Re: [PATCH v21 1/4] mm: add VM_DROPPABLE for designating always lazily freeable mappings

[David Hildenbrand <david@redhat.com>]

On 08.07.24 16:40, Jason A. Donenfeld wrote:
> Hi David, Linus,
> 
> Below is what I understand the suggestions about the UX to be. The full
> commit is in https://git.zx2c4.com/linux-rng/log/ but here's the part
> we've been discussing. I've held off on David's suggestion changing
> "DROPPABLE" to "VOLATILE" to give Linus some time to wake up on the west
> coast and voice his preference for "DROPPABLE". But the rest is in
> place.
> 
> Jason
> 
> diff --git a/include/uapi/linux/mman.h b/include/uapi/linux/mman.h
> index a246e11988d5..e89d00528f2f 100644
> --- a/include/uapi/linux/mman.h
> +++ b/include/uapi/linux/mman.h
> @@ -17,6 +17,7 @@
>   #define MAP_SHARED	0x01		/* Share changes */
>   #define MAP_PRIVATE	0x02		/* Changes are private */
>   #define MAP_SHARED_VALIDATE 0x03	/* share + validate extension flags */
> +#define MAP_DROPPABLE	0x08		/* Zero memory under memory pressure. */
>   
>   /*
>    * Huge page size encoding when MAP_HUGETLB is specified, and a huge page
> diff --git a/mm/madvise.c b/mm/madvise.c
> index a77893462b92..cba5bc652fc4 100644
> --- a/mm/madvise.c
> +++ b/mm/madvise.c
> @@ -1068,13 +1068,16 @@ static int madvise_vma_behavior(struct vm_area_struct *vma,
>   		new_flags |= VM_WIPEONFORK;
>   		break;
>   	case MADV_KEEPONFORK:
> +		if (vma->vm_flags & VM_DROPPABLE)
> +			return -EINVAL;
>   		new_flags &= ~VM_WIPEONFORK;
>   		break;
>   	case MADV_DONTDUMP:
>   		new_flags |= VM_DONTDUMP;
>   		break;
>   	case MADV_DODUMP:
> -		if (!is_vm_hugetlb_page(vma) && new_flags & VM_SPECIAL)
> +		if ((!is_vm_hugetlb_page(vma) && new_flags & VM_SPECIAL) ||
> +		    (vma->vm_flags & VM_DROPPABLE))
>   			return -EINVAL;
>   		new_flags &= ~VM_DONTDUMP;
>   		break;
> diff --git a/mm/mlock.c b/mm/mlock.c
> index 30b51cdea89d..b87b3d8cc9cc 100644
> --- a/mm/mlock.c
> +++ b/mm/mlock.c
> @@ -485,7 +485,7 @@ static int mlock_fixup(struct vma_iterator *vmi, struct vm_area_struct *vma,
>   
>   	if (newflags == oldflags || (oldflags & VM_SPECIAL) ||
>   	    is_vm_hugetlb_page(vma) || vma == get_gate_vma(current->mm) ||
> -	    vma_is_dax(vma) || vma_is_secretmem(vma))
> +	    vma_is_dax(vma) || vma_is_secretmem(vma) || (oldflags & VM_DROPPABLE))
>   		/* don't set VM_LOCKED or VM_LOCKONFAULT and don't count */
>   		goto out;
>   
> diff --git a/mm/mmap.c b/mm/mmap.c
> index 83b4682ec85c..b3d38179dd42 100644
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -1369,6 +1369,34 @@ unsigned long do_mmap(struct file *file, unsigned long addr,
>   			pgoff = 0;
>   			vm_flags |= VM_SHARED | VM_MAYSHARE;
>   			break;
> +		case MAP_DROPPABLE:
> +			/*
> +			 * A locked or stack area makes no sense to be droppable.
> +			 *
> +			 * Also, since droppable pages can just go away at any time
> +			 * it makes no sense to copy them on fork or dump them.
> +			 *
> +			 * And don't attempt to combine with hugetlb for now.
> +			 */
> +			if (flags & (MAP_LOCKED | MAP_HUGETLB))
> +			        return -EINVAL;
> +			if (vm_flags & (VM_GROWSDOWN | VM_GROWSUP))
> +			        return -EINVAL;
> +
> +			vm_flags |= VM_DROPPABLE;
> +
> +			/*
> +			 * If the pages can be dropped, then it doesn't make
> +			 * sense to reserve them.
> +			 */
> +			vm_flags |= VM_NORESERVE;

That is certainly interesting. Nothing that we might not be able to 
reclaim these pages reliably in all cases: for example when long-term 
pinning them.

In some environments (OVERCOMMIT_NEVER) MAP_NORESERE would never be 
effective. I wonder if we want to stick to the same behavior here ... 
but in theory I agree that we can set this here unconditionally, it's 
just the corner case of "there are ways to prohibit reclaim" that makes 
me wonder.

BTW, I was just trying to understand how MADV_FREE + MAP_DROPPABLE would 
behave without any swap space around.

Did you experiment with that?

I'm reading can_reclaim_anon_pages(), and I'm wondering how 
good/reliable that works when there is no swap configured.

Also, the comment in get_scan_count(): "If we have no swap space, do not 
bother scanning anon folios." makes me wonder if some work in that area 
is needed.

-- 
Cheers,

David / dhildenb