]> git.karo-electronics.de Git - karo-tx-linux.git/commitdiff
drm/radeon/kms: fix for radeon on systems >4GB without hardware iommu
authorDaniel Haid <d.haid@gogi.tv>
Wed, 8 Jun 2011 10:04:45 +0000 (20:04 +1000)
committerGreg Kroah-Hartman <gregkh@suse.de>
Thu, 23 Jun 2011 22:05:35 +0000 (15:05 -0700)
commit 62fff811d73095bd95579d72f558f03c78f7914a upstream.

On my x86_64 system with >4GB of ram and swiotlb instead of
a hardware iommu (because I have a VIA chipset), the call
to pci_set_dma_mask (see below) with 40bits returns an error.

But it seems that the radeon driver is designed to have
need_dma32 = true exactly if pci_set_dma_mask is called
with 32 bits and false if it is called with 40 bits.

I have read somewhere that the default are 32 bits. So if the
call fails I suppose that need_dma32 should be set to true.

And indeed the patch fixes the problem I have had before
and which I had described here:
http://choon.net/forum/read.php?21,106131,115940

Acked-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
drivers/gpu/drm/radeon/radeon_device.c

index 890217e678d37f7591e7004bb4253e0f06f2d998..5d6ccc3efe33be8b1bddf64624e515c03367c637 100644 (file)
@@ -752,6 +752,7 @@ int radeon_device_init(struct radeon_device *rdev,
        dma_bits = rdev->need_dma32 ? 32 : 40;
        r = pci_set_dma_mask(rdev->pdev, DMA_BIT_MASK(dma_bits));
        if (r) {
+               rdev->need_dma32 = true;
                printk(KERN_WARNING "radeon: No suitable DMA available.\n");
        }