From patchwork Thu Sep 26 01:22:42 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Helen Mae Koike Fornazier X-Patchwork-Id: 11161773 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 82321112B for ; Thu, 26 Sep 2019 01:23:15 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 69CB521655 for ; Thu, 26 Sep 2019 01:23:15 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 69CB521655 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=collabora.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=intel-gfx-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id CFE8D6EC4E; Thu, 26 Sep 2019 01:23:13 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from bhuna.collabora.co.uk (bhuna.collabora.co.uk [IPv6:2a00:1098:0:82:1000:25:2eeb:e3e3]) by gabe.freedesktop.org (Postfix) with ESMTPS id 0E3F16E0B8; Thu, 26 Sep 2019 01:23:11 +0000 (UTC) Received: from [127.0.0.1] (localhost [127.0.0.1]) (Authenticated sender: koike) with ESMTPSA id 2E9E028ED0E From: Helen Koike To: dri-devel@lists.freedesktop.org, intel-gfx@lists.freedesktop.org Date: Wed, 25 Sep 2019 22:22:42 -0300 Message-Id: <20190926012243.4893-1-helen.koike@collabora.com> X-Mailer: git-send-email 2.22.0 MIME-Version: 1.0 Subject: [Intel-gfx] [PATCH v11 1/2] drm/i915: Introduce async plane update to i915 X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: mcasas@google.com, zhenyu.z.wang@intel.com, airlied@linux.ie, daniel.vetter@ffwll.ch, linux-kernel@vger.kernel.org, tfiga@chromium.org, gustavo.padovan@collabora.com, Helen Koike , enric.balletbo@collabora.com, kernel@collabora.com Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" From: Gustavo Padovan Add implementation for async plane update callbacks Signed-off-by: Gustavo Padovan Signed-off-by: Enric Balletbo i Serra Signed-off-by: Tina Zhang Signed-off-by: Helen Koike Tested-by: Tina Zhang --- Hi, I ran the following tests and no regressions were found: ./scripts/run-tests.sh -p -s -t ".*plane_cursor_legacy.*" -t ".*kms_cursor_legacy.*" -r results-`uname -r`-v11 -v The igt test results can be found at https://people.collabora.com/~koike/results-5.3.0-rc3+-v11/html Thanks Helen Changes in v11: - v10 https://patchwork.kernel.org/project/dri-devel/list/?series=147095 - rebase on drm-intel - replace i915_gem_track_fb() by intel_frontbuffer_track() Changes in v10: - v9: https://patchwork.kernel.org/patch/11000561/ - rebase on drm-tip - call i915_sw_fence_init() in the begining of intel_atomic_commit(), otherwise drm_atomic_helper_prepare_planes() without initializing the fence object, casing errors like: ODEBUG: assert_init not available (active state 0) object type: i915_sw_fence hint: 0x0 Changes in v9: - v8: https://patchwork.kernel.org/patch/10843395/ - Added tested-by tag - submitted to intel-gfx@lists.freedesktop.org to invoke CI - rebased and fixed conflicts on top of drm-tip Changes in v8: - v7: https://lkml.org/lkml/2018/6/8/168 - v7 was splited in two, one that adds the async callbacks and another that updates the cursor. - rebase with drm-intel - allow async update in all types of planes, not only cursor - add watermark checks in async update - remove bypass of intel_prepare_plane_fb() in case of async update - add missing drm_atomic_helper_cleanup_planes(dev, state) call in intel_atomic_commit(). - use swap() function in async update to set the old_fb in the new_state object. - use helpers intel_update_plane()/intel_disable_plane() Changes in v7: - Rebase on top of drm-intel repository. Hopefully now will play nicely with autobuilders. Changes in v6: - Rework the intel_plane_atomic_async_update due driver changed from last time. - Removed the mutex_lock/unlock as causes a deadlock. Changes in v5: - Call drm_atomic_helper_async_check() from the check hook Changes in v4: - Set correct vma to new state for cleanup - Move size checks back to drivers (Ville Syrjälä) Changes in v3: - Move fb setting to core and use new state (Eric Anholt) .../gpu/drm/i915/display/intel_atomic_plane.c | 71 +++++++++++++++++++ drivers/gpu/drm/i915/display/intel_display.c | 20 +++--- .../drm/i915/display/intel_display_types.h | 6 ++ 3 files changed, 89 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_atomic_plane.c b/drivers/gpu/drm/i915/display/intel_atomic_plane.c index d1fcdf206da4..faa5d74e3ccf 100644 --- a/drivers/gpu/drm/i915/display/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/display/intel_atomic_plane.c @@ -365,8 +365,79 @@ void i9xx_update_planes_on_crtc(struct