Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Brost <matthew.brost@intel.com>
To: Tomasz Lis <tomasz.lis@intel.com>
Cc: intel-xe@lists.freedesktop.org,
	"Michał Winiarski" <michal.winiarski@intel.com>,
	"Michał Wajdeczko" <michal.wajdeczko@intel.com>,
	"Piotr Piórkowski" <piotr.piorkowski@intel.com>
Subject: Re: [PATCH v4 2/4] drm/xe: Wrappers for setting and getting LRC references
Date: Thu, 26 Feb 2026 15:01:00 -0800	[thread overview]
Message-ID: <aaDQrDdO2rxRWxoH@lstrano-desk.jf.intel.com> (raw)
In-Reply-To: <20260226212701.2937065-3-tomasz.lis@intel.com>

On Thu, Feb 26, 2026 at 10:26:59PM +0100, Tomasz Lis wrote:
> There is a small but non-zero chance that VF post migration fixups
> are running on an exec queue during teardown. The chances are
> decreased by starting the teardown by releasing guc_id, but remain
> non-zero. On the other hand the sync between fixups and EQ creation
> (wait_valid_ggtt) drastically increases the chance for such parallel
> teardown if queue creation error path is entered (err_lrc label).
> 
> The exec queue itself is not going to cause an issue, but LRCs have
> a small chance of getting freed during the fixups.
> 
> Creating a setter and a getter makes it easier to protect the fixup
> operations with a lock. For other driver activities, the original
> access method (without any protection) can still be used.
> 
> v2: Separate lock, only for LRCs. Kerneldoc fixes. Subject tag fix.
> 
> Signed-off-by: Tomasz Lis <tomasz.lis@intel.com>

