Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 7 Jul 2002 21:53:54 -0400
From:      Klaus Steden <klaus@compt.com>
To:        Randy Bush <randy@psg.com>
Cc:        FreeBSD Stable <freebsd-stable@freebsd.org>
Subject:   Re: signal 8 (fp execption) in pgp 5
Message-ID:  <20020707215354.E95654@cthulu.compt.com>
In-Reply-To: <E17RJPn-000KNc-00@rip.psg.com>; from randy@psg.com on Sun, Jul 07, 2002 at 02:17:55PM -0700
References:  <E17R4zs-000LZb-00@rip.psg.com> <20020707145020.D95654@cthulu.compt.com> <E17RJPn-000KNc-00@rip.psg.com>

next in thread | previous in thread | raw e-mail | index | archive | help
> >> yesterday's -stable and ports tree, rebuilt twice.
> >>   % pgpk -l randy
> >>   Type Bits KeyID      Created    Expires    Algorithm  Use
> >>   sec+ 1024 0xB1331439 1994-04-04 ---------- RSA        Sign & Encrypt 
> >>   uid  Randy Bush <randy@psg.com>
> >>   1 matching key found
> >>   Received signal 8.
> >> anyone else see this or have a clue?
> 
> > Yup. I tried PGP6, too, and found the same problem. Did some digging around
> > in the source, and at least with PGP6, I discovered it's a bug in
> > pgpRndUnix.c. Specifically, when loading a timer for use in an entropy
> > function (I think), there's a call to clock_getres() - using the
> > CLOCK_REALTIME clock, filling in a structure that gets returned to the caller
> > as 0, which triggers the FPE.
> > 
> > I tried using CLOCK_VIRTUAL, which stopped the FPE, but returned an error and
> > thus didn't generate any randomness.
> > 
> > Hopefully I've got my terms right - I'm no crypto expert by any stretch, but
> > that's what I found to be a problem in PGP6. I suspect something similar may
> > be at play in PGP5.
> 
> ok, interesting news.
> 
> if i move the working binary from the old system to the new system, i get
> the sig with it.  i.e. it is in the system, not the build of pgp5.  so i am
> moving this conversation to stable.
> 
Just out of curiosity, what kind of CPUs are you running this on? I'm doing
some testing with other PCs I've got access to (FreeBSD and OpenBSD), and the
machines running Intel are fine - my workstation, with AMDs, returns 0 to
clock_getres().

Klaus

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020707215354.E95654>