|
Hello,
On Thu, 6 Dec 2001, Julian Anastasov wrote:
>
> On Thu, 6 Dec 2001, Wensong Zhang wrote:
>
> > No, it is my mistake. I don't know why I suddenly thought that network
> > softirq might be interrupted by a bottom half at a CPU last night. :)
> >
> > Just had a look at the do_softirq(), it disables bottom half before
> > calling softirqs. So, there is no need to disable bh again in mod_sltimer.
> > will remove the _bh change soon.
>
> Yes, do_softirq guarantees that before the bhs are disabled
> the CPU can't be preempted from IRQ. The softirq runs with irqs enabled
> and bhs disabled (softirqs, bhs and tasklets have same interrupt level,
> only hardirqs can preempt them on same CPU). May be this is the reason
> hardirq to preempt softirq (detected in mod_sltimer) but we can't say the
When mod_sltimer is preempted by hardirqs, after hardirq finishs, no
bottom half will be called until the softirq (mod_sltimer) finishs. It
shouldn't have problems.
> same for its bh. So, for me, the report for problem in mod_sltimer is
> suspicious. But we have to perform some tests.
>
Yes, it is suspicious.
Regards,
Wensong
|
