All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dor Laor <dor.laor-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
To: Caleb Moore <cmoore-YbfuJp6tym7X/JP9YwkgDA@public.gmane.org>
Cc: kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org,
	Ghiora Drori
	<ghioradrori-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Subject: Re: Supporting Nvidia and ATI
Date: Mon, 12 Nov 2007 10:27:28 +0200	[thread overview]
Message-ID: <47380E70.30604@qumranet.com> (raw)
In-Reply-To: <1194833319.5040.33.camel-s4JzMB4ojaMt8xfdh1/dHeEGBmbFmoqDjBLMurt6cr7aRrIkyKz72Q@public.gmane.org>


[-- Attachment #1.1: Type: text/plain, Size: 2471 bytes --]

Caleb Moore wrote:
> On Sun, 2007-11-11 at 18:46 +0200, Ghiora Drori wrote:
>   
>> Hi,
>> I am interested in supporting NVDIA, ATI, Intel and presumably other 
>> display cards when running a windows guest KVM. The idea is to get 
>> Windows XP games to work properly under kvm when being hosted by
>> Linux. 
>> The screen on Linux can have more then one X windows server running on
>> different TTYs and it looks (I have not checked the code) like each is
>> running separately (aka there is a store and initialize when switching
>> between them) so when a kvm guest would get such a screen it would
>> have direct access to the display hardware. 
>> I searched Google but did not find anything significant. 
>> Any ideas where to start?
>>     
>
> For that it would probably be simplest to allow the VMM to write to the
> AGP/PCIe device.
>
> An graphics card hardware interface consists of two parts, a normal PCI
> device and a GART. The PCI interface is pretty simple, it's just a table
> containing useful information about memory regions, IRQs and IOports
> that the device will read/write to. You'll need to make an emulated PCI
> device that will map in real regions of the physical address space into
> the virtual machine. The GART is a little bit more complex, it will
> require a virtualised GART driver for your guest kernel that requests
> memory regions from the host kernel's GART interface, puts them in the
> guests address space and returns their addresses to the guest kernel.
>
> I hope that helps.
>
>
>   
Some of the work was released already by us and some will be released soon:
If you have 1-1 mapping between the guest addresses and the host you 
solve the gart mapping:
http://lkml.org/lkml/2007/10/21/125.
We also released pci passthrough and irq forwarding for qemu/kvm.
http://news.gmane.org/gmane.comp.emulators.kvm.devel/cutoff=9470
Together you can make it work (we have a passthrough NIC working).
Dor.

> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> kvm-devel mailing list
> kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/kvm-devel
>
>   


[-- Attachment #1.2: Type: text/html, Size: 3538 bytes --]

[-- Attachment #2: Type: text/plain, Size: 314 bytes --]

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[-- Attachment #3: Type: text/plain, Size: 186 bytes --]

_______________________________________________
kvm-devel mailing list
kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/kvm-devel

  parent reply	other threads:[~2007-11-12  8:27 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-11-11 16:46 Supporting Nvidia and ATI Ghiora Drori
     [not found] ` <76bfae930711110846s60e87f55n528679423df09875-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2007-11-11 21:46   ` Izik Eidus
2007-11-12  2:08   ` Caleb Moore
     [not found]     ` <1194833319.5040.33.camel-s4JzMB4ojaMt8xfdh1/dHeEGBmbFmoqDjBLMurt6cr7aRrIkyKz72Q@public.gmane.org>
2007-11-12  8:27       ` Dor Laor [this message]
     [not found]         ` <47380E70.30604-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-11-13  2:35           ` Ghiora Drori
2007-11-12  8:25   ` Amit Shah

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=47380E70.30604@qumranet.com \
    --to=dor.laor-re5jqeeqqe8avxtiumwx3w@public.gmane.org \
    --cc=cmoore-YbfuJp6tym7X/JP9YwkgDA@public.gmane.org \
    --cc=dor.laor-atKUWr5tajBWk0Htik3J/w@public.gmane.org \
    --cc=ghioradrori-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
    --cc=kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.