From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail172.messagelabs.com (mail172.messagelabs.com [216.82.254.3]) by kanga.kvack.org (Postfix) with SMTP id EFCF06B01EE for ; Wed, 7 Apr 2010 04:06:55 -0400 (EDT) Received: by pzk30 with SMTP id 30so721221pzk.12 for ; Wed, 07 Apr 2010 01:06:54 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <20100407074732.GC17892@localhost> References: <4BBA6776.5060804@mozilla.com> <20100406095135.GB5183@cmpxchg.org> <20100407022456.GA9468@localhost> <4BBBF402.70403@mozilla.com> <20100407071408.GA17892@localhost> <20100407074732.GC17892@localhost> Date: Wed, 7 Apr 2010 17:06:54 +0900 Message-ID: Subject: Re: Downsides to madvise/fadvise(willneed) for application startup From: Minchan Kim Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Sender: owner-linux-mm@kvack.org To: Wu Fengguang Cc: Taras Glek , Johannes Weiner , "linux-mm@kvack.org" , "linux-kernel@vger.kernel.org" List-ID: On Wed, Apr 7, 2010 at 4:47 PM, Wu Fengguang wrote= : > On Wed, Apr 07, 2010 at 03:33:52PM +0800, Minchan Kim wrote: >> On Wed, Apr 7, 2010 at 4:14 PM, Wu Fengguang wr= ote: >> > On Wed, Apr 07, 2010 at 12:06:07PM +0800, Minchan Kim wrote: >> >> On Wed, Apr 7, 2010 at 11:54 AM, Taras Glek wrote= : >> >> > On 04/06/2010 07:24 PM, Wu Fengguang wrote: >> >> >> >> >> >> Hi Taras, >> >> >> >> >> >> On Tue, Apr 06, 2010 at 05:51:35PM +0800, Johannes Weiner wrote: >> >> >> >> >> >>> >> >> >>> On Mon, Apr 05, 2010 at 03:43:02PM -0700, Taras Glek wrote: >> >> >>> >> >> >>>> >> >> >>>> Hello, >> >> >>>> I am working on improving Mozilla startup times. It turns out th= at page >> >> >>>> faults(caused by lack of cooperation between user/kernelspace) a= re the >> >> >>>> main cause of slow startup. I need some insights from someone wh= o >> >> >>>> understands linux vm behavior. >> >> >>>> >> >> >> >> >> >> How about improve Fedora (and other distros) to preload Mozilla (a= nd >> >> >> other apps the user run at the previous boot) with fadvise() at bo= ot >> >> >> time? This sounds like the most reasonable option. >> >> >> >> >> > >> >> > That's a slightly different usecase. I'd rather have all large apps= startup >> >> > as efficiently as possible without any hacks. Though until we get t= here, >> >> > we'll be using all of the hacks we can. >> >> >> >> >> >> As for the kernel readahead, I have a patchset to increase default >> >> >> mmap read-around size from 128kb to 512kb (except for small memory >> >> >> systems). =C2=A0This should help your case as well. >> >> >> >> >> > >> >> > Yes. Is the current readahead really doing read-around(ie does it r= ead pages >> >> > before the one being faulted)? From what I've seen, having the dyna= mic >> >> > linker read binary sections backwards causes faults. >> >> > http://sourceware.org/bugzilla/show_bug.cgi?id=3D11447 >> >> >> >> >> >> >> >> >>>> >> >> >>>> Current Situation: >> >> >>>> The dynamic linker mmap()s =C2=A0executable and data sections of= our >> >> >>>> executable but it doesn't call madvise(). >> >> >>>> By default page faults trigger 131072byte reads. To make matters= worse, >> >> >>>> the compile-time linker + gcc lay out code in a manner that does= not >> >> >>>> correspond to how the resulting executable will be executed(ie t= he >> >> >>>> layout is basically random). This means that during startup 15-4= 0mb >> >> >>>> binaries are read in basically random fashion. Even if one order= s the >> >> >>>> binary optimally, throughput is still suboptimal due to the puny >> >> >>>> readahead. >> >> >>>> >> >> >>>> IO Hints: >> >> >>>> Fortunately when one specifies madvise(WILLNEED) pagefaults trig= ger 2mb >> >> >>>> reads and a binary that tends to take 110 page faults(ie program= stops >> >> >>>> execution and waits for disk) can be reduced down to 6. This has= the >> >> >>>> potential to double application startup of large apps without an= y clear >> >> >>>> downsides. >> >> >>>> >> >> >>>> Suse ships their glibc with a dynamic linker patch to fadvise() >> >> >>>> dynamic libraries(not sure why they switched from doing madvise >> >> >>>> before). >> >> >>>> >> >> >> >> >> >> This is interesting. I wonder how SuSE implements the policy. >> >> >> Do you have the patch or some strace output that demonstrates the >> >> >> fadvise() call? >> >> >> >> >> > >> >> > glibc-2.3.90-ld.so-madvise.diff in >> >> > http://www.rpmseek.com/rpm/glibc-2.4-31.12.3.src.html?hl=3Dcom&cba= =3D0:G:0:3732595:0:15:0: >> >> > >> >> > As I recall they just fadvise the filedescriptor before accessing i= t. >> >> >> >> >> >> >> >> >>>> >> >> >>>> I filed a glibc bug about this at >> >> >>>> http://sourceware.org/bugzilla/show_bug.cgi?id=3D11431 . Uli com= mented >> >> >>>> with his concern about wasting memory resources. What is the imp= act of >> >> >>>> madvise(WILLNEED) or the fadvise equivalent on systems under mem= ory >> >> >>>> pressure? Does the kernel simply start ignoring these hints? >> >> >>>> >> >> >>> >> >> >>> It will throttle based on memory pressure. =C2=A0In idle situatio= ns it will >> >> >>> eat your file cache, however, to satisfy the request. >> >> >>> >> >> >>> Now, the file cache should be much