Re: [PATCH v8 1/2] hugetlb: don't delete vma_lock in hugetlb MADV_DONTNEED processing

[Mike Kravetz <mike.kravetz@oracle.com>]

On 11/10/22 15:56, Peter Xu wrote:
> Hi, Mike,
> 
> Sorry to be late, took me quite some time working on another bug..
> 
> On Mon, Nov 07, 2022 at 05:19:09PM -0800, Mike Kravetz wrote:
> > madvise(MADV_DONTNEED) ends up calling zap_page_range() to clear page
> > tables associated with the address range.  For hugetlb vmas,
> > zap_page_range will call __unmap_hugepage_range_final.  However,
> > __unmap_hugepage_range_final assumes the passed vma is about to be removed
> > and deletes the vma_lock to prevent pmd sharing as the vma is on the way
> > out.  In the case of madvise(MADV_DONTNEED) the vma remains, but the
> > missing vma_lock prevents pmd sharing and could potentially lead to issues
> > with truncation/fault races.
> > 
> > This issue was originally reported here [1] as a BUG triggered in
> > page_try_dup_anon_rmap.  Prior to the introduction of the hugetlb
> > vma_lock, __unmap_hugepage_range_final cleared the VM_MAYSHARE flag to
> > prevent pmd sharing.  Subsequent faults on this vma were confused as
> > VM_MAYSHARE indicates a sharable vma, but was not set so page_mapping was
> > not set in new pages added to the page table.  This resulted in pages that
> > appeared anonymous in a VM_SHARED vma and triggered the BUG.
> > 
> > Address issue by:
> > - Add a new zap flag ZAP_FLAG_UNMAP to indicate an unmap call from
> >   unmap_vmas().  This is used to indicate the 'final' unmapping of a vma.
> >   When called via MADV_DONTNEED, this flag is not set and the vm_lock is
> >   not deleted.
> > - mmu notification is removed from __unmap_hugepage_range to avoid
> >   duplication, and notification is added to the other calling routine
> >   (unmap_hugepage_range).
> > - notification range in zap_page_range_single is updated to take into
> >   account the possibility of hugetlb pmd sharing.
> > - zap_page_range_single renamed to __zap_page_range_single to be used
> >   internally within mm/memory.c
> > - zap_vma_range() interface created to zap a range within a single vma.
> > - madvise_dontneed_single_vma is updated to call zap_vma_range instead of
> >   zap_page_range as it only operates on a range within a single vma
> > 
> > [1] https://lore.kernel.org/lkml/CAO4mrfdLMXsao9RF4fUE8-Wfde8xmjsKrTNMNC9wjUb6JudD0g@mail.gmail.com/
> > Fixes: 90e7e7f5ef3f ("mm: enable MADV_DONTNEED for hugetlb mappings")
> > Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
> > Reported-by: Wei Chen <harperchen1110@gmail.com>
> > Cc: <stable@vger.kernel.org>
> > ---
> >  include/linux/mm.h |  5 +++++
> >  mm/hugetlb.c       | 45 +++++++++++++++++++++++++++------------------
> >  mm/madvise.c       |  4 ++--
> >  mm/memory.c        | 17 +++++++++++++----
> >  4 files changed, 47 insertions(+), 24 deletions(-)
> > 
> > diff --git a/include/linux/mm.h b/include/linux/mm.h
> > index 978c17df053e..d205bcd9cd2e 100644
> > --- a/include/linux/mm.h
> > +++ b/include/linux/mm.h
> > @@ -1840,6 +1840,8 @@ void zap_vma_ptes(struct vm_area_struct *vma, unsigned long address,
> >  		  unsigned long size);
> >  void zap_page_range(struct vm_area_struct *vma, unsigned long address,
> >  		    unsigned long size);
> > +void zap_vma_range(struct vm_area_struct *vma, unsigned long address,
> > +		    unsigned long size);
> >  void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt,
> >  		struct vm_area_struct *start_vma, unsigned long start,
> >  		unsigned long end);
> > @@ -3464,4 +3466,7 @@ madvise_set_anon_name(struct mm_struct *mm, unsigned long start,
> >   */
> >  #define  ZAP_FLAG_DROP_MARKER        ((__force zap_flags_t) BIT(0))
> >  
> > +/* Set in unmap_vmas() to indicate an unmap call.  Only used by hugetlb */
> > +#define  ZAP_FLAG_UNMAP              ((__force zap_flags_t) BIT(1))
> 
> It seems this is not set anywhere in the patch?

Correct.  Should be set in unmap_vmas.

