public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Jesse Barnes <jbarnes@virtuousgeek.org>
To: Jeff Garzik <jgarzik@pobox.com>
Cc: "Luck, Tony" <tony.luck@intel.com>,
	ak@suse.de, torvalds@osdl.org, Greg Edwards <edwardsg@sgi.com>,
	linux-kernel@vger.kernel.org, discuss@x86-64.org
Subject: Re: [2/3] Set compatibility flag for 4GB zone on IA64
Date: Sun, 11 Sep 2005 19:58:02 -0700	[thread overview]
Message-ID: <200509111958.02190.jbarnes@virtuousgeek.org> (raw)
In-Reply-To: <4324D70C.3090109@pobox.com>

On Sunday, September 11, 2005 6:17 pm, Jeff Garzik wrote:
> Luck, Tony wrote:
> >>Luck, Tony wrote:
> >>>ia64 isn't all that homogeneous.  SGI systems stuff *all* memory
> >>>into the DMA zone as their I/O devices have no 32-bit limits (just
> >>>as well really as there is no memory below 4G on an Altix!).
> >>
> >>SGI machines support random PCI cards, right?  If so, you
> >>cannot presume
> >>I/O devices have no 32-bit limits.
> >
> > No, SGI machines don't support random PCI cards.  The lowest
> > possible physical address in an Altix is 192GB.  Cards that
> > can only DMA to addresses below 4G aren't going to be very
> > useful, are they?
>
> Do the boxes have IOMMUs?

Yes, but they can only support devices that can do 32 bit or above 
addressing due to the fact that their IOMMU page tables are only 
referenced if the high bit of a 32 bit address is set, iirc.

Anyway, for those boxes, all memory was put in the DMA zone and the I/O 
mapping routines failed if devices that didn't support at least 32 bits 
tried to do a DMA map.

Jesse

  reply	other threads:[~2005-09-12  2:57 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-09-12  1:05 [2/3] Set compatibility flag for 4GB zone on IA64 Luck, Tony
2005-09-12  1:17 ` Jeff Garzik
2005-09-12  2:58   ` Jesse Barnes [this message]
  -- strict thread matches above, loose matches on Subject: below --
2005-09-11 20:44 Luck, Tony
2005-09-11 21:05 ` Andi Kleen
2005-09-11 23:24 ` Jeff Garzik
2005-09-12  2:09   ` Chris Wedgwood
2005-09-11 16:59 Andi Kleen

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=200509111958.02190.jbarnes@virtuousgeek.org \
    --to=jbarnes@virtuousgeek.org \
    --cc=ak@suse.de \
    --cc=discuss@x86-64.org \
    --cc=edwardsg@sgi.com \
    --cc=jgarzik@pobox.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=tony.luck@intel.com \
    --cc=torvalds@osdl.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