From: Eric Dumazet <dada1@cosmosbay.com>
To: Jeff Moyer <jmoyer@redhat.com>
Cc: Davide Libenzi <davidel@xmailserver.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Benjamin LaHaise <bcrl@kvack.org>,
Trond Myklebust <trond.myklebust@fys.uio.no>,
Andrew Morton <akpm@linux-foundation.org>,
linux-aio <linux-aio@kvack.org>,
zach.brown@oracle.com
Subject: Re: [patch] eventfd - remove fput() call from possible IRQ context (2nd rev)
Date: Wed, 18 Mar 2009 16:43:59 +0100 [thread overview]
Message-ID: <49C116BF.9080402@cosmosbay.com> (raw)
In-Reply-To: <x49ocvyhmyu.fsf@segfault.boston.devel.redhat.com>
Jeff Moyer a écrit :
> Eric Dumazet <dada1@cosmosbay.com> writes:
>
>
>>> rwfd = open("rwfile", O_RDWR|O_DIRECT); assert(rwfd != -1);
>>> if (posix_memalign((void **)&buf, getpagesize(), SIZE) < 0) {
>>> perror("posix_memalign");
>>> exit(1);
>>> }
>>> memset(buf, 0x42, SIZE);
>>>
>>> /* Write test. */
>>> res = io_queue_init(1024, &io_ctx); assert(res == 0);
>>> io_prep_pwrite(&iocb, rwfd, buf, SIZE, 0);
>>> io_set_eventfd(&iocb, efd);
>>> res = io_submit(io_ctx, 1, iocbs); assert(res == 1);
>> yes but io_submit() is blocking. so your close(efd) will come after the release in fs/aio.c
>
> I'm not sure why you think io_submit is blocking. In my setup, I
> preallocated the file, and the test code opens it with O_DIRECT. So,
> io_submit should return after the dio is issued, and the I/O size is
> large enough that it should still be outstanding when io_submit returns.
Hmm.. io_submit() is a blocking syscall, this is how I understood fs/aio.c
Then, using strace -tt -T on your program, I can confirm it is quite a long syscall (3.5 seconds,
about time needed to write a 256 MB file on my disk ;) )
16:33:15.861301 eventfd(0) = 3 <0.000010>
16:33:15.861344 open("rwfile", O_RDWR|O_DIRECT) = 4 <0.000010>
16:33:15.861444 mmap(NULL, 268443648, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f08eec55000 <0.000008>
16:33:16.100040 io_setup(1024, {139676618563584}) = 0 <0.000047>
16:33:16.100149 io_submit(139676618563584, 1, {...}) = 1 <3.558085>
16:33:19.658292 io_getevents(139676618563584, 1, 1, {...}NULL) = 1 <0.258104>
16:33:19.916486 fstat(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 0), ...}) = 0 <0.000007>
16:33:19.916559 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f08ff3b9000 <0.000009>
16:33:19.916607 write(1, "eventfd write test [SUCCESS]\n"..., 29eventfd write test [SUCCESS]
) = 29 <0.000008>
16:33:19.916653 exit_group(0) = ?
next prev parent reply other threads:[~2009-03-18 15:44 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-03-09 15:49 [patch] aio: remove aio-max-nr and instead use the memlock rlimit to limit the number of pages pinned for the aio completion ring Jeff Moyer
2009-03-09 15:54 ` [patch] factor out checks against the memlock rlimit Jeff Moyer
2009-03-09 15:59 ` [patch] man-pages: add documentation about the memlock implications of io_setup Jeff Moyer
2009-03-09 16:45 ` Michael Kerrisk
2009-03-09 16:48 ` Michael Kerrisk
2009-03-09 20:44 ` Jeff Moyer
2009-03-09 16:18 ` [patch] aio: remove aio-max-nr and instead use the memlock rlimit to limit the number of pages pinned for the aio completion ring Avi Kivity
2009-03-09 17:57 ` Jeff Moyer
2009-03-09 19:45 ` Avi Kivity
2009-03-09 20:36 ` Jamie Lokier
2009-03-10 8:36 ` Avi Kivity
2009-03-09 20:31 ` Eric Dumazet
2009-03-12 2:39 ` Eric Dumazet
2009-03-12 2:44 ` Benjamin LaHaise
2009-03-12 3:24 ` Eric Dumazet
2009-03-12 3:29 ` Benjamin LaHaise
2009-03-12 3:33 ` Eric Dumazet
2009-03-12 3:36 ` Benjamin LaHaise
2009-03-12 3:40 ` Eric Dumazet
2009-03-12 3:09 ` Eric Dumazet
2009-03-12 5:18 ` [PATCH] fs: fput() can be called from interrupt context Eric Dumazet
2009-03-12 5:42 ` [PATCH] aio: " Eric Dumazet
2009-03-12 5:47 ` [PATCH] fs: " Andrew Morton
2009-03-12 6:10 ` Eric Dumazet
2009-03-12 6:39 ` Andrew Morton
2009-03-12 13:39 ` Davide Libenzi
2009-03-13 22:34 ` Davide Libenzi
2009-03-13 22:43 ` Eric Dumazet
2009-03-13 23:28 ` Trond Myklebust
2009-03-14 1:40 ` Davide Libenzi
2009-03-14 4:02 ` Trond Myklebust
2009-03-14 14:32 ` Davide Libenzi
2009-03-15 1:36 ` [patch] eventfd - remove fput() call from possible IRQ context Davide Libenzi
2009-03-15 17:44 ` Benjamin LaHaise
2009-03-15 20:08 ` [patch] eventfd - remove fput() call from possible IRQ context (2nd rev) Davide Libenzi
2009-03-16 17:25 ` Jamie Lokier
2009-03-16 18:36 ` Davide Libenzi
2009-03-18 14:22 ` Jeff Moyer
2009-03-18 14:46 ` Davide Libenzi
2009-03-18 14:55 ` Eric Dumazet
2009-03-18 15:25 ` Jeff Moyer
2009-03-18 15:43 ` Eric Dumazet [this message]
2009-03-18 16:13 ` Jeff Moyer
2009-03-18 17:25 ` [patch] eventfd - remove fput() call from possible IRQ context (3rd rev) Davide Libenzi
2009-03-18 17:34 ` Jeff Moyer
2009-03-12 19:22 ` [PATCH] fs: fput() can be called from interrupt context Eric Dumazet
2009-03-12 20:21 ` Andrew Morton
2009-03-09 22:36 ` [patch] aio: remove aio-max-nr and instead use the memlock rlimit to limit the number of pages pinned for the aio completion ring Andrew Morton
2009-03-10 13:43 ` Jeff Moyer
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=49C116BF.9080402@cosmosbay.com \
--to=dada1@cosmosbay.com \
--cc=akpm@linux-foundation.org \
--cc=bcrl@kvack.org \
--cc=davidel@xmailserver.org \
--cc=jmoyer@redhat.com \
--cc=linux-aio@kvack.org \
--cc=linux-kernel@vger.kernel.org \
--cc=trond.myklebust@fys.uio.no \
--cc=zach.brown@oracle.com \
/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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox