Skip to contentSkip to conversation

Add support for charging battery in Nokia RX-51

PATCH Applied Pali Rohár 11 participants 83 messages Sep 8, 2013 – Jan 20, 2014

Subsystems (8)
lore
Share & tools
lore
#0 lore raw

cover Add support for charging battery in Nokia RX-51

This patch series finally bringing support for charging battery on Nokia N900 (RX-51) without any proprietary Nokia bits in userspace.

Pali Roh?r (4):
usb: musb: Call atomic_notifier_call_chain when status is changed
power: isp1704_charger: Fix driver to work with changes introduced in

    v3.5

power: isp1704_charger: Add callback function set_current
RX-51: Add platform function and data for bq24150a charger

 arch/arm/mach-omap2/board-rx51-peripherals.c |   56 +++++++++++++-
 drivers/power/isp1704_charger.c              |  107 ++++++++++++++------------
 drivers/usb/musb/omap2430.c                  |    3 +
 drivers/usb/phy/phy-twl4030-usb.c            |    2 +
 include/linux/power/isp1704_charger.h        |    1 +
 5 files changed, 117 insertions(+), 52 deletions(-)
Signature · 1 line
1.7.10.4
#1 lore raw

PATCH 1/4 usb: musb: Call atomic_notifier_call_chain when status is changed

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---
 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
drivers/usb/musb/omap2430.c +3 -0
index f44e8b5..5c40252 100644
@@ -305,6 +305,9 @@ static void omap_musb_set_mailbox(struct omap2430_glue *glue)
 	default:
 		dev_dbg(dev, "ID float\n");
 	}
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);
 }
 
 
drivers/usb/phy/phy-twl4030-usb.c +2 -0
index 8f78d2d..efe6155 100644
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct platform_device *pdev)
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 		dev_warn(&pdev->dev, "could not create sysfs file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
 	/* Our job is to use irqs and status from the power module
 	 * to keep the transceiver disabled when nothing's connected.
 	 *
Signature · 1 line
1.7.10.4
#2 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 25 lines · Pali Roh?r

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---
 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
drivers/usb/musb/omap2430.c +0 -0
index f44e8b5..5c40252 100644

drivers/usb/musb/omap2430.c static void omap_musb_set_mailbox(struct omap2430_glue *glue) · line 310

Quoted 6 lines
drivers/usb/musb/omap2430.c +3 -0
@@ -305,6 +305,9 @@ static void omap_musb_set_mailbox(struct omap2430_glue *glue)
 	default:
 		dev_dbg(dev, "ID float\n");
 	}
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);

let's add a wrapper for this:

static inline int usb_phy_notify(struct usb phy *x, unsigned val, void *v)

{
	return atomic_notifier_call_chain(&x->notifier, val, v);
}
Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130917/df17f742/attachment-0001.sig>
#3 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

In reply to Pali Rohár

drivers/usb/phy/phy-twl4030-usb.c static int twl4030_usb_probe(struct platform_device *pdev) · line 708

Quoted 9 lines · Pali Roh?r
drivers/usb/phy/phy-twl4030-usb.c +1 -0
index 8f78d2d..efe6155 100644
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct platform_device *pdev)
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 		dev_warn(&pdev->dev, "could not create sysfs file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);

BTW, this is a bugfix, send separately.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130917/dde12078/attachment.sig>
#4 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 43 lines · On Tuesday 17 September 2013 17:48:59 you wrote:
Quoted 29 lines · Pali Roh?r

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---

 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
(quoted diff) +0 -0

b/drivers/usb/musb/omap2430.c index f44e8b5..5c40252 100644

drivers/usb/musb/omap2430.c static void omap_musb_set_mailbox(struct · line 310

Quoted 10 lines
drivers/usb/musb/omap2430.c +3 -0
@@ -305,6 +305,9 @@ static void omap_musb_set_mailbox(struct
 omap2430_glue *glue)
 
 	default:
 		dev_dbg(dev, "ID float\n");
 
 	}
 
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);

let's add a wrapper for this:

Quoted 4 lines

static inline int usb_phy_notify(struct usb phy *x, unsigned

val, void *v) {
	return atomic_notifier_call_chain(&x->notifier, val, v);
}

Where to add this wrapper? To omap2430.c? or some include file?

Quoted 9 lines · On Tuesday 17 September 2013 17:49:17 Felipe Balbi wrote:
Quoted 8 lines · Pali Roh?r
(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev)
 
Quoted 7 lines

drivers/usb/phy/phy-twl4030-usb.c static int twl4030_usb_probe(struct platform_device *pdev) · line 708

Quoted 5 lines
drivers/usb/phy/phy-twl4030-usb.c +1 -0
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 
 		dev_warn(&pdev->dev, "could not create sysfs file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);

BTW, this is a bugfix, send separately.

