Re: [PATCH 03/10] mfd / platform: cros_ec: Miscellaneous character device to talk with the EC

[Lee Jones]

On Thu, 06 Jun 2019, Randy Dunlap wrote:
> > On Thu, 2019-06-06 at 16:51 +0200, Greg Kroah-Hartman wrote:
> >> Again, don't be noisy, it's not hard, and is how things have been
> >> trending for many years now.
> 
> Ack that.

Not to say that this particular print is acceptable, but there are
places where a low-level (dbg/info) print on successful probe is
helpful.  Driver initialisation is important!

There's a big difference between drivers 'being noisy', spilling all
sorts of information that may well be useful or interesting to a
driver developer, but has little value to anyone else, and providing a
single print to say that a device has been detected and successfully
initialised/probed.

I recently fell victim to a silent, but fully functional device.
Successful device initialisation should not be silent when debugging
has been set to the highest level IMHO.

And yes, of course turning on debugging for Driver Core works, but is
not practical for all cases and is certainly not the first port of
call when figuring out why initialisation seems to be failing for a
single particular device.

Truly surplus churn should absolutely be removed from the boot log, or
at the very least downgraded, leaving only truly useful information
such as highlighting a newly detected device for example.  If the user
wants an even more silent boot log, they should turn the log level
down a notch.  That is why we have log levels after all.  Simply
removing all useful prints regardless of log-level is not the way to
go IMHO.

-- 
Lee Jones [æçæ]
Linaro Services Technical Lead
Linaro.org â Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog