[PATCH] coresight: etm4x: Don't bother the user with nonessential log message

James Clark james.clark at arm.com
Thu Aug 3 10:24:14 PDT 2023



On 03/08/2023 17:34, Suzuki K Poulose wrote:
> On 03/08/2023 10:08, James Clark wrote:
>>
>>
>> On 03/08/2023 07:22, Tian Tao wrote:
>>> Each cpu will print the following log when initializing
>>> ETM "coresight etm1: CPU1: etm v4.5 initialized", if there are a lot
>>> of cpus, e.g. 128. there will be a screen full of this log. replace
>>> dev_info with dev_dbg prints only when needed.
>>>
>>> Signed-off-by: Tian Tao <tiantao6 at hisilicon.com>
>>> ---
>>>   drivers/hwtracing/coresight/coresight-etm4x-core.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c
>>> b/drivers/hwtracing/coresight/coresight-etm4x-core.c
>>> index 7e307022303a..7b51e8594fd5 100644
>>> --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c
>>> +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c
>>> @@ -2033,7 +2033,7 @@ static int etm4_add_coresight_dev(struct
>>> etm4_init_arg *init_arg)
>>>         etmdrvdata[drvdata->cpu] = drvdata;
>>>   -    dev_info(&drvdata->csdev->dev, "CPU%d: %s v%d.%d initialized\n",
>>> +    dev_dbg(&drvdata->csdev->dev, "CPU%d: %s v%d.%d initialized\n",
>>>            drvdata->cpu, type_name, major, minor);
>>>         if (boot_enable) {
>>
>> Hi Tian,
>>
>> Do you think it's worth keeping a single info printout as well as the
>> per-cpu dbg one? I'm thinking that the only way to get these messages
>> now is to boot with debug printing on, and that is quite noisy. Because
>> it's too late to use debugfs to enable just coresight debug messages
>> after booting. So something like this:
>>
>>    dev_info_once(&drvdata->csdev->dev, "%s v%d.%d initialized\n",
> 
> I would rather remove all of this. One can always check the devices
> at /sys/bus/coresight/devices/etmX.
> 

It might be nice to keep the log message if you think something can go
wrong between here and registering in sysfs as it might give a clue. If
not, yeah I suppose we can remove them.

>>
>> Also I still get this printed out per-cpu, which is from
>> coresight-syscfg.c. Do you not see this? Might be worth doing the same
>> change there.
>>
>>    CSCFG registered etm0
>>    CSCFG registered etm1
>>    CSCFG registered etm2
>>    CSCFG registered etm3
> 
> +1, Please make them debug
> 
> 
> Suzuki
> 



More information about the linux-arm-kernel mailing list