Re: [PATCH v9 3/3] net/handshake: Add Kunit tests for the handshake consumer API

[Chuck Lever III <chuck.lever@oracle.com>]

> On Apr 14, 2023, at 11:39 PM, Chuck Lever <cel@kernel.org> wrote:
> 
> On Fri, Apr 14, 2023 at 07:15:42PM -0700, Jakub Kicinski wrote:
>> On Fri, 14 Apr 2023 22:06:59 -0400 Chuck Lever wrote:
>>> On Fri, Apr 14, 2023 at 06:31:13PM -0700, Jakub Kicinski wrote:
>>>> We're getting:
>>>> 
>>>> net/handshake/.kunitconfig: warning: ignored by one of the .gitignore files
>>>> 
>>>> during allmodconfig build, any idea where that's coming from?  
>>> 
>>> As far as I know, all of the .kunitconfig files in the kernel tree
>>> are marked "ignored". I'm not sure why, nor if it's a significant
>>> problem.
>> 
>> To be clear - no idea what the problem is but I don't think all
>> of them are:
>> 
>> $ echo a > fs/fat/.kunitconfig
>> $ echo b > mm/kfence/.kunitconfig
>> $ echo c > net/sunrpc/.kunitconfig
>> $ git status
>> Changes not staged for commit:
>>  (use "git add <file>..." to update what will be committed)
>>  (use "git restore <file>..." to discard changes in working directory)
>> modified:   fs/fat/.kunitconfig
>> modified:   mm/kfence/.kunitconfig
>> modified:   net/sunrpc/.kunitconfig
> 
> The "ignored" list I got from the bot yesterday indeed included
> net/sunrpc/.kunitconfig as well as net/handshake/.kunitconfig, but
> git doesn't actually seem to ignore changes to these files, as you
> demonstrate here.
> 
> I don't see a specific pattern in the kernel's .gitconfig that
> would exclude .kunitconfig files from change tracking.
> 
> I can see where this warning might introduce false negative build
> results, but so far I haven't heard that particular complaint about
> net/sunrpc/.kunitconfig.
> 
> Again, I wasn't sure if this was a significant problem or simply
> noise, so I haven't chased it. If someone on-list has insight,
> please speak up. I can try to have a look at it tomorrow.

OK, shedding a little light on this:

[cel@oracle-102 linux]$ git check-ignore -v .kunitconfig
.gitignore:13:.* .kunitconfig
[cel@oracle-102 linux]$ git status --ignored
On branch upcall-v10
Ignored files:
  (use "git add -f <file>..." to include in what will be committed)
.kunit/
cscope.in.out
cscope.out
cscope.po.out
tools/net/ynl/lib/__pycache__/
tools/testing/kunit/__pycache__/

nothing to commit, working tree clean
[cel@oracle-102 linux]$ 

So,

a) There's a very broad wildcard pattern in .gitignore that
matches all dotfiles, and

b) git actually does not ignore .kunitconfig files.


I'm wondering if this might help the overall situation:

diff --git a/.gitignore b/.gitignore
index 70ec6037fa7a..51117ba29c88 100644
--- a/.gitignore
+++ b/.gitignore
@@ -105,6 +105,7 @@ modules.order
 !.gitignore
 !.mailmap
 !.rustfmt.toml
+!.kunitconfig
   #
 # Generated include files


--
Chuck Lever