Re: [PATCH v3 2/2] iommu/arm-smmu-v3: Replace vsmmu_size/type with get_viommu_size

[Nicolin Chen <nicolinc@nvidia.com>]

On Fri, Jul 25, 2025 at 05:11:07AM +0000, Pranjal Shrivastava wrote:
> On Thu, Jul 24, 2025 at 02:49:28PM -0700, Nicolin Chen wrote:
> I'm agreeing with all of it, it's just that the comment says something 
> was rejected in by the size op, which raises confusion as to why we're
> in the init op. The init op rejecting something due to data corruption
> is a different thing..
> 
> I totally get the point about data corruption, i.e.:
> 
> size op -> returned something valid
> <data corruption>
> init op -> rejecting corrupted type
> 
> Wheras I was just trying to understand a case where as per the comment:
> "Unsupported type was rejected in tegra241_cmdqv_get_vintf_size()", 
> i.e. ->size op returned 0, yet we ended up calling the init op

Is the updated one in v4 fine to you?

/*
 * Unsupported type should be rejected by tegra241_cmdqv_get_vintf_size.
 * Seeing one here indicates a kernel bug or some data corruption.
 */

Nicolin