From: "Christian König" <christian.koenig@amd.com> To: "Rob Clark" <robdclark@gmail.com>, "Christian König" <ckoenig.leichtzumerken@gmail.com> Cc: dri-devel <dri-devel@lists.freedesktop.org>, Daniel Vetter <daniel@ffwll.ch>, Matthew Brost <matthew.brost@intel.com>, Rob Clark <robdclark@chromium.org>, Sumit Semwal <sumit.semwal@linaro.org>, Gustavo Padovan <gustavo@padovan.org>, "open list:SYNC FILE FRAMEWORK" <linux-media@vger.kernel.org>, "moderated list:DMA BUFFER SHARING FRAMEWORK" <linaro-mm-sig@lists.linaro.org>, open list <linux-kernel@vger.kernel.org> Subject: Re: [RFC 1/4] dma-fence: Add deadline awareness Date: Wed, 28 Jul 2021 09:03:29 +0200 [thread overview] Message-ID: <9edd7083-e6b3-b230-c273-8f2fbe76ca17@amd.com> (raw) In-Reply-To: <CAF6AEGuNxi_aeYE37FT3a-atCUWgepxs-9EwxMfpiMaU7wgqdQ@mail.gmail.com> Am 27.07.21 um 16:25 schrieb Rob Clark: > On Tue, Jul 27, 2021 at 12:11 AM Christian König > <ckoenig.leichtzumerken@gmail.com> wrote: >> Am 27.07.21 um 01:38 schrieb Rob Clark: >>> From: Rob Clark <robdclark@chromium.org> >>> >>> Add a way to hint to the fence signaler of an upcoming deadline, such as >>> vblank, which the fence waiter would prefer not to miss. This is to aid >>> the fence signaler in making power management decisions, like boosting >>> frequency as the deadline approaches and awareness of missing deadlines >>> so that can be factored in to the frequency scaling. >>> >>> Signed-off-by: Rob Clark <robdclark@chromium.org> >>> --- >>> drivers/dma-buf/dma-fence.c | 39 +++++++++++++++++++++++++++++++++++++ >>> include/linux/dma-fence.h | 17 ++++++++++++++++ >>> 2 files changed, 56 insertions(+) >>> >>> diff --git a/drivers/dma-buf/dma-fence.c b/drivers/dma-buf/dma-fence.c >>> index ce0f5eff575d..2e0d25ab457e 100644 >>> --- a/drivers/dma-buf/dma-fence.c >>> +++ b/drivers/dma-buf/dma-fence.c >>> @@ -910,6 +910,45 @@ dma_fence_wait_any_timeout(struct dma_fence **fences, uint32_t count, >>> } >>> EXPORT_SYMBOL(dma_fence_wait_any_timeout); >>> >>> + >>> +/** >>> + * dma_fence_set_deadline - set desired fence-wait deadline >>> + * @fence: the fence that is to be waited on >>> + * @deadline: the time by which the waiter hopes for the fence to be >>> + * signaled >>> + * >>> + * Inform the fence signaler of an upcoming deadline, such as vblank, by >>> + * which point the waiter would prefer the fence to be signaled by. This >>> + * is intended to give feedback to the fence signaler to aid in power >>> + * management decisions, such as boosting GPU frequency if a periodic >>> + * vblank deadline is approaching. >>> + */ >>> +void dma_fence_set_deadline(struct dma_fence *fence, ktime_t deadline) >>> +{ >>> + unsigned long flags; >>> + >>> + if (dma_fence_is_signaled(fence)) >>> + return; >>> + >>> + spin_lock_irqsave(fence->lock, flags); >>> + >>> + /* If we already have an earlier deadline, keep it: */ >>> + if (test_bit(DMA_FENCE_FLAG_HAS_DEADLINE_BIT, &fence->flags) && >>> + ktime_before(fence->deadline, deadline)) { >>> + spin_unlock_irqrestore(fence->lock, flags); >>> + return; >>> + } >>> + >>> + fence->deadline = deadline; >>> + set_bit(DMA_FENCE_FLAG_HAS_DEADLINE_BIT, &fence->flags); >>> + >>> + spin_unlock_irqrestore(fence->lock, flags); >>> + >>> + if (fence->ops->set_deadline) >>> + fence->ops->set_deadline(fence, deadline); >>> +} >>> +EXPORT_SYMBOL(dma_fence_set_deadline); >>> + >>> /** >>> * dma_fence_init - Initialize a custom fence. >>> * @fence: the fence to initialize >>> diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h >>> index 6ffb4b2c6371..4e6cfe4e6fbc 100644 >>> --- a/include/linux/dma-fence.h >>> +++ b/include/linux/dma-fence.h >>> @@ -88,6 +88,7 @@ struct dma_fence { >>> /* @timestamp replaced by @rcu on dma_fence_release() */ >>> struct rcu_head rcu; >>> }; >>> + ktime_t deadline; >> Mhm, adding the flag sounds ok to me but I'm a bit hesitating adding the >> deadline as extra field here. >> >> We tuned the dma_fence structure intentionally so that it is only 64 bytes. > Hmm, then I guess you wouldn't be a fan of also adding an hrtimer? > > We could push the ktime_t (and timer) down into the derived fence > class, but I think there is going to need to be some extra storage > *somewhere*.. maybe the fence signaler could get away with just > storing the nearest upcoming deadline per fence-context instead? I would just push that into the driver instead. You most likely don't want the deadline per fence anyway in complex scenarios, but rather per frame. And a frame is usually composed from multiple fences. Regards, Christian. > > BR, > -R > >> Regards, >> Christian. >> >>> u64 context; >>> u64 seqno; >>> unsigned long flags; >>> @@ -99,6 +100,7 @@ enum dma_fence_flag_bits { >>> DMA_FENCE_FLAG_SIGNALED_BIT, >>> DMA_FENCE_FLAG_TIMESTAMP_BIT, >>> DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, >>> + DMA_FENCE_FLAG_HAS_DEADLINE_BIT, >>> DMA_FENCE_FLAG_USER_BITS, /* must always be last member */ >>> }; >>> >>> @@ -261,6 +263,19 @@ struct dma_fence_ops { >>> */ >>> void (*timeline_value_str)(struct dma_fence *fence, >>> char *str, int size); >>> + >>> + /** >>> + * @set_deadline: >>> + * >>> + * Callback to allow a fence waiter to inform the fence signaler of an >>> + * upcoming deadline, such as vblank, by which point the waiter would >>> + * prefer the fence to be signaled by. This is intended to give feedback >>> + * to the fence signaler to aid in power management decisions, such as >>> + * boosting GPU frequency. >>> + * >>> + * This callback is optional. >>> + */ >>> + void (*set_deadline)(struct dma_fence *fence, ktime_t deadline); >>> }; >>> >>> void dma_fence_init(struct dma_fence *fence, const struct dma_fence_ops *ops, >>> @@ -586,6 +601,8 @@ static inline signed long dma_fence_wait(struct dma_fence *fence, bool intr) >>> return ret < 0 ? ret : 0; >>> } >>> >>> +void dma_fence_set_deadline(struct dma_fence *fence, ktime_t deadline); >>> + >>> struct dma_fence *dma_fence_get_stub(void); >>> struct dma_fence *dma_fence_allocate_private_stub(void); >>> u64 dma_fence_context_alloc(unsigned num);
WARNING: multiple messages have this Message-ID (diff)
