NVIDIA/open-gpu-kernel-modules
1
0
Fork 0
mirror of https://github.com/NVIDIA/open-gpu-kernel-modules.git synced 2026-02-20 06:53:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

590.44.01

Browse Source
This commit is contained in:
Maneet Singh
2025-12-02 15:32:25 -08:00
parent 2af9f1f0f7
commit a5bfb10e75
954 changed files with 421883 additions and 408177 deletions
Download Patch File Download Diff File Expand all files Collapse all files

94
kernel-open/nvidia-uvm/uvm_page_tree_test.c
View File

@@ -1,5 +1,5 @@
/*******************************************************************************
Copyright (c) 2015-2024 NVIDIA Corporation
Copyright (c) 2015-2025 NVIDIA Corporation
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to
@@ -44,6 +44,10 @@
#include "clc36f.h"
#include "clc3b5.h"
// TURING_*
#include "clc46f.h"
#include "clc5b5.h"
// AMPERE_*
#include "clc56f.h"
#include "clc6b5.h"
@@ -1783,11 +1787,16 @@ static NV_STATUS entry_test_page_size_volta(uvm_gpu_t *gpu, size_t page_size)
return entry_test_page_size_pascal(gpu, page_size);
}
static NV_STATUS entry_test_page_size_ampere(uvm_gpu_t *gpu, size_t page_size)
static NV_STATUS entry_test_page_size_turing(uvm_gpu_t *gpu, size_t page_size)
{
return entry_test_page_size_volta(gpu, page_size);
}
static NV_STATUS entry_test_page_size_ampere(uvm_gpu_t *gpu, size_t page_size)
{
return entry_test_page_size_turing(gpu, page_size);
}
static NV_STATUS entry_test_page_size_hopper(uvm_gpu_t *gpu, size_t page_size)
{
uvm_mmu_mode_hal_t *hal = gpu->parent->arch_hal->mmu_mode_hal(UVM_PAGE_SIZE_64K);
@@ -1810,11 +1819,13 @@ typedef NV_STATUS (*entry_test_page_size_func)(uvm_gpu_t *gpu, size_t page_size)
static NV_STATUS entry_test_maxwell(uvm_gpu_t *gpu)
{
NV_STATUS status = NV_OK;
static const NvU64 big_page_sizes[] = {UVM_PAGE_SIZE_64K, UVM_PAGE_SIZE_128K};
NvU64 pde_bits;
uvm_mmu_page_table_alloc_t *phys_allocs[2];
uvm_mmu_page_table_alloc_t alloc_sys = fake_table_alloc(UVM_APERTURE_SYS, 0x9999999000LL);
uvm_mmu_page_table_alloc_t alloc_vid = fake_table_alloc(UVM_APERTURE_VID, 0x1BBBBBB000LL);
uvm_page_tree_t tree;
uvm_mmu_mode_hal_t *hal;
uvm_page_directory_t dir;
NvU64 big_page_size, page_size;
@@ -1895,19 +1906,30 @@ static NV_STATUS entry_test_maxwell(uvm_gpu_t *gpu)
0x1BBBBBB000LL,
UVM_PROT_READ_ONLY,
UVM_MMU_PTE_FLAGS_CACHED) == 0x80000002FBBBBBB5LL);
TEST_NV_CHECK_RET(test_page_tree_init(gpu, big_page_size, &tree));
TEST_CHECK_GOTO(tree.hal->poisoned_pte(&tree) == 0x800000011bad0007ull, cleanup_tree);
uvm_page_tree_deinit(&tree);
}
return NV_OK;
cleanup_tree:
uvm_page_tree_deinit(&tree);
return status;
}
static NV_STATUS entry_test_pascal(uvm_gpu_t *gpu, entry_test_page_size_func entry_test_page_size)
{
NV_STATUS status = NV_OK;
NvU64 page_sizes[MAX_NUM_PAGE_SIZES];
NvU64 pde_bits[2];
size_t i, num_page_sizes;
uvm_mmu_page_table_alloc_t *phys_allocs[2] = {NULL, NULL};
