Re: [RFC 07/11] mm: remove SWAP_AGAIN in ttu

2017-03-05 Thread Minchan Kim
On Fri, Mar 03, 2017 at 06:24:06PM +0530, Anshuman Khandual wrote: > On 03/02/2017 12:09 PM, Minchan Kim wrote: > > In 2002, [1] introduced SWAP_AGAIN. > > At that time, ttuo used spin_trylock(>page_table_lock) so it's > > Small nit: Please expand "ttuo" here. TTU in the first place is also > not

Re: [RFC 07/11] mm: remove SWAP_AGAIN in ttu

2017-03-05 Thread Minchan Kim
On Fri, Mar 03, 2017 at 06:24:06PM +0530, Anshuman Khandual wrote: > On 03/02/2017 12:09 PM, Minchan Kim wrote: > > In 2002, [1] introduced SWAP_AGAIN. > > At that time, ttuo used spin_trylock(>page_table_lock) so it's > > Small nit: Please expand "ttuo" here. TTU in the first place is also > not

Re: [RFC 07/11] mm: remove SWAP_AGAIN in ttu

2017-03-03 Thread Anshuman Khandual
On 03/02/2017 12:09 PM, Minchan Kim wrote: > In 2002, [1] introduced SWAP_AGAIN. > At that time, ttuo used spin_trylock(>page_table_lock) so it's Small nit: Please expand "ttuo" here. TTU in the first place is also not very clear but we have that in many places. > really easy to contend and fail

Re: [RFC 07/11] mm: remove SWAP_AGAIN in ttu

2017-03-03 Thread Anshuman Khandual
On 03/02/2017 12:09 PM, Minchan Kim wrote: > In 2002, [1] introduced SWAP_AGAIN. > At that time, ttuo used spin_trylock(>page_table_lock) so it's Small nit: Please expand "ttuo" here. TTU in the first place is also not very clear but we have that in many places. > really easy to contend and fail

[RFC 07/11] mm: remove SWAP_AGAIN in ttu

2017-03-01 Thread Minchan Kim
In 2002, [1] introduced SWAP_AGAIN. At that time, ttuo used spin_trylock(>page_table_lock) so it's really easy to contend and fail to hold a lock so SWAP_AGAIN to keep LRU status makes sense. However, now we changed it to mutex-based lock and be able to block without skip pte so there is a few of

[RFC 07/11] mm: remove SWAP_AGAIN in ttu

2017-03-01 Thread Minchan Kim
In 2002, [1] introduced SWAP_AGAIN. At that time, ttuo used spin_trylock(>page_table_lock) so it's really easy to contend and fail to hold a lock so SWAP_AGAIN to keep LRU status makes sense. However, now we changed it to mutex-based lock and be able to block without skip pte so there is a few of