From: "Christian König" <christian.koenig@amd.com> To: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>, linaro-mm-sig@lists.linaro.org, dri-devel@lists.freedesktop.org, linux-media@vger.kernel.org, intel-gfx@lists.freedesktop.org Cc: daniel@ffwll.ch Subject: Re: [PATCH 09/28] dma-buf: use the new iterator in dma_resv_poll Date: Tue, 5 Oct 2021 10:16:00 +0200 [thread overview] Message-ID: <e43bc3f9-60be-0f7f-b1a7-3cd2fe1a6289@amd.com> (raw) In-Reply-To: <ef650439-a418-979b-56fb-4cf10f91747e@linux.intel.com> Am 05.10.21 um 09:44 schrieb Tvrtko Ursulin: > > On 01/10/2021 11:05, Christian König wrote: >> Simplify the code a bit. >> >> Signed-off-by: Christian König <christian.koenig@amd.com> >> --- >> drivers/dma-buf/dma-buf.c | 36 ++++++------------------------------ >> 1 file changed, 6 insertions(+), 30 deletions(-) >> >> diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c >> index 8242b5d9baeb..beb504a92d60 100644 >> --- a/drivers/dma-buf/dma-buf.c >> +++ b/drivers/dma-buf/dma-buf.c >> @@ -209,19 +209,14 @@ static void dma_buf_poll_cb(struct dma_fence >> *fence, struct dma_fence_cb *cb) >> dma_fence_put(fence); >> } >> -static bool dma_buf_poll_shared(struct dma_resv *resv, >> +static bool dma_buf_poll_add_cb(struct dma_resv *resv, bool write, >> struct dma_buf_poll_cb_t *dcb) >> { >> - struct dma_resv_list *fobj = dma_resv_shared_list(resv); >> + struct dma_resv_iter cursor; >> struct dma_fence *fence; >> - int i, r; >> - >> - if (!fobj) >> - return false; >> + int r; >> - for (i = 0; i < fobj->shared_count; ++i) { >> - fence = rcu_dereference_protected(fobj->shared[i], >> - dma_resv_held(resv)); >> + dma_resv_for_each_fence(&cursor, resv, write, fence) { >> dma_fence_get(fence); >> r = dma_fence_add_callback(fence, &dcb->cb, dma_buf_poll_cb); >> if (!r) > > It is unchanged with this patch, but are the semantics supposed to be > like this? Signal poll event if _any_ of the shared fences has been > signaled? That had Daniel and me confused for a moment as well. We don't signal the poll when any of the shared fences has signaled, but rather install a callback on the first not-signaled fence. This callback then issues a re-test of the poll and only if we can't find any more fence the poll is considered signaled (at least that's the idea, the coding could as well be broken). Christian. > > Regards, > > Tvrtko > >> @@ -232,24 +227,6 @@ static bool dma_buf_poll_shared(struct dma_resv >> *resv, >> return false; >> } >> -static bool dma_buf_poll_excl(struct dma_resv *resv, >> - struct dma_buf_poll_cb_t *dcb) >> -{ >> - struct dma_fence *fence = dma_resv_excl_fence(resv); >> - int r; >> - >> - if (!fence) >> - return false; >> - >> - dma_fence_get(fence); >> - r = dma_fence_add_callback(fence, &dcb->cb, dma_buf_poll_cb); >> - if (!r) >> - return true; >> - dma_fence_put(fence); >> - >> - return false; >> -} >> - >> static __poll_t dma_buf_poll(struct file *file, poll_table *poll) >> { >> struct dma_buf *dmabuf; >> @@ -282,8 +259,7 @@ static __poll_t dma_buf_poll(struct file *file, >> poll_table *poll) >> spin_unlock_irq(&dmabuf->poll.lock); >> if (events & EPOLLOUT) { >> - if (!dma_buf_poll_shared(resv, dcb) && >> - !dma_buf_poll_excl(resv, dcb)) >> + if (!dma_buf_poll_add_cb(resv, true, dcb)) >> /* No callback queued, wake up any other waiters */ >> dma_buf_poll_cb(NULL, &dcb->cb); >> else >> @@ -303,7 +279,7 @@ static __poll_t dma_buf_poll(struct file *file, >> poll_table *poll) >> spin_unlock_irq(&dmabuf->poll.lock); >> if (events & EPOLLIN) { >> - if (!dma_buf_poll_excl(resv, dcb)) >> + if (!dma_buf_poll_add_cb(resv, false, dcb)) >> /* No callback queued, wake up any other waiters */ >> dma_buf_poll_cb(NULL, &dcb->cb); >> else >>
WARNING: multiple messages have this Message-ID (diff)
From: "Christian König" <christian.koenig@amd.com> To: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>, linaro-mm-sig@lists.linaro.org, dri-devel@lists.freedesktop.org, linux-media@vger.kernel.org, intel-gfx@lists.freedesktop.org Cc: daniel@ffwll.ch Subject: Re: [Intel-gfx] [PATCH 09/28] dma-buf: use the new iterator in dma_resv_poll Date: Tue, 5 Oct 2021 10:16:00 +0200 [thread overview] Message-ID: <e43bc3f9-60be-0f7f-b1a7-3cd2fe1a6289@amd.com> (raw) In-Reply-To: <ef650439-a418-979b-56fb-4cf10f91747e@linux.intel.com> Am 05.10.21 um 09:44 schrieb Tvrtko Ursulin: > > On 01/10/2021 11:05, Christian König wrote: >> Simplify the code a bit. >> >> Signed-off-by: Christian König <christian.koenig@amd.com> >> --- >> drivers/dma-buf/dma-buf.c | 36 ++++++------------------------------ >> 1 file changed, 