linux-kernel - Re: [PATCH] checkpatch: don't warn on every struct without const

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] [day] [month] [year] [list]

Date:   Tue, 10 Jan 2017 08:50:05 -0800
From:   Joe Perches <joe@...ches.com>
To:     Johannes Berg <johannes@...solutions.net>,
        Andrew Morton <akpm@...ux-foundation.org>
Cc:     linux-kernel@...r.kernel.org
Subject: Re: [PATCH] checkpatch: don't warn on every struct without
 const_structs file

On Tue, 2017-01-10 at 17:31 +0100, Johannes Berg wrote:
> > > I copied checkpatch elsewhere and ran it.
> > 
> > Why do you want to copy checkpatch "someplace else"?
> > Instead of copy, I think you should soft link it.
> 
> Well, no, I have to distribute it with that.
> 
> > > Regardless, the current code is utterly stupid - it prints a
> > > warning that it won't flag any structs, and then proceeds to flag
> > > all structs.
> > > 
> > > If you must, send a patch to abort() [whatever the perl equivalent
> > > is] when the file can't be found, but as it is, the code is just
> > > idiotic.
> > 
> > Maybe so.
> > 
> > As is, your patch description is incomplete because
> > it doesn't mention your use case.
> 
> I don't think that's relevant.

I do.  checkpatch is a script written for linux patches.

You are making use of the script outside of the linux tree.
You could just as well keep a local copy of this patch instead.

Your patch allows you to avoid this and the patch description
is relevant to why this is useful.

So your patch description is simply incorrect/incomplete:

---------------------------------------------------------------
The script says that it won't warn:
"No structs that should be const will be found [...]"

but then that doesn't work and it warns on every single struct
instead, since the regular expression ends up empty. Fix that
by checking that it's not empty first.
---------------------------------------------------------------

When used for linux, checkpatch works just fine and your
patch description describes a condition that doesn't happen.

> The script is internally inconsistent,
> as I do mention in the commit log, which is worth fixing.

Yeah, I think it's useful to fix.

Just make the commit message describe why the patch is
appropriate for non-linux uses.