diff mbox series

[3/9] iommu/dma-iommu: Add iommu_dma_copy_reserved_iova, iommu_dma_apply_resv_region to the dma-iommu api

Message ID 20190411184741.27540-4-tmurphy@arista.com (mailing list archive)
State New, archived
Headers show
Series iommu/amd: Convert the AMD iommu driver to the dma-iommu api | expand

Commit Message

Tom Murphy April 11, 2019, 6:47 p.m. UTC
To convert the AMD iommu driver to the dma-iommu we need to wrap some of
the iova reserve functions.

Signed-off-by: Tom Murphy <tmurphy@arista.com>
---
 drivers/iommu/dma-iommu.c | 27 +++++++++++++++++++++++++++
 include/linux/dma-iommu.h |  7 +++++++
 2 files changed, 34 insertions(+)

Comments

Christoph Hellwig April 15, 2019, 6:31 a.m. UTC | #1
On Thu, Apr 11, 2019 at 07:47:32PM +0100, Tom Murphy via iommu wrote:
> +
> +	WARN_ON_ONCE(iommu_dma_reserve_iova(domain, region->start, end) == NULL);

Overly long line..
Tom Murphy April 16, 2019, 1:22 p.m. UTC | #2
I hoped this could be an exception, it's easier to grok without the
line break and isn't crazy long. Because you mentioned it I'll fix it.

On Mon, Apr 15, 2019 at 7:31 AM Christoph Hellwig <hch@infradead.org> wrote:
>
> On Thu, Apr 11, 2019 at 07:47:32PM +0100, Tom Murphy via iommu wrote:
> > +
> > +     WARN_ON_ONCE(iommu_dma_reserve_iova(domain, region->start, end) == NULL);
>
> Overly long line..
Robin Murphy April 16, 2019, 1:37 p.m. UTC | #3
On 16/04/2019 14:22, Tom Murphy wrote:
> I hoped this could be an exception, it's easier to grok without the
> line break and isn't crazy long. Because you mentioned it I'll fix it.

Frankly this patch is hard to justify anyway - iommu-dma already has its 
own reserved region handling, and there should be no need for external 
callers to be poking into the innards provided the IOMMU driver reports 
the correct reserved regions in the first place.

If the iommu-dma abstraction is not quite sufficient to actually convert 
amd-iommu to use it properly, then we should improve the abstraction, 
rather than just punching holes in it to merely poke renamed parts of 
the existing amd-iommu logic into.

Robin.

> On Mon, Apr 15, 2019 at 7:31 AM Christoph Hellwig <hch@infradead.org> wrote:
>>
>> On Thu, Apr 11, 2019 at 07:47:32PM +0100, Tom Murphy via iommu wrote:
>>> +
>>> +     WARN_ON_ONCE(iommu_dma_reserve_iova(domain, region->start, end) == NULL);
>>
>> Overly long line..
diff mbox series

Patch

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index cc5da30d6e58..613d181e78a9 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -162,6 +162,33 @@  void iommu_put_dma_cookie(struct iommu_domain *domain)
 }
 EXPORT_SYMBOL(iommu_put_dma_cookie);
 
+static struct iova *iommu_dma_reserve_iova(struct iommu_domain *domain,
+		dma_addr_t iova_lo, dma_addr_t iova_hi)
+{
+	struct iommu_dma_cookie *cookie = domain->iova_cookie;
+	struct iova_domain *iovad = &cookie->iovad;
+	unsigned long order = __ffs(domain->pgsize_bitmap);
+
+	return reserve_iova(iovad, iova_lo >> order, iova_hi >> order);
+}
+
+void iommu_dma_copy_reserved_iova(struct iova_domain *from,
+		struct iommu_domain *to)
+{
+	struct iommu_dma_cookie *cookie = to->iova_cookie;
+	struct iova_domain *to_iovad = &cookie->iovad;
+
+	copy_reserved_iova(from, to_iovad);
+}
+
+void iommu_dma_apply_resv_region(struct device *dev,
+		 struct iommu_domain *domain, struct iommu_resv_region *region)
+{
+	dma_addr_t end = region->start + region->length - 1;
+
+	WARN_ON_ONCE(iommu_dma_reserve_iova(domain, region->start, end) == NULL);
+}
+
 /**
  * iommu_dma_get_resv_regions - Reserved region driver helper
  * @dev: Device from iommu_get_resv_regions()
diff --git a/include/linux/dma-iommu.h b/include/linux/dma-iommu.h
index e760dc5d1fa8..c527ded5c41c 100644
--- a/include/linux/dma-iommu.h
+++ b/include/linux/dma-iommu.h
@@ -23,6 +23,7 @@ 
 #ifdef CONFIG_IOMMU_DMA
 #include <linux/dma-mapping.h>
 #include <linux/iommu.h>
+#include <linux/iova.h>
 #include <linux/msi.h>
 
 int iommu_dma_init(void);
@@ -57,6 +58,12 @@  dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page,
 int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg,
 		int nents, int prot);
 
+/* These are wrappers around some iova functions */
+void iommu_dma_apply_resv_region(struct device *dev,
+		struct iommu_domain *domain, struct iommu_resv_region *region);
+void iommu_dma_copy_reserved_iova(struct iova_domain *from,
+		struct iommu_domain *to);
+
 /*
  * Arch code with no special attribute handling may use these
  * directly as DMA mapping callbacks for simplicity