diff mbox

[1/5] pinctrl: core: Use delayed work for hogs

Message ID 20161115170838.GK4082@atomide.com
State New
Headers show

Commit Message

Tony Lindgren Nov. 15, 2016, 5:08 p.m. UTC 
* Tony Lindgren <tony@atomide.com> [161115 07:42]:
> * Linus Walleij <linus.walleij@linaro.org> [161114 22:53]:
> > On Tue, Nov 15, 2016 at 1:47 AM, Tony Lindgren <tony@atomide.com> wrote:
> > 
> > > 8< --------------------------------
> > > From tony Mon Sep 17 00:00:00 2001
> > > From: Tony Lindgren <tony@atomide.com>
> > > Date: Tue, 25 Oct 2016 08:33:35 -0700
> > > Subject: [PATCH] pinctrl: core: Use delayed work for hogs
> > >
> > > Having the pin control framework call pin controller functions
> > > before it's probe has finished is not nice as the pin controller
> > > device driver does not yet have struct pinctrl_dev handle.
> > >
> > > Let's fix this issue by adding deferred work for late init. This is
> > > needed to be able to add pinctrl generic helper functions that expect
> > > to know struct pinctrl_dev handle. Note that we now need to call
> > > create_pinctrl() directly as we don't want to add the pin controller
> > > to the list of controllers until the hogs are claimed. We also need
> > > to pass the pinctrl_dev to the device tree parser functions as they
> > > otherwise won't find the right controller at this point.
> > >
> > > Signed-off-by: Tony Lindgren <tony@atomide.com>
> > 
> > This looks a lot better!
> > 
> > So if I understand correctly, we can guarantee that the delayed
> > work will not execute until the device driver probe() has finished,
> > and it *will* execute immediately after that?
> > 
> > So:
> > - Device driver probes
> > - Delayed work is called
> > - Next initcall
> > 
> > I'm not 100% familiar with how delayed work works... :/
> 
> Yeah well the delayed work gets scheduled for next jiffy but may
> be pre-empted as it runs in process context.
> 
> So in the worst case it could that we still may need to fix few
> drivers to support -EPROBE_DEFER. I wonder if we should check for
> hogs in probe already and only defer if hogs are defined?

Below is a version using delayed_work only if pinctrl_dt_has_hogs().

Not sure if testing only for pinctrl-0 is enough there though?

Regards,

Tony

8< --------------------------------
From tony Mon Sep 17 00:00:00 2001
From: Tony Lindgren <tony@atomide.com>
Date: Tue, 25 Oct 2016 08:33:35 -0700
Subject: [PATCH] pinctrl: core: Use delayed work for hogs

Having the pin control framework call pin controller functions
before it's probe has finished is not nice as the pin controller
device driver does not yet have struct pinctrl_dev handle.

Let's fix this issue by adding deferred work for late init. This is
needed to be able to add pinctrl generic helper functions that expect
to know struct pinctrl_dev handle. Note that we now need to call
create_pinctrl() directly as we don't want to add the pin controller
to the list of controllers until the hogs are claimed. We also need
to pass the pinctrl_dev to the device tree parser functions as they
otherwise won't find the right controller at this point.

Signed-off-by: Tony Lindgren <tony@atomide.com>
---
 drivers/pinctrl/core.c       | 90 ++++++++++++++++++++++++++++----------------
 drivers/pinctrl/core.h       |  2 +
 drivers/pinctrl/devicetree.c | 28 +++++++++++---
 drivers/pinctrl/devicetree.h | 12 +++++-
 4 files changed, 93 insertions(+), 39 deletions(-)

