From: Matthew Brost <matthew.brost@intel.com>
To: <christian.koenig@amd.com>
Cc: <phasta@kernel.org>, <simona@ffwll.ch>, <sumit.semwal@linaro.org>,
<tvrtko.ursulin@igalia.com>, <dakr@kernel.org>,
<dri-devel@lists.freedesktop.org>, <linux-media@vger.kernel.org>,
<linaro-mm-sig@lists.linaro.org>
Subject: Re: [PATCH 06/10] drm/xe: stop using dma_fence_is_signaled_locked
Date: Wed, 24 Jun 2026 17:44:05 -0700 [thread overview]
Message-ID: <ajx51f6y9t3ER0te@gsse-cloud1.jf.intel.com> (raw)
In-Reply-To: <20260624122917.2483-7-christian.koenig@amd.com>
On Wed, Jun 24, 2026 at 01:13:30PM +0200, Christian König wrote:
> This use case is a bit more complicated since the irq worker is
> actually the one signaling the fence.
>
> The patch should not indroduce any functional change, but the code can
> probably be cleaned up quite a bit after the full patch set lands.
>
Most likely - this is pretty old code largely untouched from the
original Xe implementation - will look at this after this series lands.
> Signed-off-by: Christian König <christian.koenig@amd.com>
> ---
> drivers/gpu/drm/xe/xe_hw_fence.c | 28 +++++++++++++++-------------
> 1 file changed, 15 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_hw_fence.c b/drivers/gpu/drm/xe/xe_hw_fence.c
> index 14720623ad00..a4e0278559b8 100644
> --- a/drivers/gpu/drm/xe/xe_hw_fence.c
> +++ b/drivers/gpu/drm/xe/xe_hw_fence.c
> @@ -16,6 +16,8 @@
>
> static struct kmem_cache *xe_hw_fence_slab;
>
> +static struct xe_hw_fence *to_xe_hw_fence(struct dma_fence *fence);
> +
> int __init xe_hw_fence_module_init(void)
> {
> xe_hw_fence_slab = kmem_cache_create("xe_hw_fence",
> @@ -47,6 +49,16 @@ static void fence_free(struct rcu_head *rcu)
> kmem_cache_free(xe_hw_fence_slab, fence);
> }
>
> +static bool xe_hw_fence_signaled(struct dma_fence *dma_fence)
Maybe just forward declare this function to keep git blame intact?
Matt
> +{
> + struct xe_hw_fence *fence = to_xe_hw_fence(dma_fence);
> + struct xe_device *xe = fence->xe;
> + u32 seqno = xe_map_rd(xe, &fence->seqno_map, 0, u32);
> +
> + return dma_fence->error ||
> + !__dma_fence_is_later(dma_fence, dma_fence->seqno, seqno);
> +}
> +
> static void hw_fence_irq_run_cb(struct irq_work *work)
> {
> struct xe_hw_fence_irq *irq = container_of(work, typeof(*irq), work);
> @@ -60,7 +72,9 @@ static void hw_fence_irq_run_cb(struct irq_work *work)
> struct dma_fence *dma_fence = &fence->dma;
>
> trace_xe_hw_fence_try_signal(fence);
> - if (dma_fence_is_signaled_locked(dma_fence)) {
> + if (dma_fence_test_signaled_flag(dma_fence) ||
> + xe_hw_fence_signaled(dma_fence)) {
> + dma_fence_signal_locked(dma_fence);
> trace_xe_hw_fence_signal(fence);
> list_del_init(&fence->irq_link);
> dma_fence_put(dma_fence);
> @@ -120,8 +134,6 @@ void xe_hw_fence_ctx_finish(struct xe_hw_fence_ctx *ctx)
> {
> }
>
> -static struct xe_hw_fence *to_xe_hw_fence(struct dma_fence *fence);
> -
> static struct xe_hw_fence_irq *xe_hw_fence_irq(struct xe_hw_fence *fence)
> {
> return container_of(fence->dma.extern_lock, struct xe_hw_fence_irq,
> @@ -142,16 +154,6 @@ static const char *xe_hw_fence_get_timeline_name(struct dma_fence *dma_fence)
> return fence->name;
> }
>
> -static bool xe_hw_fence_signaled(struct dma_fence *dma_fence)
> -{
> - struct xe_hw_fence *fence = to_xe_hw_fence(dma_fence);
> - struct xe_device *xe = fence->xe;
> - u32 seqno = xe_map_rd(xe, &fence->seqno_map, 0, u32);
> -
> - return dma_fence->error ||
> - !__dma_fence_is_later(dma_fence, dma_fence->seqno, seqno);
> -}
> -
> static bool xe_hw_fence_enable_signaling(struct dma_fence *dma_fence)
> {
> struct xe_hw_fence *fence = to_xe_hw_fence(dma_fence);
> --
> 2.43.0
>
next prev parent reply other threads:[~2026-06-25 0:44 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-24 11:13 dma_fence cleanup/rework Christian König
2026-06-24 11:13 ` [PATCH 01/10] dma-buf: rename dma_fence_enable_sw_signaling Christian König
2026-06-25 0:39 ` Matthew Brost
2026-06-24 11:13 ` [PATCH 02/10] dma-buf: use dma_fence_test_signaled_flag() Christian König
2026-06-24 11:13 ` [PATCH 03/10] drm/amdgpu: " Christian König
2026-06-24 11:13 ` [PATCH 04/10] drm/nouveau: " Christian König
2026-06-24 11:13 ` [PATCH 05/10] drm/sched: " Christian König
2026-06-24 11:13 ` [PATCH 06/10] drm/xe: stop using dma_fence_is_signaled_locked Christian König
2026-06-25 0:44 ` Matthew Brost [this message]
2026-06-24 11:13 ` [PATCH 07/10] dma-buf: remove dma_fence_is_signaled_locked() Christian König
2026-06-24 11:13 ` [PATCH 08/10] dma-buf: remove signaling from dma_fence_is_signaled() Christian König
2026-06-24 11:13 ` [PATCH 09/10] dma-buf: remove signaling from dma_fence_enable_signaling() Christian König
2026-06-24 11:13 ` [PATCH 10/10] dma-buf: move locking into dma_fence_ops.enable_signaling Christian König
2026-06-25 7:25 ` dma_fence cleanup/rework Philipp Stanner
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=ajx51f6y9t3ER0te@gsse-cloud1.jf.intel.com \
--to=matthew.brost@intel.com \
--cc=christian.koenig@amd.com \
--cc=dakr@kernel.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=linaro-mm-sig@lists.linaro.org \
--cc=linux-media@vger.kernel.org \
--cc=phasta@kernel.org \
--cc=simona@ffwll.ch \
--cc=sumit.semwal@linaro.org \
--cc=tvrtko.ursulin@igalia.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