Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Summers, Stuart" <stuart.summers@intel.com>
To: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
	"Nguyen, Brian3" <brian3.nguyen@intel.com>
Cc: "Roper, Matthew D" <matthew.d.roper@intel.com>
Subject: Re: [PATCH 2/2] drm/xe: Move page reclaim done_handler to own func
Date: Mon, 23 Feb 2026 22:45:12 +0000	[thread overview]
Message-ID: <a05f998d318f822c761098d6e6cc8d206d89be95.camel@intel.com> (raw)
In-Reply-To: <20260129082756.1096935-4-brian3.nguyen@intel.com>

On Thu, 2026-01-29 at 08:27 +0000, Brian Nguyen wrote:
> Originally, page reclamation is handled by the same fence as tlb
> invalidation and uses its seqno, so there was no reason to separate
> out
> the handlers. However in hindsight, for readability, and possible
> future changes, it seems more beneficial to move this all out to its
> own
> function.
> 
> Signed-off-by: Brian Nguyen <brian3.nguyen@intel.com>

Reviewed-by: Stuart Summers <stuart.summers@intel.com>

> ---
>  drivers/gpu/drm/xe/xe_guc_ct.c       | 23 ++++++++---------------
>  drivers/gpu/drm/xe/xe_page_reclaim.c | 20 ++++++++++++++++++++
>  drivers/gpu/drm/xe/xe_page_reclaim.h |  3 +++
>  3 files changed, 31 insertions(+), 15 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c
> b/drivers/gpu/drm/xe/xe_guc_ct.c
> index dfbf76037b04..3d69314170f8 100644
> --- a/drivers/gpu/drm/xe/xe_guc_ct.c
> +++ b/drivers/gpu/drm/xe/xe_guc_ct.c
> @@ -31,6 +31,7 @@
>  #include "xe_guc_submit.h"
>  #include "xe_guc_tlb_inval.h"
>  #include "xe_map.h"
> +#include "xe_page_reclaim.h"
>  #include "xe_pm.h"
>  #include "xe_sriov_vf.h"
>  #include "xe_trace_guc.h"
> @@ -1604,17 +1605,11 @@ static int process_g2h_msg(struct xe_guc_ct
> *ct, u32 *msg, u32 len)
>                 ret = xe_guc_pagefault_handler(guc, payload,
> adj_len);
>                 break;
>         case XE_GUC_ACTION_TLB_INVALIDATION_DONE:
> -       case XE_GUC_ACTION_PAGE_RECLAMATION_DONE:
> -               /*
> -                * Page reclamation is an extension of TLB
> invalidation. Both
> -                * operations share the same seqno and fence. When
> either
> -                * action completes, we need to signal the
> corresponding
> -                * fence. Since the handling logic (lookup fence by
> seqno,
> -                * fence signalling) is identical, we use the same
> handler
> -                * for both G2H events.
> -                */
>                 ret = xe_guc_tlb_inval_done_handler(guc, payload,
> adj_len);
>                 break;
> +       case XE_GUC_ACTION_PAGE_RECLAMATION_DONE:
> +               ret = xe_guc_page_reclaim_done_handler(guc, payload,
> adj_len);
> +               break;
>         case XE_GUC_ACTION_GUC2PF_RELAY_FROM_VF:
>                 ret = xe_guc_relay_process_guc2pf(&guc->relay, hxg,
> hxg_len);
>                 break;
> @@ -1822,15 +1817,13 @@ static void g2h_fast_path(struct xe_guc_ct
> *ct, u32 *msg, u32 len)
>                 ret = xe_guc_pagefault_handler(guc, payload,
> adj_len);
>                 break;
>         case XE_GUC_ACTION_TLB_INVALIDATION_DONE:
> -       case XE_GUC_ACTION_PAGE_RECLAMATION_DONE:
> -               /*
> -                * Seqno and fence handling of page reclamation and
> TLB
> -                * invalidation is identical, so we can use the same
> handler
> -                * for both actions.
> -                */
>                 __g2h_release_space(ct, len);
>                 ret = xe_guc_tlb_inval_done_handler(guc, payload,
> adj_len);
>                 break;
> +       case XE_GUC_ACTION_PAGE_RECLAMATION_DONE:
> +               __g2h_release_space(ct, len);
> +               ret = xe_guc_page_reclaim_done_handler(guc, payload,
> adj_len);
> +               break;
>         default:
>                 xe_gt_warn(gt, "NOT_POSSIBLE");
>         }
> diff --git a/drivers/gpu/drm/xe/xe_page_reclaim.c
> b/drivers/gpu/drm/xe/xe_page_reclaim.c
> index e13c71a89da2..60b0fda59ce3 100644
> --- a/drivers/gpu/drm/xe/xe_page_reclaim.c
> +++ b/drivers/gpu/drm/xe/xe_page_reclaim.c
> @@ -11,6 +11,7 @@
>  #include "xe_page_reclaim.h"
>  
>  #include "xe_gt_stats.h"
> +#include "xe_guc_tlb_inval.h"
>  #include "xe_macros.h"
>  #include "xe_pat.h"
>  #include "xe_sa.h"
> @@ -130,3 +131,22 @@ int xe_page_reclaim_list_alloc_entries(struct
> xe_page_reclaim_list *prl)
>  
>         return page ? 0 : -ENOMEM;
>  }
> +
> +/**
> + * xe_guc_page_reclaim_done_handler() - Page reclaim done handler
> + * @guc: guc
> + * @msg: message indicating page reclamation done
> + * @len: length of message
> + *
> + * Page reclamation is an extension of TLB invalidation. Both
> + * operations share the same seqno and fence. When either
> + * action completes, we need to signal the corresponding
> + * fence. Since the handling logic is currently identical, this
> + * function delegates to the TLB invalidation handler.
> + *
> + * Return: 0 on success, -EPROTO for malformed messages.
> + */
> +int xe_guc_page_reclaim_done_handler(struct xe_guc *guc, u32 *msg,
> u32 len)
> +{
> +       return xe_guc_tlb_inval_done_handler(guc, msg, len);
> +}
> diff --git a/drivers/gpu/drm/xe/xe_page_reclaim.h
> b/drivers/gpu/drm/xe/xe_page_reclaim.h
> index 3dd103e37beb..0412611f3af7 100644
> --- a/drivers/gpu/drm/xe/xe_page_reclaim.h
> +++ b/drivers/gpu/drm/xe/xe_page_reclaim.h
> @@ -20,6 +20,7 @@ struct xe_tlb_inval;
>  struct xe_tlb_inval_fence;
>  struct xe_tile;
>  struct xe_gt;
> +struct xe_guc;
>  struct xe_vma;
>  
>  struct xe_guc_page_reclaim_entry {
> @@ -122,4 +123,6 @@ static inline void
> xe_page_reclaim_entries_put(struct xe_guc_page_reclaim_entry
>                 put_page(virt_to_page(entries));
>  }
>  
> +int xe_guc_page_reclaim_done_handler(struct xe_guc *guc, u32 *msg,
> u32 len);
> +
>  #endif /* _XE_PAGE_RECLAIM_H_ */


  parent reply	other threads:[~2026-02-23 22:45 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-01-29  8:27 [PATCH 1/2] drm/xe: Skip over non leaf pte for PRL generation Brian Nguyen
2026-01-29  8:27 ` [PATCH 2/2] drm/xe: Move page reclaim done_handler to own func Brian Nguyen
2026-01-29 21:40   ` Lin, Shuicheng
2026-02-23 22:45   ` Summers, Stuart [this message]
2026-02-23 22:51   ` Matthew Brost
2026-01-29  9:14 ` ✓ CI.KUnit: success for series starting with [1/2] drm/xe: Skip over non leaf pte for PRL generation Patchwork
2026-01-29  9:48 ` ✓ Xe.CI.BAT: " Patchwork
2026-02-23 21:48 ` [PATCH 1/2] " Nguyen, Brian3
2026-02-23 22:49 ` Summers, Stuart
2026-02-23 22:59   ` Matthew Brost
2026-02-23 23:07     ` Summers, Stuart
2026-02-23 23:33       ` Nguyen, Brian3
2026-02-24  1:45         ` Matthew Brost
2026-02-24  2:02           ` Matthew Brost
2026-02-25  6:45             ` Nguyen, Brian3
2026-02-25  7:19               ` Matthew Brost

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=a05f998d318f822c761098d6e6cc8d206d89be95.camel@intel.com \
    --to=stuart.summers@intel.com \
    --cc=brian3.nguyen@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=matthew.d.roper@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox