Re: [PATCH 7.3 463/507] mm/vma: avoid spurious write locks when mm->mm_lock_seq overflows

Re: [PATCH 7.3 463/507] mm/vma: avoid spurious write locks when mm->mm_lock_seq overflows

[Miguel Ojeda]

On Fri, 13 Nov 2026 10:52:47 -0500 Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
>
> @@ -589,6 +593,16 @@ static inline void vma_end_write_all(struct mm_struct *mm)
>  {
>  	mmap_assert_write_locked(mm);
>  	mm_lock_seqcount_end(mm);
> +
> +	/*
> +	 * VMA seqcounts are not updated on unlock, so once the mm seqcount has
> +	 * wrapped, any VMA still carrying a sequence number from before the
> +	 * wrap would satisfy __is_vma_write_locked() when that number comes
> +	 * round again, and readers would not be excluded. Reset every VMA's
> +	 * seqcount so that no stale value survives the wrap.
> +	 */
> +	if (unlikely(!mm->mm_lock_seq.sequence))
> +		__mm_reset_vma_seqcounts(mm);
>  }

In a UM ~defconfig build with LLVM:

    In file included from arch/um/kernel/asm-offsets.c:8:
    In file included from ./include/linux/audit.h:13:
    In file included from ./include/linux/ptrace.h:10:
    In file included from ./include/linux/pid_namespace.h:7:
    In file included from ./include/linux/mm.h:18:
    ./include/linux/mmap_lock.h:604:20: error: no member named 'mm_lock_seq' in 'struct mm_struct'
      604 |         if (unlikely(!mm->mm_lock_seq.sequence))
          |                       ~~  ^
    ./include/linux/compiler.h:77:42: note: expanded from macro 'unlikely'
       77 | # define unlikely(x)    __builtin_expect(!!(x), 0)
          |

That exists only under `CONFIG_PER_VMA_LOCK` (commit 47e725702b0b ("mm:
make per-VMA locks available universally") from Dave's series was not
backported).

I hope this helps!

Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Liam R. Howlett <liam@infradead.org>
Cc: Lorenzo Stoakes <ljs@kernel.org>
Cc: Vlastimil Babka <vbabka@kernel.org>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: linux-mm@kvack.org

Cc: Jose A. Perez de Azpillaga <azpijr@gmail.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: David Hildenbrand (Arm) <david@kernel.org>

Cheers,
Miguel

Re: [PATCH 7.3 463/507] mm/vma: avoid spurious write locks when mm->mm_lock_seq overflows

[SJ Park]

On Sat, 14 Nov 2026 01:39:12 +0100 Miguel Ojeda <ojeda@kernel.org> wrote:

> On Fri, 13 Nov 2026 10:52:47 -0500 Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
> >
> > @@ -589,6 +593,16 @@ static inline void vma_end_write_all(struct mm_struct *mm)
> >  {
> >  	mmap_assert_write_locked(mm);
> >  	mm_lock_seqcount_end(mm);
> > +
> > +	/*
> > +	 * VMA seqcounts are not updated on unlock, so once the mm seqcount has
> > +	 * wrapped, any VMA still carrying a sequence number from before the
> > +	 * wrap would satisfy __is_vma_write_locked() when that number comes
> > +	 * round again, and readers would not be excluded. Reset every VMA's
> > +	 * seqcount so that no stale value survives the wrap.
> > +	 */
> > +	if (unlikely(!mm->mm_lock_seq.sequence))
> > +		__mm_reset_vma_seqcounts(mm);
> >  }
> 
> In a UM ~defconfig build with LLVM:
> 
>     In file included from arch/um/kernel/asm-offsets.c:8:
>     In file included from ./include/linux/audit.h:13:
>     In file included from ./include/linux/ptrace.h:10:
>     In file included from ./include/linux/pid_namespace.h:7:
>     In file included from ./include/linux/mm.h:18:
>     ./include/linux/mmap_lock.h:604:20: error: no member named 'mm_lock_seq' in 'struct mm_struct'
>       604 |         if (unlikely(!mm->mm_lock_seq.sequence))
>           |                       ~~  ^
>     ./include/linux/compiler.h:77:42: note: expanded from macro 'unlikely'
>        77 | # define unlikely(x)    __builtin_expect(!!(x), 0)
>           |
> 
> That exists only under `CONFIG_PER_VMA_LOCK` (commit 47e725702b0b ("mm:
> make per-VMA locks available universally") from Dave's series was not
> backported).

I found same thing while running kunit.  I also confirmed both applying commit
47e725702b0b from Dave and reverting this patch fix the problem on my setup.


Thanks,
SJ

[...]

Re: [PATCH 7.3 463/507] mm/vma: avoid spurious write locks when mm->mm_lock_seq overflows

[Sasha Levin]

On Fri, Nov 13, 2026 at 07:19:42PM -0800, SJ Park wrote:
> On Sat, 14 Nov 2026 01:39:12 +0100 Miguel Ojeda <ojeda@kernel.org> wrote:
>> ./include/linux/mmap_lock.h:604:20: error: no member named 'mm_lock_seq' in 'struct mm_struct'
>
> I found same thing while running kunit.  I also confirmed both applying commit
> 47e725702b0b from Dave and reverting this patch fix the problem on my setup.

Dropped from the 7.3 queue, together with the prerequisite it pulled in.

Thanks to both of you for the report and the confirmation.

-- 
Thanks,
Sasha

Re: [PATCH 7.3 463/507] mm/vma: avoid spurious write locks when mm->mm_lock_seq overflows

[Lorenzo Stoakes (ARM)]

On Sat, Nov 14, 2026 at 10:41:11AM -0500, Sasha Levin wrote:
> On Fri, Nov 13, 2026 at 07:19:42PM -0800, SJ Park wrote:
> > On Sat, 14 Nov 2026 01:39:12 +0100 Miguel Ojeda <ojeda@kernel.org> wrote:
> >> ./include/linux/mmap_lock.h:604:20: error: no member named 'mm_lock_seq' in 'struct mm_struct'
> >
> > I found same thing while running kunit.  I also confirmed both applying commit
> > 47e725702b0b from Dave and reverting this patch fix the problem on my setup.
>
> Dropped from the 7.3 queue, together with the prerequisite it pulled in.
>
> Thanks to both of you for the report and the confirmation.

I'll send a fixup.

--
Cheers, Lorenzo