> >The i2c core has per-adapter locks, so no need to protect again.
> 
>    The core's lock is unable to protect from the IRQs. So I'm proposing to
> revert this patch. It's a pity I hadn't noticed this issue when the patch
> was posted.

Yes, you are right. I noticed a while ago, too, but then got
side-tracked :(

>    I'm afraid this unlock is misplaced, the code continues to access the
> registers.
> 
> >     ret = rcar_i2c_bus_barrier(priv);
> 
>    Should probably unlock here instead?

I can check details next week earliest. If you are confident with your
suggestions, feel free to send one patch with the revert and the updates
you mentioned.

Thanks,

   Wolfram

Attachment: signature.asc
Description: Digital signature

Reply via email to