Messages in this thread Patch in this message | | | Subject | Re: [PATCH 2/4] iommu/arm-smmu: Workaround for Marvell Armada-AP806 SoC erratum #582743 | From | Robin Murphy <> | Date | Mon, 15 Oct 2018 14:00:16 +0100 |
| |
Hi Hanna,
On 15/10/18 13:00, hannah@marvell.com wrote: > From: Hanna Hawa <hannah@marvell.com> > > Due to erratum #582743, the Marvell Armada-AP806 can't access 64bit > to ARM SMMUv2 registers. > This patch split the writeq/readq to two accesses of writel/readl. > > Note that separate writes/reads to 2 is not problem regards to atomicity, > because the driver use the readq/writeq while initialize the SMMU, report > for SMMU fault, and use spinlock in one case (iova_to_phys).
In general, this doesn't work. Here's what the SMMU spec says about SMMU_CBn_TLBIVA, but others are similar:
"If SMMU_CBA2Rn.VA64 is one, then AArch64 format is selected. The programmer should use 64 bit accesses to this register. If 32-bit accesses are used then writes to the top 32 bits are ignored and writes to the lower 32 bits are zero extended."
If your interconnect won't let 64-bit transactions through, then you can't use AArch64 format at stage 1 at all, since there's no way to invalidate entries with the correct ASID, and you'll have to restrict stage 2 formats to at most 44-bit IOVAs in order for TLBIIPAS2{L} not to invalidate the wrong thing.
> Signed-off-by: Hanna Hawa <hannah@marvell.com> > --- > Documentation/arm64/silicon-errata.txt | 2 ++ > drivers/iommu/arm-smmu.c | 33 +++++++++++++++++++++++++++++---- > 2 files changed, 31 insertions(+), 4 deletions(-) > > diff --git a/Documentation/arm64/silicon-errata.txt b/Documentation/arm64/silicon-errata.txt > index 3b2f2dd..fc3f2a0 100644 > --- a/Documentation/arm64/silicon-errata.txt > +++ b/Documentation/arm64/silicon-errata.txt > @@ -67,6 +67,8 @@ stable kernels. > | Cavium | ThunderX2 SMMUv3| #74 | N/A | > | Cavium | ThunderX2 SMMUv3| #126 | N/A | > | | | | | > +| Marvell | ARM-MMU-500 | #582743 | N/A | > +| | | | |
Nit: the convention here seems to be at least alphabetically sorted by Implementer.
> | Freescale/NXP | LS2080A/LS1043A | A-008585 | FSL_ERRATUM_A008585 | > | | | | | > | Hisilicon | Hip0{5,6,7} | #161010101 | HISILICON_ERRATUM_161010101 | > diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c > index fccb1d4..d64f892 100644 > --- a/drivers/iommu/arm-smmu.c > +++ b/drivers/iommu/arm-smmu.c > @@ -119,6 +119,7 @@ enum arm_smmu_arch_version { > enum arm_smmu_implementation { > GENERIC_SMMU, > ARM_MMU500, > + MRVL_MMU500,
Is there any actually modification to the MMU-500 RTL itself here, or is the problem just in the interconnect in front of the SMMU programming interface? I would normally assume the latter, in which case treating it as a separate implementation isn't really accurate, and I'd much rather handle any workaround via smmu->options, just like the secure access workaround (which is a similar integration issue).
Robin.
> CAVIUM_SMMUV2, > }; > > @@ -276,13 +277,35 @@ static inline void smmu_writeq_relaxed(struct arm_smmu_device *smmu, > u64 val, > void __iomem *addr) > { > - writeq_relaxed(val, addr); > + /* > + * Marvell Armada-AP806 erratum #582743. > + * Split all the writeq to double writel > + */ > + if (smmu->model != MRVL_MMU500) { > + writeq_relaxed(val, addr); > + return; > + } > + > + writel_relaxed(upper_32_bits(val), addr + 4); > + writel_relaxed(lower_32_bits(val), addr); > } > > static inline u64 smmu_readq_relaxed(struct arm_smmu_device *smmu, > void __iomem *addr) > { > - return readq_relaxed(addr); > + u64 val; > + > + /* > + * Marvell Armada-AP806 erratum #582743. > + * Split all the readq to double readl > + */ > + if (smmu->model != MRVL_MMU500) > + return readq_relaxed(addr); > + > + val = (u64)readl_relaxed(addr + 4) << 32; > + val |= readl_relaxed(addr); > + > + return val; > } > > static void parse_driver_options(struct arm_smmu_device *smmu) > @@ -1611,7 +1634,7 @@ static void arm_smmu_device_reset(struct arm_smmu_device *smmu) > for (i = 0; i < smmu->num_mapping_groups; ++i) > arm_smmu_write_sme(smmu, i); > > - if (smmu->model == ARM_MMU500) { > + if (smmu->model == ARM_MMU500 || smmu->model == MRVL_MMU500) { > /* > * Before clearing ARM_MMU500_ACTLR_CPRE, need to > * clear CACHE_LOCK bit of ACR first. And, CACHE_LOCK > @@ -1640,7 +1663,7 @@ static void arm_smmu_device_reset(struct arm_smmu_device *smmu) > * Disable MMU-500's not-particularly-beneficial next-page > * prefetcher for the sake of errata #841119 and #826419. > */ > - if (smmu->model == ARM_MMU500) { > + if (smmu->model == ARM_MMU500 || smmu->model == MRVL_MMU500) { > reg = readl_relaxed(cb_base + ARM_SMMU_CB_ACTLR); > reg &= ~ARM_MMU500_ACTLR_CPRE; > writel_relaxed(reg, cb_base + ARM_SMMU_CB_ACTLR); > @@ -1923,6 +1946,7 @@ struct arm_smmu_match_data { > ARM_SMMU_MATCH_DATA(smmu_generic_v2, ARM_SMMU_V2, GENERIC_SMMU); > ARM_SMMU_MATCH_DATA(arm_mmu401, ARM_SMMU_V1_64K, GENERIC_SMMU); > ARM_SMMU_MATCH_DATA(arm_mmu500, ARM_SMMU_V2, ARM_MMU500); > +ARM_SMMU_MATCH_DATA(mrvl_mmu500, ARM_SMMU_V2, MRVL_MMU500); > ARM_SMMU_MATCH_DATA(cavium_smmuv2, ARM_SMMU_V2, CAVIUM_SMMUV2); > > static const struct of_device_id arm_smmu_of_match[] = { > @@ -1931,6 +1955,7 @@ struct arm_smmu_match_data { > { .compatible = "arm,mmu-400", .data = &smmu_generic_v1 }, > { .compatible = "arm,mmu-401", .data = &arm_mmu401 }, > { .compatible = "arm,mmu-500", .data = &arm_mmu500 }, > + { .compatible = "marvell,mmu-500", .data = &mrvl_mmu500 }, > { .compatible = "cavium,smmu-v2", .data = &cavium_smmuv2 }, > { }, > }; >
| |