Date:   Thu, 2 Feb 2023 11:46:21 -0800
From:   Jakub Kicinski <kuba@kernel.org>
To:     Moshe Shemesh <moshe@nvidia.com>
Cc:     "David S. Miller" <davem@davemloft.net>,
        Jiri Pirko <jiri@nvidia.com>, <netdev@vger.kernel.org>,
        <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH net-next 7/7] devlink: Move devlink dev selftest code to
 dev
Message-ID: <20230202114621.3f32dae1@kernel.org>
In-Reply-To: <52392558-f79e-5980-4f10-47f111d69fc0@nvidia.com>
References: <1675349226-284034-1-git-send-email-moshe@nvidia.com>
        <1675349226-284034-8-git-send-email-moshe@nvidia.com>
        <20230202101712.15a0169d@kernel.org>
        <52392558-f79e-5980-4f10-47f111d69fc0@nvidia.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Thu, 2 Feb 2023 21:33:52 +0200 Moshe Shemesh wrote:
> On 02/02/2023 20:17, Jakub Kicinski wrote:
> > On Thu, 2 Feb 2023 16:47:06 +0200 Moshe Shemesh wrote:  
> >> Move devlink dev selftest callbacks and related code from leftover.c to
> >> file dev.c. No functional change in this patch.  
> > selftest I'd put in its own file. We don't want every command which
> > doesn't have a specific sub-object to end up in dev.c, right?
> > At least that was my initial thinking. I don't see any dependencies
> > between the selftest code and the rest of the dev code either.
> > WDYT?  
> 
> I thought as it is devlink dev selftest, the sub-object is dev. 
> Otherwise, what should be the rule here ?
> 
> How do we decide if it should get its own file ?

My thinking was that it should be much easier for newcomers to grok
"what does it take to implement a devlink command" if most of the
subcommands where in their own files, like in ethtool.

The implementation could have as well made selftest a subobject.
But I don't feel strongly, if noone agrees we can apply as is and 
see if dev.c does indeed start to grow out of proportion.