Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH v2 03/12] bus: stm32-fmc2-ebi: add STM32 FMC2 EBI
 controller driver
To:     Boris Brezillon <boris.brezillon@collabora.com>
Cc:     Christophe Kerello <christophe.kerello@st.com>,
        miquel.raynal@bootlin.com, richard@nod.at, vigneshr@ti.com,
        lee.jones@linaro.org, robh+dt@kernel.org, mark.rutland@arm.com,
        tony@atomide.com, devicetree@vger.kernel.org,
        linux-kernel@vger.kernel.org, linux-mtd@lists.infradead.org,
        linux-stm32@st-md-mailman.stormreply.com
References: <1586966256-29548-1-git-send-email-christophe.kerello@st.com>
 <1586966256-29548-4-git-send-email-christophe.kerello@st.com>
 <20200416215328.53982fef@collabora.com>
 <e4a21df4-b01b-1215-7a3b-fea94f099b8f@st.com>
 <c69f86f4-5904-3f35-dac1-87080402a096@denx.de>
 <20200417174140.00456f36@collabora.com>
From:   Marek Vasut <marex@denx.de>
Message-ID: <53011403-49c4-dfcb-2f8c-11cf1ac5e2c1@denx.de>
Date:   Fri, 17 Apr 2020 17:44:41 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.5.0
MIME-Version: 1.0
In-Reply-To: <20200417174140.00456f36@collabora.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 4/17/20 5:41 PM, Boris Brezillon wrote:
> On Fri, 17 Apr 2020 17:31:49 +0200
> Marek Vasut <marex@denx.de> wrote:
> 
>> On 4/17/20 5:29 PM, Christophe Kerello wrote:
>>>
>>>
>>> On 4/16/20 9:53 PM, Boris Brezillon wrote:  
>>>> On Wed, 15 Apr 2020 17:57:27 +0200
>>>> Christophe Kerello <christophe.kerello@st.com> wrote:
>>>>  
>>>>> The driver adds the support for the STMicroelectronics FMC2 EBI
>>>>> controller
>>>>> found on STM32MP SOCs.
>>>>>
>>>>> Signed-off-by: Christophe Kerello <christophe.kerello@st.com>
>>>>> Tested-by: Marek Vasut <marex@denx.de>
>>>>> ---
>>>>> Changes in v2:
>>>>>   - call 2 APIs to manage FMC2 enable/disable instead of ops
>>>>>   - call 2 APIs to manage FMC2 NWAIT shared signal instead of ops
>>>>>
>>>>>   drivers/bus/Kconfig          |   11 +
>>>>>   drivers/bus/Makefile         |    1 +
>>>>>   drivers/bus/stm32-fmc2-ebi.c | 1091
>>>>> ++++++++++++++++++++++++++++++++++++++++++  
>>>>
>>>> Hm, I see that other memory bus controller drivers are placed under
>>>> drivers/memory/, any reason for choosing drivers/bus/? If that's where
>>>> we want to have all generic memory bus controllers to live it might be
>>>> worth moving existing drivers to the drivers/bus/ directory at some
>>>> point.
>>>>  
>>>
>>> Hi Boris,
>>>
>>> I see this controller as an external bus interface as we are able to
>>> attach different devices on it, like a PSRAM, an ethernet controller, a
>>> FPGA, a LCD display, ...
>>>
>>> When I had a look at bus/Kconfig file, I have found similar drivers
>>> (like IMX_WEIM or QCOM_EBI2 drivers). These drivers are able to connect
>>> devices like NAND Flash, SRAM, ethernet adapters, FPGAs and LCD displays
>>> as it is written in the Kconfig file.
>>>
>>> But, after checking memory/Kconfig file, it is also possible to find
>>> similar drivers (like ATMEL_EBI driver that is inspired by the WEIM bus
>>> driver).
>>>
>>> So, I will follow the recommendation and I will move it to
>>> drivers/memory folder if it is the place where this driver should be.  
>>
>> The EBI isn't a memory bus, you can very well have a non-memory device,
>> like an external parallel bus memory controller on it. So why should it
>> be in drivers/memory ?
> 
> Just like most of those so-called memory bus controllers we have under
> drivers/memory/ => they are most of the time used to interact with
> memories but can also be used for other things (Christophe listed a few
> uses cases). And I'm not pushing one way or the other, all I said was
> that we should makes things consistent instead of having half of
> those controller drivers in drivers/memory/ and the other half in
> drivers/bus/.

I agree with that.