linux-kernel - Re: [PATCH v7] regulator: fixed: Convert to use GPIO descriptor only

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: <CACRpkdbcBrjKar4CDg+yEVuCNE3RiGD61UkPasZ35SSmSNsUBA@mail.gmail.com>
Date:   Mon, 1 Oct 2018 22:16:51 +0200
From:   Linus Walleij <linus.walleij@...aro.org>
To:     leonard.crestez@....com, John Stultz <john.stultz@...aro.org>
Cc:     Mark Brown <broonie@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Andy Duan <fugang.duan@....com>,
        Fabio Estevam <fabio.estevam@....com>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        Liam Girdwood <lgirdwood@...il.com>,
        Shawn Guo <shawnguo@...nel.org>,
        Anson Huang <anson.huang@....com>,
        ext Tony Lindgren <tony@...mide.com>,
        Mike Rapoport <rppt@...ux.vnet.ibm.com>,
        Janusz Krzysztofik <jmkrzyszt@...il.com>
Subject: Re: [PATCH v7] regulator: fixed: Convert to use GPIO descriptor only

On Mon, Oct 1, 2018 at 8:53 PM Leonard Crestez <leonard.crestez@....com> wrote:

> This patch seems to break network booting on imx6sx-sdb in linux-next
> because the enet phy regulator is not on. Reverting the patch fixes
> boot.

Thanks for reporting.

John Stultz reported the same problem I'm trying to debug it.

> Here is the regulator definition:
>
> reg_enet_3v3: regulator-enet-3v3 {
>         compatible = "regulator-fixed";
>         pinctrl-names = "default";
>         pinctrl-0 = <&pinctrl_enet_3v3>;
>         regulator-name = "enet_3v3";
>         regulator-min-microvolt = <3300000>;
>         regulator-max-microvolt = <3300000>;
>         gpios = <&gpio2 6 GPIO_ACTIVE_LOW>;
> };

This is a bit odd actually, the GPIO_ACTIVE_LOW flag will
be ignored as you see:

> [    0.153171] reg_fixed_voltage_probe(179): regulator-enet-3v3 call gpiod_get_optional gflags=0x7 ena_gpio_invert
> [    0.153218] regulator-enet-3v3 GPIO handle specifies active low - ignored
> [    0.153233] of_gpio_flags_quirks(83): regulator-enet-3v3 set active low because !enable-active-high

Because regulators don't specify active high/low in the second
cell because of legacy bindings.

So this should not be in the device tree anyway, it should be
GPIO_ACTIVE_HIGH or just 0.

> That "gpio_value: 30 set 1" tracepoint is wrong, the line is set high.
>
> It seems that gpiod_set_value will check FLAG_ACTIVE_LOW and
> automatically invert
(...)
>so maybe ena_gpio_invert should not be used if a
> full gpiod is passed to regulator?
(...)
> -       cfg.ena_gpio_invert = !config->enable_high;

Indeed. I will look closer so it's the right fix and provide a patch.

> All these high/low inversions and flags are extremely confusing to me.

Yeah it's what I'm trying to get rid of with these patches,
this is just the first patch in a series that move inversion over
to the GPIO library.

Yours,
Linus Walleij