blob: 754b8473b2c06624e4098d33fc8bd027e8072ed6 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
|
From 4efa42702d14856286447eb0b2dbcb2ebb0cc44e Mon Sep 17 00:00:00 2001
From: Kevin Wang <kevin1.wang@amd.com>
Date: Fri, 10 May 2019 15:29:11 +0800
Subject: [PATCH 2371/2940] drm/amd/powerplay: fix amdgpu_pm_info show gpu load
error
due to the smu dma/RTOS restriction, the interval of catching smu
metric table should be more than 1ms. otherwise it will cause the gpu
activity data corruption.
Signed-off-by:Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/powerplay/navi10_ppt.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/amd/powerplay/navi10_ppt.c b/drivers/gpu/drm/amd/powerplay/navi10_ppt.c
index 4851222c9919..5f697ed3796b 100644
--- a/drivers/gpu/drm/amd/powerplay/navi10_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/navi10_ppt.c
@@ -816,6 +816,8 @@ static int navi10_get_current_activity_percent(struct smu_context *smu,
if (!value)
return -EINVAL;
+ msleep(1);
+
ret = smu_update_table(smu, SMU_TABLE_SMU_METRICS,
(void *)&metrics, false);
if (ret)
--
2.17.1
|