Re: [PATCH v4 50/66] mm/gup: Use maple tree navigation instead of linked list

[Vlastimil Babka]

On 12/1/21 15:30, Liam Howlett wrote:
> From: "Liam R. Howlett" <Liam.Howlett@xxxxxxxxxx>

Empty message.

> Signed-off-by: Liam R. Howlett <Liam.Howlett@xxxxxxxxxx>
> ---
>  mm/gup.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/mm/gup.c b/mm/gup.c
> index 2c51e9748a6a..60892e5df6a2 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -1570,7 +1570,6 @@ int __mm_populate(unsigned long start, unsigned long len, int ignore_errors)
>  	long ret = 0;
>  
>  	end = start + len;
> -
>  	for (nstart = start; nstart < end; nstart = nend) {
>  		/*
>  		 * We want to fault in pages for [nstart; end) address range.
> @@ -1579,10 +1578,10 @@ int __mm_populate(unsigned long start, unsigned long len, int ignore_errors)
>  		if (!locked) {
>  			locked = 1;
>  			mmap_read_lock(mm);
> -			vma = find_vma(mm, nstart);
> +			vma = find_vma_intersection(mm, start, end);

Is it ok to use start instead of nstart? At least theoretically,
populate_vma_page_range() called later can reset locked to 0, so !locked
doesn't guarantee we are in the first iteration?

>  		} else if (nstart >= vma->vm_end)
> -			vma = vma->vm_next;
> -		if (!vma || vma->vm_start >= end)
> +			vma = find_vma(mm, vma->vm_end);
> +		if (!vma)
>  			break;
>  		/*
>  		 * Set [nstart; nend) to intersection of desired address