At Tue, 7 Oct 2008 17:11:18 +0800, Wu Fengguang wrote:
On Tue, Oct 07, 2008 at 11:00:26AM +0200, Takashi Iwai wrote:
At Tue, 7 Oct 2008 16:34:29 +0800,
FYI, this is the debug patch and kernel message(lookout the first line) on HP 2230s:
diff --git a/sound/core/init.c b/sound/core/init.c index 8af467d..704d7e8 100644 --- a/sound/core/init.c +++ b/sound/core/init.c @@ -706,6 +706,8 @@ int snd_component_add(struct snd_card *card, const char *component) return 1; } if (strlen(card->components) + 1 + len + 1 > sizeof(card->components)) {
snd_printk(KERN_ERR "components = %s + %s\n",card->components, component);[ 12.590150] ALSA sound/core/init.c:710: components = HDA:11d4194a,103c3037,00100400 HDA:11c11040,103c1378,00100200 + HDA:80862802,80860101,00100000 [ 12.590163] ------------[ cut here ]------------ [ 12.590166] WARNING: at sound/core/init.c:711 snd_component_add+0xfa/0x110 [snd]() [ 12.590168] BUG?
(snip)
Argh, that's bad. Then I agree with increasing the components as a simple workaround.
How about the patch below? The control protocol number is increased for this change as well.
The modified patch looks fine, I'll test it out in the following days.
So far grep gives the following referees:
./alsa-driver/acore/ioctl32/ioctl32_new.c: MAP_COMPAT(SNDRV_CTL_IOCTL_CARD_INFO), ./alsa-driver/acore/ioctl32/ioctl32_old.c: MAP_COMPAT(SNDRV_CTL_IOCTL_CARD_INFO), ./alsa-kernel/sound/core/control_compat.c: case SNDRV_CTL_IOCTL_CARD_INFO: ./alsa-kernel/sound/core/control.c: case SNDRV_CTL_IOCTL_CARD_INFO: ./alsa-kernel/include/sound/asound.h: SNDRV_CTL_IOCTL_CARD_INFO = _IOR('U', 0x01, struct snd_ctl_card_info), ./alsa-lib/src/control/control_shm.c: ctrl->cmd = SNDRV_CTL_IOCTL_CARD_INFO; ./alsa-lib/src/control/control_hw.c: if (ioctl(hw->fd, SNDRV_CTL_IOCTL_CARD_INFO, info) < 0) { ./alsa-lib/src/control/control_hw.c: SYSERR("SNDRV_CTL_IOCTL_CARD_INFO failed"); ./alsa-lib/aserver/aserver.c: case SNDRV_CTL_IOCTL_CARD_INFO: ./alsa-lib/include/sound/asound.h: SNDRV_CTL_IOCTL_CARD_INFO = _IOR('U', 0x01, struct sndrv_ctl_card_info),
Thanks. These are all just referring to the struct, so it must be OK. We need a similar fix for alsa-lib. I'll work on it now.
Takashi