Message-ID: <97f6dc04-8a5f-c518-b804-992c7d46829b@linux.ibm.com>
Date:   Fri, 27 Jan 2023 08:40:57 -0600
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.6.0
Subject: Re: [PATCH] fsi: Add aliased device numbering
To:     Jeremy Kerr <jk@ozlabs.org>, linux-fsi@lists.ozlabs.org
Cc:     linux-kernel@vger.kernel.org, alistair@popple.id.au, joel@jms.id.au
References: <20230126223850.901302-1-eajames@linux.ibm.com>
 <8fa108c9dedf884eab6dacaf3e9680b7c66c4202.camel@ozlabs.org>
Content-Language: en-US
From:   Eddie James <eajames@linux.ibm.com>
In-Reply-To: <8fa108c9dedf884eab6dacaf3e9680b7c66c4202.camel@ozlabs.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 1/26/23 19:03, Jeremy Kerr wrote:
> Hi Eddie,
>
>> The I2C and SPI subsystems can use an aliased name to number the
>> device. Add similar support to the FSI subsystem for any device type.
> Sounds good!
>
> However, the majority of this patch seems to be the change to a (const)
> enum value, which you then convert to a string to generate the
> non-number component of the device name.
>
> Does:
>
>         dev_set_name(&scom->dev, "%s%d", fsi_get_dev_type_name(fsi_dev_scom), didx);
>
> have much of a benefit over:
>
>         dev_set_name(&scom->dev, "scom%d", didx);
>
> ?


Hi Jeremy,

I agree, it's not a benefit but my objective here is to make sure the 
device name is definitely the same as what the alias lookup is looking 
for. Perhaps it doesn't matter so much, and I can drop that part of the 
patch...

Thanks for looking!

Eddie


>
> Cheers,
>
>
> Jeremy