mm: hugetlb: fix UAF in hugetlb_handle_userfault

commit 958f32ce83 upstream.

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.com/
[2] https://lore.kernel.org/linux-mm/20220921014457.1668-1-liuzixian4@huawei.com/
Link: https://lkml.kernel.org/r/20220923042113.137273-1-liushixin2@huawei.com
Fixes: 1a1aad8a9b ("userfaultfd: hugetlbfs: add userfaultfd hugetlb hook")
Signed-off-by: Liu Shixin <liushixin2@huawei.com>
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Reported-by: syzbot+193f9cee8638750b23cf@syzkaller.appspotmail.com
Reported-by: Liu Zixian <liuzixian4@huawei.com>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Muchun Song <songmuchun@bytedance.com>
Cc: Sidhartha Kumar <sidhartha.kumar@oracle.com>
Cc: <stable@vger.kernel.org>	[4.14+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit is contained in:
Liu Shixin 2022-09-23 12:21:13 +08:00 committed by Greg Kroah-Hartman
parent c378c479c5
commit 45c3396675

View file

@ -4337,6 +4337,7 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *mm,
spinlock_t *ptl; spinlock_t *ptl;
unsigned long haddr = address & huge_page_mask(h); unsigned long haddr = address & huge_page_mask(h);
bool new_page = false; bool new_page = false;
u32 hash = hugetlb_fault_mutex_hash(mapping, idx);
/* /*
* Currently, we are forced to kill the process in the event the * Currently, we are forced to kill the process in the event the
@ -4346,7 +4347,7 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *mm,
if (is_vma_resv_set(vma, HPAGE_RESV_UNMAPPED)) { if (is_vma_resv_set(vma, HPAGE_RESV_UNMAPPED)) {
pr_warn_ratelimited("PID %d killed due to inadequate hugepage pool\n", pr_warn_ratelimited("PID %d killed due to inadequate hugepage pool\n",
current->pid); current->pid);
return ret; goto out;
} }
/* /*
@ -4365,7 +4366,6 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *mm,
* Check for page in userfault range * Check for page in userfault range
*/ */
if (userfaultfd_missing(vma)) { if (userfaultfd_missing(vma)) {
u32 hash;
struct vm_fault vmf = { struct vm_fault vmf = {
.vma = vma, .vma = vma,
.address = haddr, .address = haddr,
@ -4380,17 +4380,14 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *mm,
}; };
/* /*
* hugetlb_fault_mutex and i_mmap_rwsem must be * vma_lock and hugetlb_fault_mutex must be dropped
* dropped before handling userfault. Reacquire * before handling userfault. Also mmap_lock will
* after handling fault to make calling code simpler. * be dropped during handling userfault, any vma
* operation should be careful from here.
*/ */
hash = hugetlb_fault_mutex_hash(mapping, idx);
mutex_unlock(&hugetlb_fault_mutex_table[hash]); mutex_unlock(&hugetlb_fault_mutex_table[hash]);
i_mmap_unlock_read(mapping); i_mmap_unlock_read(mapping);
ret = handle_userfault(&vmf, VM_UFFD_MISSING); return handle_userfault(&vmf, VM_UFFD_MISSING);
i_mmap_lock_read(mapping);
mutex_lock(&hugetlb_fault_mutex_table[hash]);
goto out;
} }
page = alloc_huge_page(vma, haddr, 0); page = alloc_huge_page(vma, haddr, 0);
@ -4497,6 +4494,8 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *mm,
unlock_page(page); unlock_page(page);
out: out:
mutex_unlock(&hugetlb_fault_mutex_table[hash]);
i_mmap_unlock_read(mapping);
return ret; return ret;
backout: backout:
@ -4592,10 +4591,12 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma,
mutex_lock(&hugetlb_fault_mutex_table[hash]); mutex_lock(&hugetlb_fault_mutex_table[hash]);
entry = huge_ptep_get(ptep); entry = huge_ptep_get(ptep);
if (huge_pte_none(entry)) { if (huge_pte_none(entry))
ret = hugetlb_no_page(mm, vma, mapping, idx, address, ptep, flags); /*
goto out_mutex; * 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, flags);
ret = 0; ret = 0;