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: <6201612.znBgJCgWHB@phil>
Date:   Sat, 15 Feb 2020 12:11:19 +0100
From:   Heiko Stuebner <heiko@...ech.de>
To:     Jianqun Xu <jay.xu@...k-chips.com>
Cc:     linus.walleij@...aro.org, david.wu@...k-chips.com,
        kever.yang@...k-chips.com, linux-kernel@...r.kernel.org,
        linux-gpio@...r.kernel.org, linux-rockchip@...ts.infradead.org
Subject: Re: [PATCH v4 2/2] pinctrl: rockchip: split rockchip pinctrl driver by SoC type

Hi Jay,

Am Freitag, 17. Januar 2020, 09:13:58 CET schrieb Jianqun Xu:
> The pinctrl-rockchip driver grows larger by adding support for
> each new SoC, that make the kernel Image size too large since
> it only under one config named PINCTRL_ROCKCHIP.
> 
> This patch split driver in the form of core driver + soc driver,
> - pinctrl-rockchip.c defined an platform probe register function
> - pinctrl-rkxxxx.c init module by matching compatible name
> 
> For rockchip_defconfig, it needs to select all PINCTRL_RKxxxx to
> keep same with old driver.
> 
> For some special defconfig, it can only select one PINCTRL_RKxxxx.
> 
> Reviewed-by: Kever Yang <kever.yang@...k-chips.com>
> Signed-off-by: Jianqun Xu <jay.xu@...k-chips.com>
> ---
> changes since v3:
> - add base patch with directory change only, suggested by Robin
> - rebase patch
> 
> changes since v2:
> - remove rockchip_pinctrl_remove
> - rename rockchip_pinctrl_* to rockchip_pctrl_*
> - redule arguments for get_soc_data
> - add module author for each new driver files
> - add copyright for new driver files
> 
> changes since v1:
> - add rockchip_pinctrl_remove
> - remove unused head files in pinctrl-rockchip.h
> 
>  drivers/pinctrl/rockchip/Kconfig            |  114 +
>  drivers/pinctrl/rockchip/Makefile           |   14 +
>  drivers/pinctrl/rockchip/pinctrl-px30.c     |  224 ++
>  drivers/pinctrl/rockchip/pinctrl-rk2928.c   |   70 +
>  drivers/pinctrl/rockchip/pinctrl-rk3036.c   |   69 +
>  drivers/pinctrl/rockchip/pinctrl-rk3066a.c  |   72 +
>  drivers/pinctrl/rockchip/pinctrl-rk3066b.c  |   51 +
>  drivers/pinctrl/rockchip/pinctrl-rk3128.c   |  161 ++
>  drivers/pinctrl/rockchip/pinctrl-rk3188.c   |  147 ++
>  drivers/pinctrl/rockchip/pinctrl-rk3228.c   |  225 ++
>  drivers/pinctrl/rockchip/pinctrl-rk3288.c   |  210 ++
>  drivers/pinctrl/rockchip/pinctrl-rk3308.c   |  420 +++
>  drivers/pinctrl/rockchip/pinctrl-rk3328.c   |  272 ++
>  drivers/pinctrl/rockchip/pinctrl-rk3368.c   |  125 +
>  drivers/pinctrl/rockchip/pinctrl-rk3399.c   |  195 ++
>  drivers/pinctrl/rockchip/pinctrl-rockchip.c | 2547 ++-----------------
>  drivers/pinctrl/rockchip/pinctrl-rockchip.h |  388 +++
>  drivers/pinctrl/rockchip/pinctrl-rv1108.c   |  214 ++
>  18 files changed, 3149 insertions(+), 2369 deletions(-)

What Robin suggested, was doing this incrementally. So keep your patch1
but then do
- patch2: split out px30-pinctrl
- patch3: split out rk3288 pinctrl
- etc

Because even my mail client chokes on this massive 6000 line patch, so a
real review is actually very difficult.


> diff --git a/drivers/pinctrl/rockchip/Kconfig b/drivers/pinctrl/rockchip/Kconfig
> index 7a0077ca32dd..4873a05108f8 100644
> --- a/drivers/pinctrl/rockchip/Kconfig
> +++ b/drivers/pinctrl/rockchip/Kconfig
> @@ -5,8 +5,122 @@ if (ARCH_ROCKCHIP || COMPILE_TEST)
>  config PINCTRL_ROCKCHIP
>  	bool
>  	select PINMUX
> +	select PINCONF
>  	select GENERIC_PINCONF
> +	select GPIOLIB_IRQCHIP
>  	select GENERIC_IRQ_CHIP
>  	select MFD_SYSCON
>  
> +config PINCTRL_PX30
> +	tristate "PX30 pin controller driver"
> +	depends on GPIOLIB && OF
> +	select PINCTRL_ROCKCHIP

you might want to add a
	default y if ARM64
here
(similar default y if ARM for arm32 pinctrl drivers)

Because otherwise you're breaking peoples kernel configs and also
the default is to build a somewhat unified kernel in the default defconfigs,
so we want all matching pinctrl drivers by default and people then can
disable drivers if they really want to build a slimmed down kernel.

With the "if ARM" / "if ARM64" parts you even save some space by
default as well, as you build only the relevant drivers.


Heiko



Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