> > I think that there is a fundamental change in behavior here.
> > Previously we were taking (NVD) DB as base and only vulnerable CVEs were 
> > compared annotated with CVE_STATUS or our presence of CVE patches.
> > Now we take the CVE_STATUS and CVE patches as base and add entries from DB 
> > only if they were not annotated yet.
>
> This was a little more complicated than that. get_patched_cves() was taking a 
> part of CVE_STATUS at the beginning of the process, then applying the NVD 
> database.
> The change is to import the totality and then update the status in the 
> process. Now, the entries in CVE_STATUS had priority before, and they still 
> have.
> Now it is explicit, before it was hidden in the code. I do not see changes in 
> the end result, do you have a case in mind?

If with current master I add following to any recipe:
CVE_STATUS[2025-0001] = "not-applicable-config: test"
CVE_STATUS[2025-0002] = "fixed-version: test"
then the resulting build/tmp/log/cve/cve-summary.json which shows all CVEs for 
this recipe regardless of CVE status, it will NOT contain reference to these 
test entries.
But when I apply your patch, they will be both added to the report.
So your code changes the behavior a lot (in a good direction from my point of 
view).

>
>
> > I am not arguing against it, I actually like it much more as we will be 
> > able to insert also CVEs not in DB into our reports.
> > But I have two comments on this:
> > * this should be explicitly described in commit message
>
> Yes, we can add it here. The logic will be documented more in the standalone 
> tool, because with the CVE database, the update rule is even more complex.
> (this is also good, it forces to write the list of priorities)
> 
> > * this makes global cve includes spill into all recipe reports, so this 
> > commit series should also get rid of cve-extra-exclusions.inc file finally 
> > (or at least add a comment into it with this sideeffect)
> This is the plan. I do not want to push a series removing 
> cve-extra-inclusions.inc with this one, but maybe, as you suggest, we add a 
> warning to this file and then remove with a separate series?
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#202324): 
https://lists.openembedded.org/g/openembedded-core/message/202324
Mute This Topic: https://lists.openembedded.org/mt/107228576/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to