From: Andrew Morton <akpm@linux-foundation.org> To: Volker.Lendecke@sernet.de Cc: Milosz Tanski <milosz@adfin.com>, linux-kernel@vger.kernel.org, Christoph Hellwig <hch@infradead.org>, linux-fsdevel@vger.kernel.org, linux-aio@kvack.org, Mel Gorman <mgorman@suse.de>, Tejun Heo <tj@kernel.org>, Jeff Moyer <jmoyer@redhat.com>, "Theodore Ts'o" <tytso@mit.edu>, Al Viro <viro@zeniv.linux.org.uk>, linux-api@vger.kernel.org, Michael Kerrisk <mtk.manpages@gmail.com>, linux-arch@vger.kernel.org, Dave Chinner <david@fromorbit.com> Subject: Re: [PATCH v7 0/5] vfs: Non-blockling buffered fs read (page cache only) Date: Thu, 26 Mar 2015 23:08:33 -0700 [thread overview] Message-ID: <20150326230833.4ccfaebb.akpm@linux-foundation.org> (raw) In-Reply-To: <E1YbN1J-0084qO-3s@intern.SerNet.DE> On Fri, 27 Mar 2015 06:41:25 +0100 Volker Lendecke <Volker.Lendecke@sernet.de> wrote: > On Thu, Mar 26, 2015 at 08:28:24PM -0700, Andrew Morton wrote: > > A thing which bugs me about pread2() is that it is specifically > > tailored to applications which are able to use a partial read result. > > ie, by sending it over the network. > > Can you explain what you mean by this? Samba gets a pread > request from a client for some bytes. The client will be > confused when we send less than requested although the file > is long enough to satisfy all. Well it was my assumption that samba would be able to do something useful with a partial read - pread() is allowed to return less than requested. If it isn't the case that samba can use the partial read result then what does it do? It has to save the partial data, then do the additional IO? That's pretty clunky compared to if (it's all in cache) read it all now else ask a worker thread to read it all > > And of course fincore could be used by Samba etc to avoid blocking on > > reads. It wouldn't perform quite as well as pread2(), but I bet it's > > good enough. > > > > Bottom line: with pread2() there's still a need for fincore(), but with > > fincore() there probably isn't a need for pread2(). > > fincore would be a second syscall per pread, and it is not > atomic. I've had discussions with MIPS based vendors who > are worried about every single syscall. This is the #1 > hottest code path in Samba. Bear in mind that these operations involve physical IO and large memcpy's. Yes, a fincore() approach will consume more CPU but the additional overhead will be relatively small. Tradeoffs are involved, and it may turn out that choosing a more flexible and powerful interface which is somewhat more CPU intensive is a better decision. It's hard to say until this is quantified (ie: measured). > > And I'm doubtful about claims that it absolutely has to be non-blocking > > 100% of the time. I bet that 99.99% is good enough. A fincore() > > option to run mark_page_accessed() against present pages would help > > with the race-with-reclaim situation. > > If you can make sure that after an fincore the pages remain > in memory for x milliseconds the atomicity concern might go > away. It won't be guaranteed that the fincore()+pread() will be non-blocking. But blocking will be very rare. I don't know whether the additional expense of activating the pages within fincore() is justified - needs runtime testing.
