openwrt/target/linux/brcm2708/patches-4.9/950-0164-drm-vc4-Use-runtime-autosuspend-to-avoid-thrashing-V.patch
Rafał Miłecki fce21ae4cc brcm2708: rename all patches from raspberrypi git tree to use 950 prefix
Right now all brcm2708 patches are extracted from the non-mainline
raspberrypi/linux git tree. Many of them are hacks and/or are unneeded
in LEDE. Raspberry Pi is getting better and better mainline support so
it would be nice to finally start maintaining patches in a cleaner way:
1) Backport patches accepted in upstream tree
2) Start using upstream drivers
3) Pick only these patches that are needed for more complete support

Handling above tasks requires grouping patches - ideally using the same
prefixes as generic ones. It means we should rename existing patches to
use some high prefix. This will allow e.g. use 0xx for backported code.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Stijn Tintel <stijn@linux-ipv6.be>
2017-03-24 08:06:35 +01:00

79 lines
2.7 KiB
Diff

From 936dc8b00ccd5072c5d610b97199402275bfdf26 Mon Sep 17 00:00:00 2001
From: Eric Anholt <eric@anholt.net>
Date: Fri, 4 Nov 2016 15:58:38 -0700
Subject: [PATCH] drm/vc4: Use runtime autosuspend to avoid thrashing V3D power
state.
The pm_runtime_put() we were using immediately released power on the
device, which meant that we were generally turning the device off and
on once per frame. In many profiles I've looked at, that added up to
about 1% of CPU time, but this could get worse in the case of frequent
rendering and readback (as may happen in X rendering). By keeping the
device on until we've been idle for a couple of frames, we drop the
overhead of runtime PM down to sub-.1%.
Signed-off-by: Eric Anholt <eric@anholt.net>
(cherry picked from commit 3a62234680d86efa0239665ed8a0e908f1aef147)
---
drivers/gpu/drm/vc4/vc4_drv.c | 9 ++++++---
drivers/gpu/drm/vc4/vc4_gem.c | 6 ++++--
drivers/gpu/drm/vc4/vc4_v3d.c | 2 ++
3 files changed, 12 insertions(+), 5 deletions(-)
--- a/drivers/gpu/drm/vc4/vc4_drv.c
+++ b/drivers/gpu/drm/vc4/vc4_drv.c
@@ -61,21 +61,24 @@ static int vc4_get_param_ioctl(struct dr
if (ret < 0)
return ret;
args->value = V3D_READ(V3D_IDENT0);
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
break;
case DRM_VC4_PARAM_V3D_IDENT1:
ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev);
if (ret < 0)
return ret;
args->value = V3D_READ(V3D_IDENT1);
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
break;
case DRM_VC4_PARAM_V3D_IDENT2:
ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev);
if (ret < 0)
return ret;
args->value = V3D_READ(V3D_IDENT2);
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
break;
case DRM_VC4_PARAM_SUPPORTS_BRANCHES:
case DRM_VC4_PARAM_SUPPORTS_ETC1:
--- a/drivers/gpu/drm/vc4/vc4_gem.c
+++ b/drivers/gpu/drm/vc4/vc4_gem.c
@@ -711,8 +711,10 @@ vc4_complete_exec(struct drm_device *dev
}
mutex_lock(&vc4->power_lock);
- if (--vc4->power_refcount == 0)
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ if (--vc4->power_refcount == 0) {
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
+ }
mutex_unlock(&vc4->power_lock);
kfree(exec);
--- a/drivers/gpu/drm/vc4/vc4_v3d.c
+++ b/drivers/gpu/drm/vc4/vc4_v3d.c
@@ -222,6 +222,8 @@ static int vc4_v3d_bind(struct device *d
return ret;
}
+ pm_runtime_use_autosuspend(dev);
+ pm_runtime_set_autosuspend_delay(dev, 40); /* a little over 2 frames. */
pm_runtime_enable(dev);
return 0;