gpu: nvgpu: Make context mapping non-cacheable

Changes to the context header after the context has been loaded may
not be visible to the GPU when mapped as cacheable memory. Examples
include updating the preemption modes or boosted_ctx bits at runtime.

This patch changes the mapping to non-cacheable.

Bug 1819874
Bug 1852094
Bug 200265538

Change-Id: I3b9e87adeaf32e337ec48e01631ad9dea61cc7da
Signed-off-by: Peter Boonstoppel <pboonstoppel@nvidia.com>
Reviewed-on: http://git-master/r/1297601
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
Peter Boonstoppel
2017-02-01 16:04:31 -08:00
committed by mobile promotions
parent a58dc57282
commit bb33e25a05

View File

@@ -2941,7 +2941,7 @@ int gr_gk20a_alloc_gr_ctx(struct gk20a *g,
gr_ctx->mem.gpu_va = gk20a_gmmu_map(vm, gr_ctx->mem.gpu_va = gk20a_gmmu_map(vm,
&gr_ctx->mem.sgt, &gr_ctx->mem.sgt,
gr_ctx->mem.size, gr_ctx->mem.size,
NVGPU_MAP_BUFFER_FLAGS_CACHEABLE_TRUE, NVGPU_MAP_BUFFER_FLAGS_CACHEABLE_FALSE,
gk20a_mem_flag_none, true, gk20a_mem_flag_none, true,
gr_ctx->mem.aperture); gr_ctx->mem.aperture);
if (!gr_ctx->mem.gpu_va) if (!gr_ctx->mem.gpu_va)