Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH net] devlink: Fix devlink_param_driverinit_value_set()
 stub return code
To:     Moshe Shemesh <moshe@mellanox.com>,
        "David S. Miller" <davem@davemloft.net>
Cc:     Jiri Pirko <jiri@mellanox.com>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <1536066265-18641-1-git-send-email-moshe@mellanox.com>
 <2bf31ad0-c3f8-cdfd-8b96-1a8e3280064d@gmail.com>
 <7fd426d1-61a9-7f45-bf26-05bd9e180277@mellanox.com>
From:   David Ahern <dsahern@gmail.com>
Message-ID: <b660e196-9277-f66a-5be8-4288f73b9305@gmail.com>
Date:   Wed, 5 Sep 2018 08:26:15 -0600
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.13; rv:52.0)
 Gecko/20100101 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <7fd426d1-61a9-7f45-bf26-05bd9e180277@mellanox.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 9/5/18 6:48 AM, Moshe Shemesh wrote:
> 
> 
> On 9/4/2018 7:13 PM, David Ahern wrote:
>> On 9/4/18 7:04 AM, Moshe Shemesh wrote:
>>> The stub function returned -EOPNOTSUPP while CONFIG_NET_DEVLINK is off.
>>> It caused false warning during driver load. Driver needs to update
>>> devlink on a parameter value if devlink module is there, if not it
>>> doesn't need any error code.
>>>
>>> Fixes: ec01aeb1803e ("devlink: Add support for get/set driverinit
>>> value")
>>> Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
>>> Acked-by: Jiri Pirko <jiri@mellanox.com>
>>> ---
>>>   include/net/devlink.h | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/include/net/devlink.h b/include/net/devlink.h
>>> index b9b89d6..b467357 100644
>>> --- a/include/net/devlink.h
>>> +++ b/include/net/devlink.h
>>> @@ -781,7 +781,7 @@ static inline bool
>>> devlink_dpipe_table_counter_enabled(struct devlink *devlink,
>>>   devlink_param_driverinit_value_set(struct devlink *devlink, u32
>>> param_id,
>>>                      union devlink_param_value init_val)
>>>   {
>>> -    return -EOPNOTSUPP;
>>> +    return 0;
>>>   }
>>>     static inline void
>>>
>>
>> This should be handled by the driver -- check for -EOPNOTSUPP and not
>> log an error.
> 
> This is a stub inline function.
> The return value would be ambiguous as the non-stub function can also
> return -EOPNOTSUPP, in case the driver-init mode is not supported for a
> parameter.
>>
>> devlink is generic infrastructure. If a call is made and the operation
>> is not supported, then devlink should return an error.
> 
> In general the stub functions should take care that the driver won't
> feel the missing code as much as possible. That's why
> devlink_params_register() returns success and so should this function.
>>

A driver is accessing core infrastructure to set a value; core infra can
not because the code is not compiled in. The driver should be told that
the option is not enabled, and it is at the moment with the -EOPNOTSUPP
return code.

That is similar to devlink_dpipe_table_resource_set which returns
-EOPNOTSUPP when devlink is compiled out.