diff options
author | Sunny He <suhe@nvidia.com> | 2017-08-15 15:01:04 -0400 |
---|---|---|
committer | mobile promotions <svcmobile_promotions@nvidia.com> | 2017-09-22 15:55:24 -0400 |
commit | 17c581d75514c32d1e8c1e416beb33b3ccce22a5 (patch) | |
tree | a25d063f19b8e1f83f61af418f3aa2ac32fe0cce /drivers/gpu/nvgpu/gk20a/gk20a.h | |
parent | 0090ee5aca268a3c359f34c74b8c521df3bd8593 (diff) |
gpu: nvgpu: SGL passthrough implementation
The basic nvgpu_mem_sgl implementation provides support
for OS specific scatter-gather list implementations by
simply copying them node by node. This is inefficient,
taking extra time and memory.
This patch implements an nvgpu_mem_sgt struct to act as
a header which is inserted at the front of any scatter-
gather list implementation. This labels every struct
with a set of ops which can be used to interact with
the attached scatter gather list.
Since nvgpu common code only has to interact with these
function pointers, any sgl implementation can be used.
Initialization only requires the allocation of a single
struct, removing the need to copy or iterate through the
sgl being converted.
Jira NVGPU-186
Change-Id: I2994f804a4a4cc141b702e987e9081d8560ba2e8
Signed-off-by: Sunny He <suhe@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1541426
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Diffstat (limited to 'drivers/gpu/nvgpu/gk20a/gk20a.h')
-rw-r--r-- | drivers/gpu/nvgpu/gk20a/gk20a.h | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/drivers/gpu/nvgpu/gk20a/gk20a.h b/drivers/gpu/nvgpu/gk20a/gk20a.h index 355228db..13c62691 100644 --- a/drivers/gpu/nvgpu/gk20a/gk20a.h +++ b/drivers/gpu/nvgpu/gk20a/gk20a.h | |||
@@ -34,7 +34,7 @@ struct gk20a_debug_output; | |||
34 | struct nvgpu_clk_pll_debug_data; | 34 | struct nvgpu_clk_pll_debug_data; |
35 | struct nvgpu_nvhost_dev; | 35 | struct nvgpu_nvhost_dev; |
36 | struct nvgpu_cpu_time_correlation_sample; | 36 | struct nvgpu_cpu_time_correlation_sample; |
37 | struct nvgpu_mem_sgl; | 37 | struct nvgpu_mem_sgt; |
38 | 38 | ||
39 | #include <nvgpu/lock.h> | 39 | #include <nvgpu/lock.h> |
40 | #include <nvgpu/thread.h> | 40 | #include <nvgpu/thread.h> |
@@ -700,7 +700,7 @@ struct gpu_ops { | |||
700 | bool (*support_sparse)(struct gk20a *g); | 700 | bool (*support_sparse)(struct gk20a *g); |
701 | u64 (*gmmu_map)(struct vm_gk20a *vm, | 701 | u64 (*gmmu_map)(struct vm_gk20a *vm, |
702 | u64 map_offset, | 702 | u64 map_offset, |
703 | struct nvgpu_mem_sgl *sgl, | 703 | struct nvgpu_sgt *sgt, |
704 | u64 buffer_offset, | 704 | u64 buffer_offset, |
705 | u64 size, | 705 | u64 size, |
706 | int pgsz_idx, | 706 | int pgsz_idx, |
@@ -760,9 +760,9 @@ struct gpu_ops { | |||
760 | size_t size); | 760 | size_t size); |
761 | struct { | 761 | struct { |
762 | u32 (*enter)(struct gk20a *g, struct nvgpu_mem *mem, | 762 | u32 (*enter)(struct gk20a *g, struct nvgpu_mem *mem, |
763 | struct nvgpu_mem_sgl *sgl, u32 w); | 763 | struct nvgpu_sgt *sgt, void *sgl, u32 w); |
764 | void (*exit)(struct gk20a *g, struct nvgpu_mem *mem, | 764 | void (*exit)(struct gk20a *g, struct nvgpu_mem *mem, |
765 | struct nvgpu_mem_sgl *sgl); | 765 | void *sgl); |
766 | u32 (*data032_r)(u32 i); | 766 | u32 (*data032_r)(u32 i); |
767 | } pramin; | 767 | } pramin; |
768 | struct { | 768 | struct { |