KUnit test files are scattered across the tree and don't always match
the F: patterns in the KERNEL UNIT TESTING FRAMEWORK MAINTAINERS entry.
Add the KUnit maintainers automatically when the file path matches
common KUnit naming conventions: _kunit.c, -kunit.c, _kunit_test.c,
and -kunit-tests.c.

Signed-off-by: Matteo Croce <teknoraver@meta.com>
---
v2: added missing SOB

 scripts/get_maintainer.pl | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/scripts/get_maintainer.pl b/scripts/get_maintainer.pl
index 4414194bedcf..c07723af4132 100755
--- a/scripts/get_maintainer.pl
+++ b/scripts/get_maintainer.pl
@@ -966,6 +966,20 @@ sub get_maintainers {
 	    }
 	}
 
+	## Add KUnit maintainers for KUnit test files
+	if ($file =~ m/[_-]kunit[_-]?.*\.c$/) {
+	    my $kunit_tvi = find_first_section();
+	    while ($kunit_tvi < @typevalue) {
+		my $kunit_start = find_starting_index($kunit_tvi);
+		if ($typevalue[$kunit_start] =~ m/^KERNEL UNIT TESTING FRAMEWORK/) {
+		    $hash{$kunit_tvi} = 0;
+		    add_categories($kunit_tvi, "");
+		    last;
+		}
+		$kunit_tvi = find_ending_index($kunit_tvi) + 1;
+	    }
+	}
+
 	maintainers_in_file($file);
     }
 
-- 
2.53.0

On Sun, 2026-03-01 at 02:11 +0100, Matteo Croce wrote:
> ```
> KUnit test files are scattered across the tree and don't always match
> the F: patterns in the KERNEL UNIT TESTING FRAMEWORK MAINTAINERS entry.
> Add the KUnit maintainers automatically when the file path matches
> common KUnit naming conventions: _kunit.c, -kunit.c, _kunit_test.c,
> and -kunit-tests.c.

nack.  There's no need to modify get_maintainer.

Use F: patterns instead.

diff --git a/MAINTAINERS b/MAINTAINERS
index 96e97d25e1c2..9796e184b9d5 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -13956,6 +13956,11 @@ F:     rust/kernel/kunit.rs
 F:	rust/macros/kunit.rs
 F:	scripts/rustdoc_test_*
 F:	tools/testing/kunit/
+F:	*/*[_-]kunit*.c
+F:	*/*/*[_-]kunit*.c
+F:	*/*/*/*[_-]kunit*.c
+F:	*/*/*/*/*[_-]kunit*.c
+F:	*/*/*/*/*/*[_-]kunit*.c
 
 KERNEL USERMODE HELPER
 M:	Luis Chamberlain <mcgrof@kernel.org>

Il giorno dom 1 mar 2026 alle ore 03:36 Joe Perches <joe@perches.com>
ha scritto:
>
> On Sun, 2026-03-01 at 02:11 +0100, Matteo Croce wrote:
> > ```
> > KUnit test files are scattered across the tree and don't always match
> > the F: patterns in the KERNEL UNIT TESTING FRAMEWORK MAINTAINERS entry.
> > Add the KUnit maintainers automatically when the file path matches
> > common KUnit naming conventions: _kunit.c, -kunit.c, _kunit_test.c,
> > and -kunit-tests.c.
>
> nack.  There's no need to modify get_maintainer.
>
> Use F: patterns instead.
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 96e97d25e1c2..9796e184b9d5 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -13956,6 +13956,11 @@ F:     rust/kernel/kunit.rs
>  F:     rust/macros/kunit.rs
>  F:     scripts/rustdoc_test_*
>  F:     tools/testing/kunit/
> +F:     */*[_-]kunit*.c
> +F:     */*/*[_-]kunit*.c
> +F:     */*/*/*[_-]kunit*.c
> +F:     */*/*/*/*[_-]kunit*.c
> +F:     */*/*/*/*/*[_-]kunit*.c
>
>  KERNEL USERMODE HELPER
>  M:     Luis Chamberlain <mcgrof@kernel.org>

So many rules covering all possible depths?
Sounds good.

BTW at this point it would be useful to implement the ** glob operator,
but maybe it's not worth it.

Regards,
-- 
Matteo Croce

perl -e 'for($t=0;;$t++){print chr($t*($t>>8|$t>>13)&255)}' |aplay

On Sun, 2026-03-01 at 12:17 +0100, Matteo Croce wrote:
> BTW at this point it would be useful to implement the ** glob operator,
> but maybe it's not worth it.

