Re: [PATCH] mm: Fix the race between collapse and PT_RECLAIM under per-vma lock

[Baolin Wang <baolin.wang@linux.alibaba.com>]

On 2025/8/5 17:50, David Hildenbrand wrote:
> On 05.08.25 11:30, Qi Zheng wrote:
>>
>>
>> On 8/5/25 4:56 PM, Baolin Wang wrote:
>>>
>>>
>>> On 2025/8/5 16:17, Qi Zheng wrote:
>>>> Hi Baolin,
>>>>
>>>> On 8/5/25 3:53 PM, Baolin Wang wrote:
>>>>>
>>>>>
>>>>> On 2025/8/5 14:42, Qi Zheng wrote:
>>>>>> Hi Barry,
>>>>>>
>>>>>> On 8/5/25 11:54 AM, Barry Song wrote:
>>>>>>> From: Barry Song <v-songbaohua@oppo.com>
>>>>>>>
>>>>>>> The check_pmd_still_valid() call during collapse is currently only
>>>>>>> protected by the mmap_lock in write mode, which was sufficient when
>>>>>>> pt_reclaim always ran under mmap_lock in read mode. However, since
>>>>>>> madvise_dontneed can now execute under a per-VMA lock, this 
>>>>>>> assumption
>>>>>>> is no longer valid. As a result, a race condition can occur between
>>>>>>> collapse and PT_RECLAIM, potentially leading to a kernel panic.
>>>>>>
>>>>>> There is indeed a race condition here. And after applying this 
>>>>>> patch, I
>>>>>> can no longer reproduce the problem locally (I was able to 
>>>>>> reproduce it
>>>>>> stably locally last night).
>>>>>>
>>>>>> But I still can't figure out how this race condtion causes the
>>>>>> following panic:
>>>>>>
>>>>>> exit_mmap
>>>>>> --> mmap_read_lock()
>>>>>>       unmap_vmas()
>>>>>>       --> pte_offset_map_lock
>>>>>>           --> rcu_read_lock()
>>>>>>               check if the pmd entry is a PTE page
>>>>>>               ptl = pte_lockptr(mm, &pmdval)  <-- ptl is NULL
>>>>>>               spin_lock(ptl)                  <-- PANIC!!
>>>>>>
>>>>>> If this PTE page is freed by pt_reclaim (via RCU), then the ptl can
>>>>>> not be NULL.
>>>>>>
>>>>>> The collapse holds mmap write lock, so it is impossible to be
>>>>>> concurrent
>>>>>> with exit_mmap().
>>>>>>
>>>>>> Confusing. :(
>>>>>
>>>>> IIUC, the issue is not caused by the concurrency between exit_mmap
>>>>> and collapse, but rather by the concurrency between pt_reclaim and
>>>>> collapse.
>>>>>
>>>>> Before this patch, khugepaged might incorrectly restore a PTE
>>>>> pagetable that had already been freed.
>>>>>
>>>>> pt_reclaim has cleared the pmd entry and freed the PTE page table.
>>>>> However, due to the race condition, check_pmd_still_valid() still
>>>>> passes and continues to attempt the collapse:
>>>>>
>>>>> _pmd = pmdp_collapse_flush(vma, address, pmd); ---> returns a none
>>>>> pmd entry (the original pmd entry has been cleared)
>>>>>
>>>>> pte = pte_offset_map_lock(mm, &_pmd, address, &pte_ptl); ---> returns
>>>>> pte == NULL
>>>>>
>>>>> Then khugepaged will restore the old PTE pagetable with an invalid
>>>>> pmd entry:
>>>>>
>>>>> pmd_populate(mm, pmd, pmd_pgtable(_pmd));
>>>>>
>>>>> So when the process exits and trys to free the mapping of the
>>>>> process, traversing the invalid pmd table will lead to a crash.
>>>>
>>>> CPU0                         CPU1
>>>> ====                         ====
>>>>
>>>> collapse
>>>> --> pmd_populate(mm, pmd, pmd_pgtable(_pmd));
>>>>       mmap_write_unlock
>>>>                                exit_mmap
>>>>                                --> hold mmap lock
>>>>                                    __pte_offset_map_lock
>>>>                                    --> pte = __pte_offset_map(pmd,
>>>> addr, &pmdval);
>>>>                                        if (unlikely(!pte))
>>>>                                            return pte;   <-- will 
>>>> return
>>>
>>> __pte_offset_map() might not return NULL? Because the 'pmd_populate(mm,
>>> pmd, pmd_pgtable(_pmd))' could populate a valid page (although the
>>> '_pmd' entry is NONE), but it is not the original pagetable page.
>>
>> CPU0                          CPU1
>> ====                          ====
>>
>> collapse
>> --> check_pmd_still_valid
>>                                 vma read lock
>>                                 pt_reclaim clear the pmd entry and will
>> free the PTE page (via RCU)
>>                                 vma read unlock
>>
>>       vma write lock
>>       _pmd = pmdp_collapse_flush(vma, address, pmd) <-- pmd_none(_pmd)
>>       pte = pte_offset_map_lock(mm, &_pmd, address, &pte_ptl); <-- pte is
>> NULL
>>       pmd_populate(mm, pmd, pmd_pgtable(_pmd)); <-- populate a valid 
>> page?
>>       vma write unlock
>>
>> The above is the concurrent scenario you mentioned, right?

Yes.

>>
>> What types of this 'valid page' could be? If __pte_offset_map() returns
>> non-NULL, then it is a PTE page. Even if it is not the original one, it
>> should not cause panic. Did I miss some key information? :(

Sorry for not being clear. Let me try again.

In the race condition described above, the '_pmd' value is NONE, meaning 
that when restoring the pmd entry with ‘pmd_populate(mm, pmd, 
pmd_pgtable(_pmd))’, the 'pmd_pgtable(_pmd)' can return a struct page 
corresponding to pfn == 0 (cause the '_pmd' is NONE) to populate the pmd 
entry. Clearly, this pfn == 0 page is not a pagetable page, meaning the 
corresponding ptl lock of this page is not initialized.

Additionally, from the boot dmesg, I can see that the BIOS reports an 
address range with pfn == 0, indicating that there is a struct page 
initialized for pfn == 0 (possibly a reserved page):

[    0.000000] BIOS-provided physical RAM map:
[    0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009fbff] usable
[    0.000000] BIOS-e820: [mem 0x000000000009fc00-0x000000000009ffff] 
reserved
[    0.000000] BIOS-e820: [mem 0x00000000000f0000-0x00000000000fffff] 
reserved
[    0.000000] BIOS-e820: [mem 0x0000000000100000-0x000000007ffdffff] usable
[    0.000000] BIOS-e820: [mem 0x000000007ffe0000-0x000000007fffffff] 
reserved
[    0.000000] BIOS-e820: [mem 0x00000000feffc000-0x00000000feffffff] 
reserved
[    0.000000] BIOS-e820: [mem 0x00000000fffc0000-0x00000000ffffffff] 
reserved

Of course, this is my theoretical analysis from the code perspective. If 
there are other race conditions, I would be very surprised:)

> Wasn't the original issue all about a NULL-pointer de-reference while 
> *locking*?

Yes.

> Note that in that kernel config [1] we have CONFIG_DEBUG_SPINLOCK=y, so 
> likely we will have ALLOC_SPLIT_PTLOCKS set.
> 
> [1] https://github.com/laifryiee/syzkaller_logs/blob/ 
> main/250803_193026___pte_offset_map_lock/.config
>