Comments

Linus Walleij Dec. 2, 2016, 1:08 p.m. UTC | #1 
On Tue, Nov 15, 2016 at 6:08 PM, Tony Lindgren <tony@atomide.com> wrote:
> * Tony Lindgren <tony@atomide.com> [161115 07:42]:
>> * Linus Walleij <linus.walleij@linaro.org> [161114 22:53]:
>> > On Tue, Nov 15, 2016 at 1:47 AM, Tony Lindgren <tony@atomide.com> wrote:
>> >
>> > > 8< --------------------------------
>> > > From tony Mon Sep 17 00:00:00 2001
>> > > From: Tony Lindgren <tony@atomide.com>
>> > > Date: Tue, 25 Oct 2016 08:33:35 -0700
>> > > Subject: [PATCH] pinctrl: core: Use delayed work for hogs
>> > >
>> > > Having the pin control framework call pin controller functions
>> > > before it's probe has finished is not nice as the pin controller
>> > > device driver does not yet have struct pinctrl_dev handle.
>> > >
>> > > Let's fix this issue by adding deferred work for late init. This is
>> > > needed to be able to add pinctrl generic helper functions that expect
>> > > to know struct pinctrl_dev handle. Note that we now need to call
>> > > create_pinctrl() directly as we don't want to add the pin controller
>> > > to the list of controllers until the hogs are claimed. We also need
>> > > to pass the pinctrl_dev to the device tree parser functions as they
>> > > otherwise won't find the right controller at this point.
>> > >
>> > > Signed-off-by: Tony Lindgren <tony@atomide.com>
>> >
>> > This looks a lot better!
>> >
>> > So if I understand correctly, we can guarantee that the delayed
>> > work will not execute until the device driver probe() has finished,
>> > and it *will* execute immediately after that?
>> >
>> > So:
>> > - Device driver probes
>> > - Delayed work is called
>> > - Next initcall
>> >
>> > I'm not 100% familiar with how delayed work works... :/
>>
>> Yeah well the delayed work gets scheduled for next jiffy but may
>> be pre-empted as it runs in process context.
>>
>> So in the worst case it could that we still may need to fix few
>> drivers to support -EPROBE_DEFER. I wonder if we should check for
>> hogs in probe already and only defer if hogs are defined?
>
> Below is a version using delayed_work only if pinctrl_dt_has_hogs().
>
> Not sure if testing only for pinctrl-0 is enough there though?

Sorry for the lack of attention to this patch set on my part. :(

Do you think you could resend these last 5 patches after the
release of v4.10-rc1 so we merge it early for the next cycle
and people get a chance to test and see if it works well for
everyone?

I'm worried about adding it to the tree this late in the kernel
cycle...

However I like the look of the series overall a lot.

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Tony Lindgren Dec. 2, 2016, 4:44 p.m. UTC | #2 
* Linus Walleij <linus.walleij@linaro.org> [161202 05:08]:
> On Tue, Nov 15, 2016 at 6:08 PM, Tony Lindgren <tony@atomide.com> wrote:
> > * Tony Lindgren <tony@atomide.com> [161115 07:42]:
> >> * Linus Walleij <linus.walleij@linaro.org> [161114 22:53]:
> >> > On Tue, Nov 15, 2016 at 1:47 AM, Tony Lindgren <tony@atomide.com> wrote:
> >> >
> >> > > 8< --------------------------------
> >> > > From tony Mon Sep 17 00:00:00 2001
> >> > > From: Tony Lindgren <tony@atomide.com>
> >> > > Date: Tue, 25 Oct 2016 08:33:35 -0700
> >> > > Subject: [PATCH] pinctrl: core: Use delayed work for hogs
> >> > >
> >> > > Having the pin control framework call pin controller functions
> >> > > before it's probe has finished is not nice as the pin controller
> >> > > device driver does not yet have struct pinctrl_dev handle.
> >> > >
> >> > > Let's fix this issue by adding deferred work for late init. This is
> >> > > needed to be able to add pinctrl generic helper functions that expect
> >> > > to know struct pinctrl_dev handle. Note that we now need to call
> >> > > create_pinctrl() directly as we don't want to add the pin controller
> >> > > to the list of controllers until the hogs are claimed. We also need
> >> > > to pass the pinctrl_dev to the device tree parser functions as they
> >> > > otherwise won't find the right controller at this point.
> >> > >
> >> > > Signed-off-by: Tony Lindgren <tony@atomide.com>
> >> >
> >> > This looks a lot better!
> >> >
> >> > So if I understand correctly, we can guarantee that the delayed
> >> > work will not execute until the device driver probe() has finished,
> >> > and it *will* execute immediately after that?
> >> >
> >> > So:
> >> > - Device driver probes
> >> > - Delayed work is called
> >> > - Next initcall
> >> >
> >> > I'm not 100% familiar with how delayed work works... :/
> >>
> >> Yeah well the delayed work gets scheduled for next jiffy but may
> >> be pre-empted as it runs in process context.
> >>
> >> So in the worst case it could that we still may need to fix few
> >> drivers to support -EPROBE_DEFER. I wonder if we should check for
> >> hogs in probe already and only defer if hogs are defined?
> >
> > Below is a version using delayed_work only if pinctrl_dt_has_hogs().
> >
> > Not sure if testing only for pinctrl-0 is enough there though?
> 
> Sorry for the lack of attention to this patch set on my part. :(
> 
> Do you think you could resend these last 5 patches after the
> release of v4.10-rc1 so we merge it early for the next cycle
> and people get a chance to test and see if it works well for
> everyone?

