[PATCH] pwm: samsung: Document new member .channel in struct samsung_pwm_chip

[Krzysztof Kozlowski]

On 24/10/2023 18:46, Uwe Kleine-König wrote:
> On Tue, Oct 24, 2023 at 04:58:22PM +0200, Krzysztof Kozlowski wrote:
>> On 24/10/2023 16:55, Uwe Kleine-König wrote:
>>> Hello,
>>>
>>> On Fri, Oct 13, 2023 at 07:27:50PM +0200, Uwe Kleine-König wrote:
>>>> On Fri, Oct 13, 2023 at 03:29:35PM +0200, Thierry Reding wrote:
>>>>> On Thu, 12 Oct 2023 23:02:29 +0200, Uwe Kleine-König wrote:
>>>>>> Fixes: 4c9548d24c0d ("pwm: samsung: Put per-channel data into driver data")
>>>>>
>>>>> Applied, thanks!
>>>>>
>>>>> [1/1] pwm: samsung: Document new member .channel in struct samsung_pwm_chip
>>>>>       commit: 4bb36d126cb3147d6bbfd00242a5b846dacad595
>>>>
>>>> You might want to change 4c9548d24c0d to e3fe982b2e4e now that you
>>>> rewrote your for-next branch.
>>>
>>> This is still open. I wonder there is no automated check that warns if
>>> there is a Fixes: line in next that doesn't refer to an ancestor.
>>
>> I am using Greg's/Stephen's scripts in commit hooks:
>> https://github.com/krzk/tools/blob/master/linux/git-hooks-post-commit
>> https://github.com/krzk/tools/blob/master/linux/verify_fixes.sh
> 
> Skimming throud the verify_fixes script I'd say the following addition
> to verify_fixes should do:
> 
> diff --git a/linux/verify_fixes.sh b/linux/verify_fixes.sh
> index f22384bb6bb8..12e73da82dd4 100755
> --- a/linux/verify_fixes.sh
> +++ b/linux/verify_fixes.sh
> @@ -100,6 +100,13 @@ verify_fixes()
>  				continue
>  			fi
>  
> +			if ! git merge-base --is-ancestor "$sha" "$c"; then
> +				printf '%s%s\t\t- %s\n' "$commit_msg" "$fixes_msg" 'Target SHA should be an ancestor of your tree'
> +				commit_msg=''
> +				error=1
> +				continue
> +			fi
> +
>  			if [ "${#sha}" -lt 12 ]; then
>  				msg="${msg:+${msg}${nl}}${tab}${tab}- SHA1 should be at least 12 digits long${nl}${tab}${tab}  Can be fixed by setting core.abbrev to 12 (or more) or (for git v2.11${nl}${tab}${tab}  or later) just making sure it is not set (or set to \"auto\")."
>  			fi

Oh, thanks! It makes sense.

> 
> It has quite a few hits in next ...
> 
>> Happy to receive more ideas during:
>> https://lpc.events/event/17/contributions/1498/
> 
> I hope you pick up this idea even without me repeating this suggestion
> there :-)
> 
Best regards,
Krzysztof