[PATCH] iommu/tegra241-cmdqv: Read SMMU IDR1.CMDQS instead of hardcoding

[PATCH] iommu/tegra241-cmdqv: Read SMMU IDR1.CMDQS instead of hardcoding

[Nicolin Chen]

The hardware limitation "max=19" actually comes from SMMU Command Queue.
So, it'd be more natural for tegra241-cmdqv driver to read it out rather
than hardcoding it itself.

This is not an issue yet for a kernel on a baremetal system, but a guest
kernel setting the queue base/size in form of IPA/gPA might result in a
noncontiguous queue in the physical address space, if underlying physical
pages backing up the guest RAM aren't contiguous entirely: e.g. 2MB-page
backed guest RAM cannot guarantee a contiguous queue if it is 8MB (capped
to VCMDQ_LOG2SIZE_MAX=19). This might lead to command errors when HW does
linear-read from a noncontiguous queue memory.

Adding this extra IDR1.CMDQS cap (in the guest kernel) allows VMM to set
SMMU's IDR1.CMDQS=17 for the case mentioned above, so a guest-level queue
will be capped to maximum 2MB, ensuring a contiguous queue memory.

Fixes: a3799717b881 ("iommu/tegra241-cmdqv: Fix alignment failure at max_n_shift")
Reported-by: Ian Kalinowski <ikalinowski@nvidia.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
---
 drivers/iommu/arm/arm-smmu-v3/tegra241-cmdqv.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/iommu/arm/arm-smmu-v3/tegra241-cmdqv.c b/drivers/iommu/arm/arm-smmu-v3/tegra241-cmdqv.c
index 6e41ddaa24d6..d525ab43a4ae 100644
--- a/drivers/iommu/arm/arm-smmu-v3/tegra241-cmdqv.c
+++ b/drivers/iommu/arm/arm-smmu-v3/tegra241-cmdqv.c
@@ -79,7 +79,6 @@
 #define TEGRA241_VCMDQ_PAGE1(q)		(TEGRA241_VCMDQ_PAGE1_BASE + 0x80*(q))
 #define  VCMDQ_ADDR			GENMASK(47, 5)
 #define  VCMDQ_LOG2SIZE			GENMASK(4, 0)
-#define  VCMDQ_LOG2SIZE_MAX		19
 
 #define TEGRA241_VCMDQ_BASE		0x00000
 #define TEGRA241_VCMDQ_CONS_INDX_BASE	0x00008
@@ -505,12 +504,15 @@ static int tegra241_vcmdq_alloc_smmu_cmdq(struct tegra241_vcmdq *vcmdq)
 	struct arm_smmu_cmdq *cmdq = &vcmdq->cmdq;
 	struct arm_smmu_queue *q = &cmdq->q;
 	char name[16];
+	u32 regval;
 	int ret;
 
 	snprintf(name, 16, "vcmdq%u", vcmdq->idx);
 
-	/* Queue size, capped to ensure natural alignment */
-	q->llq.max_n_shift = min_t(u32, CMDQ_MAX_SZ_SHIFT, VCMDQ_LOG2SIZE_MAX);
+	/* Cap queue size to SMMU's IDR1.CMDQS and ensure natural alignment */
+	regval = readl_relaxed(smmu->base + ARM_SMMU_IDR1);
+	q->llq.max_n_shift =
+		min_t(u32, CMDQ_MAX_SZ_SHIFT, FIELD_GET(IDR1_CMDQS, regval));
 
 	/* Use the common helper to init the VCMDQ, and then... */
 	ret = arm_smmu_init_one_queue(smmu, q, vcmdq->page0,
-- 
2.43.0

Re: [PATCH] iommu/tegra241-cmdqv: Read SMMU IDR1.CMDQS instead of hardcoding

[Will Deacon]

On Wed, 18 Dec 2024 21:14:21 -0800, Nicolin Chen wrote:
> The hardware limitation "max=19" actually comes from SMMU Command Queue.
> So, it'd be more natural for tegra241-cmdqv driver to read it out rather
> than hardcoding it itself.
> 
> This is not an issue yet for a kernel on a baremetal system, but a guest
> kernel setting the queue base/size in form of IPA/gPA might result in a
> noncontiguous queue in the physical address space, if underlying physical
> pages backing up the guest RAM aren't contiguous entirely: e.g. 2MB-page
> backed guest RAM cannot guarantee a contiguous queue if it is 8MB (capped
> to VCMDQ_LOG2SIZE_MAX=19). This might lead to command errors when HW does
> linear-read from a noncontiguous queue memory.
> 
> [...]

Applied to will (for-joerg/arm-smmu/updates), thanks!

[1/1] iommu/tegra241-cmdqv: Read SMMU IDR1.CMDQS instead of hardcoding
      https://git.kernel.org/will/c/e94dc6ddda8d

Cheers,
-- 
Will

https://fixes.arm64.dev
https://next.arm64.dev
https://will.arm64.dev

Re: [PATCH] iommu/tegra241-cmdqv: Read SMMU IDR1.CMDQS instead of hardcoding

[Jason Gunthorpe]

On Wed, Dec 18, 2024 at 09:14:21PM -0800, Nicolin Chen wrote:
> The hardware limitation "max=19" actually comes from SMMU Command Queue.
> So, it'd be more natural for tegra241-cmdqv driver to read it out rather
> than hardcoding it itself.
> 
> This is not an issue yet for a kernel on a baremetal system, but a guest
> kernel setting the queue base/size in form of IPA/gPA might result in a
> noncontiguous queue in the physical address space, if underlying physical
> pages backing up the guest RAM aren't contiguous entirely: e.g. 2MB-page
> backed guest RAM cannot guarantee a contiguous queue if it is 8MB (capped
> to VCMDQ_LOG2SIZE_MAX=19). This might lead to command errors when HW does
> linear-read from a noncontiguous queue memory.
> 
> Adding this extra IDR1.CMDQS cap (in the guest kernel) allows VMM to set
> SMMU's IDR1.CMDQS=17 for the case mentioned above, so a guest-level queue
> will be capped to maximum 2MB, ensuring a contiguous queue memory.
> 
> Fixes: a3799717b881 ("iommu/tegra241-cmdqv: Fix alignment failure at max_n_shift")
> Reported-by: Ian Kalinowski <ikalinowski@nvidia.com>
> Cc: <stable@vger.kernel.org>
> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
> ---
>  drivers/iommu/arm/arm-smmu-v3/tegra241-cmdqv.c | 8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)

Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>

Jason