>>Why? Hardware semaphores are notorious for causing hangs. Nobody is
>>sharing the hardware under Linux, so I think we should enable access on
>>init, and not disable access until driver close. IMO the mixer should
>>be guarded by a Linux kernel semaphore... I have a patch from Thomas
>>Sailer (I think) lying around somewhere that does just that to the via
>>audio driver. Maybe we can adapt it.
>>(I cc'd this little detail, in my ali/trident.c patch review, to you)
>>
>>
>
>So add a timeout to it ?
>
>
There is a problem in via audio, that seems to be present in trident.c
too: trident_ioctl_mixdev doesn't protect the call to
codec->mixer_ioctl, which in turn can read and write to the AC97 codec.
I'm saying, (1) hardware semaphores are error prone and (2) are we using
the hardware sem to work around this lack of locking on ->mixer_ioctl?
Jeff
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/