cconf = devm_kcalloc(dev, li.conf, sizeof(*cconf), GFP_KERNEL);
- if (!dai_props || !dai_link || !dais)
+ if (!dai_props || !dai_link || !dais || !cconf)
return -ENOMEM;
It isn't mentioned on code, but, "li.conf" will be 0 if it was not DPCM case.
This means cconf NULL might be happen.
Humm, then don't call devm_kcalloc with a zero argument...
And if you keep this code, if li.conf is not zero you still have a
risk of memory allocation failure. the test should be
+ if (!dai_props || !dai_link || !dais || (!cconf && li.conf))
Yes, indeed.
Thank you for your feedback
Actually, I have few cleanup patches for simple-card/audio-graph
for v5.2, then, your patch and it will be conflict.
If you are OK, I can merge your patch on my patch-set with your
Signed-off-by (or Reported-by).
Is OK for you ? I will post it when -rc1 was released from Linus.
yes, that's fine. Thanks!
_______________________________________________
Alsa-devel mailing list
Alsa-devel@xxxxxxxxxxxxxxxx
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
