<div dir="ltr"><div><div><div><div><div><div>Alex, Christian, et al.<br><br></div>Thanks for your rapid response!<br><br></div>I can certainly make the change PPC-only (#ifdef CONFIG_PPC64...) and I think<br></div>that's the safest way to go right now. It makes the patch very specific to Cedar<br></div>on PowerPC64, so it doesn't break anything that may have come to depend on<br></div>Russell's and Alistair's changes, and my timetable for getting a fix in is very short.<br><br></div>-- Ben<br><br><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Feb 22, 2018 at 2:25 AM, Christian König <span dir="ltr"><<a href="mailto:ckoenig.leichtzumerken@gmail.com" target="_blank">ckoenig.leichtzumerken@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">Am 22.02.2018 um 03:05 schrieb Alex Deucher:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On Wed, Feb 21, 2018 at 6:41 PM, Ben Crocker <<a href="mailto:bcrocker@redhat.com" target="_blank">bcrocker@redhat.com</a>> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
In radeon_device_init, set the need_dma32 flag for Cedar chips<br>
(e.g. FirePro 2270). This fixes, or at least works around, a bug<br>
on PowerPC exposed by last year's commits<br>
<br>
8e3f1b1d8255105f31556aacf8aeb6<wbr>071b00d469 (Russell Currey)<br>
<br>
and<br>
<br>
253fd51e2f533552ae35a0c661705d<wbr>a6c4842c1b (Alistair Popple)<br>
<br>
which enabled the 64-bit DMA iommu bypass.<br>
<br>
This caused the device to freeze, in some cases unrecoverably, and is<br>
the subject of several bug reports internal to Red Hat.<br>
</blockquote>
Can we make this ppc only? 40 bit DMA works fine on x86.<br>
</blockquote>
<br></span>
Yeah and at least when the dma_coherent allocator path is used it should work fine on PPC as well.<br>
<br>
So that is not really a driver bug, but a platform bug and we should think about reverting or at least disabling those two patches instead.<br>
<br>
Christian.<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5">
<br>
Alex<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Signed-off-by: Ben Crocker <<a href="mailto:bcrocker@redhat.com" target="_blank">bcrocker@redhat.com</a>><br>
---<br>
drivers/gpu/drm/radeon/radeon_<wbr>device.c | 5 ++++-<br>
1 file changed, 4 insertions(+), 1 deletion(-)<br>
<br>
diff --git a/drivers/gpu/drm/radeon/radeo<wbr>n_device.c b/drivers/gpu/drm/radeon/radeo<wbr>n_device.c<br>
index ffc10cadcf34..02538903830d 100644<br>
--- a/drivers/gpu/drm/radeon/radeo<wbr>n_device.c<br>
+++ b/drivers/gpu/drm/radeon/radeo<wbr>n_device.c<br>
@@ -1395,7 +1395,10 @@ int radeon_device_init(struct radeon_device *rdev,<br>
if (rdev->flags & RADEON_IS_AGP)<br>
rdev->need_dma32 = true;<br>
if ((rdev->flags & RADEON_IS_PCI) &&<br>
- (rdev->family <= CHIP_RS740))<br>
+ (rdev->family <= CHIP_RS740 || rdev->family == CHIP_CEDAR))<br>
+ rdev->need_dma32 = true;<br>
+ if ((rdev->flags & RADEON_IS_PCIE) &&<br>
+ (rdev->family == CHIP_CEDAR))<br>
rdev->need_dma32 = true;<br>
<br>
dma_bits = rdev->need_dma32 ? 32 : 40;<br>
--<br>
2.13.6<br>
<br>
______________________________<wbr>_________________<br>
dri-devel mailing list<br>
<a href="mailto:dri-devel@lists.freedesktop.org" target="_blank">dri-devel@lists.freedesktop.or<wbr>g</a><br>
<a href="https://lists.freedesktop.org/mailman/listinfo/dri-devel" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/dri-devel</a><br>
</blockquote></div></div>
______________________________<wbr>_________________<br>
amd-gfx mailing list<br>
<a href="mailto:amd-gfx@lists.freedesktop.org" target="_blank">amd-gfx@lists.freedesktop.org</a><br>
<a href="https://lists.freedesktop.org/mailman/listinfo/amd-gfx" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/amd-gfx</a><br>
</blockquote>
<br>
</blockquote></div><br></div>