All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Jan Beulich" <JBeulich@novell.com>
To: kraxel@suse.de
Cc: xiaofeng.ling@intel.com, xen-devel@lists.xensource.com
Subject: Re: beginnings of allowing more than the basic 80x25 VGAscreen resolution
Date: Tue, 05 Jul 2005 08:11:31 -0600	[thread overview]
Message-ID: <s2ca4091.018@lucius.provo.novell.com> (raw)

>>> Gerd Knorr <kraxel@suse.de> 05.07.05 15:42:50 >>>
>On Tue, Jul 05, 2005 at 02:28:27PM +0100, Keir Fraser wrote:
>> 
>> On 5 Jul 2005, at 14:22, Jan Beulich wrote:
>> 
>> >>GRUB has code for switching back into real mode to do bios calls.
We
>> >>could take that, plus video.S, and put them in Xen. Then add an
info
>> >>structure for passing video info up to domain0. It's pretty ugly,
but
>> >>I think there is no nice solution here.
>
>Well, I'd simply leave that job completely to grub as it has to
>deal with that anyway and the multiboot spec have fields to pass
>on that info to the booted kernel.  Xen would simply take that
>and pass it on to the Dom0 kernel, thats it.  And maybe adjust
>it's own row/column counts in case it's vga text mode.

I'm not sure I follow here.

First of all, at least grub 0.9x (which is what I think is currently in
wide-spread use) doesn't seem to have even infrastructure to report
video information, not to think about changing the video mode. But I
agree that that'd be the best place for this to live.

Second, simply passing on the multi-boot info to the guest kernel would
still require that the interface between XEN and the guest kernel be
changed. And if a change is required, then passing the information in a
XEN-determined format seems more portable to me, especially if you
expect EFI to be used as boot loader at some point even on x86, because
then the information would perhaps come in yet another format (and
whatever Linux people might think, I believe that elilo shouldn't become
yet another intermediate layer for booting XEN then).

>> >> Would we also need vesa/vga
>> >>driver code from Linux for setting up font data in some graphics
>> >modes?
>> >
>> >Yes that would clearly be necessary, but perhaps a much more
>> >streamlined mechanism would do for XEN.
>> 
>> If that's the case, I'm sure Gerd can point out the 'moving parts'
that 
>> we actually need to get basic text output working. I was hoping
video.S 
>> alone would be sufficient... :-)
>
>Text mode is really simple, that is just another screen size
>than 80x25, but completely identical otherwise.  Font stuff is
>handled by bios and vga hardware, nothing to worry about here.

That builds on the assumption that the boot loader actually has the
necessary infrastructure. With the assumption that it's quite far away
from that, I would still think that XEN should try provide the ability
to change modes (at least until the boot loader actually does).

>In graphics mode you simply have a linear frame buffer and
>you'll have to do everything yourself.  Linux has compiled-in
>fonts for that.  Quite alot of stuff from drivers/video would
>be needed to get that going in xen, but I wouldn't even try to
>make xen render text in graphics mode and leave that completely
>to the dom0 kernel ...

Once XEN is so stable that you can afford not seeing its initial
messages, leaving the screen alone when in graphics mode is certainly an
option. But I don't think that's an option right now.

Jan

             reply	other threads:[~2005-07-05 14:11 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-07-05 14:11 Jan Beulich [this message]
     [not found] <s2ca4091.019@lucius.provo.novell.com>
2005-07-05 14:47 ` beginnings of allowing more than the basic 80x25 VGAscreen resolution Gerd Knorr
     [not found] <s2ca350f.097@lucius.provo.novell.com>
2005-07-05 13:28 ` Keir Fraser
2005-07-05 13:42   ` Gerd Knorr
  -- strict thread matches above, loose matches on Subject: below --
2005-07-05 13:22 Jan Beulich
2005-07-05 12:10 Ling, Xiaofeng
2005-07-05 12:58 ` Gerd Knorr
2005-07-05 13:07   ` Keir Fraser
2005-07-05 17:22     ` David Hopwood
2005-07-06  8:40       ` Keir Fraser

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=s2ca4091.018@lucius.provo.novell.com \
    --to=jbeulich@novell.com \
    --cc=kraxel@suse.de \
    --cc=xen-devel@lists.xensource.com \
    --cc=xiaofeng.ling@intel.com \
    /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.