lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <202407161505.A5AE57869@keescook>
Date: Tue, 16 Jul 2024 20:53:37 -0700
From: Kees Cook <kees@...nel.org>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: linux-kernel@...r.kernel.org, Alexey Dobriyan <adobriyan@...il.com>,
	"H.J. Lu" <hjl.tools@...il.com>,
	Brendan Higgins <brendanhiggins@...gle.com>,
	David Gow <davidgow@...gle.com>, Rae Moar <rmoar@...gle.com>,
	linux-kselftest@...r.kernel.org, kunit-dev@...glegroups.com
Subject: KUnit file naming conventions (was Re: [GIT PULL] execve updates for
 v6.11-rc1)

On Tue, Jul 16, 2024 at 01:10:41PM -0700, Linus Torvalds wrote:
> On Mon, 15 Jul 2024 at 09:21, Kees Cook <kees@...nel.org> wrote:
> >
> >  fs/exec.c                                   |  49 ++++++++--
> >  fs/exec_test.c                              | 141 ++++++++++++++++++++++++++++
> 
> I've pulled this, but *PLEASE* don't do this.
> 
> This screws up my workflow of just using tab-completion for filenames.
> As a result, I absolutely abhor anybody who uses the same base-name
> for different things.
> 
> No, this is not the first time it happens, and it won't be the last.
> And we had that same horrific pattern for fs/binfmt_elf_test.c from
> before, and I didn't notice because it's not a core file to me, and I
> seldom actually edit it.
> 
> I would suggest that people use the patterns from lib/, which is
> admittedly a bit schizophrenic in that you can either use
> "lib/kunit/*.c" (probably preferred) or "lib/test_xyz.c".
> 
> (Other subsystems use a "tests" subdirectory, so we do have a lot of
> different ways to deal with this).
> 
> Any of those models will keep the unit testing parts clearly separate,
> and not mess up basic command line workflows.
> 
> But do *not* use this "*_test.c" naming model. It's the worst of all
> possible worlds.
> 
> Please?

Oh, sure, no problem! I have no attachment to this convention at all;
I was trying to follow the Kunit docs:
https://docs.kernel.org/dev-tools/kunit/style.html#test-file-and-module-names

If I look at the existing naming, it's pretty scattered:

$ git grep '^static struct kunit_suite\b' | cut -d: -f1 | sort -u

/test/*		 7
/tests/*	47
*-test.[ch]	27
*_test.[ch]	27
test-*.c	 1
test_*.c	10
*-kunit.c	 1
*_kunit.c	17
kunit-*.c	 2
kunit_*.c	 1

Should we go with "put it all under a 'tests' subdirectory" ?

So for fs/exec_test.c and fs/binfmt_elf_test.c, perhaps fs/tests/exec.c
and fs/tests/binfmt_elf.c respectively?

And for the lib/*_kunit.c files, use lib/tests/*.c ?

Then we can update the docs, etc.

-- 
Kees Cook

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