[PATCH] drm/amdkfd: fix missing L2 cache info in topology

[Lazar, Lijo]

On 2/7/2025 9:40 PM, Eric Huang wrote:
> In some ASICs L2 cache info may miss in kfd topology,
> because the first bitmap may be empty, that means
> the first cu may be inactive, so to find the first
> active cu will solve the issue.
> 
> v2: Only find the first active cu in the first xcc
> 
> Signed-off-by: Eric Huang <jinhuieric.huang at amd.com>

Acked-by: Lijo Lazar <lijo.lazar at amd.com>

Thanks,
Lijo

> ---
>  drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 19 +++++++++++++++++--
>  1 file changed, 17 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
> index 4936697e6fc2..c84010427a6f 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
> @@ -1665,17 +1665,32 @@ static int fill_in_l2_l3_pcache(struct kfd_cache_properties **props_ext,
>  				int cache_type, unsigned int cu_processor_id,
>  				struct kfd_node *knode)
>  {
> -	unsigned int cu_sibling_map_mask;
> +	unsigned int cu_sibling_map_mask = 0;
>  	int first_active_cu;
>  	int i, j, k, xcc, start, end;
>  	int num_xcc = NUM_XCC(knode->xcc_mask);
>  	struct kfd_cache_properties *pcache = NULL;
>  	enum amdgpu_memory_partition mode;
>  	struct amdgpu_device *adev = knode->adev;
> +	bool found = false;
>  
>  	start = ffs(knode->xcc_mask) - 1;
>  	end = start + num_xcc;
> -	cu_sibling_map_mask = cu_info->bitmap[start][0][0];
> +
> +	/* To find the bitmap in the first active cu in the first
> +	 * xcc, it is based on the assumption that evrey xcc must
> +	 * have at least one active cu.
> +	 */
> +	for (i = 0; i < gfx_info->max_shader_engines && !found; i++) {
> +		for (j = 0; j < gfx_info->max_sh_per_se && !found; j++) {
> +			if (cu_info->bitmap[start][i % 4][j % 4]) {
> +				cu_sibling_map_mask =
> +					cu_info->bitmap[start][i % 4][j % 4];
> +				found = true;
> +			}
> +		}
> +	}
> +
>  	cu_sibling_map_mask &=
>  		((1 << pcache_info[cache_type].num_cu_shared) - 1);
>  	first_active_cu = ffs(cu_sibling_map_mask);