[PATCH] drm/amdkfd: Fix partial migrate issue
Chen, Xiaogang
xiaogang.chen at amd.com
Mon Jan 6 15:26:58 UTC 2025
On 1/2/2025 6:06 PM, Emily Deng wrote:
> For partial migrate from ram to vram, the migrate->cpages is not
> equal to migrate->npages, should use migrate->npages to check all needed
> migrate pages which could be copied or not.
>
> And only need to set those pages could be migrated to migrate->dst[i], or
> the migrate_vma_pages will migrate the wrong pages based on the migrate->dst[i].
>
> Signed-off-by: Emily Deng<Emily.Deng at amd.com>
> ---
> drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
> index 4b275937d05e..5c96c2d425e3 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
> @@ -278,7 +278,7 @@ svm_migrate_copy_to_vram(struct kfd_node *node, struct svm_range *prange,
> struct migrate_vma *migrate, struct dma_fence **mfence,
> dma_addr_t *scratch, uint64_t ttm_res_offset)
> {
> - uint64_t npages = migrate->cpages;
> + uint64_t npages = migrate->npages;
I agree this part.
> struct amdgpu_device *adev = node->adev;
> struct device *dev = adev->dev;
> struct amdgpu_res_cursor cursor;
> @@ -299,9 +299,6 @@ svm_migrate_copy_to_vram(struct kfd_node *node, struct svm_range *prange,
> struct page *spage;
>
> dst[i] = cursor.start + (j << PAGE_SHIFT);
> - migrate->dst[i] = svm_migrate_addr_to_pfn(adev, dst[i]);
> - svm_migrate_get_vram_page(prange, migrate->dst[i]);
> - migrate->dst[i] = migrate_pfn(migrate->dst[i]);
>
> spage = migrate_pfn_to_page(migrate->src[i]);
> if (spage && !is_zone_device_page(spage)) {
> @@ -345,6 +342,9 @@ svm_migrate_copy_to_vram(struct kfd_node *node, struct svm_range *prange,
> } else {
> j++;
> }
> + migrate->dst[i] = svm_migrate_addr_to_pfn(adev, dst[i]);
> + svm_migrate_get_vram_page(prange, migrate->dst[i]);
> + migrate->dst[i] = migrate_pfn(migrate->dst[i]);
I think what current code misses here is that migrate->dst[i] should
match migrate->src[i]: migrate->dst[i](vram page) got set for page that
will be migrated from system ram, otherwise migrate->dst[i] should be
zero. Your change makes migrates->dst[i] not set though its page has
been migrated by svm_migrate_copy_memory_gart for the case that cpages
!= npages, because you set migrate->dst[i] at end of loop and use
'continue' after migration.
The page migration happens at svm_migrate_copy_memory_gart, not
migrate_vma_pages that migrates struct page meta-data from source struct
page to destination struct page and has mmu notification.
Regards
Xiaogang
> }
>
> r = svm_migrate_copy_memory_gart(adev, src + i - j, dst + i - j, j,
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20250106/1d82e057/attachment-0001.htm>
More information about the amd-gfx
mailing list