Merge tag 'iommu-updates-v3.8' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu

Pull IOMMU updates from Joerg Roedel:
 "A few new features this merge-window.  The most important one is
  probably, that dma-debug now warns if a dma-handle is not checked with
  dma_mapping_error by the device driver.  This requires minor changes
  to some architectures which make use of dma-debug.  Most of these
  changes have the respective Acks by the Arch-Maintainers.

  Besides that there are updates to the AMD IOMMU driver for refactor
  the IOMMU-Groups support and to make sure it does not trigger a
  hardware erratum.

  The OMAP changes (for which I pulled in a branch from Tony Lindgren's
  tree) have a conflict in linux-next with the arm-soc tree.  The
  conflict is in the file arch/arm/mach-omap2/clock44xx_data.c which is
  deleted in the arm-soc tree.  It is safe to delete the file too so
  solve the conflict.  Similar changes are done in the arm-soc tree in
  the common clock framework migration.  A missing hunk from the patch
  in the IOMMU tree will be submitted as a seperate patch when the
  merge-window is closed."

* tag 'iommu-updates-v3.8' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu: (29 commits)
  ARM: dma-mapping: support debug_dma_mapping_error
  ARM: OMAP4: hwmod data: ipu and dsp to use parent clocks instead of leaf clocks
  iommu/omap: Adapt to runtime pm
  iommu/omap: Migrate to hwmod framework
  iommu/omap: Keep mmu enabled when requested
  iommu/omap: Remove redundant clock handling on ISR
  iommu/amd: Remove obsolete comment
  iommu/amd: Don't use 512GB pages
  iommu/tegra: smmu: Move bus_set_iommu after probe for multi arch
  iommu/tegra: gart: Move bus_set_iommu after probe for multi arch
  iommu/tegra: smmu: Remove unnecessary PTC/TLB flush all
  tile: dma_debug: add debug_dma_mapping_error support
  sh: dma_debug: add debug_dma_mapping_error support
  powerpc: dma_debug: add debug_dma_mapping_error support
  mips: dma_debug: add debug_dma_mapping_error support
  microblaze: dma-mapping: support debug_dma_mapping_error
  ia64: dma_debug: add debug_dma_mapping_error support
  c6x: dma_debug: add debug_dma_mapping_error support
  ARM64: dma_debug: add debug_dma_mapping_error support
  intel-iommu: Prevent devices with RMRRs from being placed into SI Domain
  ...

27 files changed, 477 insertions, 272 deletions

diff --git a/Documentation/DMA-API-HOWTO.txt b/Documentation/DMA-API-HOWTO.txt
index a0b6250add79..4a4fb295ceef 100644
--- a/Documentation/DMA-API-HOWTO.txt
+++ b/Documentation/DMA-API-HOWTO.txt
@@ -468,11 +468,46 @@ To map a single region, you do:
         size_t size = buffer->len;
 
         dma_handle = dma_map_single(dev, addr, size, direction);
+        if (dma_mapping_error(dma_handle)) {
+                /*
+                 * reduce current DMA mapping usage,
+                 * delay and try again later or
+                 * reset driver.
+                 */
+                goto map_error_handling;
+        }
 
 and to unmap it:
 
         dma_unmap_single(dev, dma_handle, size, direction);
 
+You should call dma_mapping_error() as dma_map_single() could fail and return
+error. Not all dma implementations support dma_mapping_error() interface.
+However, it is a good practice to call dma_mapping_error() interface, which
+will invoke the generic mapping error check interface. Doing so will ensure
+that the mapping code will work correctly on all dma implementations without
+any dependency on the specifics of the underlying implementation. Using the
+returned address without checking for errors could result in failures ranging
+from panics to silent data corruption. Couple of example of incorrect ways to
+check for errors that make assumptions about the underlying dma implementation
+are as follows and these are applicable to dma_map_page() as well.
+
+Incorrect example 1:
+        dma_addr_t dma_handle;
+
+        dma_handle = dma_map_single(dev, addr, size, direction);
+        if ((dma_handle & 0xffff != 0) || (dma_handle >= 0x1000000)) {
+                goto map_error;
+        }
+
+Incorrect example 2:
+        dma_addr_t dma_handle;
+
+        dma_handle = dma_map_single(dev, addr, size, direction);
+        if (dma_handle == DMA_ERROR_CODE) {
+                goto map_error;
+        }
+
 You should call dma_unmap_single when the DMA activity is finished, e.g.
 from the interrupt which told you that the DMA transfer is done.
 
@@ -489,6 +524,14 @@ Specifically:
         size_t size = buffer->len;
 
         dma_handle = dma_map_page(dev, page, offset, size, direction);
+        if (dma_mapping_error(dma_handle)) {
+                /*
+                 * reduce current DMA mapping usage,
+                 * delay and try again later or
+                 * reset driver.
+                 */
+                goto map_error_handling;
+        }
 
         ...
 
@@ -496,6 +539,12 @@ Specifically:
 
 Here, "offset" means byte offset within the given page.
 
+You should call dma_mapping_error() as dma_map_page() could fail and return
+error as outlined under the dma_map_single() discussion.
+
+You should call dma_unmap_page when the DMA activity is finished, e.g.
+from the interrupt which told you that the DMA transfer is done.
+
 With scatterlists, you map a region gathered from several regions by:
 
         int i, count = dma_map_sg(dev, sglist, nents, direction);
@@ -578,6 +627,14 @@ to use the dma_sync_*() interfaces.
                 dma_addr_t mapping;
 
                 mapping = dma_map_single(cp->dev, buffer, len, DMA_FROM_DEVICE);
+                if (dma_mapping_error(dma_handle)) {
+                        /*
+                         * reduce current DMA mapping usage,
+                         * delay and try again later or
+                         * reset driver.
+                         */
+                        goto map_error_handling;
+                }
 
                 cp->rx_buf = buffer;
                 cp->rx_len = len;
@@ -658,6 +715,75 @@ failure can be determined by:
                  * delay and try again later or
                  * reset driver.
                  */
+                goto map_error_handling;
+        }
+
+- unmap pages that are already mapped, when mapping error occurs in the middle
+  of a multiple page mapping attempt. These example are applicable to
+  dma_map_page() as well.
+
+Example 1:
+        dma_addr_t dma_handle1;
+        dma_addr_t dma_handle2;
+
+        dma_handle1 = dma_map_single(dev, addr, size, direction);
+        if (dma_mapping_error(dev, dma_handle1)) {
+                /*
+                 * reduce current DMA mapping usage,
+                 * delay and try again later or
+                 * reset driver.
+                 */
+                goto map_error_handling1;
+        }
+        dma_handle2 = dma_map_single(dev, addr, size, direction);
+        if (dma_mapping_error(dev, dma_handle2)) {
+                /*
+                 * reduce current DMA mapping usage,
+                 * delay and try again later or
+                 * reset driver.
+                 */
+                goto map_error_handling2;
+        }
+
+        ...
+