Code signing in OpenBSD

[new_guy]

I've searched OpenBSD.org and google for source code signing practices in OpenBSD, nothing obvious stands out. I've probably overlooked it. Just curious about this... is the process described someplace?

[Lars Hansson-5]

On Dec 5, 2007 11:16 AM, new_guy <byte8bits@...> wrote:
> I've searched OpenBSD.org and google for source code signing practices in
> OpenBSD, nothing obvious stands out. I've probably overlooked it. Just
> curious about this... is the process described someplace?

No. OpenBSD doesn't sign code.

---
Lars Hansson

[Nick Guenther]

On 12/5/07, Lars Hansson <romabysen@...> wrote:
> On Dec 5, 2007 11:16 AM, new_guy <byte8bits@...> wrote:
> > I've searched OpenBSD.org and google for source code signing practices in
> > OpenBSD, nothing obvious stands out. I've probably overlooked it. Just
> > curious about this... is the process described someplace?
>
> No. OpenBSD doesn't sign code.

Well, there's the MD5 files (e.g.
http://openbsd.arcticnetwork.ca/pub/OpenBSD/4.2/i386/MD5).
but yeah, for the most part OpenBSD doesn't need it.
-Nick

[new_guy]

Lars Hansson-5 wrote:

No. OpenBSD doesn't sign code.

---
Lars Hansson

Oh that surprises me, are OpenPGP signatures used for anything? Errata, official communication, etc... maybe this is a stupid question, by it seems everyone does it these days... even small software projects. Not being critical of OpenBSD (I love it and buy CDs) just curious as to the reasoning for not using pgp/gpg keys to sign stuff, secure communication, etc.

[Harpalus a Como]

What is the benefit of doing so? What's the point? Is the website so likely
to be hacked into, that the developers need to sign all communication just
to ensure that it comes from them? There's absolutely no need to signing
errata or official communications. Name one justifiable use for them. If the
OpenBSD developers didn't care about "secure communications", then OpenSSH
would not exist.

On Dec 5, 2007 3:03 PM, new_guy <byte8bits@...> wrote:

> Lars Hansson-5 wrote:
> >
> > No. OpenBSD doesn't sign code.
> >
> > ---
> > Lars Hansson
> >
>
> Oh that surprises me, are OpenPGP signatures used for anything? Errata,
> official communication, etc... maybe this is a stupid question, by it
> seems
> everyone does it these days... even small software projects. Not being
> critical of OpenBSD (I love it and buy CDs) just curious as to the
> reasoning
> for not using pgp/gpg keys to sign stuff, secure communication, etc.
>
>
> --
> View this message in context:
> http://www.nabble.com/Code-signing-in-OpenBSD-tf4947207.html#a14173498
> Sent from the openbsd user - misc mailing list archive at Nabble.com.

[new_guy]

Harpalus a Como wrote:

What is the benefit of doing so? What's the point? Is the website so likely
to be hacked into, that the developers need to sign all communication just
to ensure that it comes from them? There's absolutely no need to signing
errata or official communications. Name one justifiable use for them. If the
OpenBSD developers didn't care about "secure communications", then OpenSSH
would not exist.

Can you dismiss PKI and the benefits that OpenPGP signatures provide to your user community? Knowing that xyz binary is signed by OpenBSD for distribution or abc email came from an official OpenBSD source is a good thing. Trojaned binaries and forged emails happen. PKI can help mitigate this. The benefit of PKI is widely known and accepted and does not need to be rehashed here. I'm surprised that OpenBSD (the most secure OS I know of) does not use it, that's all I'm saying. I also thought there would be a real reason for not doing so and there may in fact be and I may just be unaware of it.

[Nick Guenther]

On 12/5/07, new_guy <byte8bits@...> wrote:

> Harpalus a Como wrote:
> >
> > What is the benefit of doing so? What's the point? Is the website so
> > likely
> > to be hacked into, that the developers need to sign all communication just
> > to ensure that it comes from them? There's absolutely no need to signing
> > errata or official communications. Name one justifiable use for them. If
> > the
> > OpenBSD developers didn't care about "secure communications", then OpenSSH
> > would not exist.
> >
>
> Can you dismiss PKI and the benefits that OpenPGP signatures provide to your
> user community? Knowing that xyz binary is signed by OpenBSD for
> distribution or abc email came from an official OpenBSD source is a good
> thing. Trojaned binaries and forged emails happen. PKI can help mitigate
> this. The benefit of PKI is widely known and accepted and does not need to
> be rehashed here.

