[alsa-devel] [PATCH] ASoC: soc-core: Fix null pointer dereference in soc_find_component

[Liam Girdwood]

On Sat, 2019-01-12 at 11:37 +0530, Rohit kumar wrote:
> On 1/12/2019 3:19 AM, Pierre-Louis Bossart wrote:
> > > Thanks for the overnight fix. This update fixes the issue on my 
> > > Skylake XPS13 test device (blind testing since I don't understand 
> > > what the code does).
> > > 
> > > Tested-by: Pierre-Louis Bossart <pierre-louis.bossart at linux.intel.com>
> > 
> > I need to take this back, this set of changes (initial+fix) causes an 
> > error with our HDMI support
> > 
> > [   17.437684] sof-audio sof-audio: created machine bxt-pcm512x
> > [   17.585279] bxt-pcm512x bxt-pcm512x: ASoC: failed to init link iDisp1
> > [   17.585639] bxt-pcm512x bxt-pcm512x: snd_soc_register_card failed -517
> > 
> > Removing your changes restores the functionality
> > 
> Looks like we should revert generic implementation for defering probe 
> and move to call from machine driver as done in v1.
> https://lore.kernel.org/patchwork/patch/1027560/
> https://lore.kernel.org/patchwork/patch/1027561/
> 
> @Mark, Do you have suggestion to refine current patch?
> > Adding some traces I can see that the the platform name we use doesn't 
> > seem compatible with your logic. All the Intel boards used a constant 
> > platform name matching the PCI ID, see e.g. [1], which IIRC is used to 
> > bind components. Liam, do you recall in more details if this is really 
> > required?

Sorry, I cant quite remember why the PCI ID was used for the platform name, I
think it started with the SKL generation as previous generations used "baytrail-
pcm" and "haswell-pcm" as platform names IIRC. Perhaps Vinod will know.

The platform name is only used by SOF when over writing the "legacy" platform
name e.g. "baytrail-pcm" would become "sof-audio" and this is only used for
binding DAI links (so that all legacy machine drivers can be reused without
modification). 

Liam