Re: [PATCH v2 06/12] drm/xe: Do not forward invalid TLB invalidation seqnos to upper layers

["Summers, Stuart" <stuart.summers@intel.com>]

On Tue, 2025-11-04 at 11:56 -0800, Matthew Brost wrote:
> Context-based TLB invalidations send multiple H2G messages per seqno,
> with only the final H2G containing a valid seqno — the others carry
> an
> invalid seqno. The G2H handler drops these invalid seqnos to avoid
> prematurely signaling a TLB invalidation fence.
> 
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_guc_tlb_inval.c   | 3 ++-
>  drivers/gpu/drm/xe/xe_tlb_inval_types.h | 1 +
>  2 files changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_guc_tlb_inval.c
> b/drivers/gpu/drm/xe/xe_guc_tlb_inval.c
> index 61bfa0d485f6..995789f0d31f 100644
> --- a/drivers/gpu/drm/xe/xe_guc_tlb_inval.c
> +++ b/drivers/gpu/drm/xe/xe_guc_tlb_inval.c
> @@ -235,7 +235,8 @@ int xe_guc_tlb_inval_done_handler(struct xe_guc
> *guc, u32 *msg, u32 len)
>         if (unlikely(len != 1))
>                 return -EPROTO;
>  
> -       xe_tlb_inval_done_handler(&gt->tlb_inval, msg[0]);
> +       if (msg[0] != TLB_INVALIDATION_SEQNO_INVALID)
> +               xe_tlb_inval_done_handler(&gt->tlb_inval, msg[0]);

Honestly I feel like this should be included as part of the patch that
actually uses it. No problem with the actual contents though.

Thanks,
Stuart

>  
>         return 0;
>  }
> diff --git a/drivers/gpu/drm/xe/xe_tlb_inval_types.h
> b/drivers/gpu/drm/xe/xe_tlb_inval_types.h
> index 8f8b060e9005..7a6967ce3b76 100644
> --- a/drivers/gpu/drm/xe/xe_tlb_inval_types.h
> +++ b/drivers/gpu/drm/xe/xe_tlb_inval_types.h
> @@ -80,6 +80,7 @@ struct xe_tlb_inval {
>         const struct xe_tlb_inval_ops *ops;
>         /** @tlb_inval.seqno: TLB invalidation seqno, protected by CT
> lock */
>  #define TLB_INVALIDATION_SEQNO_MAX     0x100000
> +#define TLB_INVALIDATION_SEQNO_INVALID TLB_INVALIDATION_SEQNO_MAX
>         int seqno;
>         /** @tlb_invalidation.seqno_lock: protects
> @tlb_invalidation.seqno */
>         struct mutex seqno_lock;