public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* PCI-DMA: high address but no IOMMU
@ 2005-10-27 17:47 Michael Madore
  2005-10-27 17:57 ` Nish Aravamudan
                   ` (2 more replies)
  0 siblings, 3 replies; 15+ messages in thread
From: Michael Madore @ 2005-10-27 17:47 UTC (permalink / raw)
  To: linux-kernel

Hi,

I am seeing the following errors in /var/log/messages when booting
2.6.14-rc5 on a dual Opteron nforce4 motherboard with 8GB of RAM:

Checking aperture...
CPU 0: aperture @ be8c000000 size 32 MB
Aperture from northbridge cpu 0 too small (32 MB)
No AGP bridge found
Your BIOS doesn't leave a aperture memory hole
Please enable the IOMMU option in the BIOS setup
This costs you 64 MB of RAM
Mapping aperture over 65536 KB of RAM @ 8000000

...

PCI-DMA: Disabling AGP.
PCI-DMA: More than 4GB of RAM and no IOMMU
PCI-DMA: 32bit PCI IO may malfunction.<6>PCI-DMA: Disabling IOMMU.

...

Attached scsi disk sda at scsi2, channel 0, id 0, lun 0
Kernel panic - not syncing: PCI-DMA: high address but no IOMMU.

With 2.6.13, the systems boots OK with these messages in the log:

Checking aperture...
CPU 0: aperture @ 8000000 size 32 MB
Aperture from northbridge cpu 0 too small (32 MB)
No AGP bridge found
Your BIOS doesn't leave a aperture memory hole
Please enable the IOMMU option in the BIOS setup
This costs you 64 MB of RAM
Mapping aperture over 65536 KB of RAM @ 8000000

...

PCI-DMA: Disabling AGP.
PCI-DMA: aperture base @ 8000000 size 65536 KB
PCI-DMA: Reserving 64MB of IOMMU area in the AGP aperture

Using git bisect, I narrowed the problem down to the following commit:

6142891a0c0209c91aa4a98f725de0d6e2ed4918 is first bad commit
diff-tree 6142891a0c0209c91aa4a98f725de0d6e2ed4918 (from
357e11d4cbbbb959a88a9bdbbf33a10f160b0823)
Author: Andi Kleen <ak@suse.de>
Date:   Mon Sep 12 18:49:24 2005 +0200

    [PATCH] x86-64: Avoid unnecessary double bouncing for swiotlb

    PCI_DMA_BUS_IS_PHYS has to be zero even when the GART IOMMU is disabled
    and the swiotlb is used. Otherwise the block layer does unnecessary
    double bouncing.

    Signed-off-by: Andi Kleen <ak@suse.de>
    Signed-off-by: Linus Torvalds <torvalds@osdl.org>

:040000 040000 d704d51d1e05f508f09c4388d3199cafbb9e3018
6a40204fd8b954634f2f7ea8f84861ed0a4bd88e M      include

Reverting this change allows 2.6.14-rc5 to boot.

Mike

^ permalink raw reply	[flat|nested] 15+ messages in thread
* RE: PCI-DMA: high address but no IOMMU - nForce4
@ 2005-10-31 17:50 Allen Martin
  0 siblings, 0 replies; 15+ messages in thread
From: Allen Martin @ 2005-10-31 17:50 UTC (permalink / raw)
  To: Matti Aarnio, Andi Kleen; +Cc: Michael Madore, linux-kernel

> This board has no AGP at all in it, but it does have lots
> of PCIE, and a bit of PCI-X thrown in for "legacy cards". 
> Somehow that detail breaks things when the machine really
> should use bounce-buffering, or something similar -- I don't
> know if Nvidia  nForce4 chipset does have IOMMU, though...
> 
> If Nvidia did omit such essential piece of hardware from
> a modern chipset, I do find it amazingly short-sighted...
> (Of course they don't yield documentation of the chips to
> public so that I can't quickly verify this detail...)

nForce4 does not contain an IOMMU, neither does any K8 chipset, because
the IOMMU is in the CPU.

-Allen

^ permalink raw reply	[flat|nested] 15+ messages in thread

end of thread, other threads:[~2005-12-18 20:07 UTC | newest]

Thread overview: 15+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-10-27 17:47 PCI-DMA: high address but no IOMMU Michael Madore
2005-10-27 17:57 ` Nish Aravamudan
2005-10-27 22:16 ` Andi Kleen
2005-10-27 22:39   ` Michael Madore
2005-10-28  1:59     ` Nishanth Aravamudan
2005-10-30 14:29       ` Nishanth Aravamudan
2005-10-30 14:59         ` Andi Kleen
2005-10-30 21:26           ` Nishanth Aravamudan
2005-10-30  0:27   ` PCI-DMA: high address but no IOMMU - nForce4 Matti Aarnio
2005-11-01  5:21     ` Marc Perkel
2005-11-02 15:38       ` Andreas Kleen
2005-11-06 12:35         ` Matti Aarnio
2005-11-01 16:35     ` Marc Perkel
2005-12-18 20:07 ` PCI-DMA: high address but no IOMMU (still there) Alex Bligh - linux-kernel
  -- strict thread matches above, loose matches on Subject: below --
2005-10-31 17:50 PCI-DMA: high address but no IOMMU - nForce4 Allen Martin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox