Re: [PATCH v2] mm/hugetlb: fix DEBUG_LOCKS_WARN_ON(1) when dissolve_free_hugetlb_folio()

[Andrew Morton]

On Fri, 19 Apr 2024 16:58:19 +0800 Miaohe Lin <linmiaohe@xxxxxxxxxx> wrote:

> When I did memory failure tests recently, below warning occurs:
> 
> DEBUG_LOCKS_WARN_ON(1)
> WARNING: CPU: 8 PID: 1011 at kernel/locking/lockdep.c:232 __lock_acquire+0xccb/0x1ca0
> Modules linked in: mce_inject hwpoison_inject
> CPU: 8 PID: 1011 Comm: bash Kdump: loaded Not tainted 6.9.0-rc3-next-20240410-00012-gdb69f219f4be #3
>
> ...
>
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -1773,7 +1773,7 @@ static void __update_and_free_hugetlb_folio(struct hstate *h,
>  	 * If vmemmap pages were allocated above, then we need to clear the
>  	 * hugetlb flag under the hugetlb lock.
>  	 */
> -	if (clear_flag) {
> +	if (folio_test_hugetlb(folio)) {
>  		spin_lock_irq(&hugetlb_lock);
>  		__folio_clear_hugetlb(folio);
>  		spin_unlock_irq(&hugetlb_lock);

Please let's prepare backportable fixes against current mainline, not
mm-unstable.  Because fixes against current -rcX and earlier will be
upstreamed ahead of the mm-unstable and mm-stable material.


I did this:

--- a/mm/hugetlb.c~mm-hugetlb-fix-debug_locks_warn_on1-when-dissolve_free_hugetlb_folio
+++ a/mm/hugetlb.c
@@ -1781,7 +1781,7 @@ static void __update_and_free_hugetlb_fo
 	 * If vmemmap pages were allocated above, then we need to clear the
 	 * hugetlb destructor under the hugetlb lock.
 	 */
-	if (clear_dtor) {
+	if (folio_test_hugetlb(folio)) {
 		spin_lock_irq(&hugetlb_lock);
 		__clear_hugetlb_destructor(h, folio);
 		spin_unlock_irq(&hugetlb_lock);
_