mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: John Stultz <john.stultz@linaro.org>
To: lkml <linux-kernel@vger.kernel.org>
Cc: Vincent Donnefort <vincent.donnefort@arm.com>,
	Mauro Carvalho Chehab <mchehab+huawei@kernel.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Manivannan Sadhasivam <mani@kernel.org>,
	dri-devel <dri-devel@lists.freedesktop.org>,
	Liwei Cai <cailiwei@hisilicon.com>,
	Xinliang Liu <xinliang.liu@linaro.org>,
	Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
	Sam Ravnborg <sam@ravnborg.org>,
	Sumit Semwal <sumit.semwal@linaro.org>,
	Chen Feng <puck.chen@hisilicon.com>,
	John Stultz <john.stultz@linaro.org>
Subject: [PATCH 4/3] drm: hikey9xx: remove wait for VACTIVE IRQ
Date: Thu, 20 Aug 2020 03:56:17 +0000	[thread overview]
Message-ID: <20200820035617.5836-1-john.stultz@linaro.org> (raw)
In-Reply-To: <20200820034100.128062-2-john.stultz@linaro.org>

From: Vincent Donnefort <vincent.donnefort@arm.com>

A quick add-on to my earlier fixup series as I realized what the
performance problem was.

This is against Mauro's tree here:
    https://gitlab.freedesktop.org/mchehab_kernel/hikey-970/-/commits/master/

For each display cycle, the Kirin960 display IP will generate a VACTIVE
interrupt followed by a VBLANK. During a FBIOPAN ioctl, the driver will then
wait for the first one to then wait for the second one. This is an issue when
the CPU load is too low: the wait_event() function might trigger a transition
to a deep sleep state and then, waking up from that state will take too much
time to catch the VBLANK interrupt on time, the difference between those two
interrupts being only 60 us.

* Ideal case:                   ACT                VBL
				+                  +
				v                  v
		 ---> wait(ACT) +------> wait(VBL) +-->

* Our case:                     ACT VBL        ACT VBL
				+   +          +   +
				v   v          v   v
		 ---> wait(ACT) +------> wait(VBL) +-->

The wait for VACTIVE IRQ can safely be removed: there is no hardware access
performed between the VACTIVE and the VBLANK IRQs.

This behavior has been introduced from 4.11 with the following patch:
a3fbb53f4 drm/atomic: Wait for vblank whenever a plane is added to state.

Cc: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Manivannan Sadhasivam <mani@kernel.org>
Cc: dri-devel <dri-devel@lists.freedesktop.org>
Cc: Liwei Cai <cailiwei@hisilicon.com>
Cc: Xinliang Liu <xinliang.liu@linaro.org>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Chen Feng <puck.chen@hisilicon.com>
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
[jstultz: hand-ported to Mauro's patch set]
Signed-off-by: John Stultz <john.stultz@linaro.org>
---
 .../hikey9xx/gpu/kirin9xx_drm_dpe_utils.c     |  2 +-
 .../staging/hikey9xx/gpu/kirin9xx_drm_dss.c   |  5 ---
 .../hikey9xx/gpu/kirin9xx_drm_overlay_utils.c | 36 -------------------
 3 files changed, 1 insertion(+), 42 deletions(-)

diff --git a/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dpe_utils.c b/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dpe_utils.c
index 82a0edb95953..cc80689d90c9 100644
--- a/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dpe_utils.c
+++ b/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dpe_utils.c
@@ -582,7 +582,7 @@ void dpe_interrupt_unmask(struct dss_crtc *acrtc)
 	writel(unmask, dss_base + GLB_CPU_PDP_INT_MSK);
 
 	unmask = ~0;
-	unmask &= ~(BIT_VSYNC | BIT_VACTIVE0_END | BIT_LDI_UNFLOW);
+	unmask &= ~(BIT_VSYNC | BIT_LDI_UNFLOW);
 
 	writel(unmask, dss_base + DSS_LDI0_OFFSET + LDI_CPU_ITF_INT_MSK);
 }
diff --git a/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dss.c b/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dss.c
index 7adbd924bec2..c99ce7c4d479 100644
--- a/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dss.c
+++ b/drivers/staging/hikey9xx/gpu/kirin9xx_drm_dss.c
@@ -440,11 +440,6 @@ static irqreturn_t dss_irq_handler(int irq, void *data)
 	isr_s1 &= ~(readl(dss_base + GLB_CPU_PDP_INT_MSK));
 	isr_s2 &= ~(readl(dss_base + DSS_LDI0_OFFSET + LDI_CPU_ITF_INT_MSK));
 