bigger than the amount of unne= eded >> >> >>> pages you prefault with the hint over the whole library, so I gue= ss the >> >> >>> benefit of prefaulting the right pages outweighs the downside of = evicting >> >> >>> some cache for unused library pages. >> >> >>> >> >> >>> Still, it's a workaround for deficits in the demand-paging/readah= ead >> >> >>> heuristics and thus a bit ugly, I feel. =C2=A0Maybe Wu can help. >> >> >>> >> >> >> >> >> >> Program page faults are inherently random, so the straightforward >> >> >> solution would be to increase the mmap read-around size (for deskt= ops >> >> >> with reasonable large memory), rather than to improve program layo= ut >> >> >> or readahead heuristics :) >> >> >> >> >> > >> >> > Program page faults may exhibit random behavior once they've starte= d. >> >> > >> >> > During startup page-in pattern of over-engineered OO applications i= s very >> >> > predictable. Programs are laid out based on compilation units, whic= h have no >> >> > relation to how they are executed. Another problem is that any larg= e old >> >> > application will have lots of code that is either rarely executed o= r >> >> > completely dead. Random sprinkling of live code among mostly unneed= ed code >> >> > is a problem. >> >> > I'm able to reduce startup pagefaults by 2.5x and mem usage by a fe= w MB with >> >> > proper binary layout. Even if one lays out a program wrongly, the w= orst-case >> >> > pagein pattern will be pretty similar to what it is by default. >> >> > >> >> > But yes, I completely agree that it would be awesome to increase th= e >> >> > readahead size proportionally to available memory. It's a little si= lly to be >> >> > reading tens of megabytes in 128kb increments :) =C2=A0You rock for= trying to >> >> > modernize this. >> >> >> >> Hi, Wu and Taras. >> >> >> >> I have been watched at this thread. >> >> That's because I had a experience on reducing startup latency of appl= ication >> >> in embedded system. >> >> >> >> I think sometime increasing of readahead size wouldn't good in embedd= ed. >> >> Many of embedded system has nand as storage and compression file syst= em. >> >> About nand, as you know, random read effect isn't rather big than hdd= . >> >> About compression file system, as one has a big compression, >> >> it would make startup late(big block read and decompression). >> >> We had to disable readahead of code page with kernel hacking. >> >> And it would make application slow as time goes by. >> >> But at that time we thought latency is more important than performanc= e >> >> on our application. >> >> >> >> Of course, it is different whenever what is file system and >> >> compression ratio we use . >> >> So I think increasing of readahead size might always be not good. >> >> >> >> Please, consider embedded system when you have a plan to tweak >> >> readahead, too. :) >> > >> > Minchan, glad to know that you have experiences on embedded Linux. >> > >> > While increasing the general readahead size from 128kb to 512kb, I >> > also added a limit for mmap read-around: if system memory size is less >> > than X MB, then limit read-around size to X KB. For example, do only >> > 128KB read-around for a 128MB embedded box, and 32KB ra for 32MB box. >> > >> > Do you think it a reasonable safety guard? Patch attached. >> >> Thanks for reply, Wu. >> >> I didn't have looked at the your attachment. >> That's because it's not matter of memory size in my case. > > In general, the more memory size, the less we care about the possible > readahead misses :) > >> It was alone application on system and it was first main application of = system. >> It means we had a enough memory. >> >> I guess there are such many of embedded system. >> At that time, although I could disable readahead totally with read_ahead= _kb, >> I didn't want it. That's because I don't want to disable readahead on >> the file I/O >> and data section of program. So at a loss, I hacked kernel to disable >> readahead of >> only code section. > > I would like to auto tune readahead size based on the device's > IO throughput and latency estimation, however that's not easy.. Indeed. > Other than that, if we can assert "this class of devices won't benefit > from large readahead", then we can do some static assignment. A few month ago, I saw your patch about enhancing readahead. At that time, many guys tested several size of USB and SSD which are consist of nand device. The result is good if we does readahead untile some crossover point. So I think we need readahead about file I/O in non-rotation device, too. But startup latency is important than file I/O performance in some machine. With analysis at that time, code readahead of application affected slow sta= rtup. In addition, during bootup, cache hit ratio was very small. So I hoped we can disable readahead just only code section(ie, roughly exec vma's filemap fault). :) I don't want you to solve this problem right now. Just let you understand embedded system's some problem for enhancing readahead in future. :) > Thanks, > Fengguang > --=20 Kind regards, Minchan Kim -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org