Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <db5797a5-c4e9-753d-790a-3d432ac525e3@redhat.com>
Date:   Thu, 5 Jan 2023 09:39:34 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.6.0
Subject: Re: [PATCH 1/3] mm/hugetlb: Pre-allocate pgtable pages for uffd
 wr-protects
Content-Language: en-US
To:     Peter Xu <peterx@redhat.com>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Cc:     Mike Kravetz <mike.kravetz@oracle.com>,
        Muchun Song <songmuchun@bytedance.com>,
        Nadav Amit <nadav.amit@gmail.com>,
        Andrea Arcangeli <aarcange@redhat.com>,
        James Houghton <jthoughton@google.com>,
        Axel Rasmussen <axelrasmussen@google.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-stable <stable@vger.kernel.org>
References: <20230104225207.1066932-1-peterx@redhat.com>
 <20230104225207.1066932-2-peterx@redhat.com>
From:   David Hildenbrand <david@redhat.com>
Organization: Red Hat
In-Reply-To: <20230104225207.1066932-2-peterx@redhat.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 04.01.23 23:52, Peter Xu wrote:
> Userfaultfd-wp uses pte markers to mark wr-protected pages for both shmem
> and hugetlb.  Shmem has pre-allocation ready for markers, but hugetlb path
> was overlooked.
> 
> Doing so by calling huge_pte_alloc() if the initial pgtable walk fails to
> find the huge ptep.  It's possible that huge_pte_alloc() can fail with high
> memory pressure, in that case stop the loop immediately and fail silently.
> This is not the most ideal solution but it matches with what we do with
> shmem meanwhile it avoids the splat in dmesg.
> 
> Cc: linux-stable <stable@vger.kernel.org> # 5.19+
> Fixes: 60dfaad65aa9 ("mm/hugetlb: allow uffd wr-protect none ptes")
> Reported-by: James Houghton <jthoughton@google.com>
> Signed-off-by: Peter Xu <peterx@redhat.com>
> ---
>   mm/hugetlb.c | 13 +++++++++++--
>   1 file changed, 11 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> index bf7a1f628357..017d9159cddf 100644
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -6649,8 +6649,17 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma,
>   		spinlock_t *ptl;
>   		ptep = hugetlb_walk(vma, address, psize);

if (!ptep && likely(!uffd_wp)) {
	/* Nothing to protect. */
	address |= last_addr_mask;
	continue;
} else if (!ptep) {
	...
}

Might look slightly more readable would minimize changes. This should 
work, so

Acked-by: David Hildenbrand <david@redhat.com>

>   		if (!ptep) {
> -			address |= last_addr_mask;
> -			continue;
> +			if (!uffd_wp) {
> +				address |= last_addr_mask;
> +				continue;
> +			}
> +			/*
> +			 * Userfaultfd wr-protect requires pgtable
> +			 * pre-allocations to install pte markers.
> +			 */
> +			ptep = huge_pte_alloc(mm, vma, address, psize);
> +			if (!ptep)
> +				break;
>   		}
>   		ptl = huge_pte_lock(h, mm, ptep);
>   		if (huge_pmd_unshare(mm, vma, address, ptep)) {

-- 
Thanks,

David / dhildenb