Yeah no problem, too late to do anything with them right now :)

> I'm worried about adding it to the tree this late in the kernel
> cycle...

Yup me too.

> However I like the look of the series overall a lot.

OK good to hear.

Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/pinctrl/core.c b/drivers/pinctrl/core.c
--- a/drivers/pinctrl/core.c
+++ b/drivers/pinctrl/core.c
@@ -720,7 +720,8 @@  static struct pinctrl_state *create_state(struct pinctrl *p,
 	return state;
 }
 
-static int add_setting(struct pinctrl *p, struct pinctrl_map const *map)
+static int add_setting(struct pinctrl *p, struct pinctrl_dev *pctldev,
+		       struct pinctrl_map const *map)
 {
 	struct pinctrl_state *state;
 	struct pinctrl_setting *setting;
@@ -744,7 +745,11 @@  static int add_setting(struct pinctrl *p, struct pinctrl_map const *map)
 
 	setting->type = map->type;
 
-	setting->pctldev = get_pinctrl_dev_from_devname(map->ctrl_dev_name);
+	if (pctldev)
+		setting->pctldev = pctldev;
+	else
+		setting->pctldev =
+			get_pinctrl_dev_from_devname(map->ctrl_dev_name);
 	if (setting->pctldev == NULL) {
 		kfree(setting);
 		/* Do not defer probing of hogs (circular loop) */
@@ -800,7 +805,8 @@  static struct pinctrl *find_pinctrl(struct device *dev)
 
 static void pinctrl_free(struct pinctrl *p, bool inlist);
 
-static struct pinctrl *create_pinctrl(struct device *dev)
+static struct pinctrl *create_pinctrl(struct device *dev,
+				      struct pinctrl_dev *pctldev)
 {
 	struct pinctrl *p;
 	const char *devname;
@@ -823,7 +829,7 @@  static struct pinctrl *create_pinctrl(struct device *dev)
 	INIT_LIST_HEAD(&p->states);
 	INIT_LIST_HEAD(&p->dt_maps);
 
-	ret = pinctrl_dt_to_map(p);
+	ret = pinctrl_dt_to_map(p, pctldev);
 	if (ret < 0) {
 		kfree(p);
 		return ERR_PTR(ret);
@@ -838,7 +844,7 @@  static struct pinctrl *create_pinctrl(struct device *dev)
 		if (strcmp(map->dev_name, devname))
 			continue;
 
-		ret = add_setting(p, map);
+		ret = add_setting(p, pctldev, map);
 		/*
 		 * At this point the adding of a setting may:
 		 *
@@ -899,7 +905,7 @@  struct pinctrl *pinctrl_get(struct device *dev)
 		return p;
 	}
 
-	return create_pinctrl(dev);
+	return create_pinctrl(dev, NULL);
 }
 EXPORT_SYMBOL_GPL(pinctrl_get);
 
@@ -1738,6 +1744,46 @@  static int pinctrl_check_ops(struct pinctrl_dev *pctldev)
 }
 
