git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: Ramkumar Ramachandra <artagnon@gmail.com>
Cc: Git List <git@vger.kernel.org>
Subject: Re: inotify to minimize stat() calls
Date: Fri, 08 Feb 2013 14:45:49 -0800	[thread overview]
Message-ID: <7va9rezaoy.fsf@alter.siamese.dyndns.org> (raw)
In-Reply-To: <7vehgqzc2p.fsf@alter.siamese.dyndns.org> (Junio C. Hamano's message of "Fri, 08 Feb 2013 14:15:58 -0800")

Junio C Hamano <gitster@pobox.com> writes:

> Ramkumar Ramachandra <artagnon@gmail.com> writes:
>
>> ...  Will Git ever
>> consider using inotify on Linux?  What is the downside?
>
> I think this has come up from time to time, but my understanding is
> that nobody thought things through to find a good layer in the
> codebase to interface to an external daemon that listens to inotify
> events yet.  It is not something like "somebody decreed that we
> would never consider because of such and such downsides."  We are
> not there yet.

I checked read-cache.c and preload-index.c code.  To get the
discussion rolling, I think something like the outline below may be
a good starting point and a feasible weekend hack for somebody
competent:

 * At the beginning of preload_index(), instead of spawning the
   worker thread and doing the lstat() check ourselves, we open a
   socket to our daemon (see below) that watches this repository and
   make a request for lstat update.  The request will contain:

    - The SHA1 checksum of the index file we just read (to ensure
      that we and our daemon share the same baseline to
      communicate); and

    - the pathspec data.

   Our daemon, if it already has a fresh data available, will give
   us a list of <path, lstat result>.  Our main process runs a loop
   that is equivalent to what preload_thread() runs but uses the
   lstat() data we obtained from the daemon.  If our daemon says it
   does not have a fresh data (or somehow our daemon is dead), we do
   the work ourselves.

 * Our daemon watches the index file and the working tree, and
   waits for the above consumer.  First it reads the index (and
   remembers what it read), and whenever an inotify event comes,
   does the lstat() and remembers the result.  It never writes
   to the index, and does not hold the index lock.  Whenever the
   index file changes, it needs to reload the index, and discard
   lstat() data it already has for paths that are lost from the
   updated index.

  reply	other threads:[~2013-02-08 22:46 UTC|newest]

