[PATCH 2/3] nvme: parse dsm command detailly when tracing

Chaitanya Kulkarni chaitanyak at nvidia.com
Tue Jan 30 00:05:25 PST 2024


>> I'm really not sure if we need to decode idw/idr as I've not seen those
>> fields so far
>> used by anybody maybe I'm not aware here, but can you please provide a
>> usecase where
>> you need these fields to be decoded ?
>>
>> -ck
> 
> Mostly we only use "ad" filed, but sometimes our tester will set idw/idr to
> 
> see whether any errors have been reported by our target. Well this is a
> 
> low-persuasive usecase, but I still insist it's better to follow the spec.
> 
> Best regadrs,
> 
> Guixin Liu
> 
>>

But why ? do you have a valid implementation that shows clear benefits
for idw/idr ? if not I'm not sure why we need to decode these fields,
unless others think it's a good idea let's not do that.

Following spec only makes sense where there is a valid implementation
for the fields we are printing and not sure if it applies to idw/idr ...

I'm fine with AD and NR they are useful ...

-ck




More information about the Linux-nvme mailing list