How does your "automatically scanned for possible vulnerabilites"
actually work?
Because Debian does backport security fixes, so simply checking the
version number of the software does not indicate if the vulnerability
has been fixed in Debian, or not.
Hi,
I've prepared some docker image based on Debian 12 (bookworm, fully updated) and after upload it to local registry it has been automatically scanned for possible vulnerabilities.
Then I was really surprised when discovered that according to this scan
there are 139 security vulnerabilities and 2 of them are CRITICAL (!).
I've started to dig further to find out what's going on there.
First critical on the list is "zlib1g" binary Debian package which is a part of (a result) of wider package "zlib":
https://tracker.debian.org/pkg/zlib
But it seams this is CRITICAL issue (with score 9.8 in one of its three parts):
https://www.cvedetails.com/cve/CVE-2023-45853/
Similar problem in second critical on the list: package "libaom3" which is a binary package from "aom":
https://security-tracker.debian.org/tracker/source-package/aom
The notes say:
[bookworm] - zlib <ignored> (contrib/minizip not built and src:zlib not producing binary packages)
In other words, there's no point in fixing it because Debian
doesn't build the vulnerable binary component.
Very low priority.
CVSS are often bogus.
Also a bit enigmatic explanation for me...Similar problem in second critical on the list: package "libaom3" which is a >> binary package from "aom":It could crash on invalid input. That's minor. It could crash on
https://security-tracker.debian.org/tracker/source-package/aom
invalid input. Also minor. It could potentially be used to
execute code in the privilege of the user running the software,
which is bad, but it appears to only exist in Android, so Debian
thinks it is not interesting.
On 29.01.2025 1:57 PM, David wrote:
How does your "automatically scanned for possible vulnerabilites"
actually work?
I don't know, but it does not matter in that context.
The fact is, that the
result of this "magic scan" properly found and points out the real critical security vulnerabilities in bookworm which are not fixed. Am I wrong? Please correct me then.
On 29.01.2025 2:12 PM, Dan Ritter wrote:
The notes say:
[bookworm] - zlib <ignored> (contrib/minizip not built and src:zlib not producing binary packages)
In other words, there's no point in fixing it because Debian
doesn't build the vulnerable binary component.
Very low priority.
Could you please drop a link to those notes?
If CVSS is "critical" and Debian tracking system says "bookworm - vulnerable", so why it has low priority?
Maybe I just don;t understand the process of this "Debian doesn't build the vulnerable binary component", so please clarify in more details.
CVSS are often bogus.
Hmmm... I'm not sure what you mean. All security announcements in DSAs are referring to CVSS, so... what's the source of such opinion?
You say: minor, minor, it appears to only exist in Android
Really? :-)
Well, not really what I meant in previous sentence.It does matter because you have to interpret the output of yourHow does your "automatically scanned for possible vulnerabilites"I don't know, but it does not matter in that context.
actually work?
scanner and understand it.
This strange scanner found a CVE attached to minizip. minizip is part
of zlib, but not supported. therefore, for debian it is no reason to
provide a security fix since program (minizip) is not supported by the package zlib itself.
if you use such scanner, _you_ have to understand the output of the
scanner, the CVE itself _and_ the impact on _your_ system. the scanner
can only check a version number against a CVE. but what it means _in
your situation_ is your responsibility, not debians, not the scanners.
Most recently: https://daniel.haxx.se/blog/2025/01/23/cvss-is-dead-to-us/
Yeah, another blog and opinion. Do we (debians) have some betterCVSS are often bogus.Hmmm... I'm not sure what you mean. All security announcements in DSAs are >> referring to CVSS, so... what's the source of such opinion?
Most recently:https://daniel.haxx.se/blog/2025/01/23/cvss-is-dead-to-us/
You say: minor, minor, it appears to only exist in AndroidI read the notes. You sent the links, you should read them.
Really? :-)
On 29.01.2025 2:43 PM, Dan Ritter wrote:
CVSS are often bogus.
Hmmm... I'm not sure what you mean. All security announcements in DSAs are
referring to CVSS, so... what's the source of such opinion?
