drm/amdgpu: new cache coherence change for Aldebaran

To support new cache coherence HW on A+A platform mainly in KFD.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Reviewed-by: Oak Zeng <Oak.Zeng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 3762224..a5f0a46 100644 (file)
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -31,6 +31,7 @@
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_dma_buf.h"
 #include <uapi/linux/kfd_ioctl.h>
+#include "amdgpu_xgmi.h"
 
 /* BO flag to indicate a KFD userptr BO */
 #define AMDGPU_AMDKFD_USERPTR_BO (1ULL << 63)
@@ -405,6 +406,8 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
        struct amdgpu_device *bo_adev = amdgpu_ttm_adev(mem->bo->tbo.bdev);
        bool coherent = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_COHERENT;
        uint32_t mapping_flags;
+       uint64_t pte_flags;
+       bool snoop = false;
 
        mapping_flags = AMDGPU_VM_PAGE_READABLE;
        if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE)
@@ -414,7 +417,6 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
 
        switch (adev->asic_type) {
        case CHIP_ARCTURUS:
-       case CHIP_ALDEBARAN:
                if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
                        if (bo_adev == adev)
                                mapping_flags |= coherent ?
@@ -426,12 +428,36 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
                                AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
                }
                break;
+       case CHIP_ALDEBARAN:
+               if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+                       if (bo_adev == adev) {
+                               mapping_flags |= AMDGPU_VM_MTYPE_RW;
+                               if (adev->gmc.xgmi.connected_to_cpu)
+                                       snoop = true;
+                       } else {
+                               mapping_flags |= AMDGPU_VM_MTYPE_NC;
+                               if (amdgpu_xgmi_same_hive(adev, bo_adev))
+                                       snoop = true;
+                       }
+               } else {
+                       snoop = true;
+                       if (adev->gmc.xgmi.connected_to_cpu)
+                               /* system memory uses NC on A+A */
+                               mapping_flags |= AMDGPU_VM_MTYPE_NC;
+                       else
+                               mapping_flags |= coherent ?
+                                       AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
+               }
+               break;
        default:
                mapping_flags |= coherent ?
                        AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
        }
 
-       return amdgpu_gem_va_map_flags(adev, mapping_flags);
+       pte_flags = amdgpu_gem_va_map_flags(adev, mapping_flags);
+       pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0;
+
+       return pte_flags;
 }
 
 /* add_bo_to_vm - Add a BO to a VM

diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 9455204..32b552e 100644 (file)
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1040,6 +1040,9 @@ static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev,
            !(*flags & AMDGPU_PTE_SYSTEM) &&
            mapping->bo_va->is_xgmi)
                *flags |= AMDGPU_PTE_SNOOPED;
+
+       if (adev->asic_type == CHIP_ALDEBARAN)
+               *flags |= mapping->flags & AMDGPU_PTE_SNOOPED;
 }
 
 static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)