-	if (isr_s2 & BIT_VACTIVE0_END) {
-		ctx->vactive0_end_flag++;
-		wake_up_interruptible_all(&ctx->vactive0_end_wq);
-	}
-
 	if (isr_s2 & BIT_VSYNC) {
 		ctx->vsync_timestamp = ktime_get();
 		drm_crtc_handle_vblank(crtc);
diff --git a/drivers/staging/hikey9xx/gpu/kirin9xx_drm_overlay_utils.c b/drivers/staging/hikey9xx/gpu/kirin9xx_drm_overlay_utils.c
index 5ac7f4b31d99..932bad2f428e 100644
--- a/drivers/staging/hikey9xx/gpu/kirin9xx_drm_overlay_utils.c
+++ b/drivers/staging/hikey9xx/gpu/kirin9xx_drm_overlay_utils.c
@@ -822,40 +822,6 @@ void hisifb_mctl_sw_clr(struct dss_crtc *acrtc)
 	DRM_INFO("-.\n");
 }
 
-static int hisi_dss_wait_for_complete(struct dss_crtc *acrtc)
-{
-	int ret = 0;
-	u32 times = 0;
-	u32 prev_vactive0_end = 0;
-	struct dss_hw_ctx *ctx = acrtc->ctx;
-
-	prev_vactive0_end = ctx->vactive0_end_flag;
-
-REDO:
-	ret = wait_event_interruptible_timeout(ctx->vactive0_end_wq,
-					       (prev_vactive0_end != ctx->vactive0_end_flag),
-		msecs_to_jiffies(300));
-	if (ret == -ERESTARTSYS) {
-		if (times < 50) {
-			times++;
-			mdelay(10);
-			goto REDO;
-		}
-	}
-
-	if (ret <= 0) {
-		disable_ldi(acrtc);
-		hisifb_mctl_sw_clr(acrtc);
-		DRM_ERROR("wait_for vactive0_end_flag timeout! ret=%d.\n", ret);
-
-		ret = -ETIMEDOUT;
-	} else {
-		ret = 0;
-	}
-
-	return ret;
-}
-
 void hisi_fb_pan_display(struct drm_plane *plane)
 {
 	struct drm_plane_state *state = plane->state;
@@ -932,7 +898,6 @@ void hisi_fb_pan_display(struct drm_plane *plane)
 	hisi_dss_unflow_handler(ctx, true);
 
 	enable_ldi(acrtc);
-	hisi_dss_wait_for_complete(acrtc);
 }
 
 void hisi_dss_online_play(struct kirin_fbdev *fbdev, struct drm_plane *plane,
@@ -1001,5 +966,4 @@ void hisi_dss_online_play(struct kirin_fbdev *fbdev, struct drm_plane *plane,
 	hisi_dss_unflow_handler(ctx, true);
 
 	enable_ldi(acrtc);
-	hisi_dss_wait_for_complete(acrtc);
 }
-- 
2.17.1


  reply	other threads:[~2020-08-20  3:56 UTC|newest]

