lkml.org 
[lkml]   [2016]   [Mar]   [14]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] rtc: s3c: Don't print an error on probe deferral
On 14/03/2016 at 16:59:33 -0300, Javier Martinez Canillas wrote :
> Hello Joe,
>
> On 03/14/2016 04:38 PM, Joe Perches wrote:
> > On Mon, 2016-03-14 at 16:31 -0300, Javier Martinez Canillas wrote:
> >> On 03/14/2016 04:11 PM, Joe Perches wrote:> > On Mon, 2016-03-14 at 16:05 -0300, Javier Martinez Canillas wrote:
> >>>>
> >>>> The clock and source clock looked up by the driver may not be available
> >>>> just because the clock controller driver was not probed yet so printing
> >>>> an error in this case is not correct and only adds confusion to users.
> >>>>
> >>>> However, knowing that a driver's probe was deferred may be useful so it
> >>>> can be printed as debug information.
> >>> []
> >>>>
> >>>> diff --git a/drivers/rtc/rtc-s3c.c b/drivers/rtc/rtc-s3c.c
> >>> []
> >>>>
> >>>> @@ -501,18 +501,27 @@ static int s3c_rtc_probe(struct platform_device *pdev)
> >>>>
> >>>> info->rtc_clk = devm_clk_get(&pdev->dev, "rtc");
> >>>> if (IS_ERR(info->rtc_clk)) {
> >>>> - dev_err(&pdev->dev, "failed to find rtc clock\n");
> >>>> - return PTR_ERR(info->rtc_clk);
> >>>> + ret = PTR_ERR(info->rtc_clk);
> >>>> + if (ret != -EPROBE_DEFER)
> >>>> + dev_err(&pdev->dev, "failed to find rtc clock\n");
> >>>> + else
> >>>> + dev_dbg(&pdev->dev, "probe deferred due rtc clock\n");
> >>>> + return ret;
> >>>> }
> >>>> clk_prepare_enable(info->rtc_clk);
> >>>>
> >>>> if (info->data->needs_src_clk) {
> >>>> info->rtc_src_clk = devm_clk_get(&pdev->dev, "rtc_src");
> >>>> if (IS_ERR(info->rtc_src_clk)) {
> >>>> - dev_err(&pdev->dev,
> >>>> - "failed to find rtc source clock\n");
> >>>> + ret = PTR_ERR(info->rtc_src_clk);
> >>>> + if (ret != -EPROBE_DEFER)
> >>>> + dev_err(&pdev->dev,
> >>>> + "failed to find rtc source clock\n");
> >>>> + else
> >>>> + dev_dbg(&pdev->dev,
> >>>> + "probe deferred due rtc source clock\n");
> >>>> clk_disable_unprepare(info->rtc_clk);
> >>>> - return PTR_ERR(info->rtc_src_clk);
> >>>> + return ret;
> >>>> }
> >>>> clk_prepare_enable(info->rtc_src_clk);
> >>>> }
> >>> Maybe the debug logging messages could be object->action like:
> >>>
> >>> rtc clock probe deferred
> >>> rtc source clock probe deferred
> >>>
> >> I found your suggested messages harder to read and more confusing. The
> >> action that happens is a probe function deferral and that is caused by
> >> a missing resource needed by the driver (clocks in this case).
> >>
> >> But your messages seems to imply that the probe deferred action happens
> >> to a clock, it sounds like "rtc clock disabled" and that's not correct.
> >
> > OK, then please change "due" to "due to" or "for" in your messages
> > because they make little sense now.
> >
>
> I don't think they make little sense now since even a non-native english
> speaker like me can understand it :)
>
> But yes, it's cryptic at the very least. That's the problem with long text
> and the 80 char limit to make checkpatch.pl happy. I guess I can just move
> the message a little bit even if that will make to not be properly aligned.
>

checkpatch will not complain for messages but it will definitively
complain if they are not properly aligned:)


--
Alexandre Belloni, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

\
 
 \ /
  Last update: 2016-03-14 22:01    [W:0.374 / U:0.084 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site