fixup: gpu/drm/nouveau: enable THP support for GPU memory migration
Subsystems (4)
PATCH v2 1/1 fixup: gpu/drm/nouveau: enable THP support for GPU memory migration
Kernel test robot, reported a sparse error with the patch
The sparse report states:
Quoted 1 line
Quoted 1 line
drivers/gpu/drm/nouveau/nouveau_dmem.c:259:13: sparse: sparse: incorrect type in assignment (different base types) @@ expected restrictedvm_fault_t[assigned] [usertype] ret @@ got int @@
drivers/gpu/drm/nouveau/nouveau_dmem.c:259:13: sparse: expected restricted vm_fault_t [assigned] [usertype] ret
drivers/gpu/drm/nouveau/nouveau_dmem.c:259:13: sparse: got int
ret of type vm_fault_t is reused to capture the return value of nouveau_dmem_copy_folio(), which returns an int. Use a new variable err (of type int) to get the return value from the function to fix the reported issue.
The reported issue is not new, prior to this, the function called nouveau_dmem_copy_folio() was called nouveau_dmem_copy_one() and ret was used to capture it's return value.
The reported issue also does not cause any runtime errors. The return value from the function is used as a boolean value, to check if nouveau_dmem_copy_folio() succeeded or failed.
---
This fixup should be squashed into the patch "gpu/drm/nouveau: enable THP support for GPU memory" of mm/mm-unstable
David wanted me to change the patch prefix to drm/nouveau/dmem, I can resend the patches, fixup and this one (squashed) if needed.
drivers/gpu/drm/nouveau/nouveau_dmem.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/gpu/drm/nouveau/nouveau_dmem.c +3 -2
@@ -188,6 +188,7 @@ static vm_fault_t nouveau_dmem_migrate_to_ram(struct vm_fault *vmf)struct nouveau_svmm *svmm;struct page *dpage;vm_fault_t ret =0 ; +int err;struct migrate_vma args = { .vma = vmf->vma, .pgmap_owner = drm->dev,
@@ -256,9 +257,9 @@ static vm_fault_t nouveau_dmem_migrate_to_ram(struct vm_fault *vmf) svmm =folio_zone_device_data (sfolio);mutex_lock (&svmm->mutex);nouveau_svmm_invalidate (svmm, args.start, args.end); - ret =nouveau_dmem_copy_folio (drm, sfolio, dfolio, &dma_info); + err =nouveau_dmem_copy_folio (drm, sfolio, dfolio, &dma_info);mutex_unlock (&svmm->mutex); -if (ret) { +if (err) { ret =VM_FAULT_SIGBUS ; goto done; }
Signature · 1 line
2.51.1
Quoted 8 lines · Balbir Singh
Kernel test robot, reported a sparse error with the patch
The sparse report states:
Quoted 1 line
Quoted 1 line
drivers/gpu/drm/nouveau/nouveau_dmem.c:259:13: sparse: sparse: incorrect type in assignment (different base types) @@ expected restrictedvm_fault_t[assigned] [usertype] ret @@ got int @@
drivers/gpu/drm/nouveau/nouveau_dmem.c:259:13: sparse: expected restrictedvm_fault_t[assigned] [usertype] ret
drivers/gpu/drm/nouveau/nouveau_dmem.c:259:13: sparse: got int
I would have simplified that to (and avoided the long line)
"The kernel test robot reports a sparse warning "expected restricted vm_fault_t [assigned] [usertype] ret ... got int."
Whoever applies that can massage it I am sure.
Signature · 3 lines
Cheers David