Are you *sure* of that? You might want to read
http://www.schneier.com/paper-pki-ft.txt

> I'm surprised that OpenBSD (the most secure OS I know of)
> does not use it, that's all I'm saying. I also thought there would be a real
> reason for not doing so and there may in fact be and I may just be unaware
> of it.

OpenBSD is the most secure OS, the devs know what they are doing.. and
they've rejected this as uneccessary.
You can check the MD5 files for the main distribution, and for
packages.. well the official OpenBSD mirrors are all trustworthy--if
they aren't, it will be discovered and they will no longer be official
mirrors.
This isn't a great answer, I know.

-Nick

[BOFH-5]

On Dec 5, 2007 11:46 AM, new_guy <byte8bits@...> wrote:
> Can you dismiss PKI and the benefits that OpenPGP signatures provide to your
> user community? Knowing that xyz binary is signed by OpenBSD for
> distribution or abc email came from an official OpenBSD source is a good
> thing. Trojaned binaries and forged emails happen. PKI can help mitigate
> this. The benefit of PKI is widely known and accepted and does not need to
> be rehashed here. I'm surprised that OpenBSD (the most secure OS I know of)
> does not use it, that's all I'm saying. I also thought there would be a real
> reason for not doing so and there may in fact be and I may just be unaware
> of it.

What are the risks you are trying to address?  What are the widely
known benefits of PKI?  Who downloads and installs openbsd binaries
*FROM AN EMAIL*?

Would you consider Bruce Schneier to be knowledgeable about PKI?  Have you read:
http://www.schneier.com/paper-pki.html



--
http://www.glumbert.com/media/shift
http://www.youtube.com/watch?v=tGvHNNOLnCk
"This officer's men seem to follow him merely out of idle curiosity."
-- Sandhurst officer cadet evaluation.
"Securing an environment of Windows platforms from abuse - external or
internal - is akin to trying to install sprinklers in a fireworks
factory where smoking on the job is permitted."  -- Gene Spafford

[new_guy]

Nick Guenther wrote:

Well, there's the MD5 files (e.g.
http://openbsd.arcticnetwork.ca/pub/OpenBSD/4.2/i386/MD5).
but yeah, for the most part OpenBSD doesn't need it.
-Nick

Could you explain in more detail? Why doesn't OpenBSD need to use pgp keys? Really, I'm not trying to start anything, I just want to understand. Especially since everyone else seems to do it. FreeBSD, NetBSD, Linux Kernel, etc... they all employ some sort of PKI mechanism... so how does OpenBSD handle these sort of things?

[new_guy]

BOFH-5 wrote:

Would you consider Bruce Schneier to be knowledgeable about PKI?  Have you read:
http://www.schneier.com/paper-pki.html

Yes, I've read that. He's talking about CA's. He does not ridicule PGP keys as you seem to. In fact, he has a few of his own:
   
Bruce Schneier <[e-mail]> 0x4C92D93D 2048 1997/10/16 Never
Bruce Schneier <[e-mail]> 0x7EDE4C65 1024 1995/09/26 Never

Look him and his company Counterpane up yourself:

http://keyserver.veridis.com:11371/

[Harpalus a Como]

For one thing, I think you're quite confused. Unless I'm missing something,
I'm not noticing the FreeBSD, NetBSD, Linux kernel developers "signing"
their code, or doing anything particularly differently from the OpenBSD
developers. Please explain.

You've also conveniently ignored bofh's question. Why do you see this as
being an issue? What risks does PKI mitigate? Did you just vaguely read
somewhere in an advertisement about the supposed security benefits?

On Dec 5, 2007 5:22 PM, new_guy <byte8bits@...> wrote:

> Nick Guenther wrote:
> >
> > Well, there's the MD5 files (e.g.
> > http://openbsd.arcticnetwork.ca/pub/OpenBSD/4.2/i386/MD5).
> > but yeah, for the most part OpenBSD doesn't need it.
> > -Nick
> >
>
> Could you explain in more detail? Why doesn't OpenBSD need to use pgp
> keys?
> Really, I'm not trying to start anything, I just want to understand.
> Especially since everyone else seems to do it. FreeBSD, NetBSD, Linux
> Kernel, etc... they all employ some sort of PKI mechanism... so how does
> OpenBSD handle these sort of things?
>
> --
> View this message in context:
> http://www.nabble.com/Code-signing-in-OpenBSD-tf4947207.html#a14176001
> Sent from the openbsd user - misc mailing list archive at Nabble.com.
>
>

JI

[new_guy]

Wow, my surprise grows... I shall no longer add to this thread... Bye now.

http://www.kernel.org/signature.html
http://www.freebsd.org/doc/pgpkeyring.txt

* One example of a signed Linux Kernel path... there are many others:
ftp://ftp.kernel.org/pub/linux/kernel/v2.6/patch-2.6.9.sign

