Hi Dirk, On Tue, Sep 26, 2023 at 07:53:35AM -0500, Dirk Eddelbuettel wrote: > > Hi Salvatore, > > Looks like we emailed concurrently :) (or concurrently enough for my batched > mail setup). > > On 26 September 2023 at 14:19, Salvatore Bonaccorso wrote: > | Hi Dirk, > | > | On Tue, Sep 26, 2023 at 06:54:31AM -0500, Dirk Eddelbuettel wrote: > | > > | > On 25 September 2023 at 20:58, Salvatore Bonaccorso wrote: > | > | Source: gsl > | > | Version: 2.7.1+dfsg-5 > | > ^^^^^^^^^^^^ > | > | Severity: important > | > | Tags: security upstream > | > | Forwarded: https://savannah.gnu.org/bugs/?59624 > | > | X-Debbugs-Cc: car...@debian.org, Debian Security Team > <t...@security.debian.org> > | > | Control: found -1 2.6+dfsg-2 > | > | > | > | Hi, > | > | > | > | The following vulnerability was published for gsl. > | > | > | > | CVE-2020-35357[0]: > | > | | A buffer overflow can occur when calculating the quantile value > | > | | using the Statistics Library of GSL (GNU Scientific Library), > | > | | versions 2.5 and 2.6. Processing a maliciously crafted input data > | > ^^^^^^^^^^^^ > | > > | > I presume this is still true? Is the '2020' in the CVE for the year this > is from? > | > | I did check the source and unless I did a mistake in checking then yes > | the issue is unfixed up to 2.7.1+dfsg-5 yet, and [2] applies. > > I found (thanks to your diligent links) the better upstream fix that will be > in 2.8 and used that.
Ack thanks! > | > [ I see now at [0] that is spreads 2.6 and 2.7. Out of curiousity, who > did > | > the fix for buster (security) and when ? ] > | > | For buster: > https://tracker.debian.org/news/1465169/accepted-gsl-25dfsg-6deb10u1-source-into-oldoldstable/ > > Ack. And that was only days ago so I wasn't asleep at the wheel here. > > | > | | for gsl_stats_quantile_from_sorted_data of the library may lead to > | > | | unexpected application termination or arbitrary code execution. > | > | > | > | > | > | If you fix the vulnerability please also make sure to include the > | > | CVE (Common Vulnerabilities & Exposures) id in your changelog entry. > | > > | > I'll try. I think this is only the second CVE case in my nearly 30 years > in Debian. > | > | Thanks. Note the issue does not really warrant a DSA, I had two goals > > Agreed. > > | with filling the bug: make you aware of the CVE assignment so the > | issue can be fixed first in unstable and the fix land in trixie. For > | bookworm and bullseye if you have spare cycles the fix might land in a > | point release (there is one upcoming, but the window for uploads > | closing the upcoming weekend). > > I am a bit on the fence as to whether it is needed but I suppose the change > in -6 would apply 'as is'. ok. Yes at least for stable the thing is straightforward anyway. > > | > So the debian/changelog entry needs to contain the string > 'CVE-2020-35357' -- correct? > | > | Yes that is good. > > Perfect. I used that. Great. One other thing which could have been done is to close the bug along (as the BTS can close a bug as well with multiple versions it is not needed to keep it open until as well fixed in stable and oldstable, it can simply be closed with each version in turn). Thanks for your work and swift repsonses! Regards, Salvatore
