Re: [Xen-devel] Re: ATI radeon fails with "iommu=soft swiotlb=force" (seen on RV730/RV740 and RS780/RS800) . Test commited for JF

[Boris Derzhavets <bderzhavets@xxxxxxxxx>]

Done via "git pull" for another close enough "linux-2.6-xen" clone :-
# git pull
# git checkout xen/master
# git log

commit 2a65ac0eea618596fb19a4ee36f8a4463aadad3d
Merge: 4907e6c... f68aa6d...
Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@xxxxxxxxxx>
Date:   Fri Oct 2 11:42:04 2009 -0700

    Merge branch 'xen/dom0/drm' into xen/master
   
    * xen/dom0/drm:
      drm: free vmalloc dma memory properly
      drm: make sure all pages in vmalloc area are really DMA-ready
      drm: recompute vma->vm_page_prot after changing vm_flags

commit f68aa6d4b10c84e5674aa4301623616f72e5b8a7
Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@xxxxxxxxxx>
Date:   Fri Oct 2 11:31:39 2009 -0700

    drm: free vmalloc dma memory properly
   
    Having allocated pages with drm_alloc_coherent, we need to free them with
    drm_free_coherent.  We're still not preserving the bus addresses so
    we're hoping the backend dma_ops implementation can cope with the ones
    we're making up.
   
    Derived from patch by Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
   
    Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@xxxxxxxxxx>

commit 93e7c3850b8431e19c9cba91413066bfd2360671
Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@xxxxxxxxxx>
Date:   Fri Oct 2 11:16:09 2009 -0700

    drm: make sure all pages in vmalloc area are really DMA-ready
   
    Under Xen, vmalloc_32() isn't guaranteed to return pages which are really
    under 4G in machine physical addresses (only in virtual pseudo-physical
    addresses).  To work around this, implement a vmalloc variant which
    allocates each page with dma_alloc_coherent() to guarantee that each
    page is suitable for the device in question.

. . . . . . .

Boris

--- On Fri, 10/2/09, Jeremy Fitzhardinge <jeremy@xxxxxxxx> wrote:

From: Jeremy Fitzhardinge <jeremy@xxxxxxxx>
Subject: Re: [Xen-devel] Re: ATI radeon fails with "iommu=soft swiotlb=force" (seen on RV730/RV740 and RS780/RS800)
To: "Boris Derzhavets" <bderzhavets@xxxxxxxxx>
Cc: dri-devel@xxxxxxxxxxxxxxxxxxxxx, xen-devel@xxxxxxxxxxxxxxxxxxx, JBeulich@xxxxxxxxxx, "Konrad Rzeszutek Wilk" <konrad.wilk@xxxxxxxxxx>
Date: Friday, October 2, 2009, 2:42 PM

On 10/02/09 10:23, Boris Derzhavets wrote:
> Jeremy,
> Please,  be aware of bugzilla.xensource.com [1519]  the most recent
> entries :-
>
> http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1519
>

Ah, OK.  I pushed a variant of Konrad's patches.  Could you try them out?

Thanks,
    J

Attachment: dmesg.jeremy.gz
Description: GNU Zip compressed data

Attachment: config.radeon.gz
Description: GNU Zip compressed data

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel