On Mon, 10 Aug 2020 14:22:54 +0200, Mark Brown wrote: > > On Sat, Aug 08, 2020 at 10:07:36AM +0200, Takashi Iwai wrote: > > Takashi Iwai wrote: > > > > Does the patch below fix the bug? If so, it's rather a bug in the > > > commit cf6e26c71bfd ("ASoC: soc-component: merge > > > snd_soc_component_read() and snd_soc_component_read32()"). > > > That said, the commit cf6e26c71bfd dropped the capability of returning > > an error code from snd_soc_component_read() completely, while many > > code still expect an error gets returned. The assumption mentioned in > > the patch (the error can be ignored) looks too naive. > > I did an audit of the users when the series was posted and wasn't able > to turn up any code doing anything constructive with the return values, > but then once you're past probe error handling often makes things worse > if you try. This is the first one which actually seems to have had an > impact. > > > Morimoto-san, Mark, could you address it? IMO, we may still need two > > variants in the end again: the former snd_soc_component_read32() that > > returns the value directly and snd_soc_component_read() that returns 0 > > or an error. Only once after we deal with the error handling in each > > caller side, we can unify the read functions. > > I'm not sure if that specifically is what we need but yeah we should do > something, if it fixes things your change is certainly good for the > immediate problem so could you send it with a signoff please?
OK, will do soon later. > With the new code we do now have the core code printing an error message > if the I/O fails, before they were just being ignored more often than > not. This did turn up a couple of cases where drivers were relying on > being able to do things like silently read from registers that just > don't exist or aren't currently accessible without any diagnostics which > is it's own problem :/ (especially the volatile cases). Yeah, we may need some raw access helper for such a case... thanks, Takashi