> > +
> >  #endif /* _LINUX_MM_H */
> > diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> > index ceb47c4e183a..7c8fbce4441e 100644
> > --- a/mm/hugetlb.c
> > +++ b/mm/hugetlb.c
> > @@ -5072,7 +5072,6 @@ static void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct
> >  	struct page *page;
> >  	struct hstate *h = hstate_vma(vma);
> >  	unsigned long sz = huge_page_size(h);
> > -	struct mmu_notifier_range range;
> >  	unsigned long last_addr_mask;
> >  	bool force_flush = false;
> >  
> > @@ -5087,13 +5086,6 @@ static void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct
> >  	tlb_change_page_size(tlb, sz);
> >  	tlb_start_vma(tlb, vma);
> >  
> > -	/*
> > -	 * If sharing possible, alert mmu notifiers of worst case.
> > -	 */
> > -	mmu_notifier_range_init(&range, MMU_NOTIFY_UNMAP, 0, vma, mm, start,
> > -				end);
> > -	adjust_range_if_pmd_sharing_possible(vma, &range.start, &range.end);
> > -	mmu_notifier_invalidate_range_start(&range);
> >  	last_addr_mask = hugetlb_mask_last_page(h);
> >  	address = start;
> >  	for (; address < end; address += sz) {
> > @@ -5178,7 +5170,6 @@ static void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct
> >  		if (ref_page)
> >  			break;
> >  	}
> > -	mmu_notifier_invalidate_range_end(&range);
> >  	tlb_end_vma(tlb, vma);
> >  
> >  	/*
> > @@ -5203,32 +5194,50 @@ void __unmap_hugepage_range_final(struct mmu_gather *tlb,
> >  			  unsigned long end, struct page *ref_page,
> >  			  zap_flags_t zap_flags)
> >  {
> > +	bool final = zap_flags & ZAP_FLAG_UNMAP;
> > +
> >  	hugetlb_vma_lock_write(vma);
> >  	i_mmap_lock_write(vma->vm_file->f_mapping);
> >  
> >  	__unmap_hugepage_range(tlb, vma, start, end, ref_page, zap_flags);
> >  
> >  	/*
> > -	 * Unlock and free the vma lock before releasing i_mmap_rwsem.  When
> > -	 * the vma_lock is freed, this makes the vma ineligible for pmd
> > -	 * sharing.  And, i_mmap_rwsem is required to set up pmd sharing.
> > -	 * This is important as page tables for this unmapped range will
> > -	 * be asynchrously deleted.  If the page tables are shared, there
> > -	 * will be issues when accessed by someone else.
> > +	 * When called via zap_vma_range (MADV_DONTNEED), this is not the
> > +	 * final unmap of the vma, and we do not want to delete the vma_lock.
> >  	 */
> > -	__hugetlb_vma_unlock_write_free(vma);
> > -
> > -	i_mmap_unlock_write(vma->vm_file->f_mapping);
> > +	if (final) {
> > +		/*
> > +		 * Unlock and free the vma lock before releasing i_mmap_rwsem.
> > +		 * When the vma_lock is freed, this makes the vma ineligible
> > +		 * for pmd sharing.  And, i_mmap_rwsem is required to set up
> > +		 * pmd sharing.  This is important as page tables for this
> > +		 * unmapped range will be asynchrously deleted.  If the page
> > +		 * tables are shared, there will be issues when accessed by
> > +		 * someone else.
> > +		 */
> > +		__hugetlb_vma_unlock_write_free(vma);
> > +		i_mmap_unlock_write(vma->vm_file->f_mapping);
> > +	} else {
> > +		i_mmap_unlock_write(vma->vm_file->f_mapping);
> > +		hugetlb_vma_unlock_write(vma);
> > +	}
> >  }
> >  
> >  void unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start,
> >  			  unsigned long end, struct page *ref_page,
> >  			  zap_flags_t zap_flags)
> >  {
> > +	struct mmu_notifier_range range;
> >  	struct mmu_gather tlb;
> >  
> > +	mmu_notifier_range_init(&range, MMU_NOTIFY_UNMAP, 0, vma, vma->vm_mm,
> 
> Should this be s/UNMAP/CLEAR/?  As IIUC the unmap path was only happening
> in __unmap_hugepage_range_final().

Right, unmap_hugepage_range is employed in the truncate and hole punch
path where we should be using the CLEAR notifier.

> > +				start, end);
> > +	adjust_range_if_pmd_sharing_possible(vma, &range.start, &range.end);
> >  	tlb_gather_mmu(&tlb, vma->vm_mm);
> > +
> >  	__unmap_hugepage_range(&tlb, vma, start, end, ref_page, zap_flags);
> > +
> > +	mmu_notifier_invalidate_range_end(&range);
> >  	tlb_finish_mmu(&tlb);
> >  }
> >  
> > diff --git a/mm/madvise.c b/mm/madvise.c
> > index c7105ec6d08c..9d2625b8029a 100644
> > --- a/mm/madvise.c
> > +++ b/mm/madvise.c
> > @@ -772,7 +772,7 @@ static int madvise_free_single_vma(struct vm_area_struct *vma,
> >   * Application no longer needs these pages.  If the pages are dirty,
> >   * it's OK to just throw them away.  The app will be more careful about
> >   * data it wants to keep.  Be sure to free swap resources too.  The
> > - * zap_page_range call sets things up for shrink_active_list to actually free
> > + * zap_vma_range call sets things up for shrink_active_list to actually free
> >   * these pages later if no one else has touched them in the meantime,
> >   * although we could add these pages to a global reuse list for
> >   * shrink_active_list to pick up before reclaiming other pages.
> > @@ -790,7 +790,7 @@ static int madvise_free_single_vma(struct vm_area_struct *vma,
> >  static long madvise_dontneed_single_vma(struct vm_area_struct *vma,
> >  					unsigned long start, unsigned long end)
> >  {
> > -	zap_page_range(vma, start, end - start);
> > +	zap_vma_range(vma, start, end - start);
> 
> I'd rather just call zap_page_range_single() directly with NULL passed
> over, considering that this is for stable, but no strong opinions.

As mentioned in reply to Nadav, if we expose zap_page_range_single we
need to expose 'struct zap_details'.  Any concerns there?  That was the
primary reason I did not just call zap_page_range_single directly.
-- 
Mike Kravetz