Re: CATUPDATE confusion?

From: Peter Eisentraut <peter_e(at)gmx(dot)net>
To: Stephen Frost <sfrost(at)snowman(dot)net>
Cc: Adam Brightwell <adam(dot)brightwell(at)crunchydatasolutions(dot)com>, Noah Misch <noah(at)leadboat(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: CATUPDATE confusion?
Date: 2015-03-03 21:32:27
Message-ID: 54F6286B.90802@gmx.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2/28/15 6:32 PM, Stephen Frost wrote:
> This isn't really /etc/shadow though, this is more like direct access to
> the filesystem through the device node- and you'll note that Linux
> certainly has got an independent set of permissions for that called the
> capabilities system. That's because messing with those pieces can crash
> the kernel. You're not going to crash the kernel if you goof up
> /etc/shadow.

I think this characterization is incorrect. The Linux capability system
does not exist because the actions are scary or can crash the kernel.
Capabilities exist because they are not attached to file system objects
and can therefore not be represented using the usual permission system.

Note that one can write directly to raw devices or the kernel memory
through various /dev and /proc files. No "capability" protects against
that. It's only the file permissions, possibly in combination with
mount options.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message David Fetter 2015-03-03 21:34:32 Re: Idea: GSoC - Query Rewrite with Materialized Views
Previous Message Alvaro Herrera 2015-03-03 20:49:06 Re: Idea: GSoC - Query Rewrite with Materialized Views