 /**
+ * pinctrl_late_init() - finish pin controller device registration
+ * @work: work struct
+ */
+static void pinctrl_late_init(struct work_struct *work)
+{
+	struct pinctrl_dev *pctldev;
+
+	pctldev = container_of(work, struct pinctrl_dev, late_init.work);
+
+	pctldev->p = create_pinctrl(pctldev->dev, pctldev);
+	if (!IS_ERR(pctldev->p)) {
+		kref_get(&pctldev->p->users);
+		pctldev->hog_default =
+			pinctrl_lookup_state(pctldev->p, PINCTRL_STATE_DEFAULT);
+		if (IS_ERR(pctldev->hog_default)) {
+			dev_dbg(pctldev->dev,
+				"failed to lookup the default state\n");
+		} else {
+			if (pinctrl_select_state(pctldev->p,
+						 pctldev->hog_default))
+				dev_err(pctldev->dev,
+					"failed to select default state\n");
+		}
+
+		pctldev->hog_sleep =
+			pinctrl_lookup_state(pctldev->p,
+					     PINCTRL_STATE_SLEEP);
+		if (IS_ERR(pctldev->hog_sleep))
+			dev_dbg(pctldev->dev,
+				"failed to lookup the sleep state\n");
+	}
+
+	mutex_lock(&pinctrldev_list_mutex);
+	list_add_tail(&pctldev->node, &pinctrldev_list);
+	mutex_unlock(&pinctrldev_list_mutex);
+
+	pinctrl_init_device_debugfs(pctldev);
+}
+
+/**
  * pinctrl_register() - register a pin controller device
  * @pctldesc: descriptor for this pin controller
  * @dev: parent device for this pin controller
@@ -1766,6 +1812,7 @@  struct pinctrl_dev *pinctrl_register(struct pinctrl_desc *pctldesc,
 	pctldev->driver_data = driver_data;
 	INIT_RADIX_TREE(&pctldev->pin_desc_tree, GFP_KERNEL);
 	INIT_LIST_HEAD(&pctldev->gpio_ranges);
+	INIT_DELAYED_WORK(&pctldev->late_init, pinctrl_late_init);
 	pctldev->dev = dev;
 	mutex_init(&pctldev->mutex);
 
@@ -1800,32 +1847,10 @@  struct pinctrl_dev *pinctrl_register(struct pinctrl_desc *pctldesc,
 		goto out_err;
 	}
 
-	mutex_lock(&pinctrldev_list_mutex);
-	list_add_tail(&pctldev->node, &pinctrldev_list);
-	mutex_unlock(&pinctrldev_list_mutex);
-
-	pctldev->p = pinctrl_get(pctldev->dev);
-
-	if (!IS_ERR(pctldev->p)) {
-		pctldev->hog_default =
-			pinctrl_lookup_state(pctldev->p, PINCTRL_STATE_DEFAULT);
-		if (IS_ERR(pctldev->hog_default)) {
-			dev_dbg(dev, "failed to lookup the default state\n");
-		} else {
-			if (pinctrl_select_state(pctldev->p,
-						pctldev->hog_default))
-				dev_err(dev,
-					"failed to select default state\n");
-		}
-
-		pctldev->hog_sleep =
-			pinctrl_lookup_state(pctldev->p,
-						    PINCTRL_STATE_SLEEP);
-		if (IS_ERR(pctldev->hog_sleep))
-			dev_dbg(dev, "failed to lookup the sleep state\n");
-	}
-
-	pinctrl_init_device_debugfs(pctldev);
+	if (pinctrl_dt_has_hogs(pctldev))
+		schedule_delayed_work(&pctldev->late_init, 0);
+	else
+		pinctrl_late_init(&pctldev->late_init.work);
 
 	return pctldev;
 
@@ -1848,6 +1873,7 @@  void pinctrl_unregister(struct pinctrl_dev *pctldev)
 	if (pctldev == NULL)
 		return;
 
