> On May 12, 2015, 3:49 p.m., Jan Kundrát wrote:
> > Was the old code a part of some release? If yes, this should get a CVE 
> > security announcement because it allows a local attacker to e.g. force you 
> > to overwirte some of your user's files.
> 
> Michael Palimaka wrote:
>     It looks like it was introduced in 
> 999e774b3ce117598df2029364bd10f4347be81c and released in 0.2.0 and later.
> 
> Frank Reininghaus wrote:
>     Could you elaborate on how such an attack would work? Even if we ignore 
> that the code in question is part of an autotest which is probably never 
> installed anywhere, such that systems of packagers, developers and users who 
> build from source are the only possible targets, I really don't see how an 
> attacker could use the code to cause any unintended damage. Anyone who runs 
> the test regularly creates and deletes the file /tmp/kpeople_test_db already, 
> so what other damage could a local attacker cause?

I didn't realize that it's in autotests -- I apparently noticed just the 
basename of that file, observed that there's no "test" in what I saw, and 
concluded that it's exploitable. You're right that if it's unpackaged, then 
issuing a CVE doesn't make sense.

On the other hand, if this wasn't in an autotest but instead a part of regular 
operation, something simple such as `ln -s .ssh/id_rsa /tmp/kpeople_test_db` by 
an attacker would cause any app using this library to remove user's vital file.

Sorry for noise.


- Jan


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/123724/#review80247
-----------------------------------------------------------


On May 12, 2015, 12:49 p.m., Michael Palimaka wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/123724/
> -----------------------------------------------------------
> 
> (Updated May 12, 2015, 12:49 p.m.)
> 
> 
> Review request for KDE Frameworks and KDEPIM.
> 
> 
> Repository: kpeople
> 
> 
> Description
> -------
> 
> Hardcoding files like this seems like a bad idea.
> 
> 
> Diffs
> -----
> 
>   autotests/persondatatests.h 30eeeb5cd647c713f1b438543a54516ced9f3ede 
>   autotests/persondatatests.cpp 73098d3717509ad80761bbd02000b4ce5060bbb2 
>   autotests/personsmodeltest.h 5b8879521f334459c4f73c2708b3368c543e40a3 
>   autotests/personsmodeltest.cpp b19d1baf8a2c2e617d4b6128df29fbab3b8e61a7 
> 
> Diff: https://git.reviewboard.kde.org/r/123724/diff/
> 
> 
> Testing
> -------
> 
> Tests still pass.
> 
> 
> Thanks,
> 
> Michael Palimaka
> 
>

_______________________________________________
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel

Reply via email to