linux-kernel - Re: [PATCH v2] PCI: dra7xx: Fix reset behaviour

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]

Date:   Fri, 25 Jun 2021 02:09:56 +0200
From:   Linus Walleij <linus.walleij@...aro.org>
To:     Pali Rohár <pali@...nel.org>,
        Bartosz Golaszewski <bgolaszewski@...libre.com>
Cc:     Luca Ceresoli <luca@...aceresoli.net>,
        Kishon Vijay Abraham I <kishon@...com>,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        linux-pci <linux-pci@...r.kernel.org>,
        Linux-OMAP <linux-omap@...r.kernel.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        Rob Herring <robh@...nel.org>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>
Subject: Re: [PATCH v2] PCI: dra7xx: Fix reset behaviour

On Fri, Jun 25, 2021 at 1:34 AM Pali Rohár <pali@...nel.org> wrote:

> > gpiod_set_value(gpiod, 1) == assert the line
> > gpiod_set_value(gpiod, 0) == de-assert the line
>
> Problem is that some pci controller drivers (e.g. pci-j721e.c or
> pcie-rockchip-host.c) expects that gpiod_set_value_cansleep(gpiod, 1)
> de-asserts the line and it is already used in this way.
>
> Which is opposite of the behavior which you wrote above.

I sketched a way out of the problem using a quirk in
gpiolib in another response. We have a few of these
cases where we have to code our way out of mistakes,
such things happen.

The problem is common, and due to the fact that device tree
authors ignores the flag GPIO_ACTIVE_HIGH (which has
been around since the early days of device tree on PowerPC)
instead they opt to do the inversion in code. Which violates the
contract that the DT should describe the hardware.

The ambition of the DT specifications/schemas are to be operating
system independent, and this kind of stuff creates a situation
where other operating systems can't use the specification without
also going and looking at how Linux has implemented stuff.
Which is against the ambition of the device tree work.

> I would suggest to define enum/macro with word ASSERT and DEASSERT in
> its name instead of just true/false boolean or 0/1 int.
>
> In case of this PERST# misunderstanding, having assert/deassert in name
> should really help.

Hm that looks useful, Bart &co what do you think?

#define GPIOD_ASSERTED 1
#define GPIOD_DEASSERTED 0

in consumer.h, would that be helpful for users?

Yours,
Linus Walleij