Re: [PATCH] mm/vma: refactor vma_modify_flags_name() to vma_modify_name()

[Lorenzo Stoakes <lorenzo.stoakes@oracle.com>]

On Mon, Jul 14, 2025 at 04:18:13PM +0200, David Hildenbrand wrote:
> On 14.07.25 15:58, Lorenzo Stoakes wrote:
> > The single instance in which we use this function doesn't actually need to
> > change VMA flags, so remove this parameter and update the caller
> > accordingly.
> >
> > Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
> > ---
> >   mm/madvise.c |  4 ++--
> >   mm/vma.c     |  4 +---
> >   mm/vma.h     | 15 +++++++--------
> >   3 files changed, 10 insertions(+), 13 deletions(-)
> >
> > diff --git a/mm/madvise.c b/mm/madvise.c
> > index 9de9b7c797c6..afa9e4db2adb 100644
> > --- a/mm/madvise.c
> > +++ b/mm/madvise.c
> > @@ -162,8 +162,8 @@ static int madvise_update_vma(vm_flags_t new_flags,
> >   		return 0;
> >   	if (set_new_anon_name)
> > -		vma = vma_modify_flags_name(&vmi, madv_behavior->prev, vma,
> > -			range->start, range->end, new_flags, anon_name);
> > +		vma = vma_modify_name(&vmi, madv_behavior->prev, vma,
> > +			range->start, range->end, anon_name);
> >   	else
>
> The doc of madvise_update_vma() is a bit misleading:
>
> "Update the vm_flags and/or anon_name"
>
> I assume it's xor? Do we want to sanity check that this will hold true?

Yeah should be xor, this is just a remnant of a past age :P

No need for sanity check I'd say, as the only code path that does this is very
specific - with behaviour of __MADV_SET_ANON_VMA_NAME and new_flags not updated
from vma->vm_flags.

I can send a fix-patch to update the comment for madvise_update_vma().

>
> In general, LGTM.

Cheers!

>
> --
> Cheers,
>
> David / dhildenb
>