Hi Matteo.

Given you've now touched get_maintainer, you are welcome to do it.
Maybe change "sub file_match_pattern {".
But I believe this is the first actual use case so I will not.

Il giorno dom 1 mar 2026 alle ore 17:42 Joe Perches <joe@perches.com>
ha scritto:
>
> On Sun, 2026-03-01 at 12:17 +0100, Matteo Croce wrote:
> > BTW at this point it would be useful to implement the ** glob operator,
> > but maybe it's not worth it.
>
> Hi Matteo.
>
> Given you've now touched get_maintainer, you are welcome to do it.
> Maybe change "sub file_match_pattern {".
> But I believe this is the first actual use case so I will not.

Mmm not really the first usage, I see this pattern being already used:

$ grep -F '/*/*/' MAINTAINERS
F: Documentation/devicetree/bindings/*/*/*ma35*
F: drivers/*/*/*ma35*
F: Documentation/devicetree/bindings/*/*/*npcm*
F: drivers/*/*/*npcm*
F: arch/*/*/*/*ftrace*
F: arch/*/*/*ftrace*
F: drivers/net/ethernet/*/*/*/*/*xdp*
F: drivers/net/ethernet/*/*/*xdp*

I tried to implement it and seems easier than expected, the chunk below does it
(validated with: scripts/get_maintainer.pl --self-test=patterns)

--- a/scripts/get_maintainer.pl
+++ b/scripts/get_maintainer.pl
@@ -375,7 +375,9 @@ sub read_maintainer_file {
      ##Filename pattern matching
      if ($type eq "F" || $type eq "X") {
  $value =~ s@\.@\\\.@g;       ##Convert . to \.
- $value =~ s/\*/\.\*/g;       ##Convert * to .*
+ $value =~ s/\*\*/\x00/g;     ##Convert ** to placeholder
+ $value =~ s/\*/[^\/]*/g;     ##Convert * to [^/]*
+ $value =~ s/\x00/.*/g;       ##Convert placeholder to .*
  $value =~ s/\?/\./g;         ##Convert ? to .
  ##if pattern is a directory and it lacks a trailing slash, add one
  if ((-d $value)) {
@@ -919,6 +921,8 @@ sub get_maintainers {
  if ($type eq 'F') {
      if (file_match_pattern($file, $value)) {
  my $value_pd = ($value =~ tr@/@@);
+ ## Discount slashes inside [^/] character classes
+ $value_pd -= (() = $value =~ /\[\^\/\]/g);
  my $file_pd = ($file  =~ tr@/@@);
  $value_pd++ if (substr($value,-1,1) ne "/");
  $value_pd = -1 if ($value =~ /^\.\*/);
@@ -955,7 +959,8 @@ sub get_maintainers {
  $line =~ s/([^\\])\.([^\*])/$1\?$2/g;
  $line =~ s/([^\\])\.$/$1\?/g; ##Convert . back to ?
  $line =~ s/\\\./\./g;        ##Convert \. to .
- $line =~ s/\.\*/\*/g;        ##Convert .* to *
+ $line =~ s/\[\^\/\]\*/\*/g; ##Convert [^/]* to *
+ $line =~ s/\.\*/\*\*/g;      ##Convert .* to **
      }
      my $count = $line =~ s/^([A-Z]):/$1:\t/g;
      if ($letters eq "" || (!$count || $letters =~ /$1/i)) {
@@ -1045,12 +1050,8 @@ sub file_match_pattern {
      return 1;
  }
     } else {
- if ($file =~ m@^$pattern@) {
-     my $s1 = ($file =~ tr@/@@);
-     my $s2 = ($pattern =~ tr@/@@);
-     if ($s1 == $s2) {
- return 1;
-     }
+ if ($file =~ m@^$pattern$@) {
+     return 1;
  }
     }
     return 0;

