[LTP] [PATCH v2] doc: generate CVE reproducer statistics

Sachin Sant sachinp@linux.ibm.com
Thu Apr 23 15:17:56 CEST 2026



On 23/04/26 6:01 pm, Petr Vorel wrote:
> Thanks for implementing this.
>
> I'd prefer this to be on a separate page named "CVE Reproducers"
> (OT: I'd also prefer to rename "Statistics" to "Supported syscalls", but that's
> separate thing.)
Sure, will attempt to implement it.
>
> More notes:
>
> * I suppose it shouldn't be too hard to do it similarly as what we do in the test
>    catalog [1], have each CVE id + optional binary name in parenthesis in the menu
>    on the left:
>
> 	CVE-2025-38236
> 	...
> 	CVE-2016-5195 (dirtyc0w)
> 	...
Yes, agree. I will use this format.
CVE id + binary name in parenthesis sounds better  as compared to below 
scheme.
Will wait for additional feedback.

>    Or, we could have tables by years (i.e. left menu only years with counted
>    number of tests in parenthesis)
>
> 	2025 (2)
> 	2023 (3)
>
>    I don't have a strong preference, which one is better.
>
> * Sort by CVE ID descending (newest CVE on the top, people are mostly
>    interested at current problems than about history).
Will implement this logic.
> * Maybe link to the file should be linked to the test catalog page (which has
>    also link to the test source?), e.g. [2] (link could be relative).
Sounds logical. Will make the change.
> * The description is mostly useless (often just CVE ID or "CVE reproducer
>    test"), link to the test catalog would help.
> * Year is useless (visible from CVE ID)
Will drop these columns.
> * Some of the tests don't have link now, e.g. dirtyc0w, snd_timer01.
Will fix it.
Thanks for the review.

-- 
Thanks
- Sachin




More information about the ltp mailing list