From nobody Thu Apr 2 13:15:13 2026 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 9FDC1C32771 for ; Sat, 24 Sep 2022 03:15:47 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233032AbiIXDPn (ORCPT ); Fri, 23 Sep 2022 23:15:43 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:57986 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231833AbiIXDPf (ORCPT ); Fri, 23 Sep 2022 23:15:35 -0400 Received: from szxga08-in.huawei.com (szxga08-in.huawei.com [45.249.212.255]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id B27D6D12C9 for ; Fri, 23 Sep 2022 20:15:30 -0700 (PDT) Received: from dggpemm500024.china.huawei.com (unknown [172.30.72.55]) by szxga08-in.huawei.com (SkyGuard) with ESMTP id 4MZDYy3qNcz1P6ln; Sat, 24 Sep 2022 11:11:18 +0800 (CST) Received: from dggpemm100009.china.huawei.com (7.185.36.113) by dggpemm500024.china.huawei.com (7.185.36.203) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.31; Sat, 24 Sep 2022 11:15:28 +0800 Received: from huawei.com (10.175.113.32) by dggpemm100009.china.huawei.com (7.185.36.113) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.31; Sat, 24 Sep 2022 11:15:28 +0800 From: Liu Shixin To: Liu Zixian , Mike Kravetz , Muchun Song , "Andrew Morton" , Sidhartha Kumar , John Hubbard , "David Hildenbrand" , Kefeng Wang CC: , , Liu Shixin Subject: [PATCH v4] mm: hugetlb: fix UAF in hugetlb_handle_userfault Date: Sat, 24 Sep 2022 11:49:05 +0800 Message-ID: <20220924034905.2694686-1-liushixin2@huawei.com> X-Mailer: git-send-email 2.25.1 MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-Originating-IP: [10.175.113.32] X-ClientProxiedBy: dggems704-chm.china.huawei.com (10.3.19.181) To dggpemm100009.china.huawei.com (7.185.36.113) X-CFilter-Loop: Reflected Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Type: text/plain; charset="utf-8" The vma_lock and hugetlb_fault_mutex are dropped before handling userfault and reacquire them again after handle_userfault(), but reacquire the vma_lock could lead to UAF[1,2] due to the following race, hugetlb_fault hugetlb_no_page /*unlock vma_lock */ hugetlb_handle_userfault handle_userfault /* unlock mm->mmap_lock*/ vm_mmap_pgoff do_mmap mmap_region munmap_vma_range /* clean old vma */ /* lock vma_lock again <--- UAF */ /* unlock vma_lock */ Since the vma_lock will unlock immediately after hugetlb_handle_userfault(), let's drop the unneeded lock and unlock in hugetlb_handle_userfault() to fix the issue. [1] https://lore.kernel.org/linux-mm/000000000000d5e00a05e834962e@google.co= m/ [2] https://lore.kernel.org/linux-mm/20220921014457.1668-1-liuzixian4@huawe= i.com/ Reported-by: syzbot+193f9cee8638750b23cf@syzkaller.appspotmail.com Reported-by: Liu Zixian Fixes: 1a1aad8a9b7b ("userfaultfd: hugetlbfs: add userfaultfd hugetlb hook") CC: stable@vger.kernel.org # 4.14+ Signed-off-by: Liu Shixin Signed-off-by: Kefeng Wang Reviewed-by: Mike Kravetz --- v1->v2: add reported-by and link [1]. v2->v3: add comment to explain why unlock in hugetlb_no_page. v3->v4: Rebase on mainline. mm/hugetlb.c | 37 +++++++++++++++++-------------------- 1 file changed, 17 insertions(+), 20 deletions(-) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index e070b8593b37..477e3c2369d2 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -5476,7 +5476,6 @@ static inline vm_fault_t hugetlb_handle_userfault(str= uct vm_area_struct *vma, unsigned long addr, unsigned long reason) { - vm_fault_t ret; u32 hash; struct vm_fault vmf =3D { .vma =3D vma, @@ -5494,18 +5493,14 @@ static inline vm_fault_t hugetlb_handle_userfault(s= truct vm_area_struct *vma, }; =20 /* - * hugetlb_fault_mutex and i_mmap_rwsem must be - * dropped before handling userfault. Reacquire - * after handling fault to make calling code simpler. + * vma_lock and hugetlb_fault_mutex must be dropped before handling + * userfault. Also mmap_lock will be dropped during handling + * userfault, any vma operation should be careful from here. */ hash =3D hugetlb_fault_mutex_hash(mapping, idx); mutex_unlock(&hugetlb_fault_mutex_table[hash]); i_mmap_unlock_read(mapping); - ret =3D handle_userfault(&vmf, reason); - i_mmap_lock_read(mapping); - mutex_lock(&hugetlb_fault_mutex_table[hash]); - - return ret; + return handle_userfault(&vmf, reason); } =20 static vm_fault_t hugetlb_no_page(struct mm_struct *mm, @@ -5523,6 +5518,7 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *m= m, spinlock_t *ptl; unsigned long haddr =3D address & huge_page_mask(h); bool new_page, new_pagecache_page =3D false; + u32 hash =3D hugetlb_fault_mutex_hash(mapping, idx); =20 /* * Currently, we are forced to kill the process in the event the @@ -5533,7 +5529,7 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *m= m, if (is_vma_resv_set(vma, HPAGE_RESV_UNMAPPED)) { pr_warn_ratelimited("PID %d killed due to inadequate hugepage pool\n", current->pid); - return ret; + goto out; } =20 /* @@ -5550,12 +5546,10 @@ static vm_fault_t hugetlb_no_page(struct mm_struct = *mm, page =3D find_lock_page(mapping, idx); if (!page) { /* Check for page in userfault range */ - if (userfaultfd_missing(vma)) { - ret =3D hugetlb_handle_userfault(vma, mapping, idx, + if (userfaultfd_missing(vma)) + return hugetlb_handle_userfault(vma, mapping, idx, flags, haddr, address, VM_UFFD_MISSING); - goto out; - } =20 page =3D alloc_huge_page(vma, haddr, 0); if (IS_ERR(page)) { @@ -5615,10 +5609,9 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *= mm, if (userfaultfd_minor(vma)) { unlock_page(page); put_page(page); - ret =3D hugetlb_handle_userfault(vma, mapping, idx, + return hugetlb_handle_userfault(vma, mapping, idx, flags, haddr, address, VM_UFFD_MINOR); - goto out; } } =20 @@ -5676,6 +5669,8 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *m= m, =20 unlock_page(page); out: + mutex_unlock(&hugetlb_fault_mutex_table[hash]); + i_mmap_unlock_read(mapping); return ret; =20 backout: @@ -5774,11 +5769,13 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm, stru= ct vm_area_struct *vma, =20 entry =3D huge_ptep_get(ptep); /* PTE markers should be handled the same way as none pte */ - if (huge_pte_none_mostly(entry)) { - ret =3D hugetlb_no_page(mm, vma, mapping, idx, address, ptep, + if (huge_pte_none_mostly(entry)) + /* + * hugetlb_no_page will drop vma lock and hugetlb fault + * mutex internally, which make us return immediately. + */ + return hugetlb_no_page(mm, vma, mapping, idx, address, ptep, entry, flags); - goto out_mutex; - } =20 ret =3D 0; =20 --=20 2.25.1