pcp
[Top] [All Lists]

Re: pmwebd security (was Re: [RFC] dynamic container switching)

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: pmwebd security (was Re: [RFC] dynamic container switching)
From: "Frank Ch. Eigler" <fche@xxxxxxxxxx>
Date: Wed, 11 Nov 2015 07:16:10 -0500
Cc: PCP <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <558930991.9245903.1447212148100.JavaMail.zimbra@xxxxxxxxxx>
References: <1313883527.54143616.1444783810135.JavaMail.zimbra@xxxxxxxxxx> <20151027155234.GB9303@xxxxxxxxxx> <1185678657.63582036.1446001295613.JavaMail.zimbra@xxxxxxxxxx> <20151031022337.GC28852@xxxxxxxxxx> <1634759327.3957536.1446687727569.JavaMail.zimbra@xxxxxxxxxx> <20151106202554.GE2349@xxxxxxxxxx> <319465283.6999575.1447038990025.JavaMail.zimbra@xxxxxxxxxx> <20151111023103.GA16615@xxxxxxxxxx> <558930991.9245903.1447212148100.JavaMail.zimbra@xxxxxxxxxx>
User-agent: Mutt/1.4.2.2i
Hi -


> > > [...] I'll make the needed pmwebd changes this week.
> > 
> > The local:* stuff should not be marked as deprecated nor
> > be undocumented: that still has plenty of safe & appropriate use.
> 
> If pmwebd acquires some form of explicit ACL, maybe (where people
> will still have to opt-in to exposing information) - until then no,
> there's no point risking that unnecessarily.

"unnecessarily" is in the eye of the beholder/user.  The "-P" is an
opt-in.  In another thread elsewhere we have already covered a
reasonable use case where pmwebd could run as root & expose local:,
for a single highly-privileged application.


> > "pmcd now only hangs for a few seconds after a hostile client
> > sends a few bytes of data instead of indefinitely" in code:
> 
> Yep, interesting, thanks - so not nearly as bad as those initial
> arbitrarily broad "pmcd DoS" claims.

To spell it out further ... When you are on a hostile network, you
will get not one bad guy sending those few bytes once in a while but
hundreds concurrently, and pmcd will effectively stop.


> In normal operation pmcd can be delayed for such short times under
> heavy local load anyway.  I don't think it's a realistic security
> concern in the bigger picture of uses of the service that pmcd is
> offering

OK, so now you grant my point that pmcd shouldn't be put onto a
hostile network.


> - but if its worrying you, please have at it.

You weaponized the whole issue, after I presumed to question the
wisdom & safety of building more upon pmStore, a design that still has
a number of outstanding questions/problems.  We could have hammered
those out instead of this distraction.


- FChE

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