[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Grant iomem access and map IRQs to a domU guest in Xen for ARM targets

On 06/16/2014 11:52 PM, Kapania, Ashish wrote:
>> -----Original Message-----
>> From: Dario Faggioli [mailto:dario.faggioli@xxxxxxxxxx]
>> Sent: Thursday, June 12, 2014 12:26 AM
>> To: Kapania, Ashish
>> Cc: Ian Campbell; xen-users@xxxxxxxxxxxxx; Arianna Avanzini
>> Subject: Re: [Xen-users] Grant iomem access and map IRQs to a domU
>> guest in Xen for ARM targets
>> On mer, 2014-06-11 at 22:00 +0000, Kapania, Ashish wrote:
>>> We are not tied to a particular Xen version just yet. I am working
>>> with the master branch (from a few weeks ago) i.e. xen4.5 dev branch
>> I
>>> believe.
>> That's what I was assuming, basing also on what you said yourself, but
>> I wasn't sure... Great! :-)
>>> I can try applying Arianna's patches.
>> Awesome. Let the list know how it goes.
> I tried Arianna's patches over the weekend. After applying the patches, I
> am able to use "iomem" field in xl.cfg to allow domU access to the
> specified memory ranges. So, the patches work for me.


Thank you for writing back with feedback.

> While trying the changes, I found that I need to have the device memory
> ranges declared in dom0's device tree in order to be able to map them to
> domU using "iomem". Otherwise, they are not in dom0's iomem_caps and I get
> an error. Just wanted to confirm that this is by design ?

Yes, with this patchset the I/O-memory ranges you want to make available to a
domU must be described in the dom0's device tree. This limitation of the current
implementation has been pointed out by Julien Grall in his comments to the v6
patchset ([1]).

As far as I remember, he is currently working on a follow-up to this series that
will allow to use the iomem option even on memory ranges that are not described
in the device tree ([2]). I believe that he will include it in his upcoming
patch series for device passthrough.

(I'm adding Julien Grall as Cc so that he can correct me if needed)

Thank you,

[1] http://markmail.org/message/vpfgajirmcxp66fj
[2] http://markmail.org/message/hai4bip5oeh2e3kv

> Best,
> Ashish

 * Arianna Avanzini
 * avanzini.arianna@xxxxxxxxx
 * 73628@xxxxxxxxxxxxxxxxxxx

Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.