lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <793ede4d-79e9-3615-9da1-57cfe1a44c4d@arm.com>
Date:   Wed, 15 Jul 2020 11:32:30 +0100
From:   Robin Murphy <robin.murphy@....com>
To:     Tomasz Nowicki <tn@...ihalf.com>, will@...nel.org, joro@...tes.org,
        gregory.clement@...tlin.com, robh+dt@...nel.org, hannah@...vell.com
Cc:     linux-kernel@...r.kernel.org, iommu@...ts.linux-foundation.org,
        devicetree@...r.kernel.org, catalin.marinas@....com,
        nadavh@...vell.com, linux-arm-kernel@...ts.infradead.org,
        mw@...ihalf.com
Subject: Re: [PATCH v4 2/4] iommu/arm-smmu: Workaround for Marvell
 Armada-AP806 SoC erratum #582743

On 2020-07-15 08:06, Tomasz Nowicki wrote:
> From: Hanna Hawa <hannah@...vell.com>
> 
> Due to erratum #582743, the Marvell Armada-AP806 can't access 64bit to
> ARM SMMUv2 registers.
> 
> Provide implementation relevant hooks:
> - split the writeq/readq to two accesses of writel/readl.
> - mask the MMU_IDR2.PTFSv8 fields to not use AArch64 format (but
> only AARCH32_L) since with AArch64 format 32 bits access is not supported.
> 
> Note that most 64-bit registers like TTBRn can be accessed as two 32-bit
> halves without issue, and AArch32 format ensures that the register writes
> which must be atomic (for TLBI etc.) need only be 32-bit.

Thanks Tomasz, this has ended up as clean as I'd hoped it could, and 
there's still room to come back and play more complicated games later if 
a real need for AARCH64_64K at stage 2 crops up.

Reviewed-by: Robin Murphy <robin.murphy@....com>

> Signed-off-by: Hanna Hawa <hannah@...vell.com>
> Signed-off-by: Gregory CLEMENT <gregory.clement@...tlin.com>
> Signed-off-by: Tomasz Nowicki <tn@...ihalf.com>
> ---
>   Documentation/arm64/silicon-errata.rst |  3 ++
>   drivers/iommu/arm-smmu-impl.c          | 45 ++++++++++++++++++++++++++
>   2 files changed, 48 insertions(+)
> 
> diff --git a/Documentation/arm64/silicon-errata.rst b/Documentation/arm64/silicon-errata.rst
> index 936cf2a59ca4..157214d3abe1 100644
> --- a/Documentation/arm64/silicon-errata.rst
> +++ b/Documentation/arm64/silicon-errata.rst
> @@ -125,6 +125,9 @@ stable kernels.
>   | Cavium         | ThunderX2 Core  | #219            | CAVIUM_TX2_ERRATUM_219      |
>   +----------------+-----------------+-----------------+-----------------------------+
>   +----------------+-----------------+-----------------+-----------------------------+
> +| Marvell        | ARM-MMU-500     | #582743         | N/A                         |
> ++----------------+-----------------+-----------------+-----------------------------+
> ++----------------+-----------------+-----------------+-----------------------------+

And in case anyone feels like nit-picking the order here, I think the 
current respective corporate structures perfectly justify "Marvell" 
sorting alphabetically before "NXP", to be next to "Cavium" :D

Robin.

>   | Freescale/NXP  | LS2080A/LS1043A | A-008585        | FSL_ERRATUM_A008585         |
>   +----------------+-----------------+-----------------+-----------------------------+
>   +----------------+-----------------+-----------------+-----------------------------+
> diff --git a/drivers/iommu/arm-smmu-impl.c b/drivers/iommu/arm-smmu-impl.c
> index c75b9d957b70..59422cb92488 100644
> --- a/drivers/iommu/arm-smmu-impl.c
> +++ b/drivers/iommu/arm-smmu-impl.c
> @@ -147,6 +147,48 @@ static const struct arm_smmu_impl arm_mmu500_impl = {
>   	.reset = arm_mmu500_reset,
>   };
>   
> +static u64 mrvl_mmu500_readq(struct arm_smmu_device *smmu, int page, int off)
> +{
> +	/*
> +	 * Marvell Armada-AP806 erratum #582743.
> +	 * Split all the readq to double readl
> +	 */
> +	return hi_lo_readq_relaxed(arm_smmu_page(smmu, page) + off);
> +}
> +
> +static void mrvl_mmu500_writeq(struct arm_smmu_device *smmu, int page, int off,
> +			       u64 val)
> +{
> +	/*
> +	 * Marvell Armada-AP806 erratum #582743.
> +	 * Split all the writeq to double writel
> +	 */
> +	hi_lo_writeq_relaxed(val, arm_smmu_page(smmu, page) + off);
> +}
> +
> +static int mrvl_mmu500_cfg_probe(struct arm_smmu_device *smmu)
> +{
> +
> +	/*
> +	 * Armada-AP806 erratum #582743.
> +	 * Hide the SMMU_IDR2.PTFSv8 fields to sidestep the AArch64
> +	 * formats altogether and allow using 32 bits access on the
> +	 * interconnect.
> +	 */
> +	smmu->features &= ~(ARM_SMMU_FEAT_FMT_AARCH64_4K |
> +			    ARM_SMMU_FEAT_FMT_AARCH64_16K |
> +			    ARM_SMMU_FEAT_FMT_AARCH64_64K);
> +
> +	return 0;
> +}
> +
> +static const struct arm_smmu_impl mrvl_mmu500_impl = {
> +	.read_reg64 = mrvl_mmu500_readq,
> +	.write_reg64 = mrvl_mmu500_writeq,
> +	.cfg_probe = mrvl_mmu500_cfg_probe,
> +	.reset = arm_mmu500_reset,
> +};
> +
>   
>   struct arm_smmu_device *arm_smmu_impl_init(struct arm_smmu_device *smmu)
>   {
> @@ -175,5 +217,8 @@ struct arm_smmu_device *arm_smmu_impl_init(struct arm_smmu_device *smmu)
>   	    of_device_is_compatible(np, "qcom,sc7180-smmu-500"))
>   		return qcom_smmu_impl_init(smmu);
>   
> +	if (of_device_is_compatible(np, "marvell,ap806-smmu-500"))
> +		smmu->impl = &mrvl_mmu500_impl;
> +
>   	return smmu;
>   }
> 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