This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] [PATCH 0/6] MSI-INTx interrupt translation for HVM

To: Shohei Fujiwara <fujiwara-sxa@xxxxxxxxxxxxxxx>, Qing He <qing.he@xxxxxxxxx>
Subject: Re: [Xen-devel] [PATCH 0/6] MSI-INTx interrupt translation for HVM
From: Keir Fraser <keir.fraser@xxxxxxxxxxxxx>
Date: Sun, 01 Mar 2009 14:55:21 +0000
Cc: "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Sun, 01 Mar 2009 06:55:50 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <20090227113753.6A27.CB716985@xxxxxxxxxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: AcmYhPR+5DnQj+3PS+2bfWh94pQpGQB+Mm9Z
Thread-topic: [Xen-devel] [PATCH 0/6] MSI-INTx interrupt translation for HVM
User-agent: Microsoft-Entourage/
On 27/02/2009 02:41, "Shohei Fujiwara" <fujiwara-sxa@xxxxxxxxxxxxxxx> wrote:

> I found hypervisor doesn't issue EOI before injecting the
> interrupt to guest domain, if MSI isn't maskable.
> The 2nd interrupt may be lost. What do you think about this?

The second interrupt should not be lost! It should be marked in the local
APIC's IRR bitmap and should hence cause an interrupt to fire as soon as the
in-service interrupt is EOIed (and hence cleared from the local APIC's ISR

Are you thus *sure* this is really the root-cause of your problem?

Could it be that the MSI *is* maskable, and that masking/unmasking loses
interrupt messages, for example?

 -- Keir

Xen-devel mailing list