From: "Christian König" <christian.koenig@amd.com> To: "Rob Clark" <robdclark@gmail.com>, "Christian König" <ckoenig.leichtzumerken@gmail.com> Cc: Matthew Brost <matthew.brost@intel.com>, Rob Clark <robdclark@chromium.org>, Gustavo Padovan <gustavo@padovan.org>, open list <linux-kernel@vger.kernel.org>, dri-devel <dri-devel@lists.freedesktop.org>, "moderated list:DMA BUFFER SHARING FRAMEWORK" <linaro-mm-sig@lists.linaro.org>, "open list:SYNC FILE FRAMEWORK" <linux-media@vger.kernel.org> Subject: Re: [RFC 1/4] dma-fence: Add deadline awareness Date: Wed, 28 Jul 2021 09:03:29 +0200 [thread overview] Message-ID: <9edd7083-e6b3-b230-c273-8f2fbe76ca17@amd.com> (raw) In-Reply-To: <CAF6AEGuNxi_aeYE37FT3a-atCUWgepxs-9EwxMfpiMaU7wgqdQ@mail.gmail.com> Am 27.07.21 um 16:25 schrieb Rob Clark: > On Tue, Jul 27, 2021 at 12:11 AM Christian König > <ckoenig.leichtzumerken@gmail.com> wrote: >> Am 27.07.21 um 01:38 schrieb Rob Clark: >>> From: Rob Clark <robdclark@chromium.org> >>> >>> Add a way to hint to the fence signaler of an upcoming deadline, such as >>> vblank, which the fence waiter would prefer not to miss. This is to aid >>> the fence signaler in making power management decisions, like boosting >>> frequency as the deadline approaches and awareness of missing deadlines >>> so that can be factored in to the frequency scaling. >>> >>> Signed-off-by: Rob Clark <robdclark@chromium.org> >>> --- >>> drivers/dma-buf/dma-fence.c | 39 +++++++++++++++++++++++++++++++++++++ >>> include/linux/dma-fence.h | 17 ++++++++++++++++ >>> 2 files changed, 56 insertions(+) >>> >>> diff --git a/drivers/dma-buf/dma-fence.c b/drivers/dma-buf/dma-fence.c >>> index ce0f5eff575d..2e0d25ab457e 100644 >>> --- a/drivers/dma-buf/dma-fence.c >>> +++ b/drivers/dma-buf/dma-fence.c >>> @@ -910,6 +910,45 @@ dma_fence_wait_any_timeout(struct dma_fence **fences, uint32_t count, >>> } >>> EXPORT_SYMBOL(dma_fence_wait_any_timeout); >>> >>> + >>> +/** >>> + * dma_fence_set_deadline - set desired fence-wait deadline >>> + * @fence: the fence that is to be waited on >>> + * @deadline: the time by which the waiter hopes for the fence to be >>> + * signaled >>> + * >>> + * Inform the fence signaler of an upcoming deadline, such as vblank, by >>> + * which point the waiter would prefer the fence to be signaled by. This >>> + * is intended to give feedback to the fence signaler to aid in power >>> + * management decisions, such as boosting GPU frequency if a periodic >>> + * vblank deadline is approaching. >>> + */ >>> +void dma_fence_set_deadline(struct dma_fence *fence, ktime_t deadline) >>> +{ >>> + unsigned long flags; >>> + >>> + if (dma_fence_is_signaled(fence)) >>> + return; >>> + >>> + spin_lock_irqsave(fence->lock, flags); >>> + >>> + /* If we already have an earlier deadline, keep it: */ >>> + if (test_bit(DMA_FENCE_FLAG_HAS_DEADLINE_BIT, &fence->flags) && >>> + ktime_before(fence->deadline, deadline)) { >>> + spin_unlock_irqrestore(fence->lock, flags); >>> + return; >>> + } >>> + >>> + fence->deadline = deadline; >>> + set_bit(DMA_FENCE_FLAG_HAS_DEADLINE_BIT, &fence->flags); >>> + >>> + spin_unlock_irqrestore(fence->lock, flags); >>> + >>> + if (fence->ops->set_deadline) >>> + fence->ops->set_deadline(fence, deadline); >>> +} >>> +EXPORT_SYMBOL(dma_fence_set_deadline); >>> + >>> /** >>> * dma_fence_init - Initialize a custom fence. >>> * @fence: the fence to initialize >>> diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h >>> index 6ffb4b2c6371..4e6cfe4e6fbc 100644 >>> --- a/include/linux/dma-fence.h >>> +++ b/include/linux/dma-fence.h >>> @@ -88,6 +88,7 @@ struct dma_fence { >>> /* @timestamp replaced by @rcu on dma_fence_release() */ >>> struct rcu_head