On Thu, 22 Jul 2004 21:53:05 -0400, Ben Collins <bcollins@xxxxxxxxxx> wrote:
> On Fri, Jul 23, 2004 at 02:04:03AM +0100, Matthew Wilcox wrote:
> > On Thu, Jul 22, 2004 at 07:26:53PM -0400, Ben Collins wrote:
> > > Doesn't msleep() block? If so, that's not what we want at all. We want
> > > other things to happen during that time (like maybe catch a bus reset).
> >
> > No, msleep sleeps (as its name suggests). The code that implements it is:
>
> In that case, I've got a few places that could use msleep().
The one thing is, it depends on what you mean by "like maybe catch a
bus reset" (I think), because msleep will sleep through all signals,
i.e. it provides a guaranteed minimum delay. In the case of my patch,
I didn't see you checking for any signals afterwards (or any comments
indicating you were expecting some) so I figured it would be safe to
use msleep(). Also, I think it's clearer what's going on if the time
units for delays are msecs.
-Nish
PS. Would these other places be in drivers as well?
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@xxxxxxxxxxxxxx
http://lists.osdl.org/mailman/listinfo/kernel-janitors
|
