Thanks a lot for all the responses to my question. To be honest, I can not fully understand what you are discussing but just try to understand it better - Below is part of the log when I debugging my driver code. It outputs the start sector and the sector number of every read/write request in the request queue of block layer. It seems that many of the requested sectors are contiguous. And I just curious why the block layer does not merge these contiguous sectors into one single request? For example, if the block layer generate 'start_sect: 48776, nsect: 64, rw: r' instead of below requests, I think the performance will be better. ... start_sect: 48776, nsect: 8, rw: r start_sect: 48784, nsect: 8, rw: r start_sect: 48792, nsect: 8, rw: r start_sect: 48800, nsect: 8, rw: r start_sect: 48808, nsect: 8, rw: r start_sect: 48816, nsect: 8, rw: r start_sect: 48824, nsect: 8, rw: r start_sect: 48832, nsect: 8, rw: r ... Thanks. Regards, Yunpeng >-----Original Message----- >From: James Bottomley [mailto:James.Bottomley@xxxxxxx] >Sent: 2010年4月13日 3:58 >To: Martin K. Petersen >Cc: Robert Hancock; Gao, Yunpeng; linux-ide@xxxxxxxxxxxxxxx; >linux-mmc@xxxxxxxxxxxxxxx >Subject: Re: How to make kernel block layer generate bigger request in the >request queue? > >On Mon, 2010-04-12 at 14:26 -0400, Martin K. Petersen wrote: >> >>>>> "James" == James Bottomley <James.Bottomley@xxxxxxx> writes: >> >> >> Correct. It's quite unlikely for pages to be contiguous so this is >> >> the best we can do. >> >> James> Actually, average servers do about 50% contiguous on average >> James> since we changed the mm layer to allocate in ascending physical >> James> page order ... this figure is highly sensitive to mm changes >> James> though, and can vary from release to release. >> >> Interesting. When did this happen? > >The initial work was done by Bill Irwin, years ago. For a while it was >good, but then after Mel Gorman did the page reclaim code, we became >highly sensitive to the reclaim algorithms for this, so it's fluctuated >a bit ever since. Even with all this, the efficiency is highly >dependent on the amount of free memory: once the machine starts running >to exhaustion (excluding page cache, since that usually allocates >correctly to begin with) the contiguity really drops. > >> Last time I gathered data on segment merge efficiency (1 year+ ago) I >> found that adjacent pages were quite rare for a normal fs type workload. >> Certainly not in the 50% ballpark. I'll take another look when I have a >> moment... > >I got 60% with an I/O bound test with about a gigabyte of free memory a >while ago (2.6.31, I think). Even for machines approaching memory >starvation, 30% seems achievable. > >James > ?韬{.n?????%??檩??w?{.n???{炳'^??骅w* jg????????G??⒏ ⒎?:+v???????? ????＂??????