6 insertions(+), 30 deletions(-) >> >> diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c >> index 8242b5d9baeb..beb504a92d60 100644 >> --- a/drivers/dma-buf/dma-buf.c >> +++ b/drivers/dma-buf/dma-buf.c >> @@ -209,19 +209,14 @@ static void dma_buf_poll_cb(struct dma_fence >> *fence, struct dma_fence_cb *cb) >> dma_fence_put(fence); >> } >> -static bool dma_buf_poll_shared(struct dma_resv *resv, >> +static bool dma_buf_poll_add_cb(struct dma_resv *resv, bool write, >> struct dma_buf_poll_cb_t *dcb) >> { >> - struct dma_resv_list *fobj = dma_resv_shared_list(resv); >> + struct dma_resv_iter cursor; >> struct dma_fence *fence; >> - int i, r; >> - >> - if (!fobj) >> - return false; >> + int r; >> - for (i = 0; i < fobj->shared_count; ++i) { >> - fence = rcu_dereference_protected(fobj->shared[i], >> - dma_resv_held(resv)); >> + dma_resv_for_each_fence(&cursor, resv, write, fence) { >> dma_fence_get(fence); >> r = dma_fence_add_callback(fence, &dcb->cb, dma_buf_poll_cb); >> if (!r) > > It is unchanged with this patch, but are the semantics supposed to be > like this? Signal poll event if _any_ of the shared fences has been > signaled? That had Daniel and me confused for a moment as well. We don't signal the poll when any of the shared fences has signaled, but rather install a callback on the first not-signaled fence. This callback then issues a re-test of the poll and only if we can't find any more fence the poll is considered signaled (at least that's the idea, the coding could as well be broken). Christian. > > Regards, > > Tvrtko > >> @@ -232,24 +227,6 @@ static bool dma_buf_poll_shared(struct dma_resv >> *resv, >> return false; >> } >> -static bool dma_buf_poll_excl(struct dma_resv *resv, >> - struct dma_buf_poll_cb_t *dcb) >> -{ >> - struct dma_fence *fence = dma_resv_excl_fence(resv); >> - int r; >> - >> - if (!fence) >> - return false; >> - >> - dma_fence_get(fence); >> - r = dma_fence_add_callback(fence, &dcb->cb, dma_buf_poll_cb); >> - if (!r) >> - return true; >> - dma_fence_put(fence); >> - >> - return false; >> -} >> - >> static __poll_t dma_buf_poll(struct file *file, poll_table *poll) >> { >> struct dma_buf *dmabuf; >> @@ -282,8 +259,7 @@ static __poll_t dma_buf_poll(struct file *file, >> poll_table *poll) >> spin_unlock_irq(&dmabuf->poll.lock); >> if (events & EPOLLOUT) { >> - if (!dma_buf_poll_shared(resv, dcb) && >> - !dma_buf_poll_excl(resv, dcb)) >> + if (!dma_buf_poll_add_cb(resv, true, dcb)) >> /* No callback queued, wake up any other waiters */ >> dma_buf_poll_cb(NULL, &dcb->cb); >> else >> @@ -303,7 +279,7 @@ static __poll_t dma_buf_poll(struct file *file, >> poll_table *poll) >> spin_unlock_irq(&dmabuf->poll.lock); >> if (events & EPOLLIN) { >> - if (!dma_buf_poll_excl(resv, dcb)) >> + if (!dma_buf_poll_add_cb(resv, false, dcb)) >> /* No callback queued, wake up any other waiters */ >> dma_buf_poll_cb(NULL, &dcb->cb); >> else >>
next prev parent reply other threads:[~2021-10-05 8:16 UTC|newest] Thread overview: 100+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-10-01 10:05 Deploying new iterator interface for dma-buf Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 01/28] dma-buf: add dma_resv_for_each_fence_unlocked v7 Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-04 9:29 ` Tvrtko Ursulin 2021-10-04 9:29 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-04 9:53 ` Christian König 2021-10-04 9:53 ` [Intel-gfx] " Christian König 2021-10-04 10:34 ` Tvrtko Ursulin 2021-10-04 10:34 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-04 10:44 ` Christian König 2021-10-04 10:44 ` [Intel-gfx] " Christian König 2021-10-04 10:50 ` Tvrtko Ursulin 2021-10-04 10:50 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-04 12:59 ` Christian König 2021-10-04 12:59 ` [Intel-gfx] " Christian König 2021-10-04 13:15 ` Tvrtko Ursulin 2021-10-04 13:15 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-01 10:05 ` [PATCH 02/28] dma-buf: add dma_resv_for_each_fence Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 03/28] dma-buf: add dma_resv selftest Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-04 13:11 ` Tvrtko Ursulin 2021-10-04 13:11 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-01 10:05 ` [PATCH 04/28] dma-buf: use new iterator in dma_resv_copy_fences Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 05/28] dma-buf: use new iterator in dma_resv_get_fences v3 Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 06/28] dma-buf: use new iterator in dma_resv_wait_timeout Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 07/28] dma-buf: