tests/amdgpu: fix for dispatch/draw test

1. skip test if there's no desired ring
2. clear shader buffer
3. update command buffer for gfx9

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>

diff --git a/tests/amdgpu/basic_tests.c b/tests/amdgpu/basic_tests.c
index 2d47269..8fa7088 100644 (file)
--- a/tests/amdgpu/basic_tests.c
+++ b/tests/amdgpu/basic_tests.c
@@ -329,14 +329,15 @@ static const uint32_t preamblecache_gfx9[] = {
        0xc0016900, 0x2d5, 0x10000, 0xc0016900,  0x2dc, 0x0,
        0xc0066900, 0x2de, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0xc0026900, 0x2e5, 0x0, 0x0,
        0xc0056900, 0x2f9, 0x5, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000,
-       0xc0026900, 0x311,  0x3, 0x0, 0xc0026900, 0x316, 0x1e, 0x20,
+       0xc0036900, 0x311, 0x3, 0, 0x100000, 0xc0026900, 0x316, 0x1e, 0x20,
        0xc0016900, 0x349, 0x0, 0xc0016900, 0x358, 0x0, 0xc0016900, 0x367, 0x0,
        0xc0016900, 0x376, 0x0, 0xc0016900, 0x385, 0x0, 0xc0016900, 0x19, 0x0,
        0xc0056900, 0xe8, 0x0, 0x0, 0x0, 0x0, 0x0,
        0xc0076900, 0x1e1, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
        0xc0026900, 0x204, 0x90000, 0x4, 0xc0046900, 0x20c, 0x0, 0x0, 0x0, 0x0,
        0xc0016900, 0x2b2, 0x0, 0xc0026900, 0x30e, 0xffffffff, 0xffffffff,
-       0xc0016900, 0x314, 0x0, 0xc0002f00, 0x1, 0xc0016900, 0x1, 0x1,
+       0xc0016900, 0x314, 0x0, 0xc0016900, 0x2a6, 0, 0xc0016900, 0x210, 0,
+       0xc0002f00, 0x1, 0xc0016900, 0x1, 0x1,
        0xc0016900, 0x18, 0x2, 0xc0016900, 0x206, 0x300, 0xc0017900, 0x20000243, 0x0,
        0xc0017900, 0x248, 0xffffffff, 0xc0017900, 0x249, 0x0, 0xc0017900, 0x24a, 0x0,
        0xc0017900, 0x24b, 0x0
@@ -450,7 +451,7 @@ static const uint32_t cached_cmd_gfx9[] = {
        0xc0016900, 0x0, 0x0, 0xc0026900, 0x3, 0x2a, 0x0,
        0xc0046900, 0xa, 0x0, 0x0, 0x0, 0x200020,
        0xc0016900, 0x83, 0xffff, 0xc0026900, 0x8e, 0xf, 0xf,
-       0xc0056900, 0x105, 0x0, 0x0,  0x0, 0x0, 0x12,
+       0xc0056900, 0x105, 0x0, 0x0,  0x0, 0x0, 0x1a,
        0xc0026900, 0x10b, 0x0, 0x0, 0xc0016900, 0x1e0, 0x0,
        0xc0036900, 0x200, 0x0, 0x10000, 0xcc0011,
        0xc0026900, 0x292, 0x20, 0x60201b8,
@@ -2094,10 +2095,7 @@ static int amdgpu_dispatch_init(uint32_t *ptr, uint32_t ip_type)
        ptr[i++] = PACKET3_COMPUTE(PKT3_SET_SH_REG, 3);
        ptr[i++] = 0x204;
        i += 3;
-       /* clear mmCOMPUTE_RESOURCE_LIMITS */
-       ptr[i++] = PACKET3_COMPUTE(PKT3_SET_SH_REG, 1);
-       ptr[i++] = 0x215;
-       ptr[i++] = 0;
+
        /* clear mmCOMPUTE_TMPRING_SIZE */
        ptr[i++] = PACKET3_COMPUTE(PKT3_SET_SH_REG, 1);
        ptr[i++] = 0x218;
@@ -2184,6 +2182,7 @@ static void amdgpu_memset_dispatch_test(amdgpu_device_handle device_handle,
                                        &bo_shader, &ptr_shader,
                                        &mc_address_shader, &va_shader);
        CU_ASSERT_EQUAL(r, 0);
+       memset(ptr_shader, 0, bo_shader_size);
 
        r = amdgpu_dispatch_load_cs_shader(ptr_shader, CS_BUFFERCLEAR);
        CU_ASSERT_EQUAL(r, 0);
@@ -2220,6 +2219,11 @@ static void amdgpu_memset_dispatch_test(amdgpu_device_handle device_handle,
        ptr_cmd[i++] = 0x22222222;
        ptr_cmd[i++] = 0x22222222;
 
