At Wed, 21 Dec 2011 15:23:14 +0800, Wu Fengguang wrote:
On Wed, Dec 21, 2011 at 02:46:32PM +0800, Takashi Iwai wrote:
At Wed, 21 Dec 2011 10:20:20 +0800, Wu Fengguang wrote:
Signed-off-by: Wu Fengguang fengguang.wu@intel.com
Do you really need this?
Yeah, the problem (now and future) is, our early hardware often have broken BIOS that does not get the pin-config right. However I understand that it's not sane to continue expanding the quick code.
I'm in the process removing all these model quirks.
Doesn't that risk introducing many regressions? What would be the problem if we switch to the generic HDA driver for *new* boards and leave the old working quicks/models untouched?
I've checked (manually) with hda-emu whether the new code works with the old static quirks before them. (It's also a part of the reasons why ALC880 and ALC882 model quirks still remain -- it takes time to verify all these old models.)
Basically if the BIOS provides the right pin-config, it should work with the auto-parser. If not, you can still provide the pin-config either statically or dynamically.
OK. Do you expect me to follow the example given by commit 2996bdbaa ("ALSA: hda - Remove ALC662 eeepc-p701 and ecs models")?
Yes, just add an entry with a static pin-config table for your device.
thanks,
Takashi
Thanks, Fengguang
sound/pci/hda/alc882_quirks.c | 1 + 1 file changed, 1 insertion(+)
--- linux.orig/sound/pci/hda/alc882_quirks.c 2011-12-20 14:16:51.880000058 +0800 +++ linux/sound/pci/hda/alc882_quirks.c 2011-12-20 14:16:58.740000115 +0800 @@ -2955,6 +2955,7 @@ static const struct snd_pci_quirk alc882 SND_PCI_QUIRK(0x8086, 0x0002, "DG33FBC", ALC883_3ST_6ch_INTEL), SND_PCI_QUIRK(0x8086, 0x2503, "82801H", ALC883_MITAC), SND_PCI_QUIRK(0x8086, 0x0022, "DX58SO", ALC889_INTEL),
- SND_PCI_QUIRK(0x8086, 0x7270, "Intel Los Lunas 2", ALC889_INTEL), SND_PCI_QUIRK(0x8086, 0x0021, "Intel IbexPeak", ALC889A_INTEL), SND_PCI_QUIRK(0x8086, 0x3b56, "Intel IbexPeak", ALC889A_INTEL), SND_PCI_QUIRK(0x8086, 0xd601, "D102GGC", ALC882_6ST_DIG),