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] Re: [PATCH] xen: Pack some hvmop memory structures bette

To: "Tim Deegan" <Tim.Deegan@xxxxxxxxxx>
Subject: Re: [Xen-devel] Re: [PATCH] xen: Pack some hvmop memory structures better
From: "Jan Beulich" <JBeulich@xxxxxxxxxx>
Date: Mon, 07 Feb 2011 09:22:38 +0000
Cc: George Dunlap <George.Dunlap@xxxxxxxxxxxxx>, Paolo Bonzini <pbonzini@xxxxxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Mon, 07 Feb 2011 01:24:12 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <20110204163014.GB482@xxxxxxxxxxxxxxxxxxxxxxx>
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: <dffa1a0edc8c6e2aa879.1296833298@elijah> <4D4C2834.9070204@xxxxxxxxxx> <20110204163014.GB482@xxxxxxxxxxxxxxxxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
>>> On 04.02.11 at 17:30, Tim Deegan <Tim.Deegan@xxxxxxxxxx> wrote:
> At 16:24 +0000 on 04 Feb (1296836660), Paolo Bonzini wrote:
>> Am I missing something glaring, or this is breaking the ABI between 
>> hypervisor and kernels?
> It only breaks the ABI between xen and tools, which are supposed always
> to be version-matched.  I'm not sure that it's particularly worthwhile

According to x86's do_hvm_op() I would say a HVM guest can issue
HVMOP_set_mem_type for itself. Is that perhaps a mistake?

Similarly, HVMOP_[gs]et_mem_access bail on
current->domain->domain_id == a.domid, but
rcu_lock_target_domain_by_id() happily accepts DOMID_SELF
without any further checks. At least for the "set" variant this
very much looks like a mistake to me.

> though - at first glance it's saving about ten bytes of argument space.
> Unless these operations are happening in large batches?


Xen-devel mailing list

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