linux-kernel - Re: [PATCH] checkpatch: add check for NONNETWORKING_BLOCK_COMMENT

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: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <0f037983-6dad-859c-3ded-f7a0c86953df@broadcom.com>
Date:   Wed, 10 Jun 2020 15:08:32 -0700
From:   Scott Branden <scott.branden@...adcom.com>
To:     Joe Perches <joe@...ches.com>, Andy Whitcroft <apw@...onical.com>
Cc:     BCM Kernel Feedback <bcm-kernel-feedback-list@...adcom.com>,
        linux-kernel@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH] checkpatch: add check for
 NONNETWORKING_BLOCK_COMMENT_STYLE

Hi Joe,

On 2020-06-10 2:42 p.m., Joe Perches wrote:
> On Wed, 2020-06-10 at 14:33 -0700, Scott Branden wrote:
>> Hi Joe,
>>
>> On 2020-06-10 2:16 p.m., Joe Perches wrote:
>>> On Wed, 2020-06-10 at 13:26 -0700, Scott Branden wrote:
>>>> NETWORKING_BLOCK_COMMENT_STYLE is supported by checkpatch but there
>>>> doesn't seem to be any check for the standard block comment style.
>>>> Add support for NONNETWORKING_BLOCK_COMMENT_STYLE to check for empty /*
>>>> on first line of non-networking block comments.
>>> I think there are _way_ too many instances of this form
>>> in non-networking code to enable this.
>>>
>>> $ git grep -P '^\s*/\*\s*\S.*[^\*][^\\]\s*$' -- '*.[ch]' | \
>>>     grep -v -P '^(drivers/net/|net/)' | \
>>>     wc -l
>>> 51407
>> That is true about many things that checkpatch now checks for that
>> didn't previously.
> Not in that quantity of uses it doesn't.
>
> I specifically did _not_ add this very same test
> when I added the other comment tests.
>
>> But, by adding to checkpatch the coding style clearly outlined in
>> coding-style.rst can be followed:
> Well, because there are _so_ many false positives
> that don't need change, I'm not adding this.
>
> As is, I'm nacking it.
>
> If you need it for your use, you should keep it in
> your own tree.
I think this has value for others.
>
>>>> diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
>>> []
>>>> @@ -3408,6 +3408,16 @@ sub process {
>>>>    			     "networking block comments don't use an empty /* line, use /* Comment...\n" . $hereprev);
>>>>    		}
>>>>    
>>>> +# Non-Networking with an empty initial /*
>>>> +		if ($realfile !~ m@^(drivers/net/|net/)@ &&
>>>> +		    $prevrawline =~ /^\+[ \t]*\/\*[ \t]/ &&
>>>> +		    $prevrawline !~ /\*\/[ \t]*$/ &&		#no trailing */
>>>> +		    $rawline =~ /^\+[ \t]*\*/ &&
>>>> +		    $realline > 2) {
>>>> +			WARN("NONNETWORKING_BLOCK_COMMENT_STYLE",
>>>> +			     "non-networking block comments use an empty /* on first line\n" . $hereprev);
> _Maybe_ this test _might_ be useful if it did a file/patch
> test and used CHK on file, but even then I'm very dubious.
>
> 			my $msg_level = \&WARN;
> 			$msg_level = \&CHK if ($file);
> 			&{msg_level}(etc...)
I can make such change.
As is, checkpatch basically follows the coding style documented for 
networking and has no check for outside.

I do not know the history of the 2 block coding styles but here's a 
radical thought:
Change the coding-style document to pick a single coding style for ALL 
block comment styles and
use that going forward in checkpatch instead?
>
>