linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: "Mark Powell" <medp@primagraphics.co.uk>
To: "LinuxPPC-Embedded" <linuxppc-embedded@lists.linuxppc.org>
Cc: "Chris Zimman" <chris@cryptoapps.com>
Subject: Re: PCI DMA memory problem
Date: Wed, 25 Jun 2003 13:48:34 +0100	[thread overview]
Message-ID: <0b4d01c33b18$16b65d90$3000a8c0@primag> (raw)
In-Reply-To: 20030624191935.GA9845@dingdong.cryptoapps.com


We found a problem with memory-memory DMA on 405GP. Some of our h/w runs
a standalone app directly from PPCBoot. We found that DMA didn't work
at all for physical addresses below 16MByte. This applies to the data
transfers and to the scatter/gather lists.

Chris, could you send me your PCI fixes. We tried some PCI cards in a
Walnut with the idea of building a server for native compilation, but
not much worked.

Mark
--
Mark Powell, Senior Software Engineer, Primagraphics Limited
New Cambridge House, Litlington, nr.Royston, Herts, SG8 0SS, UK
Tel. +44 1763 852222, Fax. 853324, http://www.primagraphics.co.uk

----- Original Message -----
From: "Chris Zimman" <chris@cryptoapps.com>
Sent: Tuesday, June 24, 2003 8:19 PM
Subject: Re: PCI DMA memory problem

On Tue, Jun 24, 2003 at 09:01:21PM +0200, Oliver Amft wrote:
>
> I am implementing a network driver for a DMA device on a custom
> 405GP/GPr based board. The networking device is attached over PCI bus.
> Kernel is currently 2.4.21. BIOS is PPCBoot 1.2.0.
>
> To map the data buffers from skb the driver uses pci_map_single().
> When starting the kernel with SDRAM size configurations other
> than 32MByte the device is unable to transfer data (currupted
> data) - descriptor dword read/write-back works (set up with
> pci_alloc_consistent()). For the kernel start option mem=32MB
> everything is fine (descriptor + correct data).

My first thought is -- is/are the SDRAM bank(s) configured correctly?

I don't know that it'd necessarily be related, but for the 405GP,
PPCBoot 1.2.0 doesn't have the PCI bridge setup quite right for all
PCI devices (eg. 82559 won't work).  I have code to fix the bridge
setup if you need it -- although I'm not even certain they exist in
the GPR.

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

  parent reply	other threads:[~2003-06-25 12:48 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-06-24 19:01 PCI DMA memory problem Oliver Amft
2003-06-24 19:19 ` Chris Zimman
2003-06-24 20:22   ` Wolfgang Denk
2003-06-25 12:48   ` Mark Powell [this message]
  -- strict thread matches above, loose matches on Subject: below --
2003-06-26 12:36 Oliver Amft

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='0b4d01c33b18$16b65d90$3000a8c0@primag' \
    --to=medp@primagraphics.co.uk \
    --cc=chris@cryptoapps.com \
    --cc=linuxppc-embedded@lists.linuxppc.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).