Thread overview: 105+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-19 11:45 [PATCH 00/49] DRM driver for Hikey 970 Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 01/49] staging: hikey9xx: Add hisilicon DRM driver for hikey960/970 Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 02/49] staging: hikey9xx/gpu: port it to work with Kernel v4.9 Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 03/49] staging: hikey9xx/gpu: solve tearing issue of display Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 04/49] staging: hikey9xx/gpu: resolve the performance issue by interrupt mechanism Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 05/49] staging: hikey9xx/gpu: add support to hikey970 HDMI and panel Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 06/49] staging: hikey9xx/gpu: Solve SR Cannot Display Problems Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 07/49] staging: hikey9xx/gpu: Solve HDMI compatibility Problem Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 08/49] staging: hikey9xx/gpu: Support MIPI DSI 3 lanes for hikey970 Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 09/49] staging: hikey9xx/gpu: Solve SR test reset problem " Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 10/49] staging: hikey9xx/gpu: add debug prints for this driver Mauro Carvalho Chehab
2020-08-21 20:17   ` Joe Perches
2020-08-19 11:45 ` [PATCH 11/49] staging: hikey9xx/gpu: Add support 10.1 inch special HDMI displays Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 12/49] staging: hikey9xx/gpu: get rid of adv7535 fork Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 13/49] staging: hikey9xx/gpu: rename the Kirin9xx namespace Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 14/49] staging: hikey9xx/gpu: get rid of kirin9xx_fbdev.c Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 15/49] staging: hikey9xx/gpu: get rid of some ifdefs Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 16/49] staging: hikey9xx/gpu: rename the config option for Kirin970 Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 17/49] staging: hikey9xx/gpu: change the includes to reflect upstream Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 18/49] staging: hikey9xx/gpu: port driver to upstream kAPIs Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 19/49] staging: hikey9xx/gpu: add a copy of set_reg() function there Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 20/49] staging: hikey9xx/gpu: get rid of ION headers Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 21/49] staging: hikey9xx/gpu: add support for using a reserved CMA memory Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 22/49] staging: hikey9xx/gpu: cleanup encoder attach logic Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 23/49] staging: hikey9xx/gpu: Change the logic which sets the burst mode Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 24/49] staging: hikey9xx/gpu: fix the DRM setting logic Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 25/49] staging: hikey9xx/gpu: do some code cleanups Mauro Carvalho Chehab
2020-08-20  1:53   ` John Stultz
2020-08-20  8:23     ` Mauro Carvalho Chehab
2020-08-21 20:12       ` John Stultz
2020-08-19 11:45 ` [PATCH 26/49] staging: hikey9xx/gpu: use default GEM_CMA fops Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 27/49] staging: hikey9xx/gpu: place vblank enable/disable at the right place Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 28/49] staging: hikey9xx/gpu: remove an uneeded hack Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 29/49] staging: hikey9xx/gpu: add a possible implementation for atomic_disable Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 30/49] staging: hikey9xx/gpu: register connector Mauro Carvalho Chehab
2020-08-19 11:45 ` [PATCH 31/49] staging: hikey9xx/gpu: fix driver name Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 32/49] staging: hikey9xx/gpu: get rid of iommu_format Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 33/49] staging: hikey9xx/gpu: re-work the mode validation code Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 34/49] staging: hikey9xx/gpu: add support for enable/disable ldo3 regulator Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 35/49] staging: hikey9xx/gpu: add SPMI headers Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 36/49] staging: hikey9xx/gpu: solve most coding style issues Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 37/49] staging: hikey9xx/gpu: don't use iommu code Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 38/49] staging: hikey9xx/gpu: add kirin9xx driver to the building system Mauro Carvalho Chehab