uvm_mmu_page_table_alloc_t alloc_sys = fake_table_alloc(UVM_APERTURE_SYS, 0x399999999999000LL);
uvm_mmu_page_table_alloc_t alloc_vid = fake_table_alloc(UVM_APERTURE_VID, 0x1BBBBBB000LL);
uvm_page_tree_t tree;
uvm_page_directory_t dir;
// big versions have [11:8] set as well to test the page table merging
@@ -1993,7 +2015,16 @@ static NV_STATUS entry_test_pascal(uvm_gpu_t *gpu, entry_test_page_size_func ent
for (i = 0; i < num_page_sizes; i++)
TEST_NV_CHECK_RET(entry_test_page_size(gpu, page_sizes[i]));
TEST_NV_CHECK_RET(test_page_tree_init(gpu, UVM_PAGE_SIZE_64K, &tree));
TEST_CHECK_GOTO(tree.hal->poisoned_pte(&tree) == 0x1bad000e9ull, cleanup_tree);
uvm_page_tree_deinit(&tree);
return NV_OK;
cleanup_tree:
uvm_page_tree_deinit(&tree);
return status;
}
static NV_STATUS entry_test_volta(uvm_gpu_t *gpu, entry_test_page_size_func entry_test_page_size)
@@ -2070,6 +2101,30 @@ static NV_STATUS entry_test_volta(uvm_gpu_t *gpu, entry_test_page_size_func entr
return NV_OK;
}
static NV_STATUS entry_test_turing(uvm_gpu_t *gpu, entry_test_page_size_func entry_test_page_size)
{
NV_STATUS status = NV_OK;
uvm_page_tree_t tree;
NvU64 page_sizes[MAX_NUM_PAGE_SIZES];
NvU32 i, num_page_sizes;
num_page_sizes = get_page_sizes(gpu, page_sizes);
for (i = 0; i < num_page_sizes; i++)
TEST_NV_CHECK_RET(entry_test_page_size(gpu, page_sizes[i]));
TEST_NV_CHECK_RET(test_page_tree_init(gpu, UVM_PAGE_SIZE_64K, &tree));
TEST_CHECK_GOTO(tree.hal->poisoned_pte(&tree) == 0x6000001bad000e9ull, cleanup_tree);
uvm_page_tree_deinit(&tree);
return NV_OK;
cleanup_tree:
uvm_page_tree_deinit(&tree);
return status;
}
static NV_STATUS entry_test_ampere(uvm_gpu_t *gpu, entry_test_page_size_func entry_test_page_size)
{
NvU64 page_sizes[MAX_NUM_PAGE_SIZES];
@@ -2093,6 +2148,7 @@ static NV_STATUS entry_test_hopper(uvm_gpu_t *gpu, entry_test_page_size_func ent
uvm_mmu_page_table_alloc_t *phys_allocs[2] = {NULL, NULL};
uvm_mmu_page_table_alloc_t alloc_sys = fake_table_alloc(UVM_APERTURE_SYS, 0x9999999999000LL);
uvm_mmu_page_table_alloc_t alloc_vid = fake_table_alloc(UVM_APERTURE_VID, 0xBBBBBBB000LL);
uvm_page_tree_t tree;
// Big versions have [11:8] set as well to test the page table merging
uvm_mmu_page_table_alloc_t alloc_big_sys = fake_table_alloc(UVM_APERTURE_SYS, 0x9999999999900LL);
@@ -2101,6 +2157,7 @@ static NV_STATUS entry_test_hopper(uvm_gpu_t *gpu, entry_test_page_size_func ent
uvm_mmu_mode_hal_t *hal = gpu->parent->arch_hal->mmu_mode_hal(UVM_PAGE_SIZE_64K);
memset(dirs, 0, sizeof(dirs));
// Fake directory tree.
for (i = 0; i < ARRAY_SIZE(dirs); i++) {
dirs[i] = uvm_kvmalloc_zero(sizeof(uvm_page_directory_t) + sizeof(dirs[i]->entries[0]) * 512);
@@ -2272,10 +2329,19 @@ static NV_STATUS entry_test_hopper(uvm_gpu_t *gpu, entry_test_page_size_func ent
// sked reflected
TEST_CHECK_GOTO(hal->make_sked_reflected_pte() == 0xF0F, cleanup);
// poisoned - use a fake tree as it is required by poisoned_pte's MMU HAL.
