On 01/27/2015 06:34 AM, Catalin Marinas wrote:

On Tue, Jan 27, 2015 at 11:12:32AM +0000, Robin Murphy wrote:

quoted

On 23/01/15 22:32, Murali Karicheri wrote:

quoted

Limit the dma_mask to minimum of dma_mask and dma_base + size - 1.

Also arm_iommu_create_mapping() has size parameter of size_t and
arm_setup_iommu_dma_ops() can take a value higher than that. So
limit the size to SIZE_MAX.

Signed-off-by: Murali Karicheri<redacted>
---
   arch/arm/mm/dma-mapping.c |   10 ++++++++++
   1 file changed, 10 insertions(+)

diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c
index 7864797..a1f9030 100644
--- a/arch/arm/mm/dma-mapping.c
+++ b/arch/arm/mm/dma-mapping.c

@@ -2004,6 +2004,13 @@ static bool arm_setup_iommu_dma_ops(struct device *dev, u64 dma_base, u64 size,
   	if (!iommu)
   		return false;

+	/*
+	 * currently arm_iommu_create_mapping() takes a max of size_t
+	 * for size param. So check this limit for now.
+	 */
+	if (size>  SIZE_MAX)
+		return false;
+
   	mapping = arm_iommu_create_mapping(dev->bus, dma_base, size);
   	if (IS_ERR(mapping)) {
   		pr_warn("Failed to create %llu-byte IOMMU mapping for device %s\n",

@@ -2053,6 +2060,9 @@ void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
   {
   	struct dma_map_ops *dma_ops;

+	/* limit dma_mask to the lower of the two values */
+	*dev->dma_mask = min((*dev->dma_mask), (dma_base + size - 1));
+

Is there any reason not to do this in of_dma_configure? It seems like
something everyone could benefit from - I'd cooked up a dodgy workaround
for the same issue in my arm64 IOMMU code, but handling it generically
in common code would be much nicer.

Ok Will move this to of_dma_configure().

Murali

I agree. I started something here:

http://article.gmane.org/gmane.linux.kernel/1835096

but I don't remember to have got to a clear conclusion.

-- 
Murali Karicheri
Linux Kernel, Texas Instruments