+	cancel_delayed_work_sync(&pctldev->late_init);
 	mutex_lock(&pctldev->mutex);
 	pinctrl_remove_device_debugfs(pctldev);
 	mutex_unlock(&pctldev->mutex);
diff --git a/drivers/pinctrl/core.h b/drivers/pinctrl/core.h
--- a/drivers/pinctrl/core.h
+++ b/drivers/pinctrl/core.h
@@ -33,6 +33,7 @@  struct pinctrl_gpio_range;
  * @p: result of pinctrl_get() for this device
  * @hog_default: default state for pins hogged by this device
  * @hog_sleep: sleep state for pins hogged by this device
+ * @late_init: delayed work for pin controller to finish registration
  * @mutex: mutex taken on each pin controller specific action
  * @device_root: debugfs root for this device
  */
@@ -47,6 +48,7 @@  struct pinctrl_dev {
 	struct pinctrl *p;
 	struct pinctrl_state *hog_default;
 	struct pinctrl_state *hog_sleep;
+	struct delayed_work late_init;
 	struct mutex mutex;
 #ifdef CONFIG_DEBUG_FS
 	struct dentry *device_root;
diff --git a/drivers/pinctrl/devicetree.c b/drivers/pinctrl/devicetree.c
--- a/drivers/pinctrl/devicetree.c
+++ b/drivers/pinctrl/devicetree.c
@@ -100,11 +100,12 @@  struct pinctrl_dev *of_pinctrl_get(struct device_node *np)
 	return get_pinctrl_dev_from_of_node(np);
 }
 
-static int dt_to_map_one_config(struct pinctrl *p, const char *statename,
+static int dt_to_map_one_config(struct pinctrl *p,
+				struct pinctrl_dev *pctldev,
+				const char *statename,
 				struct device_node *np_config)
 {
 	struct device_node *np_pctldev;
-	struct pinctrl_dev *pctldev;
 	const struct pinctrl_ops *ops;
 	int ret;
 	struct pinctrl_map *map;
@@ -121,7 +122,8 @@  static int dt_to_map_one_config(struct pinctrl *p, const char *statename,
 			/* OK let's just assume this will appear later then */
 			return -EPROBE_DEFER;
 		}
-		pctldev = get_pinctrl_dev_from_of_node(np_pctldev);
+		if (!pctldev)
+			pctldev = get_pinctrl_dev_from_of_node(np_pctldev);
 		if (pctldev)
 			break;
 		/* Do not defer probing of hogs (circular loop) */
@@ -166,7 +168,22 @@  static int dt_remember_dummy_state(struct pinctrl *p, const char *statename)
 	return dt_remember_or_free_map(p, statename, NULL, map, 1);
 }
 
-int pinctrl_dt_to_map(struct pinctrl *p)
+bool pinctrl_dt_has_hogs(struct pinctrl_dev *pctldev)
+{
+	struct device_node *np;
+	struct property *prop;
+	int size;
+
+	np = pctldev->dev->of_node;
+	if (!np)
+		return false;
+
+	prop = of_find_property(np, "pinctrl-0", &size);
+
+	return prop ? true : false;
+}
+
+int pinctrl_dt_to_map(struct pinctrl *p, struct pinctrl_dev *pctldev)
 {
 	struct device_node *np = p->dev->of_node;
 	int state, ret;
@@ -233,7 +250,8 @@  int pinctrl_dt_to_map(struct pinctrl *p)
 			}
 
 			/* Parse the node */
-			ret = dt_to_map_one_config(p, statename, np_config);
+			ret = dt_to_map_one_config(p, pctldev, statename,
+						   np_config);
 			of_node_put(np_config);
 			if (ret < 0)
 				goto err;
diff --git a/drivers/pinctrl/devicetree.h b/drivers/pinctrl/devicetree.h
--- a/drivers/pinctrl/devicetree.h
+++ b/drivers/pinctrl/devicetree.h
@@ -20,8 +20,10 @@  struct of_phandle_args;
 
 #ifdef CONFIG_OF
 
+bool pinctrl_dt_has_hogs(struct pinctrl_dev *pctldev);
+
 void pinctrl_dt_free_maps(struct pinctrl *p);
-int pinctrl_dt_to_map(struct pinctrl *p);
+int pinctrl_dt_to_map(struct pinctrl *p, struct pinctrl_dev *pctldev);
 
 int pinctrl_count_index_with_args(const struct device_node *np,
 				  const char *list_name);
@@ -32,7 +34,13 @@  int pinctrl_parse_index_with_args(const struct device_node *np,
 
 #else
 
-static inline int pinctrl_dt_to_map(struct pinctrl *p)
+static inline bool pinctrl_dt_has_hogs(struct pinctrl_dev *pctldev)
+{
+	return false;
+}
+
+static inline int pinctrl_dt_to_map(struct pinctrl *p,
+				    struct pinctrl_dev *pctldev)
 {
 	return 0;
 }