use new iterator in dma_resv_test_signaled Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 08/28] dma-buf: use the new iterator in dma_buf_debug_show Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-05 7:35 ` Tvrtko Ursulin 2021-10-05 7:35 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-01 10:05 ` [PATCH 09/28] dma-buf: use the new iterator in dma_resv_poll Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-05 7:44 ` Tvrtko Ursulin 2021-10-05 7:44 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-05 8:16 ` Christian König [this message] 2021-10-05 8:16 ` Christian König 2021-10-05 8:41 ` Tvrtko Ursulin 2021-10-05 8:41 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-01 10:05 ` [PATCH 10/28] drm/ttm: use the new iterator in ttm_bo_flush_all_fences Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 11/28] drm/amdgpu: use the new iterator in amdgpu_sync_resv Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 12/28] drm/amdgpu: use new iterator in amdgpu_ttm_bo_eviction_valuable Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 13/28] drm/amdgpu: use new iterator in amdgpu_vm_prt_fini Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 14/28] drm/msm: use new iterator in msm_gem_describe Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 15/28] drm/radeon: use new iterator in radeon_sync_resv Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 16/28] drm/scheduler: use new iterator in drm_sched_job_add_implicit_dependencies v2 Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:05 ` [PATCH 17/28] drm/i915: use the new iterator in i915_gem_busy_ioctl v2 Christian König 2021-10-01 10:05 ` [Intel-gfx] " Christian König 2021-10-01 10:37 ` Tvrtko Ursulin 2021-10-01 10:37 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-01 10:39 ` Christian König 2021-10-01 10:39 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 18/28] drm/i915: use the new iterator in i915_sw_fence_await_reservation v3 Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 19/28] drm/i915: use the new iterator in i915_request_await_object v2 Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 20/28] drm/i915: use new iterator in i915_gem_object_wait_reservation Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 21/28] drm/i915: use new iterator in i915_gem_object_wait_priority Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 22/28] drm/i915: use new cursor in intel_prepare_plane_fb Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 23/28] drm: use new iterator in drm_gem_fence_array_add_implicit v3 Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-05 7:48 ` Tvrtko Ursulin 2021-10-05 7:48 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-01 10:06 ` [PATCH 24/28] drm: use new iterator in drm_gem_plane_helper_prepare_fb Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-05 7:53 ` Tvrtko Ursulin 2021-10-05 7:53 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-05 10:27 ` Christian König 2021-10-05 10:27 ` [Intel-gfx] " Christian König 2021-10-05 10:47 ` Tvrtko Ursulin 2021-10-05 10:47 ` [Intel-gfx] " Tvrtko Ursulin 2021-10-05 11:23 ` Christian König 2021-10-05 11:23 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 25/28] drm/nouveau: use the new iterator in nouveau_fence_sync Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 26/28] drm/nouveau: use the new interator in nv50_wndw_prepare_fb Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 27/28] drm/etnaviv: use new iterator in etnaviv_gem_describe Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 10:06 ` [PATCH 28/28] drm/etnaviv: replace dma_resv_get_excl_unlocked Christian König 2021-10-01 10:06 ` [Intel-gfx] " Christian König 2021-10-01 17:59 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [01/28] dma-buf: add dma_resv_for_each_fence_unlocked v7 Patchwork 2021-10-01 18:01 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork 2021-10-01 18:33 ` [Intel-gfx] ✗ Fi.CI.BAT: failure " Patchwork 2021-10-05 11:37 Deploying new iterator interface for dma-buf Christian König 2021-10-05 11:37 ` [PATCH 09/28] dma-buf: use the new iterator in dma_resv_poll Christian König
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=e43bc3f9-60be-0f7f-b1a7-3cd2fe1a6289@amd.com \ --to=christian.koenig@amd.com \ --cc=daniel@ffwll.ch \ --cc=dri-devel@lists.freedesktop.org \ --cc=intel-gfx@lists.freedesktop.org \ --cc=linaro-mm-sig@lists.linaro.org \ --cc=linux-media@vger.kernel.org \ --cc=tvrtko.ursulin@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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.