From: Scott Wood <scottwood@freescale.com>
To: Liu Gang <Gang.Liu@freescale.com>
Cc: linux-gpio@vger.kernel.org, linus.walleij@linaro.org,
linuxppc-dev@lists.ozlabs.org, r61911@freescale.com,
b07421@freescale.com
Subject: Re: [PATCH] powerpc/gpio: Fix the wrong GPIO input data on MPC8572/MPC8536
Date: Wed, 20 Nov 2013 18:32:46 -0600 [thread overview]
Message-ID: <1384993966.1403.466.camel@snotra.buserror.net> (raw)
In-Reply-To: <1384916079.16677.26.camel@linux>
On Wed, 2013-11-20 at 10:54 +0800, Liu Gang wrote:
> On Tue, 2013-11-19 at 16:51 -0600, Scott Wood wrote:
> > > @@ -71,6 +71,7 @@ static int mpc8572_gpio_get(struct gpio_chip *gc, unsigned int gpio)
> > > struct mpc8xxx_gpio_chip *mpc8xxx_gc = to_mpc8xxx_gpio_chip(mm);
> > >
> > > val = in_be32(mm->regs + GPIO_DAT) & ~in_be32(mm->regs + GPIO_DIR);
> > > + mpc8xxx_gc->data &= in_be32(mm->regs + GPIO_DIR);
> > >
> > > return (val | mpc8xxx_gc->data) & mpc8xxx_gpio2mask(gpio);
> > > }
> >
> > It seems odd to update ->data in a function that's supposed to be
> > reading things... Perhaps it would be better to keep ->data in a good
> > state from the beginning.
> >
> > -Scott
>
> Yes, keeping the ->data in a good state from the beginning will be
> better. But this will need more code in different functions to cover
> all the scenarios.
> First, we should check the direct of the pin in the function
> "mpc8xxx_gpio_set", and clean the input bit in ->data after setting
> operation.
For userspace value setting, it looks like gpiolib blocks the write if
the pin if FLAG_IS_OUT is set. This suggests that this is an error
condition for other uses as well. Though, I notice that
mpc8xxx_gpio_dir_out() calls gpio_set() before actually changing the
direction. So it may be useful to avoid races where the wrong value is
output briefly after the direction is changed (especially in open drain
situations, where the signal could have a meaningful default even before
we begin outputting). But that raises the question of how you'd do that
from userspace, and it also renders the to-be-output value as write-only
data (until the direction is actually changed), since a readback would
get the input value instead.
> So maybe it's better to eliminate the effects of the ->data to the
> input pins when reading the status, regardless of the possible changes
> of the pins and the data.
> Do you think so?
Perhaps, but that doesn't require you to modify ->data in the get()
function.
-Scott
next prev parent reply other threads:[~2013-11-21 0:33 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-11-15 7:16 [PATCH] powerpc/gpio: Fix the wrong GPIO input data on MPC8572/MPC8536 Liu Gang
2013-11-19 9:49 ` Linus Walleij
2013-11-19 15:32 ` Anatolij Gustschin
2013-11-20 2:07 ` Liu Gang
2013-11-19 22:51 ` Scott Wood
2013-11-20 2:54 ` Liu Gang
2013-11-21 0:32 ` Scott Wood [this message]
2013-11-22 4:47 ` Liu Gang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1384993966.1403.466.camel@snotra.buserror.net \
--to=scottwood@freescale.com \
--cc=Gang.Liu@freescale.com \
--cc=b07421@freescale.com \
--cc=linus.walleij@linaro.org \
--cc=linux-gpio@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=r61911@freescale.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).