gpu: nvgpu: add platform atomic support

Add new variable in nvgpu_as_map_buffer_ex_args for app
to specify the platform atomic support for the page.
When platform atomic attribute flag is set, pte memory
aperture is set to be coherent type.

renamed nvgpu_aperture_mask_coh -> nvgpu_aperture_mask_raw
function.

bug 200473147

Change-Id: I18266724dafdc8dfd96a0711f23cf08e23682afc
Signed-off-by: Vinod G <vinodg@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2012679
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
Vinod G
2019-02-05 18:05:42 -08:00
committed by mobile promotions
parent 45fa0441f7
commit 9e0a9004b7
13 changed files with 69 additions and 23 deletions

View File

@@ -1927,6 +1927,7 @@ struct nvgpu_as_bind_channel_args {
#define NVGPU_AS_MAP_BUFFER_FLAGS_MAPPABLE_COMPBITS (1 << 6)
#define NVGPU_AS_MAP_BUFFER_FLAGS_L3_ALLOC (1 << 7)
#define NVGPU_AS_MAP_BUFFER_FLAGS_DIRECT_KIND_CTRL (1 << 8)
#define NVGPU_AS_MAP_BUFFER_FLAGS_PLATFORM_ATOMIC (1 << 9)
/*
* VM map buffer IOCTL
@@ -1975,6 +1976,10 @@ struct nvgpu_as_bind_channel_args {
* Set when userspace plans to pass in @compr_kind and @incompr_kind
* instead of letting the kernel work out kind fields.
*
* %NVGPU_AS_MAP_BUFFER_FLAGS_PLATFORM_ATOMIC
*
* Specify that a mapping should use platform atomics.
*
* @kind [IN]
*
* Specify the kind to use for the mapping.