linux-kernel - Re: [PATCH] pinctrl: ingenic: Garbage-collect code paths for SoCs disabled by config

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] [day] [month] [year] [list]

Message-ID: <CAHp75Vfas=H7Vb+mJor-LWckUpE8QUdD8Yauq4mdG4OY+7dfMg@mail.gmail.com>
Date:   Wed, 16 Mar 2022 16:47:24 +0200
From:   Andy Shevchenko <andy.shevchenko@...il.com>
To:     Paul Cercueil <paul@...pouillou.net>
Cc:     Linus Walleij <linus.walleij@...aro.org>, od@...ndingux.net,
        "open list:BROADCOM NVRAM DRIVER" <linux-mips@...r.kernel.org>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] pinctrl: ingenic: Garbage-collect code paths for SoCs
 disabled by config

On Wed, Mar 16, 2022 at 4:11 AM Paul Cercueil <paul@...pouillou.net> wrote:
>
> By being a bit smarter about how the SoC version checks are performed,
> it is possible to have all the code paths that correspond to SoCs
> disabled in the kernel config automatically marked as dead code by the
> compiler, and therefore garbage-collected.
>
> With this patch, when compiling a kernel that only targets the JZ4760
> for instance, the driver is now about 4.5 KiB smaller.

...

> +static const u32 enabled_socs =

If you make it unsigned long, it would be easier to switch to bitmap
APIs if needed in the future.

> +       IS_ENABLED(CONFIG_MACH_JZ4730) << ID_JZ4730 |
> +       IS_ENABLED(CONFIG_MACH_JZ4740) << ID_JZ4740 |
> +       IS_ENABLED(CONFIG_MACH_JZ4725B) << ID_JZ4725B |
> +       IS_ENABLED(CONFIG_MACH_JZ4750) << ID_JZ4750 |
> +       IS_ENABLED(CONFIG_MACH_JZ4755) << ID_JZ4755 |
> +       IS_ENABLED(CONFIG_MACH_JZ4760) << ID_JZ4760 |
> +       IS_ENABLED(CONFIG_MACH_JZ4770) << ID_JZ4770 |
> +       IS_ENABLED(CONFIG_MACH_JZ4775) << ID_JZ4775 |
> +       IS_ENABLED(CONFIG_MACH_JZ4780) << ID_JZ4780 |
> +       IS_ENABLED(CONFIG_MACH_X1000) << ID_X1000 |
> +       IS_ENABLED(CONFIG_MACH_X1500) << ID_X1500 |
> +       IS_ENABLED(CONFIG_MACH_X1830) << ID_X1830 |
> +       IS_ENABLED(CONFIG_MACH_X2000) << ID_X2000 |
> +       IS_ENABLED(CONFIG_MACH_X2100) << ID_X2100;

...

> +is_soc_or_above(const struct ingenic_pinctrl *jzpc, enum jz_version version)
> +{
> +       return (enabled_socs >> version) &&
> +               (!(enabled_socs & GENMASK((unsigned int)version - 1, 0))

Why casting? Why not use BIT()?

But these two lines seem like a very interesting way to reinvent the test_bit().

If I'm mistaken, this all version code needs a good comment.

> +                || jzpc->info->version >= version);
> +}

-- 
With Best Regards,
Andy Shevchenko