* One example of signed FreeBSD code... there are others:
http://taosecurity.blogspot.com/2007/11/updating-freebsd-70-beta2-to-70-beta3.html

Some examples of signed communications from FreeBSD & NetBSD:
http://www.freebsd.org/internal/ssh-keys.asc
http://mail-index.netbsd.org/netbsd-announce/2004/02/20/0000.html

On Dec 5, 2007 12:59 PM, Kevin Stam <harpalus.como@...> wrote:

> For one thing, I think you're quite confused. Unless I'm missing
> something, I'm not noticing the FreeBSD, NetBSD, Linux kernel developers
> "signing" their code, or doing anything particularly differently from the
> OpenBSD developers. Please explain.
>
> You've also conveniently ignored bofh's question. Why do you see this as
> being an issue? What risks does PKI mitigate? Did you just vaguely read
> somewhere in an advertisement about the supposed security benefits?

[Harpalus a Como]

Ah, my apologies. I was looking at the wrong thing. No further comment.

On Dec 5, 2007 6:18 PM, Brad Tilley <byte8bits@...> wrote:

> Wow, my surprise grows... I shall no longer add to this thread... Bye now.
>
> http://www.kernel.org/signature.html
> http://www.freebsd.org/doc/pgpkeyring.txt
>
> * One example of a signed Linux Kernel path... there are many others:
> ftp://ftp.kernel.org/pub/linux/kernel/v2.6/patch-2.6.9.sign
>
> * One example of signed FreeBSD code... there are others:
>
> http://taosecurity.blogspot.com/2007/11/updating-freebsd-70-beta2-to-70-beta3.html
>
> Some examples of signed communications from FreeBSD & NetBSD:
> http://www.freebsd.org/internal/ssh-keys.asc
> http://mail-index.netbsd.org/netbsd-announce/2004/02/20/0000.html
>
>
> On Dec 5, 2007 12:59 PM, Kevin Stam < harpalus.como@...> wrote:
>
> > For one thing, I think you're quite confused. Unless I'm missing
> > something, I'm not noticing the FreeBSD, NetBSD, Linux kernel developers
> > "signing" their code, or doing anything particularly differently from the
> > OpenBSD developers. Please explain.
> >
> > You've also conveniently ignored bofh's question. Why do you see this as
> > being an issue? What risks does PKI mitigate? Did you just vaguely read
> > somewhere in an advertisement about the supposed security benefits?

[BOFH-5]

On Dec 5, 2007 12:41 PM, new_guy <byte8bits@...> wrote:
> BOFH-5 wrote:
> >
> > Would you consider Bruce Schneier to be knowledgeable about PKI?  Have you
> > read:
> > http://www.schneier.com/paper-pki.html
> >
>
> Yes, I've read that. He's talking about CA's. He does not ridicule PGP keys
> as you seem to. In fact, he has a few of his own:

I'm not ridiculing PGP keys.  I used to run PKI (Entrust) at a fortune
100 company.  Whenever I hear people screaming about using PKI, I
always want to know - exactly what problem are you trying to solve or
prevent, or what risk you are trying to address.


--
http://www.glumbert.com/media/shift
http://www.youtube.com/watch?v=tGvHNNOLnCk
"This officer's men seem to follow him merely out of idle curiosity."
-- Sandhurst officer cadet evaluation.
"Securing an environment of Windows platforms from abuse - external or
internal - is akin to trying to install sprinklers in a fireworks
factory where smoking on the job is permitted."  -- Gene Spafford

[Rui Miguel Silva Seabra]

On Wed, Dec 05, 2007 at 11:59:31AM -0500, Nick Guenther wrote:
> > I'm surprised that OpenBSD (the most secure OS I know of)
> > does not use it, that's all I'm saying. I also thought there would be a real
> > reason for not doing so and there may in fact be and I may just be unaware
> > of it.
>
> OpenBSD is the most secure OS, the devs know what they are doing.. and
> they've rejected this as uneccessary.

I don't see what is the problem with blessing a fingerprint of the
binaries with a PKI signature, which would mean that *these* are the
binaries the devs intended to release.

Come on... twice a year and get the benefit of not being excluded from
company policies which require digital signature of software downloaded
through the internet.

> You can check the MD5 files for the main distribution, and for
> packages.. well the official OpenBSD mirrors are all trustworthy--if
> they aren't, it will be discovered and they will no longer be official
> mirrors.
> This isn't a great answer, I know.

Definitely not a great answer, as there are vectors of attack which
cover the client acessing the mirror and not the mirror in itself, like
changing on-the-fly the md5sums to match the bad binaries, etc...

