Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <d68c9137-f485-e0a0-dfc3-c72c26d0433a@gmail.com>
Date:   Tue, 20 Dec 2022 00:12:52 +0900
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Thunderbird/91.13.1
From:   Tokunori Ikegami <ikegami.t@gmail.com>
Subject: Re: [PATCH] mtd: cfi: allow building spi-intel standalone
To:     Arnd Bergmann <arnd@kernel.org>,
        Miquel Raynal <miquel.raynal@bootlin.com>,
        Richard Weinberger <richard@nod.at>,
        Vignesh Raghavendra <vigneshr@ti.com>
Cc:     Arnd Bergmann <arnd@arndb.de>, Lee Jones <lee.jones@linaro.org>,
        Mauro Lima <mauro.lima@eclypsium.com>,
        Tudor Ambarus <tudor.ambarus@microchip.com>,
        Mika Westerberg <mika.westerberg@linux.intel.com>,
        Boris Brezillon <boris.brezillon@collabora.com>,
        linux-mtd@lists.infradead.org, linux-kernel@vger.kernel.org
References: <20221215163950.649791-1-arnd@kernel.org>
Content-Language: en-US
In-Reply-To: <20221215163950.649791-1-arnd@kernel.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 2022/12/16 1:39, Arnd Bergmann wrote:
> From: Arnd Bergmann<arnd@arndb.de>
>
> When MTD or MTD_CFI_GEOMETRY is disabled, the spi-intel driver
> fails to build, as it includes the shared CFI header:
The header file linux/mtd/cfi.h is included by the head file 
linux/mtd/spi-nor.h so seems the warning caused.
Is it really necessary to inculude linux/mtd/cfi.h from linux/mtd/spi-nor.h?
(It seems that it can be resolved the issue by removing the including 
cfi.h from spi-nor.h if unnecessary.)
>
> include/linux/mtd/cfi.h:62:2: error: #warning No CONFIG_MTD_CFI_Ix selected. No NOR chip support can work. [-Werror=cpp]
>     62 | #warning No CONFIG_MTD_CFI_Ix selected. No NOR chip support can work.
>
> Use an #ifdef here to guard the settings so the header can be included
> in all configurations.
>
> Fixes: e23e5a05d1fd ("mtd: spi-nor: intel-spi: Convert to SPI MEM")
> Signed-off-by: Arnd Bergmann<arnd@arndb.de>
> ---
>   include/linux/mtd/cfi.h | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/include/linux/mtd/cfi.h b/include/linux/mtd/cfi.h
> index d88bb56c18e2..0e87b6e2faad 100644
> --- a/include/linux/mtd/cfi.h
> +++ b/include/linux/mtd/cfi.h
> @@ -15,6 +15,7 @@
>   #include <linux/mtd/cfi_endian.h>
>   #include <linux/mtd/xip.h>
>   
> +#if IS_ENABLED(CONFIG_MTD)

Is it enough to check only CONFIG_MTD here as not necessary to check if 
MTD_CFI_GEOMETRY also?
By the way it seems also that it is needed to use #ifdef only for the 
cfi_interleave warning checking as below.

#if IS_ENABLED(CONFIG_MTD) && IS_ENABLED(CONFIG_MTD_CFI_GEOMETRY)
#ifndef cfi_interleave
#warning No CONFIG_MTD_CFI_Ix selected. No NOR chip support can work.
static inline int cfi_interleave(void *cfi)
{
     BUG();
     return 0;
}
#endif /* cfi_interleave */
#endif /* IS_ENABLED(CONFIG_MTD) && IS_ENABLED(CONFIG_MTD_CFI_GEOMETRY) */

Note: The cfi_interleave warning checking was implemented by the 
following commit.
   241651d04d67 ("[MTD] Fix CFI build error when no map width or 
interleave supported")

>   #ifdef CONFIG_MTD_CFI_I1
>   #define cfi_interleave(cfi) 1
>   #define cfi_interleave_is_1(cfi) (cfi_interleave(cfi) == 1)
> @@ -88,7 +89,7 @@ static inline int cfi_interleave_supported(int i)
>   		return 0;
>   	}
>   }
> -
> +#endif
>   
>   /* NB: these values must represents the number of bytes needed to meet the
>    *     device type (x8, x16, x32).  Eg. a 32 bit device is 4 x 8 bytes.