rcu; >>> }; >>> + ktime_t deadline; >> Mhm, adding the flag sounds ok to me but I'm a bit hesitating adding the >> deadline as extra field here. >> >> We tuned the dma_fence structure intentionally so that it is only 64 bytes. > Hmm, then I guess you wouldn't be a fan of also adding an hrtimer? > > We could push the ktime_t (and timer) down into the derived fence > class, but I think there is going to need to be some extra storage > *somewhere*.. maybe the fence signaler could get away with just > storing the nearest upcoming deadline per fence-context instead? I would just push that into the driver instead. You most likely don't want the deadline per fence anyway in complex scenarios, but rather per frame. And a frame is usually composed from multiple fences. Regards, Christian. > > BR, > -R > >> Regards, >> Christian. >> >>> u64 context; >>> u64 seqno; >>> unsigned long flags; >>> @@ -99,6 +100,7 @@ enum dma_fence_flag_bits { >>> DMA_FENCE_FLAG_SIGNALED_BIT, >>> DMA_FENCE_FLAG_TIMESTAMP_BIT, >>> DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, >>> + DMA_FENCE_FLAG_HAS_DEADLINE_BIT, >>> DMA_FENCE_FLAG_USER_BITS, /* must always be last member */ >>> }; >>> >>> @@ -261,6 +263,19 @@ struct dma_fence_ops { >>> */ >>> void (*timeline_value_str)(struct dma_fence *fence, >>> char *str, int size); >>> + >>> + /** >>> + * @set_deadline: >>> + * >>> + * Callback to allow a fence waiter to inform the fence signaler of an >>> + * upcoming deadline, such as vblank, by which point the waiter would >>> + * prefer the fence to be signaled by. This is intended to give feedback >>> + * to the fence signaler to aid in power management decisions, such as >>> + * boosting GPU frequency. >>> + * >>> + * This callback is optional. >>> + */ >>> + void (*set_deadline)(struct dma_fence *fence, ktime_t deadline); >>> }; >>> >>> void dma_fence_init(struct dma_fence *fence, const struct dma_fence_ops *ops, >>> @@ -586,6 +601,8 @@ static inline signed long dma_fence_wait(struct dma_fence *fence, bool intr) >>> return ret < 0 ? ret : 0; >>> } >>> >>> +void dma_fence_set_deadline(struct dma_fence *fence, ktime_t deadline); >>> + >>> struct dma_fence *dma_fence_get_stub(void); >>> struct dma_fence *dma_fence_allocate_private_stub(void); >>> u64 dma_fence_context_alloc(unsigned num);
next prev parent reply other threads:[~2021-07-28 7:03 UTC|newest] Thread overview: 100+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-07-26 23:38 [RFC 0/4] dma-fence: Deadline awareness Rob Clark 2021-07-26 23:38 ` Rob Clark 2021-07-26 23:38 ` [RFC 1/4] dma-fence: Add deadline awareness Rob Clark 2021-07-26 23:38 ` Rob Clark 2021-07-27 7:11 ` Christian König 2021-07-27 7:11 ` Christian König 2021-07-27 14:25 ` Rob Clark 2021-07-27 14:25 ` Rob Clark 2021-07-28 7:03 ` Christian König [this message] 2021-07-28 7:03 ` Christian König 2021-07-28 11:37 ` Christian König 2021-07-28 11:37 ` Christian König 2021-07-28 15:15 ` Rob Clark 2021-07-28 15:15 ` Rob Clark 2021-07-28 17:23 ` Christian König 2021-07-28 17:23 ` Christian König 2021-07-28 17:58 ` Rob Clark 2021-07-28 17:58 ` Rob Clark 2021-07-29 7:03 ` Daniel Vetter 2021-07-29 7:03 ` Daniel Vetter 2021-07-29 15:23 ` Rob Clark 2021-07-29 15:23 ` Rob Clark 2021-07-29 16:18 ` Daniel Vetter 2021-07-29 16:18 ` Daniel