mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Robert Walsh <rjwalsh@durables.org>
To: Linux Kernel <linux-kernel@vger.kernel.org>
Subject: Question about dma_alloc_coherent and iommu
Date: Fri, 26 May 2006 11:55:18 -0700	[thread overview]
Message-ID: <1148669718.3665.11.camel@phosphene.durables.org> (raw)

[-- Attachment #1: Type: text/plain, Size: 1204 bytes --]

Hi there,

Our driver (the ipath driver in drivers/infiniband) uses
dma_alloc_coherent to allocate buffers that our device will DMA into.
We use a 64-bit DMA mask.  Although we don't set a consistent DMA mask,
it doesn't seem to matter whether we do or not for the following
problem.

Everything works fine on 2.6.16: when the device tells us it's DMA'd
into a buffer, the data is there and everyone is happy.  It also works
fine on kernels from 2.6.9 through 2.6.15, but only as long as there is
4GB or less of memory on the machine.  On our machines with > 4GB
memory, the buffers do not appear to have had anything DMA'd into them.

If we use "iommu=off" on the kernel boot line, the problem goes away and
the data starts appearing.

Has anyone seen this behaviour on other devices?  Are we missing
something in the device setup, maybe?  As I mentioned, setting a
consistent DMA mask doesn't help, no matter whether we set it to 64-bits
or 32-bits.  Setting the regular DMA mask to 32-bits doesn't help
either: the problem still crops up.

Regards,
 Robert.

-- 
Robert Walsh
Amalgamated Durables, Inc.  -  "We don't make the things you buy."
Email: rjwalsh@durables.org

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

                 reply	other threads:[~2006-05-26 18:55 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=1148669718.3665.11.camel@phosphene.durables.org \
    --to=rjwalsh@durables.org \
    --cc=linux-kernel@vger.kernel.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

all inboxes | Powered by JetHome®