From: Jonas Gorski <jogo@openwrt.org>
To: Linus Walleij <linus.walleij@linaro.org>
Cc: "linux-gpio@vger.kernel.org" <linux-gpio@vger.kernel.org>,
Alexandre Courbot <gnurou@gmail.com>,
Joachim Eastwood <manabian@gmail.com>,
Jonas Jensen <jonas.jensen@gmail.com>,
Gregory CLEMENT <gregory.clement@free-electrons.com>,
Thomas Petazzoni <thomas.petazzoni@free-electrons.com>,
James Hogan <james.hogan@imgtec.com>,
Stefan Agner <stefan@agner.ch>, Jun Nie <jun.nie@linaro.org>,
Stephen Warren <swarren@wwwdotorg.org>,
Lee Jones <lee@kernel.org>, Eric Anholt <eric@anholt.net>,
Mika Westerberg <mika.westerberg@linux.intel.com>,
Heikki Krogerus <heikki.krogerus@linux.intel.com>,
Matthias Brugger <matthias.bgg@gmail.com>,
Alessandro Rubini <rubini@unipv.it>,
Sonic Zhang <sonic.zhang@analog.com>,
Laxman Dewangan <ldewangan@nvidia.com>,
Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com>,
Baruch Siach <baruch@tkos.co.il>,
Andrew Bresticker <abrestic@chromium.org>
Subject: Re: [PATCH RFT 4/5] gpio: gpio-pl061: use the generic request/free implementations
Date: Mon, 5 Oct 2015 15:12:11 +0200 [thread overview]
Message-ID: <5612772B.8090706@openwrt.org> (raw)
In-Reply-To: <CACRpkdbVQSTW3pEPLc-34m-zihWfw8fafQxWs_=wfU29GewkYA@mail.gmail.com>
On 05.10.2015 11:19, Linus Walleij wrote:
> On Sun, Sep 13, 2015 at 3:21 PM, Jonas Gorski <jogo@openwrt.org> wrote:
>
>> spin_lock_init(&chip->lock);
>> - if (of_property_read_bool(dev->of_node, "gpio-ranges"))
>> - chip->uses_pinctrl = true;
>> + if (of_property_read_bool(dev->of_node, "gpio-ranges")) {
>> + chip->gc.request = gpiochip_generic_request;
>> + chip->gc.free = gpiochip_generic_free;
>> + }
>
> This is way better.
>
> But now this code is starting to multiply in drivers.
>
> Can we think of a way to do this even more generic:
> could gpiochip_generic_request() check if the range is
> there instead?
I thought about this, and AFAICT this would open a window in which gpio
requests could bypass the pinctrl subsystem, as ranges are only registered
after the gpio chip was registered.
pinctrl_request_gpio() -> pinctrl_get_device_gpio_range() returns -EPROBEDEFER if
it can't find a range for that reason.
To solve this we could introduce a gpiochip_add_with_range(), then we can assign
the generic _request/_free in gpiochip_add in case the callbacks aren't set yet
(and keep the way _request/_free are implemented).
If the gpio-ranges property is used, we could then check for its existence to
assign the callbacks.
So my idea is something like the following. The conversion would be then
chip->request = foo_request;
chip->free = foo_free;
gpiochip_add(chip);
gpiochip_add_pin_range(chip, "foo", 0, 0, npins);
=>
static const struct __initconst pinctrl_gpio_range foo_range = {
.base = 0,
.pinbase = 0,
.npins = npins,
};
gpiochip_add_with_ranges(chip, "foo", &foo_range, 1);
which would be a bit more verbose, but for drivers like pinctrl-cygnus-gpio,
which registers 51(!) pinranges for the same gpio chip, it would mean quite a
bit of code reduction.
diff --git a/drivers/gpio/gpiolib-of.c b/drivers/gpio/gpiolib-of.c
index fa6e3c8..b4c6119 100644
--- a/drivers/gpio/gpiolib-of.c
+++ b/drivers/gpio/gpiolib-of.c
@@ -415,6 +415,11 @@ static int of_gpiochip_add_pin_range(struct gpio_chip *chip)
return 0;
}
+bool of_gpiochip_has_pin_range(struct gpio_chip *chip)
+{
+ return !!of_find_property(chip->of_node, "gpio-ranges", NULL);
+}
+
#else
static int of_gpiochip_add_pin_range(struct gpio_chip *chip) { return 0; }
#endif
diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index e0853fb..33f79b0 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -236,14 +236,23 @@ static int gpiochip_add_to_list(struct gpio_chip *chip)
* If chip->base is negative, this requests dynamic assignment of
* a range of valid GPIOs.