intel_atomic_state *state, } } +static int intel_plane_atomic_async_check(struct drm_plane *plane, + struct drm_plane_state *state) +{ + struct drm_crtc_state *crtc_state; + + crtc_state = drm_atomic_get_existing_crtc_state(state->state, + state->crtc); + if (WARN_ON(!crtc_state)) + return -EINVAL; + + /* + * When crtc is inactive or there is a modeset pending, + * wait for it to complete in the slowpath + */ + if (!crtc_state->active || to_intel_crtc_state(crtc_state)->update_pipe) + return -EINVAL; + + /* + * If any parameters change that may affect watermarks, + * take the slowpath. Only changing fb or position should be + * in the fastpath. + */ + if (plane->state->crtc != state->crtc || + plane->state->src_w != state->src_w || + plane->state->src_h != state->src_h || + plane->state->crtc_w != state->crtc_w || + plane->state->crtc_h != state->crtc_h || + !plane->state->fb != !state->fb) + return -EINVAL; + + return 0; +} + +static void intel_plane_atomic_async_update(struct drm_plane *plane, + struct drm_plane_state *new_state) +{ + struct intel_atomic_state *intel_new_state = + to_intel_atomic_state(new_state->state); + struct intel_plane *intel_plane = to_intel_plane(plane); + struct drm_crtc *crtc = plane->state->crtc; + struct intel_crtc_state *new_crtc_state; + struct intel_crtc *intel_crtc; + int i; + + for_each_new_intel_crtc_in_state(intel_new_state, intel_crtc, + new_crtc_state, i) + WARN_ON(new_crtc_state->wm.need_postvbl_update || + new_crtc_state->update_wm_post); + + intel_frontbuffer_track(to_intel_frontbuffer(plane->state->fb), + to_intel_frontbuffer(new_state->fb), + intel_plane->frontbuffer_bit); + + plane->state->src_x = new_state->src_x; + plane->state->src_y = new_state->src_y; + plane->state->crtc_x = new_state->crtc_x; + plane->state->crtc_y = new_state->crtc_y; + + swap(plane->state->fb, new_state->fb); + + if (plane->state->visible) + intel_update_plane(intel_plane, + to_intel_crtc_state(crtc->state), + to_intel_plane_state(plane->state)); + else + intel_disable_plane(intel_plane, + to_intel_crtc_state(crtc->state)); +} + const struct drm_plane_helper_funcs intel_plane_helper_funcs = { .prepare_fb = intel_prepare_plane_fb, .cleanup_fb = intel_cleanup_plane_fb, .atomic_check = intel_plane_atomic_check, + .atomic_async_check = intel_plane_atomic_async_check, + .atomic_async_update = intel_plane_atomic_async_update, }; diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index b51d1ceb8739..8ee74cda5b45 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -3158,12 +3158,6 @@ static void intel_plane_disable_noatomic(struct intel_crtc *crtc, intel_disable_plane(plane, crtc_state); } -static struct intel_frontbuffer * -to_intel_frontbuffer(struct drm_framebuffer *fb) -{ - return fb ? to_intel_framebuffer(fb)->frontbuffer : NULL; -} - static void intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, struct intel_initial_plane_config *plane_config) @@ -14129,11 +14123,21 @@ static int intel_atomic_commit(struct drm_device *dev, struct drm_i915_private *dev_priv = to_i915(dev); int ret = 0; + i915_sw_fence_init(&state->commit_ready, + intel_atomic_commit_ready); + + if (_state->async_update) { + ret = drm_atomic_helper_prepare_planes(dev, _state); + if (ret) + return ret; + drm_atomic_helper_async_commit(dev, _state); + drm_atomic_helper_cleanup_planes(dev, _state); + return 0; + } + state->wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm); drm_atomic_state_get(&state->base); - i915_sw_fence_init(&state->commit_ready, - intel_atomic_commit_ready); /* * The intel_legacy_cursor_update() fast path takes care diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h index 449abaea619f..00ebe5b99f25 100644 --- a/drivers/gpu/drm/i915/display/intel_display_types.h +++ b/drivers/gpu/drm/i915/display/intel_display_types.h @@ -1092,6 +1092,12 @@ struct cxsr_latency { #define to_intel_plane_state(x) container_of(x, struct intel_plane_state, base) #define intel_fb_obj(x) ((x) ? to_intel_bo((x)->obj[0]) : NULL) +static inline struct intel_frontbuffer * +to_intel_frontbuffer(struct drm_framebuffer *fb) +{ + return fb ? to_intel_framebuffer(fb)->frontbuffer : NULL; +} + struct intel_hdmi { i915_reg_t hdmi_reg; int ddc_bus; From patchwork Thu Sep 26 01:22:43 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Helen Mae Koike Fornazier X-Patchwork-Id: 11161777 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 79AD6112B for ; Thu, 26 Sep 2019 01:23:19 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 6239021655 for ; Thu, 26 Sep 2019 01:23:19 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 6239021655 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=collabora.