All of lore.kernel.org
 help / color / mirror / Atom feed
From: Paul Furness <paul.furness@vil.ite.mee.com>
To: Glynn Clements <glynn.clements@virgin.net>
Cc: admin <linux-admin@vger.kernel.org>
Subject: Re: Recursive groups anyone?
Date: 14 Sep 2002 08:53:13 +0100	[thread overview]
Message-ID: <1031989993.24112.28.camel@Zebra> (raw)
In-Reply-To: <15745.64461.171180.773434@cerise.nosuchdomain.co.uk>

On Fri, 2002-09-13 at 15:53, Glynn Clements wrote:
> 
> > 
> > What actually looks up group membership when you try and read a file?
> 
> Every process has a foreground group id plus a number of supplementary
> group ids. Programs which provide user logins (e.g. "login", "xdm"
> etc) set the foreground group id from /etc/passwd, and the
> supplementary group ids from /etc/group. Child processes inherit these
> values from their parent. Filesystem accesses are checked using the
> attributes of the calling process.
> 
> You can't realistically implement your policy at the kernel level.

Hmm, now I get it. Thanks.

> 
> > Similar code already exists in sendmail and similar for mail groups.
> 
> I don't think so. I suspect that you're misunderstanding some aspect
> of mail delivery.

I was thinking of the aliases file expansion, but since I didn't
understand how the filesystem access worked, I didn't realise that it
didn't matter anyhow. :)

> 
> What you seem to be trying to acheive can't be done. If it could, it
> would probably have adverse implications for security.
 
Well, the 'instant update' of group memberships is only a minor
inconvenience - again, that came from me not really understanding how
file permissions were checked.

Groups within groups would have made the permissions much easier to
manage, but I can simply write a more complex admin tool to take care of
that and generate /etc/group after having expanded groups locally.

Thanks.

Oh, erm, what is the maximum group membership allowed? Is it alterable?
(I seem to remember it was limited to 16 on Solaris, but was
configurable. I also seem to remember that something else - possibly NFS
or NIS - broke if you set the max groups too high.)


Thanks again.

P.







 
-- 
Paul Furness

Systems Manager

2+2=5 for extremely large values of 2.


  reply	other threads:[~2002-09-14  7:53 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-09-13 11:17 Recursive groups anyone? Paul Furness
2002-09-13 14:53 ` Glynn Clements
2002-09-14  7:53   ` Paul Furness [this message]
2002-09-14 14:24     ` Glynn Clements
2002-09-13 16:10 ` Jamie Harris
2002-09-14  7:39   ` Paul Furness
2002-09-14 14:31     ` Glynn Clements

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=1031989993.24112.28.camel@Zebra \
    --to=paul.furness@vil.ite.mee.com \
    --cc=glynn.clements@virgin.net \
    --cc=linux-admin@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 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.