*/
-int gpiochip_add(struct gpio_chip *chip)
+int gpiochip_add_with_ranges(struct gpio_chip *chip, const char *pinctl_name,
+ const struct pinctrl_gpio_range *ranges,
+ unsigned int nranges)
{
unsigned long flags;
int status = 0;
- unsigned id;
+ unsigned id, i;
int base = chip->base;
struct gpio_desc *descs;
+ if ((nranges > 0) || of_gpiochip_has_pin_range(chip)) {
+ if (!chip->request)
+ chip->request = gpiochip_generic_request;
+ if (!chip->free)
+ chip->free = gpiochip_generic_free;
+ }
+
descs = kcalloc(chip->ngpio, sizeof(descs[0]), GFP_KERNEL);
if (!descs)
return -ENOMEM;
@@ -295,6 +304,16 @@ int gpiochip_add(struct gpio_chip *chip)
if (status)
goto err_remove_chip;
+ for (i = 0; i < nranges; i++) {
+ const struct pinctrl_gpio_range *range = ranges[i];
+
+ status = gpiochip_add_pin_range(chip, pinctl_name,
+ chip->base + range->base,
+ range->pinbase, range->npins);
+ if (status)
+ goto err_remove_chip;
+ }
+
acpi_gpiochip_add(chip);
status = gpiochip_sysfs_register(chip);
@@ -309,6 +328,7 @@ int gpiochip_add(struct gpio_chip *chip)
err_remove_chip:
acpi_gpiochip_remove(chip);
+ gpiochip_remove_pin_ranges(chip);
gpiochip_free_hogs(chip);
of_gpiochip_remove(chip);
spin_lock_irqsave(&gpio_lock, flags);
diff --git a/drivers/gpio/gpiolib.h b/drivers/gpio/gpiolib.h
index bf34300..05a71da 100644
--- a/drivers/gpio/gpiolib.h
+++ b/drivers/gpio/gpiolib.h
@@ -72,6 +72,15 @@ struct gpio_desc *of_get_named_gpiod_flags(struct device_node *np,
struct gpio_desc *gpiochip_get_desc(struct gpio_chip *chip, u16 hwnum);
+#if defined(CONFIG_OF_GPIO) && defined(CONFIG_PINCTRL)
+bool of_gpiochip_has_pin_range(struct gpio_chip *chip);
+#else
+static inline bool of_gpiochip_has_pin_range(struct gpio_chip *chip)
+{
+ return false;
+}
+#endif
+
extern struct spinlock gpio_lock;
extern struct list_head gpio_chips;
diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
index 0857c28..ea8a630 100644
--- a/include/linux/gpio/driver.h
+++ b/include/linux/gpio/driver.h
@@ -166,7 +166,14 @@ extern const char *gpiochip_is_requested(struct gpio_chip *chip,
unsigned offset);
/* add/remove chips */
-extern int gpiochip_add(struct gpio_chip *chip);
+static inline int gpiochip_add(struct gpio_chip *chip)
+{
+ return gpiochip_add_with_ranges(chip, NULL, NULL, 0);
+}
+
+int gpiochip_add_with_ranges(struct gpio_chip *chip, const char *pinctl_name,
+ const struct pinctrl_gpio_range *ranges,
+ unsigned int nranges);
extern void gpiochip_remove(struct gpio_chip *chip);
extern struct gpio_chip *gpiochip_find(void *data,
int (*match)(struct gpio_chip *chip, void *data));
next prev parent reply other threads:[~2015-10-05 13:12 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-13 13:21 [PATCH RFT 0/5] gpiochip: add and use generic request/free Jonas Gorski
2015-09-13 13:21 ` [PATCH RFT 1/5] gpiolib: provide generic request/free implementations Jonas Gorski
2015-10-05 9:17 ` Linus Walleij
2015-09-13 13:21 ` [PATCH RFT 2/5] gpio: replace trivial implementations of request/free with generic one Jonas Gorski
2015-09-14 14:37 ` Thomas Petazzoni
2015-09-14 17:14 ` James Hogan
2015-09-15 3:32 ` Stefan Agner
2015-09-20 13:16 ` Joachim Eastwood
2015-09-13 13:21 ` [PATCH RFT 3/5] gpio: gpio-xz: use the generic request/free implementations Jonas Gorski
2015-09-13 13:21 ` [PATCH RFT 4/5] gpio: gpio-pl061: " Jonas Gorski
2015-10-05 9:19 ` Linus Walleij
2015-10-05 13:12 ` Jonas Gorski [this message]
2015-10-06 7:23 ` Linus Walleij
2015-09-13 13:21 ` [PATCH RFT 5/5] pinctrl: replace trivial implementations of gpio_chip request/free Jonas Gorski
2015-09-13 19:07 ` Bjorn Andersson
2015-09-13 19:20 ` Heiko Stübner
2015-09-14 14:18 ` Eric Anholt
2015-09-14 14:41 ` Mika Westerberg
2015-09-14 16:53 ` Andrew Bresticker
2015-09-16 11:07 ` Baruch Siach
2015-09-20 13:40 ` Matthias Brugger
2015-10-01 12:32 ` Lee Jones
2015-10-05 10:37 ` Laxman Dewangan
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=5612772B.8090706@openwrt.org \
--to=jogo@openwrt.org \
--cc=abrestic@chromium.org \
--cc=baruch@tkos.co.il \
--cc=eric@anholt.net \
--cc=gnurou@gmail.com \
--cc=gregory.clement@free-electrons.com \
--cc=heikki.krogerus@linux.intel.com \
--cc=james.hogan@imgtec.com \
--cc=jonas.jensen@gmail.com \
--cc=jun.nie@linaro.org \
--cc=ldewangan@nvidia.com \
--cc=lee@kernel.org \
--cc=linus.walleij@linaro.org \
--cc=linux-gpio@vger.kernel.org \
--cc=manabian@gmail.com \
--cc=matthias.bgg@gmail.com \
--cc=mika.westerberg@linux.intel.com \
--cc=plagnioj@jcrosoft.com \
--cc=rubini@unipv.it \
--cc=sonic.zhang@analog.com \
--cc=stefan@agner.ch \
--cc=swarren@wwwdotorg.org \
--cc=thomas.petazzoni@free-electrons.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.