Re: [patch net-next v2 05/12] mlxsw: core_linecards: Expose HW revision and INI version

[Jiri Pirko <jiri@resnulli.us>]

Wed, Jul 20, 2022 at 02:43:17PM CEST, idosch@nvidia.com wrote:
>On Wed, Jul 20, 2022 at 02:27:40PM +0200, Jiri Pirko wrote:
>> Wed, Jul 20, 2022 at 10:56:35AM CEST, idosch@nvidia.com wrote:
>> >On Tue, Jul 19, 2022 at 08:48:40AM +0200, Jiri Pirko wrote:
>> >> +int mlxsw_linecard_devlink_info_get(struct mlxsw_linecard *linecard,
>> >> +				    struct devlink_info_req *req,
>> >> +				    struct netlink_ext_ack *extack)
>> >> +{
>> >> +	char buf[32];
>> >> +	int err;
>> >> +
>> >> +	mutex_lock(&linecard->lock);
>> >> +	if (WARN_ON(!linecard->provisioned)) {
>> >> +		err = 0;
>> >
>> >Why not:
>> >
>> >err = -EINVAL;
>> >
>> >?
>> 
>> Well, a) this should not happen. No need to push error to the user for
>> this as the rest of the info message is still fine.
>
>Not sure what you mean by "the rest of the info message is still fine".
>Which info message? If the line card is not provisioned, then it
>shouldn't even have a devlink instance and it will not appear in
>"devlink dev info" dump.
>
>I still do not understand why this error is severe enough to print a
>WARNING to the kernel log, but not emit an error code to user space.

As I wrote, WARN_ON was a leftover.

>
>> 
>> 
>> >
>> >> +		goto unlock;
>> >> +	}
>> >> +
>> >> +	sprintf(buf, "%d", linecard->hw_revision);
>> >> +	err = devlink_info_version_fixed_put(req, "hw.revision", buf);
>> >> +	if (err)
>> >> +		goto unlock;
>> >> +
>> >> +	sprintf(buf, "%d", linecard->ini_version);
>> >> +	err = devlink_info_version_running_put(req, "ini.version", buf);
>> >> +	if (err)
>> >> +		goto unlock;
>> >> +
>> >> +unlock:
>> >> +	mutex_unlock(&linecard->lock);
>> >> +	return err;
>> >> +}
>> >> +
>> >>  static int
>> >>  mlxsw_linecard_provision_set(struct mlxsw_linecard *linecard, u8 card_type,
>> >>  			     u16 hw_revision, u16 ini_version)
>> >> -- 
>> >> 2.35.3
>> >>