Re: [PATCH/RFC] SCHED: allow wait_on_bit functions to support a timeout.

From: NeilBrown
Date: Wed Apr 30 2014 - 03:52:25 EST


On Wed, 30 Apr 2014 09:31:54 +0200 Peter Zijlstra <peterz@xxxxxxxxxxxxx>
wrote:

> On Wed, Apr 30, 2014 at 12:29:26PM +1000, NeilBrown wrote:
> > If you think it is a good cleanup I'll post a proper patch with all the right
> > Cc:s.
>
> Yeah, its a good cleanup. Thanks!
>
> > +static inline int
> > +wait_on_bit(void *word, int bit, unsigned mode)
> > +{
> > + if (!test_bit(bit, word))
> > + return 0;
> > + return out_of_line_wait_on_bit(word, bit,
> > + bit_wait,
> > + mode & 65535);
> > +}
> > +
> > +static inline int
> > +wait_on_bit_io(void *word, int bit, unsigned mode)
> > +{
> > + if (!test_bit(bit, word))
> > + return 0;
> > + return out_of_line_wait_on_bit(word, bit,
> > + bit_wait_io,
> > + mode & 65535);
> > +}
>
> That actually fits on one <80 line. Also, where does the 16 bit mask
> come from? On which, I would write that in hex, 0xFFFF is slightly
> easier to recognise as (1<<16)-1.

That is a hangover from an earlier attempt which didn't work. Thanks for
catching it.

I'll refresh and do some basic testing tomorrow and send it out, including to
Oleg.

Thanks,
NeilBrown

Attachment: signature.asc
Description: PGP signature