WARNING: multiple messages have this Message-ID (diff)
From: Andrew Morton <akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org> To: Volker.Lendecke-3ekOc4rQMZmzQB+pC5nmwQ@public.gmane.org Cc: Milosz Tanski <milosz-B5zB6C1i6pkAvxtiuMwx3w@public.gmane.org>, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Christoph Hellwig <hch-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>, linux-fsdevel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-aio-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, Mel Gorman <mgorman-l3A5Bk7waGM@public.gmane.org>, Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>, Jeff Moyer <jmoyer-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>, "Theodore Ts'o" <tytso-3s7WtUTddSA@public.gmane.org>, Al Viro <viro-RmSDqhL/yNMiFSDQTTA3OLVCufUGDwFn@public.gmane.org>, linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Michael Kerrisk <mtk.manpages-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>, linux-arch-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Dave Chinner <david-FqsqvQoI3Ljby3iVrkZq2A@public.gmane.org> Subject: Re: [PATCH v7 0/5] vfs: Non-blockling buffered fs read (page cache only) Date: Thu, 26 Mar 2015 23:08:33 -0700 [thread overview] Message-ID: <20150326230833.4ccfaebb.akpm@linux-foundation.org> (raw) In-Reply-To: <E1YbN1J-0084qO-3s-dqLtpHMqGvUyWpdLl23E4A@public.gmane.org> On Fri, 27 Mar 2015 06:41:25 +0100 Volker Lendecke <Volker.Lendecke-3ekOc4rQMZmzQB+pC5nmwQ@public.gmane.org> wrote: > On Thu, Mar 26, 2015 at 08:28:24PM -0700, Andrew Morton wrote: > > A thing which bugs me about pread2() is that it is specifically > > tailored to applications which are able to use a partial read result. > > ie, by sending it over the network. > > Can you explain what you mean by this? Samba gets a pread > request from a client for some bytes. The client will be > confused when we send less than requested although the file > is long enough to satisfy all. Well it was my assumption that samba would be able to do something useful with a partial read - pread() is allowed to return less than requested. If it isn't the case that samba can use the partial read result then what does it do? It has to save the partial data, then do the additional IO? That's pretty clunky compared to if (it's all in cache) read it all now else ask a worker thread to read it all > > And of course fincore could be used by Samba etc to avoid blocking on > > reads. It wouldn't perform quite as well as pread2(), but I bet it's > > good enough. > > > > Bottom line: with pread2() there's still a need for fincore(), but with > > fincore() there probably isn't a need for pread2(). > > fincore would be a second syscall per pread, and it is not > atomic. I've had discussions with MIPS based vendors who > are worried about every single syscall. This is the #1 > hottest code path in Samba. Bear in mind that these operations involve physical IO and large memcpy's. Yes, a fincore() approach will consume more CPU but the additional overhead will be relatively small. Tradeoffs are involved, and it may turn out that choosing a more flexible and powerful interface which is somewhat more CPU intensive is a better decision. It's hard to say until this is quantified (ie: measured). > > And I'm doubtful about claims that it absolutely has to be non-blocking > > 100% of the time. I bet that 99.99% is good enough. A fincore() > > option to run mark_page_accessed() against present pages would help > > with the race-with-reclaim situation. > > If you can make sure that after an fincore the pages remain > in memory for x milliseconds the atomicity concern might go > away. It won't be guaranteed that the fincore()+pread() will be non-blocking. But blocking will be very rare. I don't know whether the additional expense of activating the pages within fincore() is justified - needs runtime testing.
