Re: [PATCH] PM / OPP: Remove confusing error message in of_cpumask_init_opp_table()

["Rafael J. Wysocki" <rjw@rjwysocki.net>]

On Wednesday, October 14, 2015 05:52:23 PM Thomas Petazzoni wrote:
> The of_cpumask_init_opp_table() function will print an error message
> (with pr_err) if it cannot find the OPP table for a certain CPU in the
> Device Tree.
> 
> There are users of the cpufreq-dt driver (which is the one calling
> of_cpumask_init_opp_table) that do not have the OPP points defined in
> the Device Tree. Instead, such users dynamically create the OPP table
> at boot time depending on the system configuration. Such a case is
> planned in the cpufreq-dt driver, which on purpose ignores the return
> value of of_cpumask_init_opp_table() with the following comment: "OPPs
> might be populated at runtime, don't check for error here".
> 
> For such platforms, the of_cpumask_init_opp_table() prints a spurious
> and confusing error message for each CPU:
> 
> [    1.749548] of_cpumask_init_opp_table: couldn't find opp table for cpu:0, -19
> [    1.756784] of_cpumask_init_opp_table: couldn't find opp table for cpu:1, -19
> [    1.764031] of_cpumask_init_opp_table: couldn't find opp table for cpu:2, -19
> [    1.771268] of_cpumask_init_opp_table: couldn't find opp table for cpu:3, -19
> 
> This is confusing because everything is working fine, cpufreq works
> and it knows the OPP table that was registered at boot time (on
> Marvell Armada XP):
> 
> $ cat /sys/devices/system/cpu/cpu0/cpufreq/stats/time_in_state
> 666500 1884
> 1333000 23333
> 
> To avoid this confusion, this patch simply deletes the error message.
> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

That's only present in linux-next now, right?

Thanks,
Rafael