Messages in this thread | | | Date | Sat, 30 Aug 2008 15:44:47 +0100 | From | Alan Cox <> | Subject | Re: [PATCH 2/2] ne.c msleep not mdelay |
| |
> There are already some msleep() in probe_irq_on() which is called just > before here. And this part is not protected by any spinlock or > preempt_disable. > > So, if rescheduling was dangerous here, we already have potential > problems, no?
Yes we do but I didn't manage to stop the other mistakes getting in when they did. If you take a schedule you get results back from the probe_irq that tend to suck in other random ISA device events (ISA being edge triggered nobody was ever careful about spurious irq)
| |