From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paolo Bonzini Subject: Re: [PATCH v2 01/14] KVM: x86: change PIT discard tick policy Date: Thu, 18 Feb 2016 17:13:29 +0100 Message-ID: <56C5EDA9.9030204@redhat.com> References: <1455736496-374-1-git-send-email-rkrcmar@redhat.com> <1455736496-374-2-git-send-email-rkrcmar@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: kvm@vger.kernel.org, Yuki Shibuya , Rik van Riel To: =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= , linux-kernel@vger.kernel.org Return-path: In-Reply-To: <1455736496-374-2-git-send-email-rkrcmar@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On 17/02/2016 20:14, Radim Kr=C4=8Dm=C3=A1=C5=99 wrote: > Discard policy uses ack_notifiers to prevent injection of PIT interru= pts > before EOI from the last one. >=20 > This patch changes the policy to always try to deliver the interrupt, > which makes a difference when its vector is in ISR. > Old implementation would drop the interrupt, but proposed one injects= to > IRR, like real hardware would. This seems like what libvirt calls the "merge" policy: Merge the missed tick(s) into one tick and inject. The guest time may be delayed, depending on how the OS reacts to the merging of ticks where the merged tick is the one placed into IRR. Unlike discard, "merge" can starve the guest through an interrupt storm. Rik, I think you originally worked on the missed tick policies in Xen. Is the above correct? If so, do you recall what would be the reason to use the merge policy instead of the discard policy? Paolo