On Thu, 26 Jul 2018 20:54:26 +0200
Ladislav Michl [off-list ref] wrote:

On Thu, Jul 26, 2018 at 09:36:58AM -0400, David Rivshin wrote:

quoted

From: David Rivshin <redacted>

If a pwm-omap-dmtimer is probed before the dmtimer it uses, the platform
data won't be set yet.

Fixes: ac30751df953 ("ARM: OMAP: pdata-quirks: Remove unused timer pdata")
Cc: <redacted> # 4.17+
Signed-off-by: David Rivshin <redacted>
Acked-by: Pavel Machek <redacted>
Tested-by: Pavel Machek <redacted>
---
Changes in v2:
* Added Pavel's Acked-by/Tested-by [1]

[1] https://lkml.org/lkml/2018/7/16/346

 drivers/pwm/pwm-omap-dmtimer.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/pwm/pwm-omap-dmtimer.c b/drivers/pwm/pwm-omap-dmtimer.c
index 665da3c8fbceb..d3d7ea7a53146 100644
--- a/drivers/pwm/pwm-omap-dmtimer.c
+++ b/drivers/pwm/pwm-omap-dmtimer.c

@@ -264,8 +264,9 @@ static int pwm_omap_dmtimer_probe(struct platform_device *pdev)
 
 	timer_pdata = dev_get_platdata(&timer_pdev->dev);
 	if (!timer_pdata) {
-		dev_err(&pdev->dev, "dmtimer pdata structure NULL\n");
-		ret = -EINVAL;
+		dev_info(&pdev->dev,
+			 "dmtimer pdata structure NULL, deferring probe\n");  

This seems to be a bit verbose for EPROBE_DEFER case. Could we either remove
it as it is done later in pdata->request_by_node(timer) failure case or at
least make it dev_dbg? Otherwise thank you and with mentioned change
Acked-by: Ladislav Michl <redacted>

Hi Ladislav, thanks for the review.

I had grepped through other drivers and found no consistent pattern. Some
places used dev_err still, others reduced to one of dev_{warn,info,dbg}, 
and others no message at all. Some messages mentioned they are deferring 
the probe, other didn't. I was already getting a couple of dev_info from 
the pinctrl core code, so I went that way. I figured the message might be
useful to someone, but I don't feel strongly.

I personally would lean to dev_dbg if you think dev_info is too harsh, 
just in case someone's board suddenly isn't working after upgrade. But 
I'm certainly willing to remove the message entirely if you feel strongly,
or anyone else cares to weigh in.

quoted

+		ret = -EPROBE_DEFER;
 		goto put;
 	}
 

base-commit: d72e90f33aa4709ebecc5005562f52335e106a60
-- 
2.17.1