Bug 238699

Summary: Tests fail, probably due to wrong detection of my custom kernel features
Product: Base System Reporter: Lorenzo Salvadore <salvadore>
Component: testsAssignee: Enji Cooper <ngie>
Status: Closed Overcome By Events    
Severity: Affects Only Me CC: emaste, markj, ngie
Priority: ---    
Version: 12.0-STABLE   
Hardware: Any   
OS: Any   
Attachments:
Description Flags
kyua report
none
src.conf
none
make.conf
none
custom kernel none

Description Lorenzo Salvadore freebsd_committer freebsd_triage 2019-06-18 22:13:55 UTC
Created attachment 205212 [details]
kyua report

I run the tests suite in /usr/tests with kyua on 12-STABLE r349136 and many tests fail.
On the stable mailing list it was suggested that this was due to wong detection of the features built in my custom kernel features.

I attach the relevant files.

I was also encouraged to run the tests suite again with the MINIMAL kernel, however the MINIMAL kernel is too minimal to let me start the system: I figured out what I needed to add to get my disk drive detected, but I missed what I needed to get the zfs module loaded correctly, while adding "device zfs" into MINIMAL kernel configuration resulted into the build complaining about missing symbols such as xdr_char, xdr_short, xdr_int etc.
If it is indeed useful that I run the tests suite with MINIMAL kernel, I would need help with this.
Comment 1 Lorenzo Salvadore freebsd_committer freebsd_triage 2019-06-18 22:14:54 UTC
Created attachment 205213 [details]
src.conf
Comment 2 Lorenzo Salvadore freebsd_committer freebsd_triage 2019-06-18 22:15:39 UTC
Created attachment 205214 [details]
make.conf
Comment 3 Lorenzo Salvadore freebsd_committer freebsd_triage 2019-06-18 22:16:11 UTC
Created attachment 205215 [details]
custom kernel
Comment 4 Mark Johnston freebsd_committer freebsd_triage 2020-06-11 19:00:11 UTC
I suspect it would be useful to try again on a new HEAD.  I know that some of these issues (testing assuming certain kernel features are present without checking) have been fixed.
Comment 5 Enji Cooper freebsd_committer freebsd_triage 2020-06-13 15:27:53 UTC
* 12.0-STABLE is pretty old (I acknowledge this bug is over a year old though, so this is an issue with triage not being done in a timely manner :/..).
* There are a lot of issues with a number of tests spanning multiple functional areas. It's better to have individual bugs for groups of failures so they can be appropriately triaged/fixed by developers (this is one instance where I think Jira subtasks would be awesome).

I'd be inclined to close this as Closed -> Overcome by Events, but I want to get some clarity around source versions first, etc.

Thank you for the report :)!