mailarchive of the ptxdist mailing list
 help / color / mirror / Atom feed
From: "Sven Püschel" <s.pueschel@pengutronix.de>
To: ptxdist@pengutronix.de
Cc: Alexander Dahl <ada@thorsis.com>
Subject: Re: [ptxdist] [PATCH] xz: version bump 5.8.1 -> 5.8.4
Date: Tue, 22 Sep 2026 12:10:50 +0200	[thread overview]
Message-ID: <29a27425-8323-4047-91f7-b62ad1d1b493@pengutronix.de> (raw)
In-Reply-To: <20260921-granola-stimulus-1fd65d7980e8@thorsis.com>

Hi Alexander,

On 9/21/26 08:18, Alexander Dahl via ptxdist wrote:
> Hello Michael,
>
> Am Sat, Sep 19, 2026 at 09:05:28AM +0200 schrieb Michael Olbrich:
>> On Mon, Sep 14, 2026 at 05:00:55PM +0200, Alexander Dahl via ptxdist wrote:
> […]
>>> Notes:
>>>      Not sure what security gain we have from tarballs created by GitHub?
>> It helps detecting supply chain attacks. The "make dist" archive is
>> basically impossible to reproduce, so we need to trust the uploader that it
>> was not modified. For the git tarball, all changes are visible in the
>> commit history. While someone still needs to look, it's much easier to spot
>> something.
>> And that is not theoretical: The xz backdoor did just that with a modified
>> build-to-host.m4.
>>
>>>      This just makes packaging harder.
>> I'm not sure what you mean with that.
> Usually upstream publishes sha256 sums and gpg signatures along a
> tarball release.  I usually check those, and compare with the sums
> other distributions (e.g. buildroot) have in their packages.  If we
> take the generated tarball from git, those checks are impossible.

I can totally understand both sides. In the end the question is what 
thread we want to protect against?
Your thread model is that GitHub or someone with access to GitHub does 
bad things. Michael's thread model is against a bad/compromised 
maintainer trying to hide the actual attack.
Both are correct and neither guarantees security, as a bad actor could 
also simply hide his stuff in some commit. For best security one would 
have to review the whole (commit) diff between the releases.

But given that xz also has PGP signed commit, you could also verify it 
in the following way:

(clone xz repo and verify signature of the given tag)
$ git archive --format=tar --prefix=xz-5.8.4/ v5.8.4 | sha256sum
ac47844373932e3fed48d7d791fdea426005894592578eaba3f878394d62bfa3 -
$ wcurl 
https://github.com/tukaani-project/xz/archive/refs/tags/v5.8.4.tar.gz
(...)
$ gzip --decompress --stdout v5.8.4.tar.gz | sha256sum
ac47844373932e3fed48d7d791fdea426005894592578eaba3f878394d62bfa3 -


Sidenote: From an operational point of view GitHub btw. also disagrees 
from using their auto generated archives, as they cannot guarantee that 
they are reproducible [1], but they do their best to not break anything [2].

[1] https://github.com/github/docs/issues/31141#issuecomment-1899101631
[2] https://github.com/github/docs/issues/31141#issuecomment-1899166346

Sincerely
     Sven

>
> Greets
> Alex



      reply	other threads:[~2026-09-22 10:11 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-14 15:00 Alexander Dahl via ptxdist
2026-09-19  7:05 ` Michael Olbrich
2026-09-21  6:18   ` Alexander Dahl via ptxdist
2026-09-22 10:10     ` Sven Püschel [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=29a27425-8323-4047-91f7-b62ad1d1b493@pengutronix.de \
    --to=s.pueschel@pengutronix.de \
    --cc=ada@thorsis.com \
    --cc=ptxdist@pengutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox