Commit 298a96c1 authored by Ville Syrjälä's avatar Ville Syrjälä Committed by Thomas Gleixner

x86/dma-mapping: Fix arch_dma_alloc_attrs() oops with NULL dev

Commit 6894258e broke drivers that pass NULL as the device pointer
to dma_alloc. The reason is that arch_dma_alloc_attrs() now calls
dma_alloc_coherent_gfp_flags() which in turn calls
dma_alloc_coherent_mask(), where the device pointer is dereferenced
unconditionally.

Fix things by moving the ISA DMA fallback device assignment before the
call to dma_alloc_coherent_gfp_flags().

Fixes: 6894258e ("dma-mapping: consolidate dma_{alloc,free}_{attrs,coherent}")
Reported-and-tested-by: default avatarMeelis Roos <mroos@linux.ee>
Signed-off-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
Cc: Christoph Hellwig <hch@lst.de>
Link: http://lkml.kernel.org/r/1445807503-8920-1-git-send-email-ville.syrjala@linux.intel.comSigned-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
parent 32b88194
......@@ -131,11 +131,12 @@ void dma_generic_free_coherent(struct device *dev, size_t size, void *vaddr,
bool arch_dma_alloc_attrs(struct device **dev, gfp_t *gfp)
{
if (!*dev)
*dev = &x86_dma_fallback_dev;
*gfp &= ~(__GFP_DMA | __GFP_HIGHMEM | __GFP_DMA32);
*gfp = dma_alloc_coherent_gfp_flags(*dev, *gfp);
if (!*dev)
*dev = &x86_dma_fallback_dev;
if (!is_device_dma_capable(*dev))
return false;
return true;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment