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;
MIME-Version: 1.0
References: <f31b818cf8d682de61c74b133beffcc8a8202478.1660041358.git.christophe.leroy@csgroup.eu>
 <CACRpkdY53c0qXx24Am1TMivXr-MV+fQ8B0CDjtGi6=+2tn4-7A@mail.gmail.com> <CAK8P3a1Vh1Uehuin-u5QrTO5qh+t0aK_hA-QZwqc00Db_+MKcw@mail.gmail.com>
In-Reply-To: <CAK8P3a1Vh1Uehuin-u5QrTO5qh+t0aK_hA-QZwqc00Db_+MKcw@mail.gmail.com>
From:   Linus Walleij <linus.walleij@linaro.org>
Date:   Thu, 18 Aug 2022 13:13:30 +0200
Message-ID: <CACRpkdbhbwBe=jU5prifXCYUXPqULhst0se3ZRH+sWOh9XeoLQ@mail.gmail.com>
Subject: Re: [PATCH] gpio: Allow user to customise maximum number of GPIOs
To:     Arnd Bergmann <arnd@arndb.de>
Cc:     Christophe Leroy <christophe.leroy@csgroup.eu>,
        Bartosz Golaszewski <brgl@bgdev.pl>,
        Jonathan Corbet <corbet@lwn.net>,
        Russell King <linux@armlinux.org.uk>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        "maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT)" <x86@kernel.org>,
        "H. Peter Anvin" <hpa@zytor.com>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@vger.kernel.org>,
        "open list:DOCUMENTATION" <linux-doc@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>,
        "moderated list:ARM PORT" <linux-arm-kernel@lists.infradead.org>,
        "open list:GENERIC INCLUDE/ASM HEADER FILES" 
        <linux-arch@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Aug 18, 2022 at 11:48 AM Arnd Bergmann <arnd@arndb.de> wrote:

> As I understood, the problem that Christophe ran into is that the
> dynamic registration of additional gpio chips is broken because
> it unregisters the chip if the number space is exhausted:
>
>                 base = gpiochip_find_base(gc->ngpio);
>                 if (base < 0) {
>                         ret = base;
>                         spin_unlock_irqrestore(&gpio_lock, flags);
>                         goto err_free_label;
>                 }
>
> From the git history, it looks like this error was never handled gracefully
> even if the intention was to keep going without a number assignment,
> so there are probably other bugs one runs into after changing this.

Hm that should be possible to get rid of altogether? I suppose it is only
there to satisfy

static inline bool gpio_is_valid(int number)
{
        return number >= 0 && number < ARCH_NR_GPIOS;
}

?

If using GPIO descriptors, any descriptor != NULL is valid,
this one is just used with legacy GPIOs. Maybe we should just
delete gpio_is_valid() everywhere and then drop the cap?

I think there may be systems and users that still depend on GPIO base
numbers being assigned from ARCH_NR_GPIOS and
downwards (userspace GPIO numbers in sysfs will also change...)
otherwise we could assign from 0 and up.

Right now the safest would be:
Assign from 512 and downwards until we hit 0 then assign
from something high, like U32_MAX and downward.

That requires dropping gpio_is_valid() everywhere.

If we wanna be bold, just delete gpio_is_valid() and assign
bases from 0 and see what happens. But I think that will
lead to regressions.

Yours,
Linus Walleij