drm/amdkfd: Add SVM range mapped_to_gpu flag

To avoid unnecessary unmap SVM range from GPUs if range is not mapped on
GPUs when migrating the range. This flag will also be used to flush TLB
when updating the existing mapping on GPUs.

It is protected by prange->migrate_mutex and mmap read lock in MMU
notifier callback.

Signed-off-by: Philip Yang <[email protected]>
Reviewed-by: Felix Kuehling <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>

2 files changed, 17 insertions, 1 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index aed2f6f3ce5e..fff17e747a71 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -950,6 +950,7 @@ svm_range_split_adjust(struct svm_range *new, struct svm_range *old,
 	new->prefetch_loc = old->prefetch_loc;
 	new->actual_loc = old->actual_loc;
 	new->granularity = old->granularity;
+	new->mapped_to_gpu = old->mapped_to_gpu;
 	bitmap_copy(new->bitmap_access, old->bitmap_access, MAX_GPU_INSTANCE);
 	bitmap_copy(new->bitmap_aip, old->bitmap_aip, MAX_GPU_INSTANCE);
 
@@ -1203,6 +1204,17 @@ svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start,
 	uint32_t gpuidx;
 	int r = 0;
 
+	if (!prange->mapped_to_gpu) {
+		pr_debug("prange 0x%p [0x%lx 0x%lx] not mapped to GPU\n",
+			 prange, prange->start, prange->last);
+		return 0;
+	}
+
+	if (prange->start == start && prange->last == last) {
+		pr_debug("unmap svms 0x%p prange 0x%p\n", prange->svms, prange);
+		prange->mapped_to_gpu = false;
+	}
+
 	bitmap_or(bitmap, prange->bitmap_access, prange->bitmap_aip,
 		  MAX_GPU_INSTANCE);
 	p = container_of(prange->svms, struct kfd_process, svms);
@@ -1587,8 +1599,10 @@ unlock_out:
 		addr = next;
 	}
 
-	if (addr == end)
+	if (addr == end) {
 		prange->validated_once = true;
+		prange->mapped_to_gpu = true;
+	}
 
 unreserve_out:
 	svm_range_unreserve_bos(&ctx);
@@ -1819,6 +1833,7 @@ static struct svm_range *svm_range_clone(struct svm_range *old)
 	new->prefetch_loc = old->prefetch_loc;
 	new->actual_loc = old->actual_loc;
 	new->granularity = old->granularity;
+	new->mapped_to_gpu = old->mapped_to_gpu;
 	bitmap_copy(new->bitmap_access, old->bitmap_access, MAX_GPU_INSTANCE);
 	bitmap_copy(new->bitmap_aip, old->bitmap_aip, MAX_GPU_INSTANCE);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
index 66c77f00ac3e..2d54147b4dda 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
@@ -133,6 +133,7 @@ struct svm_range {
 	DECLARE_BITMAP(bitmap_access, MAX_GPU_INSTANCE);
 	DECLARE_BITMAP(bitmap_aip, MAX_GPU_INSTANCE);
 	bool				validated_once;
+	bool				mapped_to_gpu;
 };
 
 static inline void svm_range_lock(struct svm_range *prange)