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] Xen on machines with non-contiguous io-apics

To: "Steven Smith" <steven.smith@xxxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
Subject: Re: [Xen-devel] Xen on machines with non-contiguous io-apics
From: "Jan Beulich" <JBeulich@xxxxxxxxxx>
Date: Mon, 17 Aug 2009 10:44:45 +0100
Delivery-date: Mon, 17 Aug 2009 02:45:13 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <20090817091728.GA26646@xxxxxxxxxxxxxxxxxxxxxxxxxx>
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>
References: <20090817091728.GA26646@xxxxxxxxxxxxxxxxxxxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
>>> Steven Smith <steven.smith@xxxxxxxxxx> 17.08.09 11:17 >>>
>I have a test box with two IOAPICS which present non-contiguous GSIs
>according to the ACPI tables (the first offers 0-24 and the second
>32-56).  The latest xen-unstable with current pvops dom0 fails to boot
>on this machine, because the dom0 kernel can't claim any GSIs above
>48.  This is because the PHYSDEVOP_alloc_irq_vector implementation
>won't let it get at anything above nr_irqs, which is just the sum of
>the number of registers across all APICs.
>The patch below fixes this by just setting nr_irqs to the highest GSI
>offered by any IOAPIC.  I don't understand this stuff well enough to
>say whether that's the right fix (and the patch isn't suitable for
>applying, anyway), but it does make the machine boot.

This seems to mostly be in line with upstream Linux, except that it's lacking
the fallback for when ACPI is disabled. See acpi_probe_gsi() and its use
in 2.6.31-rcX.


Xen-devel mailing list

<Prev in Thread] Current Thread [Next in Thread>