Reviewed-by: Matthew Brost <matthew.brost@intel.com>

> ---
>  drivers/gpu/drm/xe/xe_exec_queue.c       | 73 ++++++++++++++++++------
>  drivers/gpu/drm/xe/xe_exec_queue.h       |  1 +
>  drivers/gpu/drm/xe/xe_exec_queue_types.h |  5 ++
>  3 files changed, 60 insertions(+), 19 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c
> index b4ef725a682d..c0e95f1a9911 100644
> --- a/drivers/gpu/drm/xe/xe_exec_queue.c
> +++ b/drivers/gpu/drm/xe/xe_exec_queue.c
> @@ -231,6 +231,7 @@ static struct xe_exec_queue *__xe_exec_queue_alloc(struct xe_device *xe,
>  	INIT_LIST_HEAD(&q->hw_engine_group_link);
>  	INIT_LIST_HEAD(&q->pxp.link);
>  	spin_lock_init(&q->multi_queue.lock);
> +	spin_lock_init(&q->lrc_lookup_lock);
>  	q->multi_queue.priority = XE_MULTI_QUEUE_PRIORITY_NORMAL;
>  
>  	q->sched_props.timeslice_us = hwe->eclass->sched_props.timeslice_us;
> @@ -270,6 +271,56 @@ static struct xe_exec_queue *__xe_exec_queue_alloc(struct xe_device *xe,
>  	return q;
>  }
>  
> +static void xe_exec_queue_set_lrc(struct xe_exec_queue *q, struct xe_lrc *lrc, u16 idx)
> +{
> +	xe_assert(gt_to_xe(q->gt), idx < q->width);
> +
> +	scoped_guard(spinlock, &q->lrc_lookup_lock)
> +		q->lrc[idx] = lrc;
> +}
> +
> +/**
> + * xe_exec_queue_get_lrc() - Get the LRC from exec queue.
> + * @q: The exec queue instance.
> + * @idx: Index within multi-LRC array.
> + *
> + * Retrieves LRC of given index for the exec queue under lock
> + * and takes reference.
> + *
> + * Return: Pointer to LRC on success, error on failure, NULL on
> + * lookup failure.
> + */
> +struct xe_lrc *xe_exec_queue_get_lrc(struct xe_exec_queue *q, u16 idx)
> +{
> +	struct xe_lrc *lrc;
> +
> +	xe_assert(gt_to_xe(q->gt), idx < q->width);
> +
> +	scoped_guard(spinlock, &q->lrc_lookup_lock) {
> +		lrc = q->lrc[idx];
> +		if (lrc)
> +			xe_lrc_get(lrc);
> +	}
> +
> +	return lrc;
> +}
> +
> +/**
> + * xe_exec_queue_lrc() - Get the LRC from exec queue.
> + * @q: The exec queue instance.
> + *
> + * Retrieves the primary LRC for the exec queue. Note that this function
> + * returns only the first LRC instance, even when multiple parallel LRCs
> + * are configured. This function does not increment reference count,
> + * so the reference can be just forgotten after use.
> + *
> + * Return: Pointer to LRC on success, error on failure
> + */
> +struct xe_lrc *xe_exec_queue_lrc(struct xe_exec_queue *q)
> +{
> +	return q->lrc[0];
> +}
> +
>  static void __xe_exec_queue_fini(struct xe_exec_queue *q)
>  {
>  	int i;
> @@ -327,8 +378,7 @@ static int __xe_exec_queue_init(struct xe_exec_queue *q, u32 exec_queue_flags)
>  			goto err_lrc;
>  		}
>  
> -		/* Pairs with READ_ONCE to xe_exec_queue_contexts_hwsp_rebase */
> -		WRITE_ONCE(q->lrc[i], lrc);
> +		xe_exec_queue_set_lrc(q, lrc, i);
>  	}
>  
>  	return 0;
> @@ -1288,21 +1338,6 @@ int xe_exec_queue_get_property_ioctl(struct drm_device *dev, void *data,
>  	return ret;
>  }
>  
> -/**
> - * xe_exec_queue_lrc() - Get the LRC from exec queue.
> - * @q: The exec_queue.
> - *
> - * Retrieves the primary LRC for the exec queue. Note that this function
> - * returns only the first LRC instance, even when multiple parallel LRCs
> - * are configured.
> - *
> - * Return: Pointer to LRC on success, error on failure
> - */
> -struct xe_lrc *xe_exec_queue_lrc(struct xe_exec_queue *q)
> -{
> -	return q->lrc[0];
> -}
> -
>  /**
>   * xe_exec_queue_is_lr() - Whether an exec_queue is long-running
>   * @q: The exec_queue
> @@ -1662,14 +1697,14 @@ int xe_exec_queue_contexts_hwsp_rebase(struct xe_exec_queue *q, void *scratch)
>  	for (i = 0; i < q->width; ++i) {
>  		struct xe_lrc *lrc;
>  
> -		/* Pairs with WRITE_ONCE in __xe_exec_queue_init  */
> -		lrc = READ_ONCE(q->lrc[i]);
> +		lrc = xe_exec_queue_get_lrc(q, i);
>  		if (!lrc)
>  			continue;
>  
>  		xe_lrc_update_memirq_regs_with_address(lrc, q->hwe, scratch);
>  		xe_lrc_update_hwctx_regs_with_address(lrc);
>  		err = xe_lrc_setup_wa_bb_with_scratch(lrc, q->hwe, scratch);
> +		xe_lrc_put(lrc);
>  		if (err)
>  			break;
>  	}
> diff --git a/drivers/gpu/drm/xe/xe_exec_queue.h b/drivers/gpu/drm/xe/xe_exec_queue.h
> index c9e3a7c2d249..a82d99bd77bc 100644
> --- a/drivers/gpu/drm/xe/xe_exec_queue.h
> +++ b/drivers/gpu/drm/xe/xe_exec_queue.h
> @@ -160,6 +160,7 @@ void xe_exec_queue_update_run_ticks(struct xe_exec_queue *q);
>  int xe_exec_queue_contexts_hwsp_rebase(struct xe_exec_queue *q, void *scratch);
>  
>  struct xe_lrc *xe_exec_queue_lrc(struct xe_exec_queue *q);
> +struct xe_lrc *xe_exec_queue_get_lrc(struct xe_exec_queue *q, u16 idx);
>  
>  /**
>   * xe_exec_queue_idle_skip_suspend() - Can exec queue skip suspend
> diff --git a/drivers/gpu/drm/xe/xe_exec_queue_types.h b/drivers/gpu/drm/xe/xe_exec_queue_types.h
> index 3791fed34ffa..a1f3938f4173 100644
> --- a/drivers/gpu/drm/xe/xe_exec_queue_types.h
> +++ b/drivers/gpu/drm/xe/xe_exec_queue_types.h
> @@ -257,6 +257,11 @@ struct xe_exec_queue {
>  	u64 tlb_flush_seqno;
>  	/** @hw_engine_group_link: link into exec queues in the same hw engine group */
>  	struct list_head hw_engine_group_link;
> +	/**
> +	 * @lrc_lookup_lock: Lock for protecting lrc array access. Only used when
> +	 * running in parallel to queue creation is possible.
> +	 */
> +	spinlock_t lrc_lookup_lock;
>  	/** @lrc: logical ring context for this exec queue */
>  	struct xe_lrc *lrc[] __counted_by(width);
>  };
> -- 
> 2.25.1
> 

  reply	other threads:[~2026-02-26 23:01 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-02-26 21:26 [PATCH v4 0/4] drm/xe/vf: Fix exec queue creation during post-migration recovery Tomasz Lis
2026-02-26 21:26 ` [PATCH v4 1/4] drm/xe/queue: Call fini on exec queue creation fail Tomasz Lis
2026-02-26 21:26 ` [PATCH v4 2/4] drm/xe: Wrappers for setting and getting LRC references Tomasz Lis
2026-02-26 23:01   ` Matthew Brost [this message]
2026-02-26 21:27 ` [PATCH v4 3/4] drm/xe/vf: Wait for all fixups before using default LRCs Tomasz Lis
2026-02-26 21:27 ` [PATCH v4 4/4] drm/xe/vf: Redo LRC creation while in VF fixups Tomasz Lis
2026-02-26 23:14   ` Matthew Brost
2026-02-26 21:32 ` ✓ CI.KUnit: success for drm/xe/vf: Fix exec queue creation during post-migration recovery (rev4) Patchwork
2026-02-26 22:21 ` ✓ Xe.CI.BAT: " Patchwork
2026-02-27  3:43 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-02-27 13:25   ` Lis, Tomasz

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=aaDQrDdO2rxRWxoH@lstrano-desk.jf.intel.com \
    --to=matthew.brost@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=michal.wajdeczko@intel.com \
    --cc=michal.winiarski@intel.com \
    --cc=piotr.piorkowski@intel.com \
    --cc=tomasz.lis@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