diff options
author | Alex Waterman <alexw@nvidia.com> | 2017-06-09 14:42:50 -0400 |
---|---|---|
committer | mobile promotions <svcmobile_promotions@nvidia.com> | 2017-07-06 17:44:16 -0400 |
commit | 583704620db88e391f6b14acc57af859a70127de (patch) | |
tree | 8fc3becf2850b724e87011b0e0250c52d0efb7ee /drivers/gpu/nvgpu/gk20a/mm_gk20a.c | |
parent | c1393d5b68e63c992f4c689cb788139fdf8c2f1a (diff) |
gpu: nvgpu: Implement PD packing
In some cases page directories require less than a full page of memory.
For example, on Pascal, the final PD level for large pages is only 256 bytes;
thus 16 PDs can fit in a single page. To allocate an entire page for each of
these 256 B PDs is extremely wasteful. This patch aims to alleviate the
wasted DMA memory from having small PDs in a full page by packing multiple
small PDs into a single page.
The packing is implemented as a slab allocator - each page is a slab and
from each page multiple PD instances can be allocated. Several modifications
to the nvgpu_gmmu_pd struct also needed to be made to support this. The
nvgpu_mem is now a pointer and there's an explicit offset into the nvgpu_mem
struct so that each nvgpu_gmmu_pd knows what portion of the memory it's
using.
The nvgpu_pde_phys_addr() function and the pd_write() functions also require
some changes since the PD no longer is always situated at the start of the
nvgpu_mem.
Initialization and cleanup of the page tables for each VM was slightly
modified to work through the new pd_cache implementation. Some PDs (i.e
the PDB), despite not being a full page, still require a full page for
alignment purposes (HW requirements). Thus a direct allocation method for
PDs is still provided. This is also used when a PD that could in principle
be cached is greater than a page in size.
Lastly a new debug flag was added for the pd_cache code.
JIRA NVGPU-30
Change-Id: I64c8037fc356783c1ef203cc143c4d71bbd5d77c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master/r/1506610
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
GVS: Gerrit_Virtual_Submit
Diffstat (limited to 'drivers/gpu/nvgpu/gk20a/mm_gk20a.c')
-rw-r--r-- | drivers/gpu/nvgpu/gk20a/mm_gk20a.c | 9 |
1 files changed, 5 insertions, 4 deletions
diff --git a/drivers/gpu/nvgpu/gk20a/mm_gk20a.c b/drivers/gpu/nvgpu/gk20a/mm_gk20a.c index 558a1b06..0a84cabb 100644 --- a/drivers/gpu/nvgpu/gk20a/mm_gk20a.c +++ b/drivers/gpu/nvgpu/gk20a/mm_gk20a.c | |||
@@ -478,6 +478,7 @@ static void gk20a_remove_mm_support(struct mm_gk20a *mm) | |||
478 | 478 | ||
479 | gk20a_semaphore_sea_destroy(g); | 479 | gk20a_semaphore_sea_destroy(g); |
480 | gk20a_vidmem_destroy(g); | 480 | gk20a_vidmem_destroy(g); |
481 | nvgpu_pd_cache_fini(g); | ||
481 | } | 482 | } |
482 | 483 | ||
483 | static int gk20a_alloc_sysmem_flush(struct gk20a *g) | 484 | static int gk20a_alloc_sysmem_flush(struct gk20a *g) |
@@ -1560,7 +1561,7 @@ static inline u32 big_valid_pde0_bits(struct gk20a *g, | |||
1560 | struct nvgpu_gmmu_pd *pd, u64 addr) | 1561 | struct nvgpu_gmmu_pd *pd, u64 addr) |
1561 | { | 1562 | { |
1562 | u32 pde0_bits = | 1563 | u32 pde0_bits = |
1563 | nvgpu_aperture_mask(g, &pd->mem, | 1564 | nvgpu_aperture_mask(g, pd->mem, |
1564 | gmmu_pde_aperture_big_sys_mem_ncoh_f(), | 1565 | gmmu_pde_aperture_big_sys_mem_ncoh_f(), |
1565 | gmmu_pde_aperture_big_video_memory_f()) | | 1566 | gmmu_pde_aperture_big_video_memory_f()) | |
1566 | gmmu_pde_address_big_sys_f( | 1567 | gmmu_pde_address_big_sys_f( |
@@ -1573,7 +1574,7 @@ static inline u32 small_valid_pde1_bits(struct gk20a *g, | |||
1573 | struct nvgpu_gmmu_pd *pd, u64 addr) | 1574 | struct nvgpu_gmmu_pd *pd, u64 addr) |
1574 | { | 1575 | { |
1575 | u32 pde1_bits = | 1576 | u32 pde1_bits = |
1576 | nvgpu_aperture_mask(g, &pd->mem, | 1577 | nvgpu_aperture_mask(g, pd->mem, |
1577 | gmmu_pde_aperture_small_sys_mem_ncoh_f(), | 1578 | gmmu_pde_aperture_small_sys_mem_ncoh_f(), |
1578 | gmmu_pde_aperture_small_video_memory_f()) | | 1579 | gmmu_pde_aperture_small_video_memory_f()) | |
1579 | gmmu_pde_vol_small_true_f() | /* tbd: why? */ | 1580 | gmmu_pde_vol_small_true_f() | /* tbd: why? */ |
@@ -2173,14 +2174,14 @@ static int gk20a_init_ce_vm(struct mm_gk20a *mm) | |||
2173 | void gk20a_mm_init_pdb(struct gk20a *g, struct nvgpu_mem *inst_block, | 2174 | void gk20a_mm_init_pdb(struct gk20a *g, struct nvgpu_mem *inst_block, |
2174 | struct vm_gk20a *vm) | 2175 | struct vm_gk20a *vm) |
2175 | { | 2176 | { |
2176 | u64 pdb_addr = nvgpu_mem_get_base_addr(g, &vm->pdb.mem, 0); | 2177 | u64 pdb_addr = nvgpu_mem_get_base_addr(g, vm->pdb.mem, 0); |
2177 | u32 pdb_addr_lo = u64_lo32(pdb_addr >> ram_in_base_shift_v()); | 2178 | u32 pdb_addr_lo = u64_lo32(pdb_addr >> ram_in_base_shift_v()); |
2178 | u32 pdb_addr_hi = u64_hi32(pdb_addr); | 2179 | u32 pdb_addr_hi = u64_hi32(pdb_addr); |
2179 | 2180 | ||
2180 | gk20a_dbg_info("pde pa=0x%llx", pdb_addr); | 2181 | gk20a_dbg_info("pde pa=0x%llx", pdb_addr); |
2181 | 2182 | ||
2182 | nvgpu_mem_wr32(g, inst_block, ram_in_page_dir_base_lo_w(), | 2183 | nvgpu_mem_wr32(g, inst_block, ram_in_page_dir_base_lo_w(), |
2183 | nvgpu_aperture_mask(g, &vm->pdb.mem, | 2184 | nvgpu_aperture_mask(g, vm->pdb.mem, |
2184 | ram_in_page_dir_base_target_sys_mem_ncoh_f(), | 2185 | ram_in_page_dir_base_target_sys_mem_ncoh_f(), |
2185 | ram_in_page_dir_base_target_vid_mem_f()) | | 2186 | ram_in_page_dir_base_target_vid_mem_f()) | |
2186 | ram_in_page_dir_base_vol_true_f() | | 2187 | ram_in_page_dir_base_vol_true_f() | |