From: ebiederm@xmission.com (Eric W. Biederman)
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Jiri Slaby <jslaby@suse.cz>, Jiri Slaby <jirislaby@gmail.com>,
Alan Cox <alan@lxorguk.ukuu.org.uk>,
LKML <linux-kernel@vger.kernel.org>,
Al Viro <viro@zeniv.linux.org.uk>,
Maciej Rutecki <maciej.rutecki@gmail.com>
Subject: Re: [PATCH] sysfs: Optionally count subdirectories to support buggy applications
Date: Thu, 08 Mar 2012 14:18:47 -0800 [thread overview]
Message-ID: <m1zkbqu46w.fsf@fess.ebiederm.org> (raw)
In-Reply-To: <CA+55aFy1PXiNkXw08tDM7G9AO0_92xm2mEFF961-4Wpizo+6KQ@mail.gmail.com> (Linus Torvalds's message of "Mon, 5 Mar 2012 08:47:27 -0800")
Linus Torvalds <torvalds@linux-foundation.org> writes:
> On Mon, Mar 5, 2012 at 8:09 AM, Greg Kroah-Hartman
> <gregkh@linuxfoundation.org> wrote:
>>
>> I don't remember. I thought there was a proposed patch for this issue
>> from Eric, but I don't see it in my queue anywhere.
>
> That patch was an abortion. Adding a config option for behavior like
> this is totally bogus, and the only reason for that config option was
> that sysfs did silly things.
The biggest reason it is bogus is that it doesn't get properly tested
or reviewed. Sigh. My first patch to fix things had a bad typo
that everyone missed.
> It's only in -next, though, I was assuming that the whole "Kill nlink
> counting" commit never makes it to me. Because I won't take it.
>
> I outlined how the counting could easily be done without actually
> having to maintain an explicit count in the sysfs.
And if you had bothered to look you would have seen how we used to
have that code and it was removed because it was a performance
bottleneck.
> Or we should just keep doing the counting.
The current counting that we do gives the wrong numbers, in the
edge cases. To my knowledge a deleted sysfs directory has never
returned nlink == 0.
Keeping compatibility is easy enough that it looks like it is worth
doing, but maintaining 30+ years of backwards compatibility is what
nlink >1 in unix filesystem directories is. I don't see any practical
sense in keeping . and .. directories on disk or upping the unix
nlink directory count because of them. To me it looks like just one
of those things you do. Like hash directory entries so you can
have a big directory and still be able to have a 32bit offset you
can pass to lseek that is stable across renames and deletes.
>From the point of view of maintaining sysfs a 32bit nlink_t in sysfs is
too small. It is wrong for sysfs to refuse to represent devices that
exist and I have heard of machines that have enough memory it possible
to create more than 2^32 network devices. So sysfs must handle overflow
and sysfs must use the nlink == 1 in some cases. I was just thinking
we would get better userspace test coverage if we don't bother to handle
the other cases.
Eric
next prev parent reply other threads:[~2012-03-08 22:15 UTC|newest]
Thread overview: 75+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-01-30 21:56 sysfs regression: wrong link counts Jiri Slaby
2012-01-30 22:06 ` Greg KH
2012-01-30 22:10 ` Alan Cox
2012-01-30 22:27 ` Greg KH
2012-01-30 22:43 ` Al Viro
2012-01-30 22:56 ` Al Viro
2012-01-31 1:27 ` Eric W. Biederman
2012-01-31 10:48 ` Jiri Slaby
2012-01-31 12:44 ` Eric W. Biederman
2012-01-31 16:45 ` Linus Torvalds
2012-01-31 19:18 ` Al Viro
