Messages in this thread | | | Date | Thu, 3 Dec 2009 00:22:38 -0800 | From | Dmitry Torokhov <> | Subject | Re: Should we use preempt_disable() in sleep_on_common()? |
| |
On Thu, Dec 03, 2009 at 04:15:49PM +0800, liu pf wrote: > On Thu, Dec 3, 2009 at 3:12 PM, Dmitry Torokhov > <dmitry.torokhov@gmail.com> wrote: > > On Wed, Dec 02, 2009 at 11:07:46PM -0800, Arjan van de Ven wrote: > >> On Thu, 3 Dec 2009 14:49:14 +0800 > >> liu pf <kernelfans@gmail.com> wrote: > >> > >> > Hi: > >> > > >> > I am puzzled with the following scenario. Could anyone enlighten me? > >> > >> > >> sleep_on family of APIs is very racy and just cannot be used correctly; > >> I'm not surprised that there's a preempt race in it, but trust me, it's > >> not the biggest race... never ever use these APIs!!! > >> > >> > > > > BTW, why do we still have them? I checked couple and they don't seem to > > be used... > > > > -- > > Dmitry > > > > Hi, what is the substitution for sleep_on family of APIs? Any sample code? >
wait_event() and friends. Just look up any non-ancient driver.
-- Dmitry
| |