xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: "Pasi Kärkkäinen" <pasik@iki.fi>
To: Ian Campbell <Ian.Campbell@citrix.com>
Cc: "xen-devel@lists.xen.org" <xen-devel@lists.xen.org>
Subject: Re: [PATCH] xl.cfg: gfx_passthru documentation improvements
Date: Tue, 4 Sep 2012 22:12:29 +0300	[thread overview]
Message-ID: <20120904191229.GH8912@reaktio.net> (raw)
In-Reply-To: <1346667053.25864.21.camel@zakaz.uk.xensource.com>

On Mon, Sep 03, 2012 at 11:10:53AM +0100, Ian Campbell wrote:
> On Sun, 2012-09-02 at 06:35 +0100, Pasi Kärkkäinen wrote:
> > Hello,
> > 
> > xl.cfg.pod.5 documentation improvements:
> > 
> > - gfx_passthru: Document gfx_passthru makes the GPU become primary in the guest
> >   and other generic info about gfx_passthru.
> > 
> > 
> > Signed-off-by: Pasi Kärkkäinen <pasik@iki.fi>
> > --- docs/man/xl.cfg.pod.5.orig1 2012-09-02 07:16:39.316782158 +0300
> > +++ docs/man/xl.cfg.pod.5       2012-09-02 08:26:40.081639087 +0300
> > @@ -968,7 +968,43 @@
> >  
> >  =item B<gfx_passthru=BOOLEAN>
> >  
> > -Enable graphics device PCI passthrough. XXX which device is passed through ?
> > +Enable graphics device PCI passthrough. This option makes the passthru
> > +graphics card become primary graphics card in the VM,
> 
> The option is slightly misnamed then I guess, a better name would have
> been gfx_passthru_primary? oh well, we are stuck with it now.
> 

Yeah.. I think earlier gfx_passthru wasn't a boolean, but an integer, to choose the mode.

> (Am I alone in thinking that "thru" is a horrible alternative to
> through?)
> 

:) 

> >  so the Qemu emulated 
> > +graphics adapter is disabled, and the VNC console for the VM won't have
> > +any graphics output. All graphics output, including boot time Qemu BIOS
> > +messages from the VM, will go to the physical outputs of the passed thru
> > +physical graphics card.
> > +
> > +Graphics card PCI device to passthru is chosen with B<pci> option, 
> > +exactly in the same way as normal Xen PCI device passthru/assignment is done. 
> > +Note that gfx_passthru doesn't do any kind of sharing
> > +of the GPU, so you can only assign the GPU to one single VM at a time.
> > +
> > +gfx_passthru also enables various legacy VGA memory ranges, BARs, MMIOs, 
> > +and ioports to be passed thru to the VM, since those are required
> > +for correct operation of things like VGA BIOS, text mode, VBE, etc.
> > +
> > +Enabling gfx_passthru option also copies the physical graphics card 
> > +video bios to the guest memory, and executes the vbios in the guest 
> 
> "BIOS" and (I expect) "VBIOS"?
> 

video bios == vbios, or did you ask something else? 


> > +to get the graphics card initialized.
> > +
> > +Most graphics adapters require vendor specific tweaks for properly
> > +working graphics passthru. Xen currently includes necessary tweaks
> > +for Intel IGD (Integrated Graphics Device) GPUs.  
> > +
> > +Support for AMD/Nvidia gfx_passthru is not yet merged to Xen,
> > +but there are out-of-tree patches available.
> 
> If there is a comprehensive list of supported/not-supported devices on
> the wiki then I think a reference here would be more useful than these
> two brief paragraphs.
> 

We have a list, but it's not up-to-date unfortunately:
http://wiki.xen.org/wiki/XenVGAPassthroughTestedAdapters


> In particular the second isn't very useful without links and those tend
> to become out of date in docs IME, I would tend to omit this from here
> and instead include them in the wiki where they can be kept up to date
>

Ok. I'll fix that, and add a link to:
http://wiki.xen.org/wiki/XenVGAPassthrough


> (or better yet somebody could finally submit the patches for inclusion).
> 

Indeed. I was talking to both AMD and Nvidia engineers at XenSummit NA 2012,
and there's some hope for them to do some work on this stuff :) 

And I volunteered earlier for this aswell, but so far I've been way too busy with other things :(


> > +gfx_passthru is currently only supported with the qemu-xen-traditional
> > +device-model. Upstream qemu-xen device-model currently doesn't have
> > +support for gfx_passthru.
> > +
> > +Note that some graphics adapters (AMD/ATI cards, for example) don't
> > +necessarily require gfx_passthru option, so you can use the normal
> > +Xen PCI passthru to assign the graphics card as a secondary graphics card
> > +to the VM. Qemu emulated graphics card stays as the primary graphics card, 
> > +and you get VNC output from the Qemu-emulated primary adapter.
> > +
> >  
> >  =item B<nomigrate=BOOLEAN>
> > 
> 


-- Pasi

  reply	other threads:[~2012-09-04 19:12 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-09-02  5:35 [PATCH] xl.cfg: gfx_passthru documentation improvements Pasi Kärkkäinen
2012-09-03 10:10 ` Ian Campbell
2012-09-04 19:12   ` Pasi Kärkkäinen [this message]
2012-09-05  8:33     ` Ian Campbell
2012-09-05 19:54       ` Pasi Kärkkäinen
2012-09-05 20:08       ` [PATCH v2] " Pasi Kärkkäinen
2012-09-10 10:15         ` Ian Campbell
2012-09-12 17:19           ` Pasi Kärkkäinen

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=20120904191229.GH8912@reaktio.net \
    --to=pasik@iki.fi \
    --cc=Ian.Campbell@citrix.com \
    --cc=xen-devel@lists.xen.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).