A digital signature would enable the non-repudiation of the fingerprints
file (at least), giving a moderate level of assurance that attack
vectors would have to concentrate on upstream development servers (where
the devs *really* know what they are doing).

Rui

--
Hail Eris!
Today is Prickle-Prickle, the 47th day of The Aftermath in the YOLD 3173
+ No matter how much you do, you never do enough -- unknown
+ Whatever you do will be insignificant,
| but it is very important that you do it -- Gandhi
+ So let's do it...?

[Ted Unangst-2]

On 12/5/07, new_guy <byte8bits@...> wrote:
> Can you dismiss PKI and the benefits that OpenPGP signatures provide to your
> user community?

yes.

[Dave Ewart]

On Wednesday, 05.12.2007 at 17:59 +0000, Kevin Stam wrote:

> For one thing, I think you're quite confused. Unless I'm missing
> something, I'm not noticing the FreeBSD, NetBSD, Linux kernel
> developers "signing" their code, or doing anything particularly
> differently from the OpenBSD developers. Please explain.

I'm guessing that he's referring to the fact that some Linux
*distributions* (not the kernel developers or necessarily any of the
components) sign their binary packages: for example Debian do this.

I believe one of the supposed benefits of this is that it allows anyone
to set up a public Debian mirror and, after checking the signatures
during download, one can be sure that they are 'real' Debian packages.

I believe that in some circumstances this may lead to a false sense of
security:

- Said mirror could have old (vulnerable) versions of packages.  Just
  because they're signed doesn't mean they're safe;

- The signing relates only to the packaging: if the underlying source
  code is compromised, then all bets are off.

Would signing help for OpenBSD?  I don't particular see that it would,
given that you are trading off the hassle of implementing it,
maintaining it and so on, against the benefits of doing so, which are
probably small or non-existent.

Dave.

--
Dave Ewart davee@..., jabber:davee@..., freenode:davee
All email from me is now digitally signed, http://www.sungate.co.uk/
Fingerprint: AEC5 9360 0A35 7F66 66E9 82E4 9E10 6769 CD28 DA92

[demime 1.01d removed an attachment of type application/pgp-signature which had a name of signature.asc]

[Ted Unangst-2]

On 12/5/07, Rui Miguel Silva Seabra <rms@...> wrote:
> Come on... twice a year and get the benefit of not being excluded from
> company policies which require digital signature of software downloaded
> through the internet.

sign it yourself, then download it.  problem solved.

[Harpalus a Como]

Yes, that's what I gathered was meant. Going into PKI and code signing,
however, I assumed he meant signing and verifying the underlying source
code, and navigating the trees, I haven't noticed that.

Evidently he meant signing binary packages. In that case, I can kind of
understand the requirement - particularly for business - but whether it's
worth it is up to the OpenBSD team, not me. :) I'm having trouble seeing how
somebody could easily manage to get a compromised binary onto OpenBSD
servers. Seems more trouble to implement then it's worth.

On Dec 5, 2007 7:13 PM, Dave Ewart <davee@...> wrote:

> On Wednesday, 05.12.2007 at 17:59 +0000, Kevin Stam wrote:
>
> > For one thing, I think you're quite confused. Unless I'm missing
> > something, I'm not noticing the FreeBSD, NetBSD, Linux kernel
> > developers "signing" their code, or doing anything particularly
> > differently from the OpenBSD developers. Please explain.
>
> I'm guessing that he's referring to the fact that some Linux
> *distributions* (not the kernel developers or necessarily any of the
> components) sign their binary packages: for example Debian do this.
>
> I believe one of the supposed benefits of this is that it allows anyone
> to set up a public Debian mirror and, after checking the signatures
> during download, one can be sure that they are 'real' Debian packages.
>
> I believe that in some circumstances this may lead to a false sense of
> security:
>
> - Said mirror could have old (vulnerable) versions of packages.  Just
>  because they're signed doesn't mean they're safe;
>
> - The signing relates only to the packaging: if the underlying source
>  code is compromised, then all bets are off.
>
> Would signing help for OpenBSD?  I don't particular see that it would,
> given that you are trading off the hassle of implementing it,
> maintaining it and so on, against the benefits of doing so, which are
> probably small or non-existent.
>
> Dave.
>
> --
> Dave Ewart davee@..., jabber:davee@..., freenode:davee
> All email from me is now digitally signed, http://www.sungate.co.uk/
> Fingerprint: AEC5 9360 0A35 7F66 66E9 82E4 9E10 6769 CD28 DA92
>
> [demime 1.01d removed an attachment of type application/pgp-signature
> which had a name of signature.asc]