/lib/libglib-1.2.so.0
#1 0x0804c67a in block_new () at block.c:51
#2 0x0804ebd8 in source_default_read_block_impl (s=0x807d138)
at source-default.c:118
#3 0x0804b1ae in _source_process (s=0x807d138) at source.c:238
#4 0x0804b32b in _source_thread (arg=0x807d138) at source.c:288
#5 0x40050c8e in pthread_start_thread_event
(arg=0xbbfffc00) at manager.c:274
I have glib-1.2.8-4 from Redhat 7.0
Logged In: YES
user_id=4197
Here is the backtrace with ALSA but this time its from the core.
#0 0x40148e75 in ?? ()
#1 0x400537e9 in ?? ()
#2 0x4004fb8c in ?? ()
#3 0x08051788 in notify_wait (n=0x807a1a8) at notify.c:28
#4 0x0804d358 in block_queue_push (b=0x807a178, e=0x806c108)
at block-queue.c:156
#5 0x0804af2f in source_output_write_block (s=0x8079e90,
b=0x806c108)
at source.c:133
#6 0x0804b18f in _source_process (s=0x8079e90) at source.c:235
#7 0x0804b32b in _source_thread (arg=0x8079e90) at source.c:288
#8 0x4005187d in ?? ()
Logged In: YES
user_id=4197
Another try (I've downgraded ALSA drivers to 0.5.11).
It segfaults only after playing many samples but with the
same backtrace as the ALSA 0.9.4 beta.
Also it plays wrong the short samples. It seems to me as it
plays something behind the sample memory because it
sometimes plays parts of longer samples which I've played
before.