Re: [PATCH] mm/swap: fix swap_cache_get_folio() returning a folio that left the swap cache

[Jiayuan Chen <jiayuan.chen@linux.dev>]

Hi Andrew,

On 10/16/26 10:06 AM, Andrew Morton wrote:
> On Thu, 15 Oct 2026 14:05:27 +0800 Jiayuan Chen <jiayuan.chen@linux.dev> wrote:
>
>> From: Jiayuan Chen <jiayuan.chen@shopee.com>
>>
>> swap_cache_get_folio() reads the folio from the swap table and takes a
>> reference with folio_try_get(), without any lock. If the folio is
>> removed from the swap cache and freed in between, and its memory gets
>> reused, folio_try_get() still succeeds and we return a page that has
>> nothing to do with the swap entry:
>>
>>    CPU 0                                  CPU 1
>>    do_swap_page()
>>      swap_cache_get_folio()
>>        swp_tb = swap_table_get()
>>        folio = swp_tb_to_folio(swp_tb)
>>                                           madvise(MADV_DONTNEED)
>>                                             zap_nonpresent_ptes()
>>                                               swap_put_entries_direct()
>>                                                 swap_put_entries_cluster()
>>                                                   __try_to_reclaim_swap()
>>                                                     swap_cache_del_folio()
>>                                                     folio_put()   <- freed
>>                                           page reused, e.g. as an anon
>>                                           folio of another task
>>        folio_try_get(folio)               <- succeeds
>>        return folio
>>      swap_update_readahead()              <- may clear PG_readahead on it
>>      folio_lock_or_retry()                <- locks it
>>      folio_matches_swap_entry() fails, unlock and put
>>
>> filemap_get_entry() does the same lockless lookup for the page cache,
>> and after folio_try_get() it rechecks the xarray, because the folio may
>> have been freed and reused in between. swap_cache_get_folio() lost
>> that recheck when the swap cache moved to the swap table.
>>
>> So CPU 0 writes to a folio it doesn't own: it may clear PG_readahead and
>> takes the folio lock of someone else's page, before
>> folio_matches_swap_entry() finds out the folio is not ours.
>>
>> Recheck the slot after folio_try_get() and retry if it no longer holds
>> this folio.
> LLM says to me:
>
> : The original folio could conceivably be freed, its physical page
> : reused, and the same PFN inserted into the same swap-table slot before
> : the second lookup.  That would pass the check.  However, the returned
> : reference would then belong to the newly inserted folio, and the later
> : swap-entry validation would still be required.  I don't see that as
> : invalidating this fix.



I had the same worry at first, data corruption or something worse, but
after thinking it through I agree. Passing the check there is fine:
the folio we return is the one currently in the slot, which is what
the function promises, and callers still check folio_matches_swap_entry()
under the folio lock before using it. That's why I didn't mention it in 
the changelog.


>> --- a/mm/swap_state.c
>> +++ b/mm/swap_state.c
>> @@ -92,17 +92,27 @@ void show_swap_cache_info(void)
>>    */
>>   struct folio *swap_cache_get_folio(swp_entry_t entry)
>>   {
>> +	struct swap_cluster_info *ci = __swap_entry_to_cluster(entry);
>> +	unsigned int ci_off = swp_cluster_offset(entry);
>>   	unsigned long swp_tb;
>>   	struct folio *folio;
>>   
>>   	for (;;) {
>> -		swp_tb = swap_table_get(__swap_entry_to_cluster(entry),
>> -					swp_cluster_offset(entry));
>> +		swp_tb = swap_table_get(ci, ci_off);
>>   		if (!swp_tb_is_folio(swp_tb))
>>   			return NULL;
>>   		folio = swp_tb_to_folio(swp_tb);
>> -		if (likely(folio_try_get(folio)))
>> +		if (!folio_try_get(folio))
>> +			continue;
>> +		/*
>> +		 * The folio may have been freed and its memory reused before
>> +		 * we got the reference. Make sure it is still in the slot.
>> +		 */
>> +		swp_tb = swap_table_get(ci, ci_off);
>> +		if (likely(swp_tb_is_folio(swp_tb) &&
>> +			   swp_tb_to_folio(swp_tb) == folio))
>>   			return folio;
>> +		folio_put(folio);
>>   	}
>>   
>>   	return NULL;
> This return is now unreachable...


Right, and it never was reachable, the loop has no break. It should have 
been dropped earlier...