diff options
| author | Deepak Nibade <dnibade@nvidia.com> | 2016-01-06 07:54:48 -0500 |
|---|---|---|
| committer | Terje Bergstrom <tbergstrom@nvidia.com> | 2016-01-21 11:36:41 -0500 |
| commit | f28526bb727679de8016c66607a204f5c294b014 (patch) | |
| tree | 8d466fcbe4beaab7705d0c33256b6e82b5d998e7 /include/linux/tegra_vgpu.h | |
| parent | cd09ac26c7e10fbc4ca96b54be3b1578614e3680 (diff) | |
gpu: nvgpu: move pmu_load_update() to get_dev_status()
We currently call gk20a_pmu_load_update() before calling
update_devfreq()
But it is possible to disable governor and set a
constant/max frequency.
In that case we will unnecessarily keep executing
gk20a_pmu_load_update() for each submit
Hence. move gk20a_pmu_load_update() to gk20a_scale_get_dev_status()
so that we call gk20a_pmu_load_update() only when we really
have to scale the frequency
Bug 200161377
Change-Id: Ifac5a659a3a2d088b636f048213c2fbec801bdb9
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/929509
(cherry picked from commit f857a1b31400dfc0c35c58c6424aaac36bc09e7c)
Reviewed-on: http://git-master/r/933704
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Diffstat (limited to 'include/linux/tegra_vgpu.h')
0 files changed, 0 insertions, 0 deletions