I'm pretty sure without this, and with DMAR, things explode
spectacularly. I got to this patch via bisecting broken i-g-t on my
machine.
If I had wired ethernet, I would figure out exactly what's failing with
netconsole. To be honest, nothing particularly stands out to me looking
over the code as to why it should fail though. If someone is more
inquisitive than I, I would be interested to know why we actually need
this bit of code. Not sure if things are broken without DMAR.
Daniel noticed this was accidentally dropped after I rebased my branch,
but figured it wasn't really needed. Seems it is.
Given that my patch:
commit 2ca466ae28f25e62090e4b57c90bcfee080a47a9
Author: Ben Widawsky <ben at bwidawsk.net>
Date: Fri Jan 18 12:30:33 2013 -0800
drm/i915: Needs_dmar, not
I'd suggest just squashing this into 2ca466ae28 since this was never
intentional anyway.
Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
Signed-off-by: Ben Widawsky <ben at bwidawsk.net>
---
drivers/gpu/drm/i915/i915_gem_gtt.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index f8fb7e3..a0ba4a9 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -721,6 +721,9 @@ int i915_gem_gtt_init(struct drm_device *dev)
return 0;
}
+ if (!pci_set_dma_mask(dev->pdev, DMA_BIT_MASK(40)))
+ pci_set_consistent_dma_mask(dev->pdev, DMA_BIT_MASK(40));
+
dev_priv->mm.gtt = kzalloc(sizeof(*dev_priv->mm.gtt), GFP_KERNEL);
if (!dev_priv->mm.gtt)
return -ENOMEM;
--
1.8.1.1