// The tests above manually set the MMU HAL but used functions that don't
// have a uvm_page_tree_t argument.
TEST_NV_CHECK_GOTO(test_page_tree_init(gpu, UVM_PAGE_SIZE_64K, &tree), cleanup);
TEST_CHECK_GOTO(tree.hal->poisoned_pte(&tree) == 0x2bad0006f9ull, cleanup_tree);
num_page_sizes = get_page_sizes(gpu, page_sizes);
for (i = 0; i < num_page_sizes; i++)
TEST_NV_CHECK_GOTO(entry_test_page_size(gpu, page_sizes[i]), cleanup);
TEST_NV_CHECK_GOTO(entry_test_page_size(gpu, page_sizes[i]), cleanup_tree);
cleanup_tree:
uvm_page_tree_deinit(&tree);
cleanup:
for (i = 0; i < ARRAY_SIZE(dirs); i++)
@@ -2286,6 +2352,9 @@ cleanup:
static NV_STATUS entry_test_blackwell(uvm_gpu_t *gpu, entry_test_page_size_func entry_test_page_size)
{
// We use entry_test_ampere() because we only want to check for an
// additional page size, no MMU page table format changes between Hopper and
// Blackwell.
return entry_test_ampere(gpu, entry_test_page_size_blackwell);
}
@@ -2523,6 +2592,15 @@ static NV_STATUS fake_gpu_init_volta(uvm_gpu_t *fake_gpu)
fake_gpu);
}
static NV_STATUS fake_gpu_init_turing(uvm_gpu_t *fake_gpu)
{
return fake_gpu_init(TURING_CHANNEL_GPFIFO_A,
TURING_DMA_COPY_A,
NV2080_CTRL_MC_ARCH_INFO_ARCHITECTURE_TU100,
0,
fake_gpu);
}
static NV_STATUS fake_gpu_init_ampere(uvm_gpu_t *fake_gpu)
{
return fake_gpu_init(AMPERE_CHANNEL_GPFIFO_A,
@@ -2641,6 +2719,15 @@ static NV_STATUS volta_test_page_tree(uvm_gpu_t *volta)
return NV_OK;
}
static NV_STATUS turing_test_page_tree(uvm_gpu_t *turing)
{
TEST_CHECK_RET(fake_gpu_init_turing(turing) == NV_OK);
MEM_NV_CHECK_RET(entry_test_turing(turing, entry_test_page_size_turing), NV_OK);
return NV_OK;
}
static NV_STATUS ampere_test_page_tree(uvm_gpu_t *ampere)
{
NvU32 i, tlb_batch_saved_max_pages;
@@ -2803,6 +2890,7 @@ NV_STATUS uvm_test_page_tree(UVM_TEST_PAGE_TREE_PARAMS *params, struct file *fil
TEST_NV_CHECK_GOTO(maxwell_test_page_tree(gpu), done);
TEST_NV_CHECK_GOTO(pascal_test_page_tree(gpu), done);
TEST_NV_CHECK_GOTO(volta_test_page_tree(gpu), done);
TEST_NV_CHECK_GOTO(turing_test_page_tree(gpu), done);
TEST_NV_CHECK_GOTO(ampere_test_page_tree(gpu), done);
TEST_NV_CHECK_GOTO(hopper_test_page_tree(gpu), done);
TEST_NV_CHECK_GOTO(blackwell_test_page_tree(gpu), done);