next prev parent reply other threads:[~2015-03-27 6:07 UTC|newest] Thread overview: 94+ messages / expand[flat|nested] mbox.gz Atom feed top 2015-03-16 18:27 [PATCH v7 0/5] vfs: Non-blockling buffered fs read (page cache only) Milosz Tanski 2015-03-16 18:27 ` [PATCH v7 1/5] vfs: Prepare for adding a new preadv/pwritev with user flags Milosz Tanski 2015-03-16 18:27 ` Milosz Tanski 2015-03-16 21:05 ` Andreas Dilger 2015-03-16 21:05 ` Andreas Dilger 2015-03-16 18:27 ` [PATCH v7 2/5] vfs: Define new syscalls preadv2,pwritev2 Milosz Tanski 2015-03-16 18:27 ` Milosz Tanski 2015-03-16 18:27 ` [PATCH v7 3/5] x86: wire up preadv2 and pwritev2 Milosz Tanski 2015-03-16 18:27 ` Milosz Tanski 2015-03-16 18:27 ` [PATCH v7 4/5] vfs: RWF_NONBLOCK flag for preadv2 Milosz Tanski 2015-03-16 18:27 ` Milosz Tanski 2015-03-16 18:27 ` [PATCH v7 5/5] xfs: add RWF_NONBLOCK support Milosz Tanski 2015-03-16 18:27 ` Milosz Tanski 2015-03-16 22:04 ` Dave Chinner 2015-03-16 18:32 ` [PATCH] Add preadv2/pwritev2 documentation Milosz Tanski 2015-03-27 16:49 ` Andrew Morton 2015-03-30 7:33 ` Christoph Hellwig 2015-03-30 7:33 ` Christoph Hellwig 2015-03-16 18:34 ` [PATCH] fstests: generic test for preadv2 behavior on linux Milosz Tanski 2015-03-16 18:34 ` Milosz Tanski 2015-03-16 21:07 ` Andreas Dilger 2015-03-16 21:07 ` Andreas Dilger 2015-03-16 22:03 ` Milosz Tanski 2015-03-16 22:02 ` Dave Chinner 2015-03-16 22:02 ` Dave Chinner 2015-03-16 22:11 ` Milosz Tanski 2015-03-16 22:56 ` Dave Chinner 2015-03-16 22:56 ` Dave Chinner 2015-03-26 11:55 ` [PATCH v7 0/5] vfs: Non-blockling buffered fs read (page cache only) Christoph Hellwig 2015-03-26 11:55 ` Christoph Hellwig 2015-03-26 19:12 ` Milosz Tanski 2015-03-26 19:12 ` Milosz Tanski 2015-03-27 2:26 ` Milosz Tanski 2015-03-27 2:29 ` Milosz Tanski 2015-03-27 2:29 ` Milosz Tanski 2015-03-27 3:28 ` Andrew Morton 2015-03-27 3:28 ` Andrew Morton 2015-03-27 5:41 ` Volker Lendecke 2015-03-27 5:41 ` Volker Lendecke 2015-03-27 6:08 ` Andrew Morton [this message] 2015-03-27 6:08 ` Andrew Morton 2015-03-27 8:02 ` Volker Lendecke 2015-03-27 8:02 ` Volker Lendecke 2015-03-27 8:12 ` Christoph Hellwig 2015-03-27 8:18 ` Christoph Hellwig 2015-03-27 8:18 ` Christoph Hellwig 2015-03-27 8:35 ` Andrew Morton 2015-03-27 8:35 ` Andrew Morton 2015-03-27 8:48 ` Christoph Hellwig 2015-03-27 9:01 ` Andrew Morton 2015-03-27 9:01 ` Andrew Morton 2015-03-27 9:44 ` Volker Lendecke 2015-03-27 15:58 ` Jeremy Allison 2015-03-27 15:58 ` Jeremy Allison 2015-03-27 16:30 ` Andrew Morton 2015-03-27 16:30 ` Andrew Morton 2015-03-27 16:30 ` Andrew Morton 2015-03-27 16:30 ` Andrew Morton 2015-03-27 16:39 ` Jeremy Allison 2015-03-27 16:39 ` Jeremy Allison 2015-03-27 16:39 ` Andrew Morton 2015-03-27 16:45 ` Milosz Tanski 2015-03-31 1:27 ` Milosz Tanski 2015-03-27 16:38 ` Milosz Tanski 2015-03-27 16:38 ` Milosz Tanski 2015-03-30 7:36 ` Christoph Hellwig 2015-03-30 17:19 ` Jeremy Allison 2015-03-30 17:19 ` Jeremy Allison 2015-03-30 22:51 ` Milosz Tanski 2015-03-30 20:26 ` Andrew Morton 2015-03-30 20:26 ` Andrew Morton 2015-03-30 20:32 ` Jeremy Allison 2015-03-30 20:37 ` Andrew Morton 2015-03-30 20:49 ` Jeremy Allison 2015-03-30 21:33 ` Andrew Morton 2015-03-30 22:35 ` Milosz Tanski 2015-03-30 22:49 ` Milosz Tanski 2015-03-30 22:57 ` Andrew Morton 2015-03-30 23:06 ` Milosz Tanski 2015-03-30 23:06 ` Milosz Tanski 2015-03-30 23:25 ` Milosz Tanski 2015-04-04 3:42 ` Andrew Morton 2015-04-06 3:53 ` Milosz Tanski 2015-04-06 3:53 ` Milosz Tanski 2015-03-30 23:09 ` Milosz Tanski 2015-03-27 15:21 ` Milosz Tanski 2015-03-27 15:21 ` Milosz Tanski 2015-03-27 17:04 ` Andrew Morton 2015-03-30 7:40 ` Christoph Hellwig 2015-03-30 7:40 ` Christoph Hellwig 2015-03-30 18:54 ` Andrew Morton 2015-03-30 22:40 ` Milosz Tanski 2015-03-30 22:50 ` Andrew Morton 2015-03-30 22:50 ` Andrew Morton
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20150326230833.4ccfaebb.akpm@linux-foundation.org \ --to=akpm@linux-foundation.org \ --cc=Volker.Lendecke@sernet.de \ --cc=david@fromorbit.com \ --cc=hch@infradead.org \ --cc=jmoyer@redhat.com \ --cc=linux-aio@kvack.org \ --cc=linux-api@vger.kernel.org \ --cc=linux-arch@vger.kernel.org \ --cc=linux-fsdevel@vger.kernel.org \ --cc=linux-kernel@vger.kernel.org \ --cc=mgorman@suse.de \ --cc=milosz@adfin.com \ --cc=mtk.manpages@gmail.com \ --cc=tj@kernel.org \ --cc=tytso@mit.edu \ --cc=viro@zeniv.linux.org.uk \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.