On Tue, Sep 23, 2025 at 02:17:04PM -0700, Andrew Morton wrote:
> On Tue, 23 Sep 2025 13:52:09 +0200 Sumanth Korikkar <[email protected]>
> wrote:
>
> > > --- a/fs/hugetlbfs/inode.c
> > > +++ b/fs/hugetlbfs/inode.c
> > > @@ -96,8 +96,15 @@ static const struct fs_parameter_spec
> > > hugetlb_fs_parameters[] = {
> > > #define PGOFF_LOFFT_MAX \
> > > (((1UL << (PAGE_SHIFT + 1)) - 1) << (BITS_PER_LONG - (PAGE_SHIFT + 1)))
> > >
> > > -static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct
> > > *vma)
> > > +static int hugetlb_file_mmap_prepare_success(const struct vm_area_struct
> > > *vma)
> > > {
> > > + /* Unfortunate we have to reassign vma->vm_private_data. */
> > > + return hugetlb_vma_lock_alloc((struct vm_area_struct *)vma);
> > > +}
> >
> > Hi Lorenzo,
> >
> > The following tests causes the kernel to enter a blocked state,
> > suggesting an issue related to locking order. I was able to reproduce
> > this behavior in certain test runs.
>
> Thanks. I pulled this series out of mm.git's mm-stable branch, put it
> back into mm-unstable.
Hi all,
The issue is reproducible again in linux-next with the following commit:
5fdb155933fa ("mm/hugetlbfs: update hugetlbfs to use mmap_prepare")
Thanks,
Sumanth
