On Wed, 09 Dec 2015 04:46:38 +0100, Valdis Kletnieks wrote:
I'm experiencing a spot of trouble with this commit:
commit aad730d0704545ad97654bd929b0aba05adb1436 Author: Takashi Iwai tiwai@suse.de Date: Mon Dec 2 13:33:57 2013 +0100
ALSA: hda - Always do delayed probes for HD-audio devicesThat adds some code:
#ifndef CONFIG_SND_HDA_I915 if (chip->driver_caps & AZX_DCAPS_I915_POWERWELL) dev_err(card->dev, "Haswell must build in CONFIG_SND_HDA_I915\n");
"Oh, bother" -- Winnie the Pooh
So I go looking how to fix this, and 'make menuconfig' tells me:
Symbol: SND_HDA_I915 [=n] x x Type : boolean x x Defined at sound/hda/Kconfig:8 x x Depends on: SOUND [=y] && !M68K && !UML && SND [=y] && DRM_I915 [=n] && SND_HDA_CORE [=y]
And why is DRM_I915=n?
# lspci | grep NV 01:00.0 VGA compatible controller: NVIDIA Corporation GF108GLM [NVS 5200M] (rev a1)
Because this Dell laptop has NVidia graphics onboard, and no sign of I915 graphics that needs I915 DRM.
Right, and this is exactly the purpose of this warning.
I think the infrastructure that you're trying to re-use from DRM_I915 needs to be moved to another symbol which can be SELECT/DEPEND by both DRM_I915 and SND_HSA_I915.
The HD-audio driver does soft-binding with i915, and the check above was for the case where i915 is *not* included although the audio driver tries binding. This check was mostly for Haswell/Broadwell HDMI/DP. Now we see this often on linux-next due to the recent change to allow i915 binding even on old PCHs.
So, the check itself is correct, per se, but it becomes superfluous in many cases now. I already addressed this, but forgot to submit / merge. The patches will be seen shortly later on alsa-devel ML.
thanks,
Takashi