On 20.11.18 12:33, Cornelia Huck wrote: > On Mon, 19 Nov 2018 18:25:25 +0100 > Michael Mueller <mimu@xxxxxxxxxxxxx> wrote: > >> Do not call __deliver_io() for adapter interruptions already >> pending in the IPM. That is a double effort. They will >> be processed as soon the vcpu control is given to SIE. >> >> Signed-off-by: Michael Mueller <mimu@xxxxxxxxxxxxx> >> --- >> arch/s390/kvm/interrupt.c | 54 ++++++++++++++++++++++------------------------- >> 1 file changed, 25 insertions(+), 29 deletions(-) > > I think this patch does what it says on the tin, but I'm a bit lost as > to the why. (Might make more sense with the gib.) > > Currently, we are trying to process any I/O interrupts, even if we'd > get them delivered via the gisa, when we're out of the SIE anyway. > IIRC, while this looks a bit like a belt-and-suspenders approach, it > also prevented performance problems when the vcpu did not go back into > the SIE immediately (it even may exit to userspace). > > Also, if you're ignoring the I/O interrupts pending in the ipm, you may > end up delivering interrupts with a lower priority (higher isc) first. > I'm not sure that's what we want. > > But maybe I'm just missing another bit of the code that makes this > safe. Can you elaborate a bit? > For interrupt priorities to work at least somewhat predictable, we should always try to inject all interrupts even if the HW would be doing it for us. In the order of priority. But we don't have the same thing for external calls injected via SCA. I remember that I once had a patch lying around a couple of years ago to fix that ... it went missing :) IO interrupt almost have lowest priority, so letting the HW inject them could be problematic when mixing IO interrupt priorities between SW and HW injected ones (hat Conny described). There are other corner cases if a e.g. a RESTART interrupt is pending for that CPU. We would deliver eventually the RESTART interrupt before delivering the IO interrupt, which would be wrong. -- Thanks, David / dhildenb
