pcp
[Top] [All Lists]

Re: pcp updates: lberk/marko merges, build+docs updates

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: pcp updates: lberk/marko merges, build+docs updates
From: "Frank Ch. Eigler" <fche@xxxxxxxxxx>
Date: Tue, 5 Jan 2016 18:32:44 -0500
Cc: pcp developers <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <177508323.4781358.1452036255000.JavaMail.zimbra@xxxxxxxxxx>
References: <805278639.3777112.1451891182627.JavaMail.zimbra@xxxxxxxxxx> <780203978.3777208.1451891272970.JavaMail.zimbra@xxxxxxxxxx> <y0mio37n8i3.fsf@xxxxxxxx> <404092777.4770722.1452032875540.JavaMail.zimbra@xxxxxxxxxx> <20160105225138.GA5695@xxxxxxxxxx> <177508323.4781358.1452036255000.JavaMail.zimbra@xxxxxxxxxx>
User-agent: Mutt/1.4.2.2i
Hi -

On Tue, Jan 05, 2016 at 06:24:15PM -0500, Nathan Scott wrote:
> [...]
> > Here are two realistic cases, not of "circumvention" but of normal
> 
> First case does not seem very realistic, but hmmm, *shrug*, maybe.

Thanks!

> The second case is just the pmdaproc case rehashed (i.e. an opt-in
> pmdapapi -A option would be a far better approach, localising that
> privilege elevation).

Except that -A does the opposite of localization, security-wise: it
disables authentication pmcd-wide, so now all other pcp clients get
more privileges than they should.  -A seems like a crutch.

- FChE

<Prev in Thread] Current Thread [Next in Thread>