2012-02-01 5:06 ` Eric W. Biederman
2012-02-01 22:21 ` [PATCH] sysfs: Optionally count subdirectories to support buggy applications Eric W. Biederman
2012-02-01 22:24 ` Greg Kroah-Hartman
2012-02-01 22:44 ` Eric W. Biederman
2012-02-01 22:49 ` Greg Kroah-Hartman
2012-02-01 22:31 ` Dave Jones
2012-02-01 22:35 ` Jiri Slaby
2012-02-01 23:15 ` Linus Torvalds
2012-02-01 23:18 ` Linus Torvalds
2012-02-02 1:22 ` Al Viro
2012-02-02 21:24 ` [RFC] killing boilerplate checks in ->link/->mkdir/->rename Al Viro
2012-02-02 23:46 ` Linus Torvalds
2012-02-03 1:16 ` Al Viro
2012-02-03 1:45 ` Al Viro
2012-02-03 2:00 ` Linus Torvalds
2012-02-03 14:57 ` Chris Mason
2012-02-03 17:08 ` Al Viro
2012-02-03 19:34 ` Artem Bityutskiy
2012-02-06 8:50 ` Artem Bityutskiy
2012-02-06 13:56 ` Al Viro
2012-02-06 17:05 ` Artem Bityutskiy
2012-02-06 17:11 ` Al Viro
2012-02-07 7:21 ` Artem Bityutskiy
2012-02-06 22:49 ` Dave Chinner
2012-02-03 8:25 ` Andreas Dilger
2012-02-03 17:03 ` Al Viro
2012-02-04 7:42 ` Andreas Dilger
2012-03-05 13:30 ` [PATCH] sysfs: Optionally count subdirectories to support buggy applications Jiri Slaby
2012-03-05 16:09 ` Greg Kroah-Hartman
2012-03-05 16:47 ` Linus Torvalds
2012-03-08 21:05 ` Greg Kroah-Hartman
2012-03-08 22:18 ` Eric W. Biederman [this message]
2012-03-08 23:40 ` Linus Torvalds
2012-03-08 21:28 ` Eric W. Biederman
2012-03-08 21:34 ` [PATCH 1/3] sysfs: Compact sysfs_dirent s_flags into a byte Eric W. Biederman
2012-03-08 21:36 ` [PATCH 2/3] sysfs: Maintain usable nlink directory counts Eric W. Biederman
2012-03-08 21:37 ` [PATCH 3/3] sysfs: Remove SYSFS_FLAG_REMOVED and use sd->s_nlink == 0 instead Eric W. Biederman
2012-03-09 3:40 ` Linus Torvalds
2012-03-08 22:28 ` [PATCH 1/3] sysfs: Compact sysfs_dirent s_flags into a byte Greg Kroah-Hartman
2012-03-09 2:49 ` Eric W. Biederman
2012-01-31 3:45 ` sysfs regression: wrong link counts Eric W. Biederman
2012-01-31 11:54 ` Alan Cox
2012-01-30 22:52 ` Kay Sievers
2012-01-31 10:41 ` network regression: cannot rename netdev twice Jiri Slaby
2012-01-31 10:52 ` Kay Sievers
2012-01-31 11:00 ` Jiri Slaby
2012-01-31 11:13 ` Kay Sievers
2012-01-31 11:17 ` Jiri Slaby
2012-01-31 11:58 ` Kay Sievers
2012-01-31 14:18 ` Eric W. Biederman
2012-01-31 14:40 ` [PATCH] sysfs: Update the name hash when renaming sysfs entries Eric W. Biederman
2012-01-31 14:41 ` Jiri Slaby
2012-01-31 14:55 ` Greg KH
2012-02-04 2:14 ` network regression: cannot rename netdev twice Henrique de Moraes Holschuh
2012-02-06 20:03 ` Kay Sievers
2012-02-08 2:00 ` Henrique de Moraes Holschuh
2012-02-08 3:50 ` Kay Sievers
2012-02-08 6:42 ` Valdis.Kletnieks
2012-02-08 10:57 ` Kay Sievers
2012-02-08 20:06 ` Valdis.Kletnieks
2012-02-08 20:27 ` Stephen Hemminger
2012-02-08 23:48 ` Kay Sievers
2012-01-31 1:32 ` sysfs regression: wrong link counts Eric W. Biederman
2012-02-01 18:29 ` Maciej Rutecki
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=m1zkbqu46w.fsf@fess.ebiederm.org \
--to=ebiederm@xmission.com \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=gregkh@linuxfoundation.org \
--cc=jirislaby@gmail.com \
--cc=jslaby@suse.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=maciej.rutecki@gmail.com \
--cc=torvalds@linux-foundation.org \
--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: 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).