| Message ID | 0484444b6257bfb6adb68405a72c64fc4fc98142.1460391217.git.robin.murphy@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, Apr 13, 2016 at 06:12:59PM +0100, Robin Murphy wrote: > MMU-500 erratum #841119 is tickled by a particular set of circumstances > interacting with the next-page prefetcher. Since said prefetcher is > quite dumb and actually detrimental to performance in some cases (by > causing unwanted TLB evictions for non-sequential access patterns), we > lose very little by turning it off, and what we gain is a guarantee that > the erratum is never hit. > > As a bonus, the same workaround will also prevent erratum #826419 once > v7 short descriptor support is implemented. > > CC: Catalin Marinas <catalin.marinas@arm.com> > CC: Will Deacon <will.deacon@arm.com> > Signed-off-by: Robin Murphy <robin.murphy@arm.com> > --- > Documentation/arm64/silicon-errata.txt | 1 + > drivers/iommu/arm-smmu.c | 16 +++++++++++++++- > 2 files changed, 16 insertions(+), 1 deletion(-) Looks fine, but can you update silicon-errata.txt too, please? It's a handy reference for "what is Linux actively working around" and this one is somewhat hidden away. Will
On Wed, Apr 13, 2016 at 06:12:59PM +0100, Robin Murphy wrote: > MMU-500 erratum #841119 is tickled by a particular set of circumstances > interacting with the next-page prefetcher. Since said prefetcher is > quite dumb and actually detrimental to performance in some cases (by > causing unwanted TLB evictions for non-sequential access patterns), we > lose very little by turning it off, and what we gain is a guarantee that > the erratum is never hit. > > As a bonus, the same workaround will also prevent erratum #826419 once > v7 short descriptor support is implemented. > > CC: Catalin Marinas <catalin.marinas@arm.com> > CC: Will Deacon <will.deacon@arm.com> > Signed-off-by: Robin Murphy <robin.murphy@arm.com> > --- > Documentation/arm64/silicon-errata.txt | 1 + > drivers/iommu/arm-smmu.c | 16 +++++++++++++++- > 2 files changed, 16 insertions(+), 1 deletion(-) > > diff --git a/Documentation/arm64/silicon-errata.txt b/Documentation/arm64/silicon-errata.txt > index 806f91c..c6938e5 100644 > --- a/Documentation/arm64/silicon-errata.txt > +++ b/Documentation/arm64/silicon-errata.txt > @@ -53,6 +53,7 @@ stable kernels. > | ARM | Cortex-A57 | #832075 | ARM64_ERRATUM_832075 | > | ARM | Cortex-A57 | #852523 | N/A | > | ARM | Cortex-A57 | #834220 | ARM64_ERRATUM_834220 | > +| ARM | MMU-500 | #841119,#826419 | N/A | Found it :) Will
diff --git a/Documentation/arm64/silicon-errata.txt b/Documentation/arm64/silicon-errata.txt index 806f91c..c6938e5 100644 --- a/Documentation/arm64/silicon-errata.txt +++ b/Documentation/arm64/silicon-errata.txt @@ -53,6 +53,7 @@ stable kernels. | ARM | Cortex-A57 | #832075 | ARM64_ERRATUM_832075 | | ARM | Cortex-A57 | #852523 | N/A | | ARM | Cortex-A57 | #834220 | ARM64_ERRATUM_834220 | +| ARM | MMU-500 | #841119,#826419 | N/A | | | | | | | Cavium | ThunderX ITS | #22375, #24313 | CAVIUM_ERRATUM_22375 | | Cavium | ThunderX GICv3 | #23154 | CAVIUM_ERRATUM_23154 | diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c index d8bc20a..085fc8d 100644 --- a/drivers/iommu/arm-smmu.c +++ b/drivers/iommu/arm-smmu.c @@ -203,6 +203,7 @@ #define ARM_SMMU_CB(smmu, n) ((n) * (1 << (smmu)->pgshift)) #define ARM_SMMU_CB_SCTLR 0x0 +#define ARM_SMMU_CB_ACTLR 0x4 #define ARM_SMMU_CB_RESUME 0x8 #define ARM_SMMU_CB_TTBCR2 0x10 #define ARM_SMMU_CB_TTBR0 0x20 @@ -234,6 +235,8 @@ #define SCTLR_M (1 << 0) #define SCTLR_EAE_SBOP (SCTLR_AFE | SCTLR_TRE) +#define ARM_MMU500_ACTLR_CPRE (1 << 1) + #define CB_PAR_F (1 << 0) #define ATSR_ACTIVE (1 << 0) @@ -280,6 +283,7 @@ enum arm_smmu_arch_version { enum arm_smmu_implementation { GENERIC_SMMU, + ARM_MMU500, CAVIUM_SMMUV2, }; @@ -1517,6 +1521,15 @@ static void arm_smmu_device_reset(struct arm_smmu_device *smmu) cb_base = ARM_SMMU_CB_BASE(smmu) + ARM_SMMU_CB(smmu, i); writel_relaxed(0, cb_base + ARM_SMMU_CB_SCTLR); writel_relaxed(FSR_FAULT, cb_base + ARM_SMMU_CB_FSR); + /* + * Disable MMU-500's not-particularly-beneficial next-page + * prefetcher for the sake of errata #841119 and #826419. + */ + if (smmu->model == ARM_MMU500) { + reg = readl_relaxed(cb_base + ARM_SMMU_CB_ACTLR); + reg &= ~ARM_MMU500_ACTLR_CPRE; + writel_relaxed(reg, cb_base + ARM_SMMU_CB_ACTLR); + } } /* Invalidate the TLB, just in case */ @@ -1762,6 +1775,7 @@ static struct arm_smmu_match_data name = { .version = ver, .model = imp } ARM_SMMU_MATCH_DATA(smmu_generic_v1, ARM_SMMU_V1, GENERIC_SMMU); ARM_SMMU_MATCH_DATA(smmu_generic_v2, ARM_SMMU_V2, GENERIC_SMMU); +ARM_SMMU_MATCH_DATA(arm_mmu500, ARM_SMMU_V2, ARM_MMU500); ARM_SMMU_MATCH_DATA(cavium_smmuv2, ARM_SMMU_V2, CAVIUM_SMMUV2); static const struct of_device_id arm_smmu_of_match[] = { @@ -1769,7 +1783,7 @@ static const struct of_device_id arm_smmu_of_match[] = { { .compatible = "arm,smmu-v2", .data = &smmu_generic_v2 }, { .compatible = "arm,mmu-400", .data = &smmu_generic_v1 }, { .compatible = "arm,mmu-401", .data = &smmu_generic_v1 }, - { .compatible = "arm,mmu-500", .data = &smmu_generic_v2 }, + { .compatible = "arm,mmu-500", .data = &arm_mmu500 }, { .compatible = "cavium,smmu-v2", .data = &cavium_smmuv2 }, { }, };
MMU-500 erratum #841119 is tickled by a particular set of circumstances interacting with the next-page prefetcher. Since said prefetcher is quite dumb and actually detrimental to performance in some cases (by causing unwanted TLB evictions for non-sequential access patterns), we lose very little by turning it off, and what we gain is a guarantee that the erratum is never hit. As a bonus, the same workaround will also prevent erratum #826419 once v7 short descriptor support is implemented. CC: Catalin Marinas <catalin.marinas@arm.com> CC: Will Deacon <will.deacon@arm.com> Signed-off-by: Robin Murphy <robin.murphy@arm.com> --- Documentation/arm64/silicon-errata.txt | 1 + drivers/iommu/arm-smmu.c | 16 +++++++++++++++- 2 files changed, 16 insertions(+), 1 deletion(-)