Re: [f2fs-dev] [PATCH] f2fs: split bio cache

2017-05-12 Thread Jaegeuk Kim
On 05/12, Chao Yu wrote: > On 2017/5/11 10:35, Jaegeuk Kim wrote: > > On 05/11, Chao Yu wrote: > >> On 2017/5/11 7:50, Jaegeuk Kim wrote: > >>> On 05/09, Chao Yu wrote: > Hi Jaegeuk, > > On 2017/5/9 5:23, Jaegeuk Kim wrote: > > Hi Chao, > > > > I can't see a strong

Re: [f2fs-dev] [PATCH] f2fs: split bio cache

2017-05-12 Thread Jaegeuk Kim
On 05/12, Chao Yu wrote: > On 2017/5/11 10:35, Jaegeuk Kim wrote: > > On 05/11, Chao Yu wrote: > >> On 2017/5/11 7:50, Jaegeuk Kim wrote: > >>> On 05/09, Chao Yu wrote: > Hi Jaegeuk, > > On 2017/5/9 5:23, Jaegeuk Kim wrote: > > Hi Chao, > > > > I can't see a strong

Re: [PATCH] f2fs: split bio cache

2017-05-11 Thread Chao Yu
On 2017/5/11 10:35, Jaegeuk Kim wrote: > On 05/11, Chao Yu wrote: >> On 2017/5/11 7:50, Jaegeuk Kim wrote: >>> On 05/09, Chao Yu wrote: Hi Jaegeuk, On 2017/5/9 5:23, Jaegeuk Kim wrote: > Hi Chao, > > I can't see a strong reason to split meta from data/node and rename the

Re: [PATCH] f2fs: split bio cache

2017-05-11 Thread Chao Yu
On 2017/5/11 10:35, Jaegeuk Kim wrote: > On 05/11, Chao Yu wrote: >> On 2017/5/11 7:50, Jaegeuk Kim wrote: >>> On 05/09, Chao Yu wrote: Hi Jaegeuk, On 2017/5/9 5:23, Jaegeuk Kim wrote: > Hi Chao, > > I can't see a strong reason to split meta from data/node and rename the

Re: [PATCH] f2fs: split bio cache

2017-05-10 Thread Jaegeuk Kim
On 05/11, Chao Yu wrote: > On 2017/5/11 7:50, Jaegeuk Kim wrote: > > On 05/09, Chao Yu wrote: > >> Hi Jaegeuk, > >> > >> On 2017/5/9 5:23, Jaegeuk Kim wrote: > >>> Hi Chao, > >>> > >>> I can't see a strong reason to split meta from data/node and rename the > >>> existing > >>> function names.

Re: [PATCH] f2fs: split bio cache

2017-05-10 Thread Jaegeuk Kim
On 05/11, Chao Yu wrote: > On 2017/5/11 7:50, Jaegeuk Kim wrote: > > On 05/09, Chao Yu wrote: > >> Hi Jaegeuk, > >> > >> On 2017/5/9 5:23, Jaegeuk Kim wrote: > >>> Hi Chao, > >>> > >>> I can't see a strong reason to split meta from data/node and rename the > >>> existing > >>> function names.

Re: [PATCH] f2fs: split bio cache

2017-05-10 Thread Chao Yu
On 2017/5/11 7:50, Jaegeuk Kim wrote: > On 05/09, Chao Yu wrote: >> Hi Jaegeuk, >> >> On 2017/5/9 5:23, Jaegeuk Kim wrote: >>> Hi Chao, >>> >>> I can't see a strong reason to split meta from data/node and rename the >>> existing >>> function names. Instead, how about keeping the existing one

Re: [PATCH] f2fs: split bio cache

2017-05-10 Thread Chao Yu
On 2017/5/11 7:50, Jaegeuk Kim wrote: > On 05/09, Chao Yu wrote: >> Hi Jaegeuk, >> >> On 2017/5/9 5:23, Jaegeuk Kim wrote: >>> Hi Chao, >>> >>> I can't see a strong reason to split meta from data/node and rename the >>> existing >>> function names. Instead, how about keeping the existing one

