From: Jesse Barnes <jbarnes@virtuousgeek.org>
To: linux-pci@atrey.karlin.mff.cuni.cz
Subject: Re: Weirdness in pci_read_bases()
Date: Mon, 3 Mar 2008 13:02:14 -0800
User-Agent: KMail/1.9.6 (enterprise 0.20071204.744707)
Cc: benh@kernel.crashing.org, Linux Kernel list <linux-kernel@vger.kernel.org>,
       Matthew Wilcox <matthew@wil.cx>
References: <1204252667.15052.402.camel@pasglop> <1204576954.21545.10.camel@pasglop> <200803031257.33249.jbarnes@virtuousgeek.org>
In-Reply-To: <200803031257.33249.jbarnes@virtuousgeek.org>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200803031302.14234.jbarnes@virtuousgeek.org>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1057
Lines: 21

On Monday, March 03, 2008 12:57 pm Jesse Barnes wrote:
> > Not only iommu's. On lots of platforms, we have the PCI MMIO space
> > mapped 1:N, that is, an access at MMIO location N turns into a PCI cycle
> > with address 0. So outbound PCI is effectively remapped, thus allowing
> > access to things like 0 etc... without punching holes in RAM and without
> > other dirty x86-like tricks.
>
> Right, bus address space is really a separate entity; I was just using
> IOMMUs as an example of that...

Err, wasn't really clear here.  CPU->PCI space and PCI->host space are really 
separate entities... IOMMUs will definitely affect the PCI->host mapping, and 
CPU->PCI mappings can be arbitrarily complex as well, though as you say on 
PCs it's typically just flatly mapped into CPU physical space somewhere.

Jesse
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/