Vetter 2021-07-29 17:32 ` Rob Clark 2021-07-29 17:32 ` Rob Clark 2021-07-26 23:38 ` [RFC 2/4] drm/vblank: Add helper to get next vblank time Rob Clark 2021-07-26 23:38 ` Rob Clark 2021-07-26 23:38 ` [RFC 3/4] drm/atomic-helper: Set fence deadline for vblank Rob Clark 2021-07-26 23:38 ` Rob Clark 2021-07-27 10:44 ` Michel Dänzer 2021-07-27 10:44 ` Michel Dänzer 2021-07-27 14:33 ` Rob Clark 2021-07-27 14:33 ` Rob Clark 2021-07-26 23:38 ` [RFC 4/4] drm/scheduler: Add fence deadline support Rob Clark 2021-07-26 23:38 ` Rob Clark 2021-07-26 23:51 ` [RFC 0/4] dma-fence: Deadline awareness Rob Clark 2021-07-26 23:51 ` Rob Clark 2021-07-27 14:41 ` Michel Dänzer 2021-07-27 14:41 ` Michel Dänzer 2021-07-27 15:12 ` Rob Clark 2021-07-27 15:12 ` Rob Clark 2021-07-27 15:19 ` Michel Dänzer 2021-07-27 15:37 ` Rob Clark 2021-07-27 15:37 ` Rob Clark 2021-07-28 11:36 ` Christian König 2021-07-28 11:36 ` Christian König 2021-07-28 13:08 ` Michel Dänzer 2021-07-28 13:13 ` Christian König 2021-07-28 13:13 ` Christian König 2021-07-28 13:24 ` Michel Dänzer 2021-07-28 13:24 ` Michel Dänzer 2021-07-28 13:31 ` Christian König 2021-07-28 13:31 ` Christian König 2021-07-28 13:57 ` Pekka Paalanen 2021-07-28 13:57 ` Pekka Paalanen 2021-07-28 14:30 ` Christian König 2021-07-28 14:30 ` Christian König 2021-07-29 8:08 ` Michel Dänzer 2021-07-29 8:23 ` Pekka Paalanen 2021-07-29 8:23 ` Pekka Paalanen 2021-07-29 8:43 ` Christian König 2021-07-29 8:43 ` Christian König 2021-07-29 9:15 ` Pekka Paalanen 2021-07-29 9:15 ` Pekka Paalanen 2021-07-29 10:14 ` Christian König 2021-07-29 10:14 ` Christian König 2021-07-29 10:28 ` Michel Dänzer 2021-07-29 10:28 ` Michel Dänzer 2021-07-29 11:00 ` Pekka Paalanen 2021-07-29 11:00 ` Pekka Paalanen 2021-07-29 11:43 ` Christian König 2021-07-29 11:43 ` Christian König 2021-07-29 12:49 ` Pekka Paalanen 2021-07-29 12:49 ` Pekka Paalanen 2021-07-29 13:41 ` Christian König 2021-07-29 13:41 ` Christian König 2021-07-29 14:10 ` Pekka Paalanen 2021-07-29 14:10 ` Pekka Paalanen 2021-07-28 15:27 ` Rob Clark 2021-07-28 15:27 ` Rob Clark 2021-07-28 17:20 ` Christian König 2021-07-28 15:34 ` Rob Clark 2021-07-28 15:34 ` Rob Clark 2021-07-29 7:09 ` Daniel Vetter 2021-07-29 7:09 ` Daniel Vetter 2021-07-29 8:17 ` Michel Dänzer 2021-07-29 8:17 ` Michel Dänzer 2021-07-29 9:03 ` Daniel Vetter 2021-07-29 9:03 ` Daniel Vetter 2021-07-29 9:37 ` Pekka Paalanen 2021-07-29 9:37 ` Pekka Paalanen 2021-07-29 12:18 ` Daniel Vetter 2021-07-29 12:18 ` Daniel Vetter 2021-07-29 12:59 ` Pekka Paalanen 2021-07-29 12:59 ` Pekka Paalanen 2021-07-29 14:05 ` Daniel Vetter 2021-07-29 14:05 ` Daniel Vetter 2021-07-27 21:17 ` [RFC 5/4] drm/msm: Add deadline based boost support Rob Clark 2021-07-27 21:17 ` Rob Clark
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=9edd7083-e6b3-b230-c273-8f2fbe76ca17@amd.com \ --to=christian.koenig@amd.com \ --cc=ckoenig.leichtzumerken@gmail.com \ --cc=daniel@ffwll.ch \ --cc=dri-devel@lists.freedesktop.org \ --cc=gustavo@padovan.org \ --cc=linaro-mm-sig@lists.linaro.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-media@vger.kernel.org \ --cc=matthew.brost@intel.com \ --cc=robdclark@chromium.org \ --cc=robdclark@gmail.com \ --cc=sumit.semwal@linaro.org \ /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.