Re: [PATCH] f2fs: split bio cache

2017-05-10 Thread Jaegeuk Kim
On 05/09, Chao Yu wrote: > Hi Jaegeuk, > > On 2017/5/9 5:23, Jaegeuk Kim wrote: > > Hi Chao, > > > > I can't see a strong reason to split meta from data/node and rename the > > existing > > function names. Instead, how about keeping the existing one while adding > > some > > page types to deal

Re: [PATCH] f2fs: split bio cache

2017-05-10 Thread Jaegeuk Kim
On 05/09, Chao Yu wrote: > Hi Jaegeuk, > > On 2017/5/9 5:23, Jaegeuk Kim wrote: > > Hi Chao, > > > > I can't see a strong reason to split meta from data/node and rename the > > existing > > function names. Instead, how about keeping the existing one while adding > > some > > page types to deal

Re: [PATCH] f2fs: split bio cache

2017-05-09 Thread Chao Yu
On 2017/5/9 9:49, Chao Yu wrote: > Hi Jaegeuk, > > On 2017/5/9 5:23, Jaegeuk Kim wrote: >> Hi Chao, >> >> I can't see a strong reason to split meta from data/node and rename the >> existing >> function names. Instead, how about keeping the existing one while adding some >> page types to deal

Re: [PATCH] f2fs: split bio cache

2017-05-09 Thread Chao Yu
On 2017/5/9 9:49, Chao Yu wrote: > Hi Jaegeuk, > > On 2017/5/9 5:23, Jaegeuk Kim wrote: >> Hi Chao, >> >> I can't see a strong reason to split meta from data/node and rename the >> existing >> function names. Instead, how about keeping the existing one while adding some >> page types to deal

Re: [PATCH] f2fs: split bio cache

2017-05-08 Thread Chao Yu
Hi Jaegeuk, On 2017/5/9 5:23, Jaegeuk Kim wrote: > Hi Chao, > > I can't see a strong reason to split meta from data/node and rename the > existing > function names. Instead, how about keeping the existing one while adding some > page types to deal with log types? Hmm.. before write this patch,

Re: [PATCH] f2fs: split bio cache

2017-05-08 Thread Chao Yu
Hi Jaegeuk, On 2017/5/9 5:23, Jaegeuk Kim wrote: > Hi Chao, > > I can't see a strong reason to split meta from data/node and rename the > existing > function names. Instead, how about keeping the existing one while adding some > page types to deal with log types? Hmm.. before write this patch,

Re: [PATCH] f2fs: split bio cache

2017-05-08 Thread Jaegeuk Kim
Hi Chao, I can't see a strong reason to split meta from data/node and rename the existing function names. Instead, how about keeping the existing one while adding some page types to deal with log types? Thanks, On 05/08, Chao Yu wrote: > Split DATA/NODE type bio cache according to different

Re: [PATCH] f2fs: split bio cache

2017-05-08 Thread Jaegeuk Kim
Hi Chao, I can't see a strong reason to split meta from data/node and rename the existing function names. Instead, how about keeping the existing one while adding some page types to deal with log types? Thanks, On 05/08, Chao Yu wrote: > Split DATA/NODE type bio cache according to different

[PATCH] f2fs: split bio cache

2017-05-08 Thread Chao Yu
Split DATA/NODE type bio cache according to different temperature, so write IOs with the same temperature can be merged in corresponding bio cache as much as possible, otherwise, different temperature write IOs submitting into one bio cache will always cause split of bio. Signed-off-by: Chao Yu

[PATCH] f2fs: split bio cache

2017-05-08 Thread Chao Yu
Split DATA/NODE type bio cache according to different temperature, so write IOs with the same temperature can be merged in corresponding bio cache as much as possible, otherwise, different temperature write IOs submitting into one bio cache will always cause split of bio. Signed-off-by: Chao Yu