Regards,
-- 
Matteo Croce

perl -e 'for($t=0;;$t++){print chr($t*($t>>8|$t>>13)&255)}' |aplay

On Mon, 2026-03-02 at 01:50 +0100, Matteo Croce wrote:
> ```
> Il giorno dom 1 mar 2026 alle ore 17:42 Joe Perches <[joe@perches.com](mailto:joe@perches.com)>
> ha scritto:
> 
> > On Sun, 2026-03-01 at 12:17 +0100, Matteo Croce wrote:
> BTW at this point it would be useful to implement the ** glob operator,
> > > but maybe it's not worth it.
> 
> > 
> > Hi Matteo.
> > 
> > Given you've now touched get_maintainer, you are welcome to do it.
> > Maybe change "sub file_match_pattern {".
> > But I believe this is the first actual use case so I will not.
> 
> 
> Mmm not really the first usage, I see this pattern being already used:
> 
> $ grep -F '/*/*/' MAINTAINERS
> F: Documentation/devicetree/bindings/*/*/*ma35*
> F: drivers/*/*/*ma35*
> F: Documentation/devicetree/bindings/*/*/*npcm*
> F: drivers/*/*/*npcm*
> F: arch/*/*/*/*ftrace*
> F: arch/*/*/*ftrace*
> F: drivers/net/ethernet/*/*/*/*/*xdp*
> F: drivers/net/ethernet/*/*/*xdp*

Nope.

Those are specific directory depth matches.
Your use case is _all_ directory depth matches.

> I tried to implement it and seems easier than expected, the chunk below does it
> (validated with: scripts/get_maintainer.pl --self-test=patterns)
> 
> --- a/scripts/get_maintainer.pl
> +++ b/scripts/get_maintainer.pl

This doesn't apply as there are tab/space problems.

> @@ -375,7 +375,9 @@ sub read_maintainer_file {
[]
> @@ -1045,12 +1050,8 @@ sub file_match_pattern {
>       return 1;
>   }
>      } else {
> - if ($file =~ m@^$pattern@) {
> -     my $s1 = ($file =~ tr@/@@);
> -     my $s2 = ($pattern =~ tr@/@@);
> -     if ($s1 == $s2) {
> - return 1;
> -     }
> + if ($file =~ m@^$pattern$@) {
> +     return 1;

and I think the count of / characters matters
but if you send a patch that applies, I'll verify.

cheers, 
>   }
>      }
>      return 0;
> 
> Regards,
> 
> ```

Il giorno lun 2 mar 2026 alle ore 03:05 Joe Perches <joe@perches.com>
ha scritto:
>
> and I think the count of / characters matters
> but if you send a patch that applies, I'll verify.
>

I see that it's used in file_match_pattern.
Anyway I made another simpler patch, I'll send in another standalone
patch, so it won't be corrupted

> cheers,

cheers,

-- 
Matteo Croce

perl -e 'for($t=0;;$t++){print chr($t*($t>>8|$t>>13)&255)}' |aplay

On 2026-03-01 12:17:17+0100, Matteo Croce wrote:
> Il giorno dom 1 mar 2026 alle ore 03:36 Joe Perches <joe@perches.com>
> ha scritto:
> >
> > On Sun, 2026-03-01 at 02:11 +0100, Matteo Croce wrote:
> > > ```
> > > KUnit test files are scattered across the tree and don't always match
> > > the F: patterns in the KERNEL UNIT TESTING FRAMEWORK MAINTAINERS entry.
> > > Add the KUnit maintainers automatically when the file path matches
> > > common KUnit naming conventions: _kunit.c, -kunit.c, _kunit_test.c,
> > > and -kunit-tests.c.
> >
> > nack.  There's no need to modify get_maintainer.
> >
> > Use F: patterns instead.
> >
> > diff --git a/MAINTAINERS b/MAINTAINERS
> > index 96e97d25e1c2..9796e184b9d5 100644
> > --- a/MAINTAINERS
> > +++ b/MAINTAINERS
> > @@ -13956,6 +13956,11 @@ F:     rust/kernel/kunit.rs
> >  F:     rust/macros/kunit.rs
> >  F:     scripts/rustdoc_test_*
> >  F:     tools/testing/kunit/
> > +F:     */*[_-]kunit*.c
> > +F:     */*/*[_-]kunit*.c
> > +F:     */*/*/*[_-]kunit*.c
> > +F:     */*/*/*/*[_-]kunit*.c
> > +F:     */*/*/*/*/*[_-]kunit*.c
> >
> >  KERNEL USERMODE HELPER
> >  M:     Luis Chamberlain <mcgrof@kernel.org>
> 
> So many rules covering all possible depths?
> Sounds good.
> 
> BTW at this point it would be useful to implement the ** glob operator,
> but maybe it's not worth it.

K: or N: patterns would be cleaner than both proposals here.


Thomas