[PATCH v2] iommu/dma: Rationalise types for DMA masks

[PATCH v2] iommu/dma: Rationalise types for DMA masks

[Robin Murphy]

Since iommu_dma_alloc_iova() combines incoming masks with the u64 bus
limit, it makes more sense to pass them around in their native u64
rather than converting to dma_addr_t early. Do that, and resolve the
remaining type discrepancy against the domain geometry with a cheeky
cast to keep things simple.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---

v2: Reworked from "iommu/dma: Use a better type for dma_limit"

 drivers/iommu/dma-iommu.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 0cc702a70a96..6e573d1cb8bf 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -399,7 +399,7 @@ static int dma_info_to_prot(enum dma_data_direction dir, bool coherent,
 }
 
 static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain,
-		size_t size, dma_addr_t dma_limit, struct device *dev)
+		size_t size, u64 dma_limit, struct device *dev)
 {
 	struct iommu_dma_cookie *cookie = domain->iova_cookie;
 	struct iova_domain *iovad = &cookie->iovad;
@@ -424,7 +424,7 @@ static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain,
 	dma_limit = min_not_zero(dma_limit, dev->bus_dma_limit);
 
 	if (domain->geometry.force_aperture)
-		dma_limit = min(dma_limit, domain->geometry.aperture_end);
+		dma_limit = min(dma_limit, (u64)domain->geometry.aperture_end);
 
 	/* Try to get PCI devices a SAC address */
 	if (dma_limit > DMA_BIT_MASK(32) && dev_is_pci(dev))
@@ -477,7 +477,7 @@ static void __iommu_dma_unmap(struct device *dev, dma_addr_t dma_addr,
 }
 
 static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
-		size_t size, int prot, dma_addr_t dma_mask)
+		size_t size, int prot, u64 dma_mask)
 {
 	struct iommu_domain *domain = iommu_get_dma_domain(dev);
 	struct iommu_dma_cookie *cookie = domain->iova_cookie;
-- 
2.23.0.dirty

_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [PATCH v2] iommu/dma: Rationalise types for DMA masks

[Christoph Hellwig]

On Wed, Dec 11, 2019 at 06:33:26PM +0000, Robin Murphy wrote:
> Since iommu_dma_alloc_iova() combines incoming masks with the u64 bus
> limit, it makes more sense to pass them around in their native u64
> rather than converting to dma_addr_t early. Do that, and resolve the
> remaining type discrepancy against the domain geometry with a cheeky
> cast to keep things simple.
> 
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>

Looks good to me:

Reviewed-by: Christoph Hellwig <hch@lst.de>

Joerg, let me know if you want to pick this up through the iommu tree as
it touches the iommu code, or through the dma-mapping tree that
introduced the warning.
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [PATCH v2] iommu/dma: Rationalise types for DMA masks

[Nathan Chancellor]

On Wed, Dec 11, 2019 at 06:33:26PM +0000, Robin Murphy wrote:
> Since iommu_dma_alloc_iova() combines incoming masks with the u64 bus
> limit, it makes more sense to pass them around in their native u64
> rather than converting to dma_addr_t early. Do that, and resolve the
> remaining type discrepancy against the domain geometry with a cheeky
> cast to keep things simple.
> 
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>

Tested-by: Nathan Chancellor <natechancellor@gmail.com> # build
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [PATCH v2] iommu/dma: Rationalise types for DMA masks

[Nicolas Saenz Julienne]

[-- Attachment #1.1: Type: text/plain, Size: 509 bytes --]

On Wed, 2019-12-11 at 18:33 +0000, Robin Murphy wrote:
> Since iommu_dma_alloc_iova() combines incoming masks with the u64 bus
> limit, it makes more sense to pass them around in their native u64
> rather than converting to dma_addr_t early. Do that, and resolve the
> remaining type discrepancy against the domain geometry with a cheeky
> cast to keep things simple.
> 
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>

Thanks!

Reviewed-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>


[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[-- Attachment #2: Type: text/plain, Size: 156 bytes --]

_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [PATCH v2] iommu/dma: Rationalise types for DMA masks

[Joerg Roedel]

On Wed, Dec 11, 2019 at 08:02:35PM +0100, Christoph Hellwig wrote:
> On Wed, Dec 11, 2019 at 06:33:26PM +0000, Robin Murphy wrote:
> > Since iommu_dma_alloc_iova() combines incoming masks with the u64 bus
> > limit, it makes more sense to pass them around in their native u64
> > rather than converting to dma_addr_t early. Do that, and resolve the
> > remaining type discrepancy against the domain geometry with a cheeky
> > cast to keep things simple.
> > 
> > Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> 
> Looks good to me:
> 
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> 
> Joerg, let me know if you want to pick this up through the iommu tree as
> it touches the iommu code, or through the dma-mapping tree that
> introduced the warning.

I'll take it through my tree, as I am about to collect fixes anyway.

Patch is now applied, thanks everyone.
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu