From: Francois Dugast <francois.dugast@intel.com>
To: Matthew Brost <matthew.brost@intel.com>
Cc: <intel-xe@lists.freedesktop.org>,
<thomas.hellstrom@linux.intel.com>,
<himal.prasad.ghimiray@intel.com>, <michal.mrozek@intel.com>
Subject: Re: [PATCH 04/11] drm/xe: Implement xe_pagefault_handler
Date: Thu, 28 Aug 2025 13:26:43 +0200 [thread overview]
Message-ID: <aLA88yLhaBk4acE8@fdugast-desk> (raw)
In-Reply-To: <20250806062242.1090416-5-matthew.brost@intel.com>
On Tue, Aug 05, 2025 at 11:22:35PM -0700, Matthew Brost wrote:
> Enqueue (copy) the input struct xe_pagefault into a queue (i.e., into a
> memory buffer) and schedule a worker to service it.
>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Francois Dugast <francois.dugast@intel.com>
> ---
> drivers/gpu/drm/xe/xe_pagefault.c | 32 +++++++++++++++++++++++++++++--
> 1 file changed, 30 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_pagefault.c b/drivers/gpu/drm/xe/xe_pagefault.c
> index aef389e51612..98be3203a9df 100644
> --- a/drivers/gpu/drm/xe/xe_pagefault.c
> +++ b/drivers/gpu/drm/xe/xe_pagefault.c
> @@ -3,6 +3,8 @@
> * Copyright © 2025 Intel Corporation
> */
>
> +#include <linux/circ_buf.h>
> +
> #include <drm/drm_managed.h>
>
> #include "xe_device.h"
> @@ -156,6 +158,14 @@ void xe_pagefault_reset(struct xe_device *xe, struct xe_gt *gt)
> xe_pagefault_queue_reset(xe, gt, xe->usm.pf_queue + i);
> }
>
> +static bool xe_pagefault_queue_full(struct xe_pagefault_queue *pf_queue)
> +{
> + lockdep_assert_held(&pf_queue->lock);
> +
> + return CIRC_SPACE(pf_queue->head, pf_queue->tail, pf_queue->size) <=
> + xe_pagefault_entry_size();
> +}
> +
> /**
> * xe_pagefault_handler() - Page fault handler
> * @xe: xe device instance
> @@ -168,6 +178,24 @@ void xe_pagefault_reset(struct xe_device *xe, struct xe_gt *gt)
> */
> int xe_pagefault_handler(struct xe_device *xe, struct xe_pagefault *pf)
> {
> - /* TODO - implement */
> - return 0;
> + struct xe_pagefault_queue *pf_queue = xe->usm.pf_queue +
> + (pf->consumer.asid % XE_PAGEFAULT_QUEUE_COUNT);
> + unsigned long flags;
> + bool full;
> +
> + spin_lock_irqsave(&pf_queue->lock, flags);
> + full = xe_pagefault_queue_full(pf_queue);
> + if (!full) {
> + memcpy(pf_queue->data + pf_queue->head, pf, sizeof(*pf));
> + pf_queue->head = (pf_queue->head + xe_pagefault_entry_size()) %
> + pf_queue->size;
> + queue_work(xe->usm.pf_wq, &pf_queue->worker);
> + } else {
> + drm_warn(&xe->drm,
> + "PageFault Queue (%d) full, shouldn't be possible\n",
> + pf->consumer.asid % XE_PAGEFAULT_QUEUE_COUNT);
> + }
> + spin_unlock_irqrestore(&pf_queue->lock, flags);
> +
> + return full ? -ENOSPC : 0;
> }
> --
> 2.34.1
>
next prev parent reply other threads:[~2025-08-28 11:26 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-08-06 6:22 [PATCH 00/11] Pagefault refactor, fine grained fault locking, threaded prefetch Matthew Brost
2025-08-06 6:22 ` [PATCH 01/11] drm/xe: Stub out new pagefault layer Matthew Brost
2025-08-06 23:01 ` Summers, Stuart
2025-08-06 23:53 ` Matthew Brost
2025-08-07 17:20 ` Summers, Stuart
2025-08-07 18:10 ` Matthew Brost
2025-08-28 20:18 ` Summers, Stuart
2025-08-28 20:20 ` Matthew Brost
2025-08-27 15:29 ` Francois Dugast
2025-08-27 16:03 ` Matthew Brost
2025-08-27 16:25 ` Francois Dugast
2025-08-27 16:40 ` Matthew Brost
2025-08-27 18:00 ` Matthew Brost
2025-08-28 20:08 ` Summers, Stuart
2025-08-06 6:22 ` [PATCH 02/11] drm/xe: Implement xe_pagefault_init Matthew Brost
2025-08-06 23:08 ` Summers, Stuart
2025-08-06 23:59 ` Matthew Brost
2025-08-07 18:22 ` Summers, Stuart
2025-08-27 16:30 ` Francois Dugast
2025-08-27 16:49 ` Matthew Brost
2025-08-28 20:10 ` Summers, Stuart
2025-08-28 20:14 ` Matthew Brost
2025-08-28 20:19 ` Summers, Stuart
2025-08-06 6:22 ` [PATCH 03/11] drm/xe: Implement xe_pagefault_reset Matthew Brost
2025-08-06 23:16 ` Summers, Stuart
2025-08-07 0:12 ` Matthew Brost
2025-08-07 18:29 ` Summers, Stuart
2025-08-06 6:22 ` [PATCH 04/11] drm/xe: Implement xe_pagefault_handler Matthew Brost
2025-08-28 11:26 ` Francois Dugast [this message]
2025-08-28 20:24 ` Summers, Stuart
2025-08-06 6:22 ` [PATCH 05/11] drm/xe: Implement xe_pagefault_queue_work Matthew Brost
2025-08-28 12:29 ` Francois Dugast
2025-08-28 18:39 ` Matthew Brost
2025-08-28 22:04 ` Summers, Stuart
2025-08-29 0:51 ` Matthew Brost
2025-08-06 6:22 ` [PATCH 06/11] drm/xe: Add xe_guc_pagefault layer Matthew Brost
2025-08-28 13:27 ` Francois Dugast
2025-08-28 18:38 ` Matthew Brost
2025-08-28 22:11 ` Summers, Stuart
2025-08-29 0:54 ` Matthew Brost
2025-08-06 6:22 ` [PATCH 07/11] drm/xe: Remove unused GT page fault code Matthew Brost
2025-08-28 19:13 ` Summers, Stuart
2025-08-06 6:22 ` [PATCH 08/11] drm/xe: Fine grained page fault locking Matthew Brost
2025-08-06 6:22 ` [PATCH 09/11] drm/xe: Allow prefetch-only VM bind IOCTLs to use VM read lock Matthew Brost
2025-08-06 6:22 ` [PATCH 10/11] drm/xe: Thread prefetch of SVM ranges Matthew Brost
2025-08-28 22:55 ` Summers, Stuart
2025-08-29 1:06 ` Matthew Brost
2025-08-06 6:22 ` [PATCH 11/11] drm/xe: Add num_pf_queue modparam Matthew Brost
2025-08-28 22:58 ` Summers, Stuart
2025-08-06 6:36 ` ✗ CI.checkpatch: warning for Pagefault refactor, fine grained fault locking, threaded prefetch Patchwork
2025-08-06 6:36 ` ✗ CI.KUnit: failure " Patchwork
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=aLA88yLhaBk4acE8@fdugast-desk \
--to=francois.dugast@intel.com \
--cc=himal.prasad.ghimiray@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.brost@intel.com \
--cc=michal.mrozek@intel.com \
--cc=thomas.hellstrom@linux.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