On 7/7/07, Bhanu Kalyan Chetlapalli <chbhanukalyan@xxxxxxxxx> wrote:
On 7/7/07, Rajat Jain <Rajat.Jain@xxxxxxxxxxxx> wrote: > > > > Hi, > > > And if you don't want the kernel to preempt you you can disable > > preemption while doing your sleeping... not very well behaved kernel > > code, but you can do it... > > Huh? So something like following is allowed? > > preempt_disable(); > udelay(5); > preempt_enable(); > > ????? > > I must admit, this is a shock to me. Hmmm even I think this is not possible, because these functions call the schedule() function, which should not be done in atomic context. It does a check for in_atomic() which checks the preempt_counter().
"these functions"? udelay() is a busy-loop, no sleeping, so it is fine in atomic contexts. Fine in the sense of allowed, not fine in the sense of good code (although there can be justified uses). Thanks, Nish -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
