Konsta Hölttä d660a5d0e8 gpu: nvgpu: ensure coherency in sema wait ioctl ...

The semaphore dmabuf supplied in NVGPU_IOCTL_CHANNEL_WAIT is not
necessarily always cache coherent with the GPU. Call
dma_buf_begin_cpu_access() and dma_buf_end_cpu_access() around the sema
read to make sure we see updated values after the interrupt.

Jira NVGPU-5387
Bug 3028497

Change-Id: I09d23c8a679621c86bdfe609d454199e05fa2987
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2359002
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

2020-12-15 14:13:28 -06:00

..

nvgpu

gpu: nvgpu: ensure coherency in sema wait ioctl

2020-12-15 14:13:28 -06:00

Makefile

nvgpu: use kernel overlay features

2017-10-29 11:00:55 -07:00