All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Korsgaard <peter@korsgaard.com>
To: Arnout Vandecappelle <arnout@mind.be>
Cc: Robert Smigielski <ptdropper@gmail.com>, buildroot@buildroot.org
Subject: Re: [Buildroot] CycloneDX SBOM support
Date: Mon, 28 Aug 2023 21:57:51 +0200	[thread overview]
Message-ID: <87msyb548w.fsf@48ers.dk> (raw)
In-Reply-To: <71359643-6efe-4a83-55e3-8eb3d87edfe5@mind.be> (Arnout Vandecappelle's message of "Mon, 28 Aug 2023 21:38:11 +0200")

>>>>> "Arnout" == Arnout Vandecappelle <arnout@mind.be> writes:

Hello,

 >  I mentioned PURL vs CPE in my talk at ELC this year. You can look it
 >  up on youtube. It was near the end of the talk.

OK, I'll have a look.


 >> Conceptually they seem quite similar, with PURL being more generic, but
 >> I fail to see how we could use PURLS in Buildroot, E.G. how to do the
 >> equivalent of the CPE matching we use to figure out if the version of a
 >> package in Buildroot is vulnerable to a specific CVE?

 >  I think Robert is not necessarily primarily concerned with finding
 >  vulnerabilities, but rather with constructing a meaningful and
 > accurate SBoM (which is what dependencytrack does).

True. The monitoring stuff seems quite interesting for vulnerabilities
though.


 >  That said, it you want to use PURLs for vulnerabilities, you have to
 >  use a vulnerability database that uses PURLs. To my knowledge, there
 > is just one "open source" one: https://osv.dev. (There's also Sonatype
 > which can be used gratis but is not free.) Since there are many, many
 > CVEs that don't make it into OSV, using _only_ PURLs is certainly not
 > enough. But we could combine the two.

OK. It doesn't sound like it will bring a lot of advantages for the
effort to maintain PURL identifiers :/


 >  Another issue with PURLs is: which one do we use? PURLs are organised
 >  around ecosystems. For PyPI it's clear, but for your typical C
 > library/application it's less so. E.g. openssl appears in the Debian,
 > Alpine, AlmaLinux, RPM, RockyLinux ecosystems (and possibly more),
 > each with a distinct PURL. We could start our own namespace, but
 > that's kind of pointless unless we also issue advisories...

I guess we should use the one matching where we get the source code from
(if any). The cyclonedx tool uses a "generic"
pkg:generic/$name?download_url=$site/$tarball, so we could default to
that and just use pypi/github/whatever for the special cases where there
is a more accurate one.


 >  There's by the way another issue (which also exists for the CPE-based
 >  approach): our "BoM" for the cargo and go packages is not correct: we
 > vendor the dependencies, but they're not taken into account in the
 > BoM. The tarball we put in legal-info does include the vendored
 > dependencies, but they're not mentioned in the manifest, and we don't
 > scan their vulnerabilities.

True. I am not sure of a good way how to fix that though. That shouldn't
stop us from generating a good SBOM for all the other packages.

-- 
Bye, Peter Korsgaard
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

  reply	other threads:[~2023-08-28 19:58 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-10 12:55 [Buildroot] CycloneDX SBOM support Robert Smigielski
2023-08-28  6:00 ` Peter Korsgaard
2023-08-28 11:38   ` Michael Nosthoff via buildroot
2023-08-28 12:11     ` Robert Smigielski
2023-08-28 14:57       ` Peter Korsgaard
2023-08-28 19:38         ` Arnout Vandecappelle via buildroot
2023-08-28 19:57           ` Peter Korsgaard [this message]
2023-08-28 23:19             ` Robert Smigielski
2023-08-29  6:55               ` Arnout Vandecappelle via buildroot
2023-08-29 13:10               ` Peter Korsgaard
2023-08-29 13:38               ` Robert Smigielski
2023-09-13 15:12                 ` Robert Smigielski
2023-08-28 23:12         ` Robert Smigielski

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=87msyb548w.fsf@48ers.dk \
    --to=peter@korsgaard.com \
    --cc=arnout@mind.be \
    --cc=buildroot@buildroot.org \
    --cc=ptdropper@gmail.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.