Ok. I'll work on rebasing this evening, and then around midnight I'll update stable and start pumping changes into the head. There are still a handful of changes to, for example, the params stuff that we need to agree on, plus the warn with character array vs. warn with character pointer issue from a while ago.
Gabe Quoting nathan binkert <[EMAIL PROTECTED]>: > I actually have some outstanding diffs to this area of the code that > centralize the wakeup code. It would probably be best for us to > coordinate some of this. It probably makes sense for you to get all > of your o3 generalization diffs committed first though. > > Nate > > On Wed, Oct 8, 2008 at 12:52 AM, Gabe Black <[EMAIL PROTECTED]> wrote: > > I'm at the point where a timer interrupt needs to wake the CPU out > > of its idle thread. The idle thread uses a hlt instruction to suspend > > itself, and that ends up as a call to halt() on the exec context. I'm > > not sending interrupts through the older post, clear, etc., so I don't > > know if the CPU realizes it has some work to do and should stop being > > halted. I added a call to activate() in the interrupt object invoke > > method, but I don't think that'll help because the CPU's not checking > > its interrupts. What I probably need to do is to make the interrupts > > object call that function, but for that I'd need a way to get at the CPU > > it's part of, and know what thread to wake up. What's the right way to > > go about this? > > > > Gabe > > _______________________________________________ > > m5-dev mailing list > > [email protected] > > http://m5sim.org/mailman/listinfo/m5-dev > > > > > _______________________________________________ > m5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/m5-dev > _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
