linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: will.deacon@arm.com (Will Deacon)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH] iommu/arm-smmu-v2: ThunderX(errata-23399) mis-extends 64bit registers
Date: Thu, 30 Jul 2015 19:45:23 +0100	[thread overview]
Message-ID: <20150730184523.GC16663@arm.com> (raw)
In-Reply-To: <1438278906-29627-1-git-send-email-tchalamarla@caviumnetworks.com>

Hello,

On Thu, Jul 30, 2015 at 06:55:06PM +0100, tchalamarla at caviumnetworks.com wrote:
> From: Tirumalesh Chalamarla <tchalamarla@caviumnetworks.com>
> 
> The SMMU architecture defines two different behaviors when 64-bit
> registers are written with 32-bit writes.  The first behavior causes
> zero extension into the upper 32-bits.  The second behavior splits a
> 64-bit register into "normal" 32-bit register pairs.
> 
> On some passes of ThunderX,
> the following registers incorrectly zero extended when they should
> instead behave as normal 32-bit register pairs:
> 
>   SMMU()_(S)GFAR
>   SMMU()_NSGFAR
>   SMMU()_CB()_TTBR0
>   SMMU()_CB()_TTBR1
>   SMMU()_CB()_FAR
> 
> Signed-off-by: Tirumalesh Chalamarla <tchalamarla@caviumnetworks.com>
> ---
>  drivers/iommu/arm-smmu.c | 51 ++++++++++++++++++++++++++++++++++--------------
>  1 file changed, 36 insertions(+), 15 deletions(-)

[...]

> @@ -762,22 +766,39 @@ static void arm_smmu_init_context_bank(struct arm_smmu_domain *smmu_domain,
>  
>  	/* TTBRs */
>  	if (stage1) {
> -		reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[0];
> -		writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR0_LO);
> -		reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[0] >> 32;
> -		reg |= ARM_SMMU_CB_ASID(cfg) << TTBRn_HI_ASID_SHIFT;
> -		writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR0_HI);
> -
> -		reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[1];
> -		writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR1_LO);
> -		reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[1] >> 32;
> -		reg |= ARM_SMMU_CB_ASID(cfg) << TTBRn_HI_ASID_SHIFT;
> -		writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR1_HI);
> +		if (smmu->options & ARM_SMMU_OPT_64BIT_WRITES_ONLY) {
> +			reg64 = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[0];
> +			reg64 |= ((u64) ARM_SMMU_CB_ASID(cfg)) <<
> +						(TTBRn_HI_ASID_SHIFT + 32);
> +			writeq_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR0_LO);
> +
> +			reg64 = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[1];
> +			reg64 |= ((u64)ARM_SMMU_CB_ASID(cfg)) <<
> +						(TTBRn_HI_ASID_SHIFT + 32);
> +			writeq_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR1_LO);
> +		} else {
> +			reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[0];
> +			writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR0_LO);
> +			reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[0] >> 32;
> +			reg |= ARM_SMMU_CB_ASID(cfg) << TTBRn_HI_ASID_SHIFT;
> +			writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR0_HI);
> +
> +			reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[1];
> +			writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR1_LO);
> +			reg = pgtbl_cfg->arm_lpae_s1_cfg.ttbr[1] >> 32;
> +			reg |= ARM_SMMU_CB_ASID(cfg) << TTBRn_HI_ASID_SHIFT;
> +			writel_relaxed(reg, cb_base + ARM_SMMU_CB_TTBR1_HI);
> +		}

I'm fine with making this sort of change if you need it, but this is pretty
ugly. Worse, it won't compile for 32-bit ARM.

How about we add a wrapper around these, say smmu_writeq(...), which can
then either expand to 2x writel_relaxed or 1x writeq_relaxed depending on
CONFIG_64BIT and your erratum workaround?

Don't forgot to update the ATOS code too (so you need to write the high word
first).

Will

  reply	other threads:[~2015-07-30 18:45 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-07-30 17:55 [PATCH] iommu/arm-smmu-v2: ThunderX(errata-23399) mis-extends 64bit registers tchalamarla at caviumnetworks.com
2015-07-30 18:45 ` Will Deacon [this message]
2015-07-30 19:08   ` Robin Murphy
     [not found]   ` <96AD8EF9-D416-43AB-BA6F-2C25098BB8BB@caviumnetworks.com>
     [not found]     ` <91236962-23EF-4306-B0FA-0F7DAE4F06CD@caviumnetworks.com>
2015-07-31 11:27       ` Will Deacon
2015-07-31 11:48       ` Russell King - ARM Linux

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20150730184523.GC16663@arm.com \
    --to=will.deacon@arm.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).