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=intel-gfx-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 35FEA6ED2A; Thu, 26 Sep 2019 01:23:18 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from bhuna.collabora.co.uk (bhuna.collabora.co.uk [46.235.227.227]) by gabe.freedesktop.org (Postfix) with ESMTPS id C628F6ED2A; Thu, 26 Sep 2019 01:23:16 +0000 (UTC) Received: from [127.0.0.1] (localhost [127.0.0.1]) (Authenticated sender: koike) with ESMTPSA id 0959E28ED13 From: Helen Koike To: dri-devel@lists.freedesktop.org, intel-gfx@lists.freedesktop.org Date: Wed, 25 Sep 2019 22:22:43 -0300 Message-Id: <20190926012243.4893-2-helen.koike@collabora.com> X-Mailer: git-send-email 2.22.0 In-Reply-To: <20190926012243.4893-1-helen.koike@collabora.com> References: <20190926012243.4893-1-helen.koike@collabora.com> MIME-Version: 1.0 Subject: [Intel-gfx] [PATCH v11 2/2] drm/i915: update cursors asynchronously through atomic X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: mcasas@google.com, zhenyu.z.wang@intel.com, airlied@linux.ie, daniel.vetter@ffwll.ch, linux-kernel@vger.kernel.org, tfiga@chromium.org, gustavo.padovan@collabora.com, Helen Koike , enric.balletbo@collabora.com, kernel@collabora.com Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" From: Gustavo Padovan Replace the legacy cursor implementation by the async callbacks Signed-off-by: Gustavo Padovan Signed-off-by: Enric Balletbo i Serra Signed-off-by: Helen Koike --- Changes in v11: None Changes in v10: None Changes in v9: - v8: https://patchwork.kernel.org/patch/10843397/ - rebased and fixed conflicts on top of drm-tip Changes in v8: - v7: https://lkml.org/lkml/2018/6/8/168 - v7 was splited in two, one that adds the async callbacks and another that updates the cursor. - Update comment in intel_pm.c that was referencing intel_plane_atomic_async_update() Changes in v7: None Changes in v6: None Changes in v5: None Changes in v4: None Changes in v3: None drivers/gpu/drm/i915/display/intel_display.c | 163 +------------------ drivers/gpu/drm/i915/intel_pm.c | 2 +- 2 files changed, 6 insertions(+), 159 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 8ee74cda5b45..11540aae76a9 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -13643,6 +13643,10 @@ static int intel_atomic_check(struct drm_device *dev, if (ret) goto fail; + if (_state->legacy_cursor_update) + _state->async_update = !drm_atomic_helper_async_check(dev, + _state); + intel_fbc_choose_crtc(dev_priv, state); ret = calc_watermark_data(state); if (ret) @@ -14139,34 +14143,6 @@ static int intel_atomic_commit(struct drm_device *dev, drm_atomic_state_get(&state->base); - /* - * The intel_legacy_cursor_update() fast path takes care - * of avoiding the vblank waits for simple cursor - * movement and flips. For cursor on/off and size changes, - * we want to perform the vblank waits so that watermark - * updates happen during the correct frames. Gen9+ have - * double buffered watermarks and so shouldn't need this. - * - * Unset state->legacy_cursor_update before the call to - * drm_atomic_helper_setup_commit() because otherwise - * drm_atomic_helper_wait_for_flip_done() is a noop and - * we get FIFO underruns because we didn't wait - * for vblank. - * - * FIXME doing watermarks and fb cleanup from a vblank worker - * (assuming we had any) would solve these problems. - */ - if (INTEL_GEN(dev_priv) < 9 && state->base.legacy_cursor_update) { - struct intel_crtc_state *new_crtc_state; - struct intel_crtc *crtc; - int i; - - for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) - if (new_crtc_state->wm.need_postvbl_update || - new_crtc_state->update_wm_post) - state->base.legacy_cursor_update = false; - } - ret = intel_atomic_prepare_commit(state); if (ret) { DRM_DEBUG_ATOMIC("Preparing state failed with %i\n", ret); @@ -14658,137 +14634,8 @@ static const struct drm_plane_funcs i8xx_plane_funcs = { .format_mod_supported = i8xx_plane_format_mod_supported, }; -static int -intel_legacy_cursor_update(struct drm_plane *plane, - struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - u32 src_x, u32 src_y, - u32 src_w, u32 src_h, - struct drm_modeset_acquire_ctx *ctx) -{ - struct drm_i915_private *dev_priv = to_i915(crtc->dev); - struct drm_plane_state *old_plane_state, *new_plane_state; - struct