On 29. 09. 22 15:38, Mark Brown wrote:
On Thu, Sep 29, 2022 at 10:06:54AM +0200, Jaroslav Kysela wrote:
The driver field in the struct snd_ctl_card_info is a valid
user space identifier. Actually, many ASoC drivers do not care
and let to initialize this field using a standard wrapping method.
This breaks at least an arm multi_v7_defconfig build:
/build/stage/linux/sound/soc/soc-core.c: In function ‘snd_soc_bind_card’:
/build/stage/linux/sound/soc/soc-core.c:2055:36: error: ‘struct snd_soc_card’ ha
s no member named ‘dmi_longname’
2055 | if (card->long_name == card->dmi_longname)
| ^~
Also, it does not make sense to set the driver field to the
card name composed from DMI. This card name is longer in most
(all?) cases. Use a generic "ASoC-DMI" string here.
This should be a separate change, and DMI is a term specific to the
ACPI/EFI so I don't think we should be using it as a generic here, this
seems like a step back. If we want to make a change there I'd expect it
to be more picking the actual card driver name.
Thanks for the review. Yes, I made a mistake here. I wrongly mixed name and
long name strings in my head. I removed the DMI check and posted v2 of the patch.
Jaroslav
--
Jaroslav Kysela <perex@xxxxxxxx>
Linux Sound Maintainer; ALSA Project; Red Hat, Inc.
