linux-kernel - Re: [PATCH v4 6/6] hwmon: (spd5118) Add configuration option for auto-detection

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <452386bd-8238-4fac-ad6d-6a8f096ecc35@t-8ch.de>
Date: Tue, 4 Jun 2024 06:37:47 +0200
From: Thomas Weißschuh <linux@...ssschuh.net>
To: Guenter Roeck <linux@...ck-us.net>
Cc: linux-hwmon@...r.kernel.org, linux-i2c@...r.kernel.org, 
	linux-kernel@...r.kernel.org, devicetree@...r.kernel.org, 
	Krzysztof Kozlowski <krzk+dt@...nel.org>, Wolfram Sang <wsa+renesas@...g-engineering.com>, 
	René Rebe <rene@...ctcode.de>, Armin Wolf <W_Armin@....de>, 
	Stephen Horvath <s.horvath@...look.com.au>
Subject: Re: [PATCH v4 6/6] hwmon: (spd5118) Add configuration option for
 auto-detection

On 2024-06-03 21:02:37+0000, Guenter Roeck wrote:
> With SPD5118 chip detection for the most part handled by the i2c-smbus
> core using DMI information, the spd5118 driver no longer needs to
> auto-detect spd5118 compliant chips.
> 
> Auto-detection by the driver is still needed on systems with no DMI support
> or on systems with more than eight DIMMs and can not be removed entirely.
> However, it affects boot time and introduces the risk of mis-identifying
> chips. Add configuration option to be able to disable it on systems where
> chip detection is handled outside the driver.
> 
> Cc: Wolfram Sang <wsa+renesas@...g-engineering.com>
> Signed-off-by: Guenter Roeck <linux@...ck-us.net>
> ---
> v4: New patch
> 
>  drivers/hwmon/Kconfig   | 18 ++++++++++++++++++
>  drivers/hwmon/spd5118.c |  4 +++-
>  2 files changed, 21 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
> index 7a84e7637b51..0bb1bdee3e43 100644
> --- a/drivers/hwmon/Kconfig
> +++ b/drivers/hwmon/Kconfig
> @@ -2185,6 +2185,7 @@ config SENSORS_SPD5118
>  	tristate "SPD5118 Compliant Temperature Sensors"
>  	depends on I2C
>  	select REGMAP_I2C
> +	select SENSORS_SPD5118_DETECT if !DMI
>  	help
>  	  If you say yes here you get support for SPD5118 (JEDEC JESD300)
>  	  compliant temperature sensors. Such sensors are found on DDR5 memory
> @@ -2193,6 +2194,23 @@ config SENSORS_SPD5118
>  	  This driver can also be built as a module. If so, the module
>  	  will be called spd5118.
>  
> +config SENSORS_SPD5118_DETECT
> +	bool "Enable detect function"
> +	depends on SENSORS_SPD5118
> +	default y
> +	help
> +	  If enabled, the driver auto-detects if a chip in the SPD address
> +	  range is compliant to the SPD51888 standard and auto-instantiates
> +	  if that is the case. If disabled, SPD5118 compliant devices have
> +	  to be instantiated by other means. On systems with DMI support
> +	  this will typically be done from DMI DDR detection code in the
> +	  I2C SMBus subsystem.
> +	  Disabling the detect function will speed up boot time and reduce
> +	  the risk of mis-detecting SPD5118 compliant devices. In general
> +	  it should only be enabled if necessary.
> +
> +	  If unsure, say Y.

The combination of

"In general it should only be enabled if necessary."

and

"default y" / "If unsure, say Y."

looks weird.


Also right now it is not possible to disable detection on non-DMI
configurations. But when using OF, custom kernel code or userspace
instantiation then neither DMI nor CONFIG_DETECT are necessary.

The following would support those usecases, too:

config SENSORS_SPD5118_DETECT
	bool "Enable detect function"
	depends on SENSORS_SPD5118
	default !DMI

(And no "select SENSORS_SPD5118_DETECT if !DMI")

> +
>  config SENSORS_TC74
>  	tristate "Microchip TC74"
>  	depends on I2C
> diff --git a/drivers/hwmon/spd5118.c b/drivers/hwmon/spd5118.c
> index 5cb5e52c0a38..19d203283a21 100644
> --- a/drivers/hwmon/spd5118.c
> +++ b/drivers/hwmon/spd5118.c
> @@ -313,7 +313,7 @@ static bool spd5118_vendor_valid(u8 bank, u8 id)
>  }
>  
>  /* Return 0 if detection is successful, -ENODEV otherwise */
> -static int spd5118_detect(struct i2c_client *client, struct i2c_board_info *info)
> +static int __maybe_unused spd5118_detect(struct i2c_client *client, struct i2c_board_info *info)
>  {
>  	struct i2c_adapter *adapter = client->adapter;
>  	int regval;
> @@ -647,7 +647,9 @@ static struct i2c_driver spd5118_driver = {
>  	},
>  	.probe		= spd5118_probe,
>  	.id_table	= spd5118_id,
> +#ifdef CONFIG_SENSORS_SPD5118_DETECT
>  	.detect		= spd5118_detect,
> +#endif
>  	.address_list	= normal_i2c,

.address_list is also only needed with CONFIG_SENSORS_SPD5118_DETECT.


If you use

.detect         = IS_ENABLED(CONFIG_SENSORS_SPD5118_DETECT) ?  spd5118_detect : NULL,
.address_list   = IS_ENABLED(CONFIG_SENSORS_SPD5118_DETECT) ?  normal_i2c : NULL,

then the need for __maybe_unused goes away and type checking is a tiny
bit better.