Thread overview: 88+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-02-08 21:10 inotify to minimize stat() calls Ramkumar Ramachandra
2013-02-08 22:15 ` Junio C Hamano
2013-02-08 22:45   ` Junio C Hamano [this message]
2013-02-09  2:10     ` Duy Nguyen
2013-02-09  2:37       ` Junio C Hamano
2013-02-09  2:56     ` Junio C Hamano
2013-02-09  3:36       ` Robert Zeh
2013-02-09 12:05         ` Ramkumar Ramachandra
2013-02-09 12:11           ` Ramkumar Ramachandra
2013-02-09 12:53           ` Ramkumar Ramachandra
2013-02-09 12:59             ` Duy Nguyen
2013-02-09 17:10               ` Ramkumar Ramachandra
2013-02-09 18:56                 ` Ramkumar Ramachandra
2013-02-10  5:24                 ` Duy Nguyen
2013-02-10 11:17                   ` Duy Nguyen
2013-02-10 11:22                     ` Duy Nguyen
2013-02-10 20:16                       ` Junio C Hamano
2013-02-11  2:56                         ` Duy Nguyen
2013-02-11 11:12                           ` Duy Nguyen
2013-03-07 22:16                           ` Torsten Bögershausen
2013-03-08  0:04                             ` Junio C Hamano
2013-03-08  7:01                               ` Torsten Bögershausen
2013-03-08  8:15                                 ` Junio C Hamano
2013-03-08  9:24                                   ` Torsten Bögershausen
2013-03-08 10:53                                   ` Duy Nguyen
2013-03-10  8:23                                     ` Ramkumar Ramachandra
2013-03-13 12:59                                     ` [PATCH] status: hint the user about -uno if read_directory takes too long Nguyễn Thái Ngọc Duy
2013-03-13 15:21                                       ` Torsten Bögershausen
2013-03-13 16:16                                       ` Junio C Hamano
2013-03-14 10:22                                         ` Duy Nguyen
2013-03-14 15:05                                           ` Junio C Hamano
2013-03-15 12:30                                             ` Duy Nguyen
2013-03-15 15:52                                               ` Torsten Bögershausen
2013-03-15 15:57                                                 ` Ramkumar Ramachandra
2013-03-15 16:53                                                 ` Junio C Hamano
2013-03-15 17:41                                                   ` Torsten Bögershausen
2013-03-15 20:06                                                     ` Junio C Hamano
2013-03-15 21:14                                                       ` Torsten Bögershausen
2013-03-15 21:59                                                         ` Junio C Hamano
2013-03-16  7:21                                                           ` Torsten Bögershausen
2013-03-17  4:47                                                             ` Junio C Hamano
2013-03-16  1:51                                           ` Duy Nguyen
2013-02-10 13:26                     ` inotify to minimize stat() calls demerphq
2013-02-10 15:35                       ` Duy Nguyen
2013-02-14 14:36                       ` Magnus Bäck
2013-02-10 16:45                     ` Ramkumar Ramachandra
2013-02-11  3:03                       ` Duy Nguyen
2013-02-10 16:58                     ` Erik Faye-Lund
2013-02-11  3:53                       ` Duy Nguyen
2013-02-12 20:48                         ` Karsten Blees
2013-02-13 10:06                           ` Duy Nguyen
2013-02-13 12:15                           ` Duy Nguyen
2013-02-13 18:18                             ` Jeff King
2013-02-13 19:47                               ` Jeff King
2013-02-13 20:25                               ` Karsten Blees
2013-02-13 22:55                                 ` Jeff King
2013-02-14  0:48                                   ` Karsten Blees
2013-02-27 14:45                                     ` [PATCH] name-hash.c: fix endless loop with core.ignorecase=true Karsten Blees
2013-02-27 16:53                                       ` Junio C Hamano
2013-02-27 21:52                                         ` Karsten Blees
2013-02-27 23:57                                           ` [PATCH v2] " Karsten Blees
2013-02-28  0:27                                             ` Junio C Hamano
2013-02-19  9:49                           ` inotify to minimize stat() calls Ramkumar Ramachandra
2013-02-19 14:25                             ` Karsten Blees
2013-02-19 13:16                   ` Drew Northup
2013-02-19 13:47                     ` Duy Nguyen
2013-02-09 19:35           ` Junio C Hamano
2013-02-10 19:03             ` Robert Zeh
2013-02-10 19:26               ` Martin Fick
2013-02-10 20:18                 ` Robert Zeh
2013-02-11  3:21               ` Duy Nguyen
2013-02-11 14:13                 ` Robert Zeh
2013-02-19  9:57                   ` Ramkumar Ramachandra
2013-04-24 17:20               ` [PATCH] " Robert Zeh
2013-04-24 21:32                 ` Duy Nguyen
2013-04-25 19:44                   ` Robert Zeh
2013-04-25 21:20                     ` Duy Nguyen
2013-04-26 15:35                       ` Robert Zeh
2013-04-25  8:18                 ` Thomas Rast
2013-04-25 19:37                   ` Robert Zeh
2013-04-25 19:59                     ` Thomas Rast
2013-04-27 13:51                       ` Thomas Rast
2013-04-27 23:56                 ` Duy Nguyen
     [not found]                   ` <CAKXa9=r2A7UeBV2s2H3wVGdPkS1zZ9huNJhtvTC-p0S5Ed12xA@mail.gmail.com>
2013-04-30  0:27                     ` Duy Nguyen
2013-02-09 11:32       ` Ramkumar Ramachandra
2013-02-14 15:16 ` Ævar Arnfjörð Bjarmason
2013-02-14 16:31   ` Junio C Hamano
2013-02-19  9:40   ` Ramkumar Ramachandra

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=7va9rezaoy.fsf@alter.siamese.dyndns.org \
    --to=gitster@pobox.com \
    --cc=artagnon@gmail.com \
    --cc=git@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).