intel_plane *intel_plane = to_intel_plane(plane); - struct intel_crtc_state *crtc_state = - to_intel_crtc_state(crtc->state); - struct intel_crtc_state *new_crtc_state; - int ret; - - /* - * When crtc is inactive or there is a modeset pending, - * wait for it to complete in the slowpath - */ - if (!crtc_state->base.active || needs_modeset(crtc_state) || - crtc_state->update_pipe) - goto slow; - - old_plane_state = plane->state; - /* - * Don't do an async update if there is an outstanding commit modifying - * the plane. This prevents our async update's changes from getting - * overridden by a previous synchronous update's state. - */ - if (old_plane_state->commit && - !try_wait_for_completion(&old_plane_state->commit->hw_done)) - goto slow; - - /* - * If any parameters change that may affect watermarks, - * take the slowpath. Only changing fb or position should be - * in the fastpath. - */ - if (old_plane_state->crtc != crtc || - old_plane_state->src_w != src_w || - old_plane_state->src_h != src_h || - old_plane_state->crtc_w != crtc_w || - old_plane_state->crtc_h != crtc_h || - !old_plane_state->fb != !fb) - goto slow; - - new_plane_state = intel_plane_duplicate_state(plane); - if (!new_plane_state) - return -ENOMEM; - - new_crtc_state = to_intel_crtc_state(intel_crtc_duplicate_state(crtc)); - if (!new_crtc_state) { - ret = -ENOMEM; - goto out_free; - } - - drm_atomic_set_fb_for_plane(new_plane_state, fb); - - new_plane_state->src_x = src_x; - new_plane_state->src_y = src_y; - new_plane_state->src_w = src_w; - new_plane_state->src_h = src_h; - new_plane_state->crtc_x = crtc_x; - new_plane_state->crtc_y = crtc_y; - new_plane_state->crtc_w = crtc_w; - new_plane_state->crtc_h = crtc_h; - - ret = intel_plane_atomic_check_with_state(crtc_state, new_crtc_state, - to_intel_plane_state(old_plane_state), - to_intel_plane_state(new_plane_state)); - if (ret) - goto out_free; - - ret = mutex_lock_interruptible(&dev_priv->drm.struct_mutex); - if (ret) - goto out_free; - - ret = intel_plane_pin_fb(to_intel_plane_state(new_plane_state)); - if (ret) - goto out_unlock; - - intel_frontbuffer_flush(to_intel_frontbuffer(fb), ORIGIN_FLIP); - intel_frontbuffer_track(to_intel_frontbuffer(old_plane_state->fb), - to_intel_frontbuffer(fb), - intel_plane->frontbuffer_bit); - - /* Swap plane state */ - plane->state = new_plane_state; - - /* - * We cannot swap crtc_state as it may be in use by an atomic commit or - * page flip that's running simultaneously. If we swap crtc_state and - * destroy the old state, we will cause a use-after-free there. - * - * Only update active_planes, which is needed for our internal - * bookkeeping. Either value will do the right thing when updating - * planes atomically. If the cursor was part of the atomic update then - * we would have taken the slowpath. - */ - crtc_state->active_planes = new_crtc_state->active_planes; - - if (plane->state->visible) - intel_update_plane(intel_plane, crtc_state, - to_intel_plane_state(plane->state)); - else - intel_disable_plane(intel_plane, crtc_state); - - intel_plane_unpin_fb(to_intel_plane_state(old_plane_state)); - -out_unlock: - mutex_unlock(&dev_priv->drm.struct_mutex); -out_free: - if (new_crtc_state) - intel_crtc_destroy_state(crtc, &new_crtc_state->base); - if (ret) - intel_plane_destroy_state(plane, new_plane_state); - else - intel_plane_destroy_state(plane, old_plane_state); - return ret; - -slow: - return drm_atomic_helper_update_plane(plane, crtc, fb, - crtc_x, crtc_y, crtc_w, crtc_h, - src_x, src_y, src_w, src_h, ctx); -} - static const struct drm_plane_funcs intel_cursor_plane_funcs = { - .update_plane = intel_legacy_cursor_update, + .update_plane = drm_atomic_helper_update_plane, .disable_plane = drm_atomic_helper_disable_plane, .destroy = intel_plane_destroy, .atomic_duplicate_state = intel_plane_duplicate_state, diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 75ee027abb80..14fc28a29cd2 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -826,7 +826,7 @@ static bool intel_wm_plane_visible(const struct intel_crtc_state *crtc_state, * can happen faster than the vrefresh rate, and the current * watermark code doesn't handle that correctly. Cursor updates * which set/clear the fb or change the cursor size are going - * to get throttled by intel_legacy_cursor_update() to work + * to get throttled by intel_plane_atomic_async_update() to work * around this problem with the watermark code. */ if (plane->id == PLANE_CURSOR)