linux-kernel - Re: [PATCH] mfd: motorola-cpcap: Do not hardcode SPI mode flags

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: <20200107183706.GH5885@atomide.com>
Date:   Tue, 7 Jan 2020 10:37:06 -0800
From:   Tony Lindgren <tony@...mide.com>
To:     Linus Walleij <linus.walleij@...aro.org>
Cc:     Lee Jones <lee.jones@...aro.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Mark Brown <broonie@...nel.org>,
        Sebastian Reichel <sre@...nel.org>, linux-omap@...r.kernel.org
Subject: Re: [PATCH] mfd: motorola-cpcap: Do not hardcode SPI mode flags

* Tony Lindgren <tony@...mide.com> [200107 18:24]:
> * Linus Walleij <linus.walleij@...aro.org> [200107 09:58]:
> > On Thu, Dec 19, 2019 at 6:04 PM Tony Lindgren <tony@...mide.com> wrote:
> > 
> > > * Linus Walleij <linus.walleij@...aro.org> [700101 00:00]:
> > > > The current use of mode flags to us SPI_MODE_0 and
> > > > SPI_CS_HIGH is fragile: it overwrites anything already
> > > > assigned by the SPI core. Change it thusly:
> > > >
> > > > - Just |= the SPI_MODE_0 so we keep other flags
> > > > - Assign ^= SPI_CS_HIGH since we might be active high
> > > >   already, and that is usually the case with GPIOs used
> > > >   for chip select, even if they are in practice active low.
> > > >
> > > > Add a comment clarifying why ^= SPI_CS_HIGH is the right
> > > > choice here.
> > >
> > > Looks like this breaks booting for droid4 with a cpcap
> > > PMIC, probably as regulators won't work. There's no GPIO
> > > controller involved in this case for the chip select, the
> > > pins are directly controlled by the spi-omap2-mcspi.c
> > > driver.
> > >
> > > From the pin muxing setup we see there's a pull-down on
> > > mcspi1_cs0 pin meaning it's active high:
> > >
> > > /* 0x4a100138 mcspi1_cs0.mcspi1_cs0 ae23 */
> > > OMAP4_IOPAD(0x138, PIN_INPUT_PULLDOWN | MUX_MODE0)
> > >
> > > My guess a similar issue is with similar patches for
> > > all non-gpio spi controllers?
> > >
> > > Let me know if you want me to test some other changes,
> > > or if this patch depends on some other changes.
> > 
> > So this must mean that something else is setting SPI_CS_HIGH
> > for this driver, such as the device tree, right?
> 
> Hmm yes we have "spi-cs-high" property set in the dts.
> 
> But looking at drivers/spi/spi-omap2-mcspi.c, it also
> provides an option to use a GPIO for chip select in
> omap2_mcspi_setup(). That does not seem to be used though
> based on a quick grep though.
> 
> > And the |= SPI_CS_HIGH assignment in the driver is just
> > surplus and we should just delete this code instead.
> > 
> > Would that be right?
> 
> Sorry I don't know, maybe. If you have some test patch to
> try I can easily test.

Oh sorry, I guess you're suggesting we drop SPI_CS_HIGH from
drivers/mfd/motorola-cpcap.c, and not from spi-omap2-mcspi.c..
Too much glög over the holidays probably :)

Anyways, removing SPI_CS_HIGH motorola-cpcap.c also causes
mounting MMC to fail looks like. I wonder if we're missing
of_property_read_bool() for spi-cs-high in spi-omap2-mcspi.c?

Regards,

Tony