mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* Call Trace from snd_emu10k1
@ 2002-10-19 11:46 Jan Dittmer
  2002-10-19 14:30 ` Jaroslav Kysela
  0 siblings, 1 reply; 2+ messages in thread
From: Jan Dittmer @ 2002-10-19 11:46 UTC (permalink / raw)
  To: Linux Kernel List; +Cc: alsa-devel

Hello!

Got a couple of these with 2.5.44, while playing with mplayer with ac3 
passthrough (on the sb live).
Let me know if I can provide you with anything else helpful. Perhaps I should 
add, that the system isn't hanging, halting or anything else. Sound is 
flawless (at least with mplayer).

jan

Complete system info at http://lx.sfhq.hn.org/

Debug: sleeping function called from illegal context at mm/slab.c:1374
Call Trace:
 [<c0118d34>] __might_sleep+0x54/0x60
 [<c01359ae>] kmalloc+0x4e/0x120
 [<e0900910>] snd_emu10k1_fx8010_playback_irq+0x0/0x10 [snd-emu10k1]
 [<e09007bc>] snd_emu10k1_fx8010_register_irq_handler+0x6c/0x110 [snd-emu10k1]
 [<e0900d83>] snd_emu10k1_fx8010_playback_trigger+0x73/0x110 [snd-emu10k1]
 [<e0900910>] snd_emu10k1_fx8010_playback_irq+0x0/0x10 [snd-emu10k1]
 [<e08dfda9>] snd_pcm_start_R492990f9+0xa9/0x130 [snd-pcm]
 [<e08e807f>] snd_pcm_lib_write1+0x43f/0x4b0 [snd-pcm]
 [<e08e8216>] snd_pcm_lib_write_R61c057e4+0x126/0x130 [snd-pcm]
 [<e08e7b90>] snd_pcm_lib_write_transfer+0x0/0xb0 [snd-pcm]
 [<e08e2b86>] snd_pcm_playback_ioctl1+0x1a6/0x410 [snd-pcm]
 [<e08e31c0>] snd_pcm_playback_ioctl+0x20/0x30 [snd-pcm]
 [<c01522d9>] sys_ioctl+0x279/0x2d0
 [<c010709b>] syscall_call+0x7/0xb

00:07.5 Multimedia audio controller: VIA Technologies, Inc. VT82C686 AC97 
Audio Controller (rev 50)
	Subsystem: Micro-star International Co Ltd: Unknown device 3300
	Control: I/O+ Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- 
SERR- FastB2B-
	Status: Cap+ 66Mhz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- SERR-

00:0a.0 Multimedia audio controller: Creative Labs SB Live! EMU10k1 (rev 08)
	Subsystem: Creative Labs CT4832 SBLive! Value
	Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- 
SERR- FastB2B-
	Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- SERR-


#
# Advanced Linux Sound Architecture
#
CONFIG_SND=m
CONFIG_SND_SEQUENCER=m
# CONFIG_SND_SEQ_DUMMY is not set
CONFIG_SND_OSSEMUL=y
CONFIG_SND_MIXER_OSS=m
CONFIG_SND_PCM_OSS=m
CONFIG_SND_SEQUENCER_OSS=y
CONFIG_SND_RTCTIMER=m
CONFIG_SND_VERBOSE_PRINTK=y
CONFIG_SND_DEBUG=y
# CONFIG_SND_DEBUG_MEMORY is not set
# CONFIG_SND_DEBUG_DETECT is not set
CONFIG_SND_EMU10K1=m
CONFIG_SND_VIA82XX=m

Module                  Size  Used by    Not tainted
btaudio                12260   0
snd-rtctimer            2344   0 (unused)
snd-pcm-oss            51620   0
snd-mixer-oss          15256   0 [snd-pcm-oss]
snd-emu10k1            74064   4
snd-rawmidi            19296   0 [snd-emu10k1]
snd-pcm                90112   3 [snd-pcm-oss snd-emu10k1]
snd-timer              16728   0 [snd-rtctimer snd-pcm]
snd-hwdep               5600   0 [snd-emu10k1]
snd-ac97-codec         31652   0 [snd-emu10k1]
snd-util-mem            3468   0 [snd-emu10k1]
snd-seq-device          5968   0 [snd-emu10k1 snd-rawmidi]



^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Call Trace from snd_emu10k1
  2002-10-19 11:46 Call Trace from snd_emu10k1 Jan Dittmer
@ 2002-10-19 14:30 ` Jaroslav Kysela
  0 siblings, 0 replies; 2+ messages in thread
From: Jaroslav Kysela @ 2002-10-19 14:30 UTC (permalink / raw)
  To: Jan Dittmer; +Cc: Linux Kernel List, alsa-devel

On Sat, 19 Oct 2002, Jan Dittmer wrote:

> Hello!
> 
> Got a couple of these with 2.5.44, while playing with mplayer with ac3 
> passthrough (on the sb live).
> Let me know if I can provide you with anything else helpful. Perhaps I should 
> add, that the system isn't hanging, halting or anything else. Sound is 
> flawless (at least with mplayer).
> 
> jan
> 
> Complete system info at http://lx.sfhq.hn.org/
> 
> Debug: sleeping function called from illegal context at mm/slab.c:1374
> Call Trace:
>  [<c0118d34>] __might_sleep+0x54/0x60
>  [<c01359ae>] kmalloc+0x4e/0x120
>  [<e0900910>] snd_emu10k1_fx8010_playback_irq+0x0/0x10 [snd-emu10k1]
>  [<e09007bc>] snd_emu10k1_fx8010_register_irq_handler+0x6c/0x110 [snd-emu10k1]

This patch will help you:

--- emufx.c     11 Oct 2002 13:29:36 -0000      1.23
+++ emufx.c     19 Oct 2002 14:29:43 -0000
@@ -413,7 +413,7 @@

        snd_runtime_check(emu, return -EINVAL);
        snd_runtime_check(handler, return -EINVAL);
-       irq = kmalloc(sizeof(*irq), GFP_KERNEL);
+       irq = kmalloc(sizeof(*irq), GFP_ATOMIC);
        if (irq == NULL)
                return -ENOMEM;
        irq->handler = handler;


						Jaroslav

-----
Jaroslav Kysela <perex@suse.cz>
Linux Kernel Sound Maintainer
ALSA Project  http://www.alsa-project.org
SuSE Linux    http://www.suse.com


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2002-10-19 14:24 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-10-19 11:46 Call Trace from snd_emu10k1 Jan Dittmer
2002-10-19 14:30 ` Jaroslav Kysela

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®