Re: [PATCH] mm: centralize and fix max map count limit checking

[Mike Rapoport <rppt@kernel.org>]

On Thu, Sep 04, 2025 at 03:11:46PM -0400, Liam R. Howlett wrote:
> * David Hildenbrand <david@redhat.com> [250904 15:02]:
> > On 04.09.25 20:49, Liam R. Howlett wrote:
> > > * Kalesh Singh <kaleshsingh@google.com> [250904 13:51]:
> > > > On Thu, Sep 4, 2025 at 10:42 AM David Hildenbrand <david@redhat.com> wrote:
> > > > > 
> > > > > On 04.09.25 19:33, Lorenzo Stoakes wrote:
> > > > > > On Thu, Sep 04, 2025 at 01:22:51PM -0400, Liam R. Howlett wrote:
> > > > > > > > > diff --git a/mm/mremap.c b/mm/mremap.c
> > > > > > > > > index e618a706aff5..793fad58302c 100644
> > > > > > > > > --- a/mm/mremap.c
> > > > > > > > > +++ b/mm/mremap.c
> > > > > > > > > @@ -1040,7 +1040,7 @@ static unsigned long prep_move_vma(struct vma_remap_struct *vrm)
> > > > > > > > >      * We'd prefer to avoid failure later on in do_munmap:
> > > > > > > > >      * which may split one vma into three before unmapping.
> > > > > > > > >      */
> > > > > > > > > -  if (current->mm->map_count >= sysctl_max_map_count - 3)
> > > > > > > > > +  if (exceeds_max_map_count(current->mm, 4))
> > > > > > > > >             return -ENOMEM;
> > > > > > > > 
> > > > > > > > In my version this would be:
> > > > > > > > 
> > > > > > > >      if (map_count_capacity(current->mm) < 4)
> > > > > > > >              return -ENOMEM;
> > > > > > > > 
> > > > > > > 
> > > > > > > Someone could write map_count_capacity(current->mm) <= 4 and reintroduce
> > > > > > > what this is trying to solve.  And with the way it is written in this
> > > > > > > patch, someone could pass in the wrong number.
> > > > > > 
> > > > > > Right, but I think 'capacity' is pretty clear here, if the caller does something
> > > > > > silly then that's on them...
> > > > > > 
> > > > > > > 
> > > > > > > I'm not sure this is worth doing.  There are places we allow the count
> > > > > > > to go higher.
> > > > > > 
> > > > > > ...But yeah, it's kinda borderline as to how useful this is.
> > > > > > 
> > > > > > I _do_ however like the 'put map count in one place statically' rather than
> > > > > > having a global, so a minimal version of this could be to just have a helper
> > > > > > function that gets the sysctl_max_map_count, e.g.:
> > > > > > 
> > > > > > if (current->mm->mmap_count >= max_map_count() - 3)
> > > > > 
> > > > > I enjoy seeing sysctl_max_map_count hidden. But map_count_capacity() is
> > > > > even more readable, so I like it.
> > > > > 
> > > > > I don't complete like the "capacity" term, but I cannot think of
> > > > > something better right now. Maybe something around "free" or
> > > > > "remaining", not sure.
> > > > > 
> > > > > I also don't completely like "map_count" (I know, I know, we call it
> > > > > like that in structures), because it reminds me of the mapcount ...
> > > > > talking somehow about "vmas" would be quite clear.
> > > > 
> > > > Thanks David, my original implementation started with vma_limit() :).
> > > > Maybe something like vma_count_remaining() ?
> > > 
> > > Yes, reducing this confusion would very much be helpful.  In fact, if
> > > you put it in its own function we could change the actual name with
> > > lower impact.  map_count vs mapcount is annoying.
> > > 
> > > vma_headroom() ?
> > > additional_vma_space() ?
> > 
> > VMA space might be interpreted as VA space.
> 
> Fair enough.
> 
> > 
> > I think basing it on "vma_count" would be good.
> > 
> > vma_count_capacity()
> > 
> > vma_count_headroom()
> > 
> > vma_count_remaining()
> 
> headroom or remaining have my vote as the clearest.

I like 'remaining' more
 
> > vma_count_avail()
> > 
> > vma_count_left()
> > 

-- 
Sincerely yours,
Mike.