Re: Regression with wait_event_timeout in next-20140226

[Stephen Rothwell]

Hi Peter,

On Wed, 26 Feb 2014 17:50:43 +0100 Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
>
> On Wed, Feb 26, 2014 at 05:35:19PM +0100, Gregory CLEMENT wrote:
> > 
> > while testing next-20140226 I got an issue with the function
> > wait_event_timeout. When this function timed out instead of returning
> > 0, it returned the value of the timeout passed in parameter. I found
> > that reverting "sched/wait: Suppress Sparse 'variable shadowing'
> > warning" fixed this regression.
> > 
> > I got this issue in the driver drivers/i2c/busses/i2c-mv64xxx.c.
> 
> Ah indeed. We actually rely on the shadowing for ___wait_cond_timeout().
> 
> We further used the __ret variable in __wait_event_timeout()'s cmd
> argument: __ret = schedule_timeout(__ret). That now explicitly uses the
> wrong __ret.
> 
> Yeah, we need to pull that patch.

I have reverted that commit from linux-next for today pending some other
resolution.

-- 
Cheers,
Stephen Rothwell                    sfr@xxxxxxxxxxxxxxxx

Attachment: pgpXGJ0wqVZ66.pgp
Description: PGP signature