On Fri, Mar 31, 2017 at 12:03:33PM +0100, Robin Murphy wrote:
quoted hunk
The current SMR masking support using a 2-cell iommu-specifier is
primarily intended to handle individual masters with large and/or
complex Stream ID assignments; it quickly gets a bit clunky in other SMR
use-cases where we just want to consistently mask out the same part of
every Stream ID (e.g. for MMU-500 configurations where the appended TBU
number gets in the way unnecessarily). Let's add a new property to allow
a single global mask value to better fit the latter situation.
Tested-by: Nipun Gupta <redacted>
Signed-off-by: Robin Murphy <redacted>
---
v2: Rewrite the documentation to actually make sense, and add an
explicit example per Mark's suggestion.
.../devicetree/bindings/iommu/arm,smmu.txt | 27 ++++++++++++++++++++++
drivers/iommu/arm-smmu.c | 4 +++-
2 files changed, 30 insertions(+), 1 deletion(-)
diff --git a/Documentation/devicetree/bindings/iommu/arm,smmu.txt b/Documentation/devicetree/bindings/iommu/arm,smmu.txt
index 6cdf32d037fc..b744b231766f 100644
--- a/Documentation/devicetree/bindings/iommu/arm,smmu.txt
+++ b/Documentation/devicetree/bindings/iommu/arm,smmu.txt
@@ -60,6 +60,16 @@ conditions.
aliases of secure registers have to be used during
SMMU configuration.
+- stream-match-mask : For SMMUs supporting stream matching and using
+ #iommu-cells = <1>, specifies a fixed value to set in
+ the SMRn.MASK field of every stream match register
+ used, for cases where it is desirable to ignore some
+ portion of every Stream ID (e.g. for certain MMU-500
+ configurations given globally unique input IDs). This
+ property is not valid for SMMUs using stream indexing,
+ or using stream matching with #iommu-cells = <2>, and
+ may be ignored if present in such cases.
This generally sounds fine, but I'd like it to be a little more explicit
about what this means for matching. Can we change the first sentence to:
For SMMUs supporting stream matching and using
#iommu-cells = <1>, specifies a mask of bits to
ignore when matching stream IDs (e.g. this may be
programmed into the SMRn.MASK field of every
stream match register used).
Otherwise, this looks good to me; thanks for respinning this.
If you're happy to make the above change:
Acked-by: Mark Rutland <redacted>
Thanks,
Mark.
quoted hunk
+
** Deprecated properties:
- mmu-masters (deprecated in favour of the generic "iommus" binding) :
@@ -109,3 +119,20 @@ conditions.
master3 {
iommus = <&smmu2 1 0x30>;
};
+
+
+ /* ARM MMU-500 with 10-bit stream ID input configuration */
+ smmu3: iommu {
+ compatible = "arm,mmu-500", "arm,smmu-v2";
+ ...
+ #iommu-cells = <1>;
+ /* always ignore appended 5-bit TBU number */
+ stream-match-mask = 0x7c00;
+ };
+
+ bus {
+ /* bus whose child devices emit one unique 10-bit stream
+ ID each, but may master through multiple SMMU TBUs */
+ iommu-map = <0 &smmu3 0 0x400>;
+ ...
+ };diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c
index abf6496843a6..e394d55146a6 100644
--- a/drivers/iommu/arm-smmu.c
+++ b/drivers/iommu/arm-smmu.c
@@ -1590,13 +1590,15 @@ static int arm_smmu_domain_set_attr(struct iommu_domain *domain,
static int arm_smmu_of_xlate(struct device *dev, struct of_phandle_args *args)
{
- u32 fwid = 0;
+ u32 mask, fwid = 0;
if (args->args_count > 0)
fwid |= (u16)args->args[0];
if (args->args_count > 1)
fwid |= (u16)args->args[1] << SMR_MASK_SHIFT;
+ else if (!of_property_read_u32(args->np, "stream-match-mask", &mask))
+ fwid |= (u16)mask << SMR_MASK_SHIFT;
return iommu_fwspec_add_ids(dev, &fwid, 1);
}--
2.11.0.dirty