David Gibson's on April 17, 2019 10:47 am: > On Mon, Apr 15, 2019 at 03:06:43PM +1000, Nicholas Piggin wrote: >> It needs to be cleared at all vCPU dispatch points to SPEC, not just >> when calling H_CEDE as Ben's patch had. I think complexity would be >> significant for questionable benefit. Like the dispatch sequence, it >> seems like the test is trying to cover some race condition for the >> client but does not really do it well (and for Linux not necessary). >> >> prod bit is cleared after vCPU returns from preemption, so it can >> clear at any time and you can't rely on it, unless you look at >> dispatch sequence numbers to decipher if it was reset or not. >> >> KVM does implement something like the prodded flag as Ben's patch did >> but that's not to spec AFAIKS. > > Hm, ok. Maybe include this rationale in the comment here.
Done (hopefully) >> > I don't see any sign that H_JOIN is implemented in KVM, although >> > H_CONFER and H_PROD certainly are. >> >> H_JOIN is not. > > Right, so we shouldn't be trying to enable it. What will happen if we > use a KVM H_CONFGER and H_PROD along with a userspace H_JOIN? Yeah good question I'll have to test before sending the H_JOIN part again. May have to modify KVM to make that work properly. Thanks, Nick