Skip to contentSkip to conversation

Problem with arecord on Audigy 2 NX

DISCUSSION Yitao Duan 2 participants 4 messages Mar 26, 2004 – Mar 30, 2004

lore
Share & tools
lore

Yitao Duan

author
#0 lore raw

Hi,

I am having trouble recording using Creative Labs Sound Blaster Audigy 2 NX. The linux rec program worked fine. But whenever I use arecord (e.g. arecord test.wav), I got Segmentation fault. I pinpointed the problem to the function pcm_read in aplay.c. Specifically it was the line:

r = readi_func(handle, data, count);

that caused the fault. readi_func was called 3 times without problem. But the 4th call will cause the fault. gdb showed that a NULL pointer was encountered in function snd_pcm_rate_grab_next_period: rate->slave was NULL. The following is a stack trace when the fault happened. However, rate->slave was not NULL during the 3 normal readi_func calls and I checked that the handle passed to readi_func was always the same. What could happen that made the pointer NULL for the 4th call? I am running 2.4.21-4.EL kernel and gcc 3.2.3. Any information will be highly appreciated.

Thanks.

Yitao

Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1220721088 (LWP 7171)]
0xb758de5d in snd_pcm_rate_grab_next_period (pcm=0x8052d10, hw_offset=0) at pcm_rate.c:1047
1047 cont = rate->slave->buffer_size - slave_offset;
(gdb) bt
#0 0xb758de5d in snd_pcm_rate_grab_next_period (pcm=0x8052d10, hw_offset=0) at pcm_rate.c:1047
#1 0xb758e430 in snd_pcm_rate_avail_update (pcm=0x8052d10) at pcm_rate.c:1205
#2 0xb75782fc in snd_pcm_avail_update (pcm=0x0) at pcm.c:2087 #3 0xb757c236 in snd_pcm_read_areas (pcm=0x8052d10, areas=0xbffff3b0, offset=0, size=1000,

    func=0xb759159c <snd_pcm_mmap_read_areas>) at pcm.c:5968

#4 0xb759178f in snd_pcm_mmap_readi (pcm=0x8052d10, buffer=0x805b0a8, size=1000) at pcm_mmap.c:230
#5 0xb7576c21 in snd_pcm_readi (pcm=0x8052d10, buffer=0x805b0a8, size=4) at pcm_local.h:378
#6 0x0804d8d7 in pcm_read (

    data=0x805b0a8 "\200\200\200\200", '\177' <repeats 11 times>, '\200'

<repeats 24 times>, '\177' <repeats 12 times>, '\200' <repeats 29 times>, "\177\177\200\200\200\200\177\177\177\200\200\200\200\200", '\177' <repeats 30 times>, "\200", '\177' <repeats 75 times>..., rcount=1000) at aplay.c:1411
#7 0x0804eed5 in capture_go (fd=5, count=4294967294, rtype=2, name=0xbffffb50 "test") at aplay.c:2096
#8 0x0804f257 in capture (name=0xbffffb50 "test") at aplay.c:2191 #9 0x0804bc5e in main (argc=2, argv=0xbffff664) at aplay.c:801 (gdb) print rate
$1 = (snd_pcm_rate_t *) 0x8052d10
(gdb) print *rate
$2 = {slave = 0x0, close_slave = 0, watom = {begin = 12, end = 1}, appl_ptr = 0, hw_ptr = 6,
type = RATE_TYPE_BANDLIMIT, get_idx = 1, put_idx = 3, pitch = 1, func
= 0,
sformat = SND_PCM_FORMAT_U8, srate = 8000, states = 0x3e8, pareas =
0x1e848, sareas = 0x0}
(gdb) print rate->slave
$3 = (snd_pcm_t *) 0x0
(gdb) print slave_offset
$4 = 16000
(gdb) print rate->slave
$5 = (snd_pcm_t *) 0x0
(gdb) print rate->slave->buf

buf                buffer.0           buffer_size
buffer_time        bufsize.1
buf.0              buffer_frames      buffer_size.0

buffered_vfprintf
buffer buffer_pos buffer_size.1 buflen.1
(gdb) print rate->slave->buffer_size
Cannot access memory at address 0x88

-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click

#1 lore raw
Quoted 9 lines · Yitao Duan

Hi,

I am having trouble recording using Creative Labs Sound Blaster Audigy 2 NX. The linux rec program worked fine. But whenever I use arecord (e.g. arecord test.wav), I got Segmentation fault. I pinpointed the problem to the function pcm_read in aplay.c. Specifically it was the line:

r = readi_func(handle, data, count);

Quoted 9 lines

that caused the fault. readi_func was called 3 times without problem. But the 4th call will cause the fault. gdb showed that a NULL pointer was encountered in function snd_pcm_rate_grab_next_period: rate->slave was NULL. The following is a stack trace when the fault happened. However, rate->slave was not NULL during the 3 normal readi_func calls and I checked that the handle passed to readi_func was always the same. What could happen that made the pointer NULL for the 4th call? I am running 2.4.21-4.EL kernel and gcc 3.2.3. Any information will be highly appreciated.

Try alsa-lib-1.0.4rc1.

Jaroslav

-----
Jaroslav Kysela <[email protected]>
Linux Kernel Sound Maintainer
ALSA Project, SuSE Labs

-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click

author
#2 lore raw

Thanks for the info, Jaroslav. I upgraded all of the lib, driver, tools, and utils to 1.0.4rc1. But I got the same results: linux rec command worked fine but arecord caused Segmentation fault after readi_func was called 3 times (aplay also worked alright). Any idea what can be happening? Or is there anybody using the same card (Audigy 2 NX) without problem?

Thanks.

Yitao

Quoted 21 lines · Jaroslav Kysela
Quoted 9 lines · Yitao Duan

Hi,

I am having trouble recording using Creative Labs Sound Blaster Audigy 2 NX. The linux rec program worked fine. But whenever I use arecord (e.g. arecord test.wav), I got Segmentation fault. I pinpointed the problem to the function pcm_read in aplay.c. Specifically it was the line:

r = readi_func(handle, data, count);

Quoted 9 lines

that caused the fault. readi_func was called 3 times without problem. But the 4th call will cause the fault. gdb showed that a NULL pointer was encountered in function snd_pcm_rate_grab_next_period: rate->slave was NULL. The following is a stack trace when the fault happened. However, rate->slave was not NULL during the 3 normal readi_func calls and I checked that the handle passed to readi_func was always the same. What could happen that made the pointer NULL for the 4th call? I am running 2.4.21-4.EL kernel and gcc 3.2.3. Any information will be highly appreciated.

Quoted 8 lines

Try alsa-lib-1.0.4rc1.

Jaroslav

-----
Jaroslav Kysela <[email protected]>
Linux Kernel Sound Maintainer
ALSA Project, SuSE Labs

-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click

#3 lore raw
Quoted 5 lines · Yitao Duan

Thanks for the info, Jaroslav. I upgraded all of the lib, driver, tools, and utils to 1.0.4rc1. But I got the same results: linux rec command worked fine but arecord caused Segmentation fault after readi_func was called 3 times (aplay also worked alright). Any idea what can be happening? Or is there anybody using the same card (Audigy 2 NX) without problem?

Could you send me 'arecord -v' output?

Jaroslav

-----
Jaroslav Kysela <[email protected]>
Linux Kernel Sound Maintainer
ALSA Project, SuSE Labs

-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click