[PATCH v2] iommu/arm-smmu-v3: Restore prints in arm_smmu_attach_dev()

[PATCH v2] iommu/arm-smmu-v3: Restore prints in arm_smmu_attach_dev()

[Nicolin Chen]

The prints for incompatible failures in arm_smmu_attach_dev() might be
still useful for error diagnosis. As VFIO and IOMMUFD would do testing
attach() call, having previous dev_err() isn't ideal either.

Add the prints back using dev_dbg().

Fixes: f4a147735793 ("iommu: Use EINVAL for incompatible device/domain in ->attach_dev")
Suggested-by: Will Deacon <will@kernel.org>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
---
Changelog
v1->v2:
 * Add "Fixes" line and Reviewed-by from Jason

 drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
index ab160198edd6..076d3710b1fe 100644
--- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
+++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
@@ -2433,14 +2433,21 @@ static int arm_smmu_attach_dev(struct iommu_domain *domain, struct device *dev)
 			goto out_unlock;
 		}
 	} else if (smmu_domain->smmu != smmu) {
+		dev_dbg(dev, "cannot attach to SMMU %s (upstream of %s)\n",
+			dev_name(smmu_domain->smmu->dev), dev_name(smmu->dev));
 		ret = -EINVAL;
 		goto out_unlock;
 	} else if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1 &&
 		   master->ssid_bits != smmu_domain->s1_cfg.s1cdmax) {
+		dev_dbg(dev,
+			"cannot attach to incompatible domain (%u SSID bits != %u)\n",
+			smmu_domain->s1_cfg.s1cdmax, master->ssid_bits);
 		ret = -EINVAL;
 		goto out_unlock;
 	} else if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1 &&
 		   smmu_domain->stall_enabled != master->stall_enabled) {
+		dev_dbg(dev, "cannot attach to stall-%s domain\n",
+			smmu_domain->stall_enabled ? "enabled" : "disabled");
 		ret = -EINVAL;
 		goto out_unlock;
 	}
-- 
2.38.1

Re: [PATCH v2] iommu/arm-smmu-v3: Restore prints in arm_smmu_attach_dev()

[Will Deacon]

On Wed, Nov 09, 2022 at 10:46:36AM -0800, Nicolin Chen wrote:
> The prints for incompatible failures in arm_smmu_attach_dev() might be
> still useful for error diagnosis. As VFIO and IOMMUFD would do testing
> attach() call, having previous dev_err() isn't ideal either.
> 
> Add the prints back using dev_dbg().
> 
> Fixes: f4a147735793 ("iommu: Use EINVAL for incompatible device/domain in ->attach_dev")
> Suggested-by: Will Deacon <will@kernel.org>
> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
> ---
> Changelog
> v1->v2:
>  * Add "Fixes" line and Reviewed-by from Jason
> 
>  drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> index ab160198edd6..076d3710b1fe 100644
> --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> @@ -2433,14 +2433,21 @@ static int arm_smmu_attach_dev(struct iommu_domain *domain, struct device *dev)
>  			goto out_unlock;
>  		}
>  	} else if (smmu_domain->smmu != smmu) {
> +		dev_dbg(dev, "cannot attach to SMMU %s (upstream of %s)\n",
> +			dev_name(smmu_domain->smmu->dev), dev_name(smmu->dev));
>  		ret = -EINVAL;
>  		goto out_unlock;
>  	} else if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1 &&
>  		   master->ssid_bits != smmu_domain->s1_cfg.s1cdmax) {
> +		dev_dbg(dev,
> +			"cannot attach to incompatible domain (%u SSID bits != %u)\n",
> +			smmu_domain->s1_cfg.s1cdmax, master->ssid_bits);
>  		ret = -EINVAL;
>  		goto out_unlock;
>  	} else if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1 &&
>  		   smmu_domain->stall_enabled != master->stall_enabled) {
> +		dev_dbg(dev, "cannot attach to stall-%s domain\n",
> +			smmu_domain->stall_enabled ? "enabled" : "disabled");
>  		ret = -EINVAL;
>  		goto out_unlock;

Thanks!

Acked-by: Will Deacon <will@kernel.org>

Will