Most recently: [1]https://daniel.haxx.se/blog/2025/01/23/cvss-is-dead-to-us/
Yeah, another blog and opinion.
Do we (debians) have some better
alternatives?
Are there plans to switch to other solution? Or maybe just discussion
about such switch?
You say: minor, minor, it appears to only exist in Android
Really? :-)
I read the notes. You sent the links, you should read them.
Another misunderstanding - sorry maybe that's my "language side-effect"
;-)
I sent the links, but it seems I don't fully understand them, so I ask for
explanation.
Then you cite some parts form that links in plain text, so I guess you
understand them better and (again - I guess) you fully agree with those
statements.
So could you please explain me what's wrong with my understanding?
So now I suppose I just don't fully understand those information I found, so that's why I ask you guys for help on this Debian user mailing list.
The notes say:
[bookworm] - zlib <ignored> (contrib/minizip not built and src:zlib not producing binary packages)
In other words, there's no point in fixing it because Debian doesn't build the vulnerable binary component.
Very low priority.
Yes. But I'm not asking for "responsibility", but a bit more explanation without blaming anyone.
The notes say:
[bookworm] - zlib <ignored> (contrib/minizip not built and src:zlib not producing binary packages)
In other words, there's no point in fixing it because Debian doesn't build the vulnerable binary component.
Very low priority.
so, this CVE is telling you about a bug which is not affecting Debians
zlib1g since it doesn't build minizip.
that is what your job is: finding out wether the bug is really
affecting you and if so, how to mitigate it.
But still don;t understand "Debian itself does *not* build the affected
component" as I can find "minizip" (and maybe other) package based on that
vulnerable library - see my previous post above as Re- to Hanno.
Anyway thank you for trying to explain me things that are not obvious to
me.
On Wed, Jan 29, 2025 at 03:22:02PM +0100, Rafał Lichwała wrote:
On 29.01.2025 2:43 PM, Dan Ritter wrote:Did you actually read and understand the entire article?
CVSS are often bogus.
Hmmm... I'm not sure what you mean. All security announcements in DSAs are >> referring to CVSS, so... what's the source of such opinion?
Most recently: [1]https://daniel.haxx.se/blog/2025/01/23/cvss-is-dead-to-us/
A blog by the author of cURL. I would submit that his opinion is
extremely relevant, if for no other reason that there is hardly a more important/commonly used piece of network software out there.
Do we (debians) have some betterYes, you read the CVE, you look at how the CVSS score was derived, you
alternatives?
adjust as need for your specific use case, and then you make a decision
based on that.
Are there plans to switch to other solution? Or maybe just discussionMany alternatives are under discussion, but the industry is largely
about such switch?
driven by people who have a vested interested in making every
vulnerability seem as critical as possible. Then they can sell security scanning and remediation solutions for a lot of money. If every
vulnerability was basically "this might be a problem for 0.1% of users
and a minor problem at that" then they would have a hard time selling
their products and services.
What is happening here is that Debian tracks this CVE as affecting itsMaybe that's the explanation I was asking for - thank you.
zlib package because in theory someone could take the source of zlib and modify it to produce the vulnerable binary. This is something that
people should know about, since taking and modifying/rebuilding Debian
source packages is rather common.
However, Debian itself does *not* build the affected component. So, it
makes no sense for Debian as a project to put limited effort into fixing
such a vulnerability.
Yes, it still means that. The minizip binary package you are seeing
comes from a different source package, also called minizip:
https://packages.debian.org/source/bookworm/minizip
On 29.01.2025 4:16 PM, Roberto C. Sánchez wrote:
Yes, it still means that. The minizip binary package you are seeing
comes from a different source package, also called minizip:
https://packages.debian.org/source/bookworm/minizip
Aha! Got it :-)
And there are no binary components in Debian based on vulnerable zlib1g >library in bookworm?
But I have to be aware of this if I want to build some package by myself >which depends on zlib1g, right?
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 546 |
Nodes: | 16 (0 / 16) |
Uptime: | 169:18:51 |
Calls: | 10,385 |
Calls today: | 2 |
Files: | 14,057 |
Messages: | 6,416,551 |