Message-ID: <4FDEF293.9080305@antcom.de>
Date: Mon, 18 Jun 2012 11:19:15 +0200
From: Roland Stigge <stigge@antcom.de>
Organization: ANTCOM IT Research & Development
User-Agent: Mozilla/5.0 (X11; Linux i686 on x86_64; rv:10.0.4) Gecko/20120510 Icedove/10.0.4
MIME-Version: 1.0
To: Stephen Warren <swarren@wwwdotorg.org>
CC: cjb@laptop.org, grant.likely@secretlab.ca, rob.herring@calxeda.com,
        rmk+kernel@arm.linux.org.uk, ulf.hansson@stericsson.com,
        linus.walleij@linaro.org, linux-mmc@vger.kernel.org,
        linux-kernel@vger.kernel.org, aletes.xgr@gmail.com,
        linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH] gpio: of_get_named_gpio_flags() return -EPROBE_DEFER
 if GPIO not yet available
References: <1339927893-8842-1-git-send-email-stigge@antcom.de> <4FDE8D27.6030508@wwwdotorg.org>
In-Reply-To: <4FDE8D27.6030508@wwwdotorg.org>
OpenPGP: url=subkeys.pgp.net
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2487
Lines: 71

On 06/18/2012 04:06 AM, Stephen Warren wrote:
> On 06/17/2012 04:11 AM, Roland Stigge wrote:
>> of_get_named_gpio_flags() and of_get_named_gpio() return -EPROBE_DEFER if the
>> respective GPIO is not (yet) available. This is useful if driver's probe()
>> functions try to get a GPIO whose controller isn't probed yet. Thus, the driver
>> can be probed again later on.
>>
>> The function still returns -EINVAL on other errors (parse error or node doesn't
>> exist). This way, the case of an optional/intentionally missing GPIO is handled
>> appropriately.
> 
> While I agree this is a correct change, it is going to break some
> existing code - at least sound/soc/tegra/tegra_{wm8903.c,alc5632.c}.

Can you please tell in which way the patch breaks those drivers?
However, I can see that those drivers solved the same problem in a
different way (deferring of_get_named_gpio(), via the sound init()). So
they could be adjusted to take advantage of new -EPROBE_DEFER.

> I'm
> happy to send patches for those files though (is this going into 3.5 or
> 3.6?).

For 3.6 would be best, IMO.

> However, have you audited all existing callers (including
> indirect, e.g. through plain of_get_named_gpio()) for issues this will
> cause?

Thanks for the hint, I searched the code and found

drivers/spi/spi-pl022.c

to be using -ENODEV as indication to return -EPROBE_DEFER from probe().
Will send a patch that adjusts to our of_get_named_gpio_flags() patch
and if it's good, we should join the two.

Some drivers possibly suffer from the same issue that the patch adresses:

drivers/mfd/twl6040-core.c
drivers/staging/nvec/nvec.c
drivers/usb/host/ohci-at91.c
drivers/usb/host/ehci-tegra.c
drivers/usb/gadget/at91_udc.c
drivers/spi/spi-imx.c
drivers/spi/spi-sirf.c
drivers/mmc/host/sdhci-tegra.c
drivers/mmc/host/omap_hsmmc.c
drivers/mmc/host/mxs-mmc.c
drivers/mmc/host/sdhci-esdhc-imx.c
drivers/regulator/tps62360-regulator.c
drivers/regulator/fixed.c

But the breakage therefore exists even before the patch.

Other drivers don't handle the result of of_get_named_gpio*() correctly
at all:

drivers/power/sbs-battery.c

I can help the respective maintainers to work out this issue when we
have the EPROBE_DEFER ready in gpiolib-of.c.

Thanks,

Roland
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/