> There are two types of iterators mas and vmi in the current code. If the
> maple tree comes from the mm struct, we can use the vma iterator. Avoid
> using mas directly as possible.
>
Helge,
Yajun has changed code that affects parisc with this patch, which hasn't
been tested.
* Yajun Deng <yajun.deng@xxxxxxxxx> [240226 01:16]:
> Keep using mas for the mt_detach tree, since it doesn't come from the mm
> struct.
>
> Remove as many mas uses as feasible, but we will still have a few that
> must be passed through in unmap_vmas() and free_pgtables().
>
> Also introduce vma_iter_reset, vma_iter_{prev, next}_range_limit and
> vma_iter_area_{lowest, highest} helper functions for using vma interator.
>
> Signed-off-by: Yajun Deng <yajun.deng@xxxxxxxxx>
> ---
> v2: Add vma_iter_reset() and vma_iter_{prev, next}_range_limit() helper functions.
> v1: https://lore.kernel.org/all/9f76ce23-67b1-ccbe-a722-1db0e8f0a408@xxxxxxxxx/
> ---
...
>
> @@ -1975,7 +1974,7 @@ static int expand_upwards(struct vm_area_struct *vma, unsigned long address)
> struct vm_area_struct *next;
> unsigned long gap_addr;
> int error = 0;
> - MA_STATE(mas, &mm->mm_mt, vma->vm_start, address);
> + VMA_ITERATOR(vmi, mm, vma->vm_start);
>
> if (!(vma->vm_flags & VM_GROWSUP))
> return -EFAULT;
> @@ -2001,15 +2000,15 @@ static int expand_upwards(struct vm_area_struct *vma, unsigned long address)
> }
>
> if (next)
> - mas_prev_range(&mas, address);
> + vma_iter_prev_range_limit(&vmi, address);
>
> - __mas_set_range(&mas, vma->vm_start, address - 1);
> - if (mas_preallocate(&mas, vma, GFP_KERNEL))
> + vma_iter_config(&vmi, vma->vm_start, address);
> + if (vma_iter_prealloc(&vmi, vma))
> return -ENOMEM;
>
> /* We must make sure the anon_vma is allocated. */
> if (unlikely(anon_vma_prepare(vma))) {
> - mas_destroy(&mas);
> + vma_iter_free(&vmi);
> return -ENOMEM;
> }
>
> @@ -2049,7 +2048,7 @@ static int expand_upwards(struct vm_area_struct *vma, unsigned long address)
> anon_vma_interval_tree_pre_update_vma(vma);
> vma->vm_end = address;
> /* Overwrite old entry in mtree. */
> - mas_store_prealloc(&mas, vma);
> + vma_iter_store(&vmi, vma);
> anon_vma_interval_tree_post_update_vma(vma);
> spin_unlock(&mm->page_table_lock);
>
> @@ -2058,7 +2057,7 @@ static int expand_upwards(struct vm_area_struct *vma, unsigned long address)
> }
> }
> anon_vma_unlock_write(vma->anon_vma);
> - mas_destroy(&mas);
> + vma_iter_free(&vmi);
> validate_mm(mm);
> return error;
> }
Can you make sure that the expand_upwards() still functions on parisc?
Thanks,
Liam
...