+       /* clear mmCOMPUTE_RESOURCE_LIMITS */
+       ptr_cmd[i++] = PACKET3_COMPUTE(PKT3_SET_SH_REG, 1);
+       ptr_cmd[i++] = 0x215;
+       ptr_cmd[i++] = 0;
+
        /* dispatch direct command */
        ptr_cmd[i++] = PACKET3_COMPUTE(PACKET3_DISPATCH_DIRECT, 3);
        ptr_cmd[i++] = 0x10;
@@ -2321,6 +2325,7 @@ static void amdgpu_memcpy_dispatch_test(amdgpu_device_handle device_handle,
                                        &bo_shader, &ptr_shader,
                                        &mc_address_shader, &va_shader);
        CU_ASSERT_EQUAL(r, 0);
+       memset(ptr_shader, 0, bo_shader_size);
 
        r = amdgpu_dispatch_load_cs_shader(ptr_shader, CS_BUFFERCOPY );
        CU_ASSERT_EQUAL(r, 0);
@@ -2365,6 +2370,11 @@ static void amdgpu_memcpy_dispatch_test(amdgpu_device_handle device_handle,
        ptr_cmd[i++] = 0x400;
        ptr_cmd[i++] = 0x74fac;
 
+       /* clear mmCOMPUTE_RESOURCE_LIMITS */
+       ptr_cmd[i++] = PACKET3_COMPUTE(PKT3_SET_SH_REG, 1);
+       ptr_cmd[i++] = 0x215;
+       ptr_cmd[i++] = 0;
+
        /* dispatch direct command */
        ptr_cmd[i++] = PACKET3_COMPUTE(PACKET3_DISPATCH_DIRECT, 3);
        ptr_cmd[i++] = 0x10;
@@ -2438,6 +2448,8 @@ static void amdgpu_dispatch_test(void)
 
        r = amdgpu_query_hw_ip_info(device_handle, AMDGPU_HW_IP_COMPUTE, 0, &info);
        CU_ASSERT_EQUAL(r, 0);
+       if (!info.available_rings)
+               printf("SKIP ... as there's no compute ring\n");
 
        for (ring_id = 0; (1 << ring_id) & info.available_rings; ring_id++) {
                amdgpu_memset_dispatch_test(device_handle, AMDGPU_HW_IP_COMPUTE, ring_id);
@@ -2446,6 +2458,8 @@ static void amdgpu_dispatch_test(void)
 
        r = amdgpu_query_hw_ip_info(device_handle, AMDGPU_HW_IP_GFX, 0, &info);
        CU_ASSERT_EQUAL(r, 0);
+       if (!info.available_rings)
+               printf("SKIP ... as there's no graphics ring\n");
 
        for (ring_id = 0; (1 << ring_id) & info.available_rings; ring_id++) {
                amdgpu_memset_dispatch_test(device_handle, AMDGPU_HW_IP_GFX, ring_id);
@@ -2901,12 +2915,14 @@ static void amdgpu_memset_draw_test(amdgpu_device_handle device_handle,
                                        &bo_shader_ps, &ptr_shader_ps,
                                        &mc_address_shader_ps, &va_shader_ps);
        CU_ASSERT_EQUAL(r, 0);
+       memset(ptr_shader_ps, 0, bo_shader_size);
 
        r = amdgpu_bo_alloc_and_map(device_handle, bo_shader_size, 4096,
                                        AMDGPU_GEM_DOMAIN_VRAM, 0,
                                        &bo_shader_vs, &ptr_shader_vs,
                                        &mc_address_shader_vs, &va_shader_vs);
        CU_ASSERT_EQUAL(r, 0);
+       memset(ptr_shader_vs, 0, bo_shader_size);
 
        r = amdgpu_draw_load_ps_shader(ptr_shader_ps, PS_CONST);
        CU_ASSERT_EQUAL(r, 0);
@@ -2996,7 +3012,7 @@ static void amdgpu_memcpy_draw(amdgpu_device_handle device_handle,
        ptr_cmd[i++] = 0x92;
        i += 3;
 
-       ptr_cmd[i++] = PACKET3(PKT3_SET_SH_REG, 1);
+       ptr_cmd[i++] = PACKET3(PACKET3_SET_CONTEXT_REG, 1);
        ptr_cmd[i++] = 0x191;
        ptr_cmd[i++] = 0;
 
@@ -3074,12 +3090,14 @@ static void amdgpu_memcpy_draw_test(amdgpu_device_handle device_handle, uint32_t
                                        &bo_shader_ps, &ptr_shader_ps,
                                        &mc_address_shader_ps, &va_shader_ps);
        CU_ASSERT_EQUAL(r, 0);
+       memset(ptr_shader_ps, 0, bo_shader_size);
 
        r = amdgpu_bo_alloc_and_map(device_handle, bo_shader_size, 4096,
                                        AMDGPU_GEM_DOMAIN_VRAM, 0,
                                        &bo_shader_vs, &ptr_shader_vs,
                                        &mc_address_shader_vs, &va_shader_vs);
        CU_ASSERT_EQUAL(r, 0);
+       memset(ptr_shader_vs, 0, bo_shader_size);
 
        r = amdgpu_draw_load_ps_shader(ptr_shader_ps, PS_TEX);
        CU_ASSERT_EQUAL(r, 0);
@@ -3105,6 +3123,8 @@ static void amdgpu_draw_test(void)
 
        r = amdgpu_query_hw_ip_info(device_handle, AMDGPU_HW_IP_GFX, 0, &info);
        CU_ASSERT_EQUAL(r, 0);
+       if (!info.available_rings)
+               printf("SKIP ... as there's no graphics ring\n");
 
        for (ring_id = 0; (1 << ring_id) & info.available_rings; ring_id++) {
                amdgpu_memset_draw_test(device_handle, ring_id);