Re: [PATCH 1/5] mm, swap: Fix comment in __read_swap_cache_async
- Date: Mon, 20 Mar 2017 10:07:24 +0800
- From: "Huang\, Ying" <ying.huang@xxxxxxxxx>
- Subject: Re: [PATCH 1/5] mm, swap: Fix comment in __read_swap_cache_async
Rafael Aquini <aquini@xxxxxxxxxx> writes:
> On Fri, Mar 17, 2017 at 02:46:19PM +0800, Huang, Ying wrote:
>> From: Huang Ying <ying.huang@xxxxxxxxx>
>> The commit cbab0e4eec29 ("swap: avoid read_swap_cache_async() race to
>> deadlock while waiting on discard I/O completion") fixed a deadlock in
>> read_swap_cache_async(). Because at that time, in swap allocation
>> path, a swap entry may be set as SWAP_HAS_CACHE, then wait for
>> discarding to complete before the page for the swap entry is added to
>> the swap cache. But in the commit 815c2c543d3a ("swap: make swap
>> discard async"), the discarding for swap become asynchronous, waiting
>> for discarding to complete will be done before the swap entry is set
>> as SWAP_HAS_CACHE. So the comments in code is incorrect now. This
>> patch fixes the comments.
>> The cond_resched() added in the commit cbab0e4eec29 is not necessary
>> now too. But if we added some sleep in swap allocation path in the
>> future, there may be some hard to debug/reproduce deadlock bug. So it
>> is kept.
> ^ this is a rather disconcerting way to describe why you left that part
> behind, and I recollect telling you about it in a private discussion.
> The fact is that __read_swap_cache_async() still races against get_swap_page()
> with a way narrower window due to the async fashioned SSD wear leveling
> done for swap nowadays and other changes made within __read_swap_cache_async()'s
> while loop thus making that old deadlock scenario very improbable to strike again.
Thanks for your comments! Could you tell me which kind of race
> All seems legit, apart from that last paragraph in the commit log
> Acked-by: Rafael Aquini <aquini@xxxxxxxxxx>
>> Cc: Shaohua Li <shli@xxxxxxxxxx>
>> Cc: Rafael Aquini <aquini@xxxxxxxxxx>
>> Signed-off-by: "Huang, Ying" <ying.huang@xxxxxxxxx>
>> mm/swap_state.c | 12 +-----------
>> 1 file changed, 1 insertion(+), 11 deletions(-)
>> diff --git a/mm/swap_state.c b/mm/swap_state.c
>> index 473b71e052a8..7bfb9bd1ca21 100644
>> --- a/mm/swap_state.c
>> +++ b/mm/swap_state.c
>> @@ -360,17 +360,7 @@ struct page *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
>> * We might race against get_swap_page() and stumble
>> * across a SWAP_HAS_CACHE swap_map entry whose page
>> - * has not been brought into the swapcache yet, while
>> - * the other end is scheduled away waiting on discard
>> - * I/O completion at scan_swap_map().
>> - *
>> - * In order to avoid turning this transitory state
>> - * into a permanent loop around this -EEXIST case
>> - * if !CONFIG_PREEMPT and the I/O completion happens
>> - * to be waiting on the CPU waitqueue where we are now
>> - * busy looping, we just conditionally invoke the
>> - * scheduler here, if there are some more important
>> - * tasks to run.
>> + * has not been brought into the swapcache yet.