2020-08-19 15:30   ` Randy Dunlap
2020-08-19 11:46 ` [PATCH 39/49] staging: hikey9xx/gpu: get rid of typedefs Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 40/49] staging: hikey9xx/gpu: get rid of input/output macros Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 41/49] staging: hikey9xx/gpu: get rid of some unused data Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 42/49] staging: hikey9xx/gpu: place common definitions at kirin9xx_dpe.h Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 43/49] staging: hikey9xx/gpu: get rid of DRM_HISI_KIRIN970 Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 44/49] dts: hisilicon: hi3670.dtsi: add I2C settings Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 45/49] dts: hikey970-pinctrl.dtsi: add missing pinctrl settings Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 46/49] dt: hisilicon: add support for the PMIC found on Hikey 970 Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 47/49] dts: add support for Hikey 970 DRM Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 48/49] staging: hikey9xx/gpu: drop kirin9xx_pwm Mauro Carvalho Chehab
2020-08-19 11:46 ` [PATCH 49/49] dt: display: Add binds for the DPE and DSI controller for Kirin 960/970 Mauro Carvalho Chehab
2020-08-19 21:02   ` Sam Ravnborg
2020-08-19 22:36   ` Rob Herring
2020-08-19 15:21 ` [PATCH 00/49] DRM driver for Hikey 970 Sam Ravnborg
2020-08-19 15:30   ` Laurent Pinchart
2020-08-19 19:52     ` John Stultz
2020-08-20  7:03       ` Mauro Carvalho Chehab
2020-08-20 10:02         ` Laurent Pinchart
2020-08-24 19:29           ` Dave Airlie
2020-08-25 11:30             ` Mauro Carvalho Chehab
2020-08-25 11:38               ` Laurent Pinchart
2020-08-25 12:31               ` Daniel Stone
2020-08-26 14:44               ` Nicolas Dufresne
2020-08-19 15:40   ` Mauro Carvalho Chehab
2020-08-19 17:35     ` Sam Ravnborg
2020-08-20 14:06       ` Mauro Carvalho Chehab
2020-08-20 14:48         ` Sam Ravnborg
2020-08-20 15:13           ` Mauro Carvalho Chehab
2020-08-20 15:36             ` Sam Ravnborg
2020-08-21 13:37       ` Mauro Carvalho Chehab
2020-08-21 13:58       ` Mauro Carvalho Chehab
2020-08-21 15:55         ` Sam Ravnborg
2020-08-24 16:02           ` Mauro Carvalho Chehab
2020-08-24 18:45             ` Sam Ravnborg
2020-08-24 21:10             ` Sam Ravnborg
2020-08-24 21:24             ` Sam Ravnborg
2020-08-25 18:11             ` Sam Ravnborg
2020-08-25 20:21             ` Sam Ravnborg
2020-08-21 14:41       ` Mauro Carvalho Chehab
2020-08-21 15:56         ` Sam Ravnborg
2020-08-24 16:06           ` Mauro Carvalho Chehab
2020-08-19 20:48     ` Sam Ravnborg
2020-08-20  7:21       ` Michel Dänzer
2020-08-20  7:21       ` Mauro Carvalho Chehab
2020-08-21 16:09       ` Joe Perches
2020-08-19 21:13 ` John Stultz
2020-08-19 21:25   ` Sam Ravnborg
2020-08-20  6:40     ` Mauro Carvalho Chehab
2020-08-20  6:34   ` Mauro Carvalho Chehab
2020-08-19 21:36 ` John Stultz
2020-08-20  2:01   ` John Stultz
2020-08-20  3:28     ` John Stultz
2020-08-20  7:48       ` Mauro Carvalho Chehab
2020-08-24  6:49       ` Mauro Carvalho Chehab
2020-08-24 13:18         ` Mauro Carvalho Chehab
2020-08-20  8:04   ` Mauro Carvalho Chehab
2020-08-20  3:39 ` [PATCH 1/3] drm: hikey9xx: Fix inconsistent compat string John Stultz
2020-08-20  3:40   ` [PATCH 2/3] drm: kirin9xxx: Add additional color formats John Stultz
2020-08-20  3:41     ` [PATCH 3/3] drm: hikey9xxx: Fix null pointer crash on reboot John Stultz
2020-08-20  3:56       ` John Stultz [this message]
2020-08-20  9:04   ` [PATCH 1/3] drm: hikey9xx: Fix inconsistent compat string Mauro Carvalho Chehab

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=20200820035617.5836-1-john.stultz@linaro.org \
    --to=john.stultz@linaro.org \
    --cc=cailiwei@hisilicon.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mani@kernel.org \
    --cc=mchehab+huawei@kernel.org \
    --cc=puck.chen@hisilicon.com \
    --cc=sam@ravnborg.org \
    --cc=sumit.semwal@linaro.org \
    --cc=vincent.donnefort@arm.com \
    --cc=xinliang.liu@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: 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

Powered by JetHome