On 21.11.18 13:19, Cornelia Huck wrote:
On Wed, 21 Nov 2018 09:30:34 +0100
Michael Mueller <mimu@xxxxxxxxxxxxx> wrote:
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?
Function kvm_s390_deliver_pending_interrupts() is called in the
beginning of vcpu_pre_run() and we are about to enter the SIE anyway.
SIE will also grant the right ISC priority for adapter interruptions.
I basically want to avoid that a GISA that is part of the alert list
will get its IPM bits cleared outside the GIB alert interruption
context. process_gib_alert_list()
Unfinished sentence?
Anyway, I think the patch description would benefit from adding a
sentence or two describing possible performance benefits and ease of
alert handling when the gib is introduced (?)
The patch is independent from the GIB related code. It shall be
active with the introduction of the GISA already. Yes, the GIB code
will benefit from this change as well but I don't like the forward
reference to upcoming patches if not really required.
I will update the commit message to:
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 or by
another SIE instance of the guest currently running.
