Re: [PATCH 1/1] scripts/checkpatch.pl: to give more detailed warning message in case printk is used in any patch

[Levente Kurusa <levex@linux.com>]

Hi,

[+CC LKML, Joe]
[Leaving full copy for LKML, Joe]

On 03/02/2014 04:29 PM, Yogesh Chaudhari wrote:
> On 2 March 2014 20:50, Levente Kurusa <levex@linux.com> wrote:
>> Hi,
>>
>> On 03/02/2014 04:01 PM, Yogesh Chaudhari wrote:
>>> Based on the discussion here:
>>> https://lkml.org/lkml/2014/3/2/17
>>>
>>> I would like to propose this patch to improve the warning message in
>>> checkpatch.pl.  Comments/Suggestions on possible improvements are
>>> welcome.
>>>
>>>
>>> =========================================================
>>>
>>> This patch modifies scripts/checkpatch.pl to give more detailed
>>> warning message in case printk is used in any patch.
>>>
>>> Signed-off-by: Yogesh Chaudhari <mr.yogesh@gmail.com>
>>> ---
>>>  scripts/checkpatch.pl | 5 ++++-
>>>  1 file changed, 4 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
>>> index 464dcef..526f33aa 100755
>>> --- a/scripts/checkpatch.pl
>>> +++ b/scripts/checkpatch.pl
>>> @@ -2799,7 +2799,10 @@ sub process {
>>>              my $level2 = $level;
>>>              $level2 = "dbg" if ($level eq "debug");
>>>              WARN("PREFER_PR_LEVEL",
>>> -                 "Prefer netdev_$level2(netdev, ... then
>>> dev_$level2(dev, ... then pr_$level(...  to printk(KERN_$orig ...\n" .
>>> $herecurr);
>>
>> Whoops, that's a word-wrap!
>> Try using git-format-patch+git-send-email to send the patch.
> Ack, my bad, I will make this change.
> 
>>
>>> +                 "Order of preference for printing debug messages:
>>> +                 1. [subsystem]_$level2([subsystem]dev, ... eg
>>> netdev_$level2(netdev, ... for netdevice object
>>> +                 2. dev_$level2(dev, ... for drivers with struct device
>>> +                 3. pr_$level(...  to printk(KERN_$orig ...\n" . $herecurr);
>>>          }
>>>
>>>          if ($line =~ /\bpr_warning\s*\(/) {
>>>
>>
>> IMHO, this message is too big. The one we already have is nice and clean.
>> I would simply do: s/netdev/[subsystem]/ or something among the lines.
> 
> Seems proper way to go about. I agree that this makes it a bit too
> long, however, I was wondering, if there is a detailed
> documentation/information file about checkpatch (or patching in
> general where a detailed message would be accurate). Particularly,
> where we can make a note of proper way of using such debug calls.

Yes there is. Documentation/CodingStyle#Chapter13 is where you should
do stuff like that.

-- 
Regards,
Levente Kurusa