Revert "gpu: nvgpu: cache gpu clk rate"

This reverts commit e9a6d179a4 ("gpu: nvgpu: cache gpu clk rate")

 - Real clock rate doesn't always equal clock rate requested by caller
 - call of clk_set_rate() and update of cached_rate are not atomic
 - Real root cause for Bug 2051688 is in bpmp and gboost design

Bug 2538692

Change-Id: I9248e0c69e2271ed2d0070587db59afa6f8160f2
Signed-off-by: Peng Liu <pengliu@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2109708
(cherry picked from commit cc70f89bb4)
Reviewed-on: https://git-master.nvidia.com/r/2113647
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Aaron Tian <atian@nvidia.com>
Tested-by: Aaron Tian <atian@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
Peng Liu
2019-03-13 10:30:54 -07:00
committed by mobile promotions
parent af95d14bb0
commit 27625718c4
5 changed files with 9 additions and 28 deletions

View File

@@ -1,7 +1,7 @@
/*
* GK20A Platform (SoC) Interface
*
* Copyright (c) 2014-2018, NVIDIA CORPORATION. All rights reserved.
* Copyright (c) 2014-2019, NVIDIA CORPORATION. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify it
* under the terms and conditions of the GNU General Public License,
@@ -271,9 +271,6 @@ struct gk20a_platform {
/* stream id to use */
u32 ltc_streamid;
/* scaling rate */
unsigned long cached_rate;
/* synchronized access to platform->clk_get_freqs */
struct nvgpu_mutex clk_get_freq_lock;
};