What to send separately?

This full patch 1/4 is bugfix. And I did not understand what you want. Maybe it could be easier for you to apply this small 3+2 lines patch how you need.

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130917/b9164414/attachment.sig>
#5 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 44 lines · Pali Roh?r
Quoted 43 lines · On Tuesday 17 September 2013 17:48:59 you wrote:
Quoted 29 lines · Pali Roh?r

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---

 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
(quoted diff) +0 -0

b/drivers/usb/musb/omap2430.c index f44e8b5..5c40252 100644

drivers/usb/musb/omap2430.c static void omap_musb_set_mailbox(struct · line 310

Quoted 10 lines
drivers/usb/musb/omap2430.c +3 -0
@@ -305,6 +305,9 @@ static void omap_musb_set_mailbox(struct
 omap2430_glue *glue)
 
 	default:
 		dev_dbg(dev, "ID float\n");
 
 	}
 
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);

let's add a wrapper for this:

Quoted 6 lines
Quoted 4 lines

static inline int usb_phy_notify(struct usb phy *x, unsigned

val, void *v) {
	return atomic_notifier_call_chain(&x->notifier, val, v);
}

Where to add this wrapper? To omap2430.c? or some include file?

<linux/usb/phy.h>

Quoted 18 lines · On Tuesday 17 September 2013 17:49:17 Felipe Balbi wrote:
Quoted 9 lines · Pali Roh?r
Quoted 8 lines
(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev)
 
Quoted 7 lines

drivers/usb/phy/phy-twl4030-usb.c static int twl4030_usb_probe(struct platform_device *pdev) · line 708

Quoted 5 lines
drivers/usb/phy/phy-twl4030-usb.c +1 -0
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 
 		dev_warn(&pdev->dev, "could not create sysfs file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);

BTW, this is a bugfix, send separately.

Quoted 5 lines

What to send separately?

This full patch 1/4 is bugfix. And I did not understand what you want. Maybe it could be easier for you to apply this small 3+2 lines patch how you need.

This hunk here (initializaing notifier head) is a separate bug fix and needs its own patch.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130917/a7e83fbb/attachment.sig>
#6 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 76 lines · On Tuesday 17 September 2013 18:08:35 Felipe Balbi wrote:
Quoted 47 lines · Pali Roh?r
Quoted 46 lines · On Tuesday 17 September 2013 17:48:59 you wrote:
Quoted 29 lines · Pali Roh?r

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---

 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
(quoted diff) +0 -0

b/drivers/usb/musb/omap2430.c index f44e8b5..5c40252 100644 --- a/drivers/usb/musb/omap2430.c
+++ b/drivers/usb/musb/omap2430.c

drivers/usb/musb/omap2430.c static void · line 310

Quoted 10 lines
drivers/usb/musb/omap2430.c +3 -0
@@ -305,6 +305,9 @@ static void
 omap_musb_set_mailbox(struct omap2430_glue *glue)
 
 	default:
 		dev_dbg(dev, "ID float\n");
 
 	}
 
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);

let's add a wrapper for this:

static inline int usb_phy_notify(struct usb phy *x, unsigned val, void *v) {

Quoted 6 lines
Quoted 3 lines
	return atomic_notifier_call_chain(&x->notifier, val, v);

}

Where to add this wrapper? To omap2430.c? or some include file?

<linux/usb/phy.h>

Quoted 19 lines · On Tuesday 17 September 2013 17:49:17 Felipe Balbi wrote:
Quoted 17 lines · Pali Roh?r
Quoted 6 lines
(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index
8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c static int twl4030_usb_probe(struct

Quoted 8 lines
drivers/usb/phy/phy-twl4030-usb.c +1 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev)
 
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 
 		dev_warn(&pdev->dev, "could not create sysfs
 		file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);

BTW, this is a bugfix, send separately.

Quoted 8 lines
Quoted 5 lines

What to send separately?

This full patch 1/4 is bugfix. And I did not understand what you want. Maybe it could be easier for you to apply this small 3+2 lines patch how you need.

This hunk here (initializaing notifier head) is a separate bug fix and needs its own patch.

So will you do that? Or it is needed to resend this one line hunk again in new email again?

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130917/2aec70d8/attachment.sig>
#7 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 83 lines · Pali Roh?r
Quoted 82 lines · On Tuesday 17 September 2013 18:08:35 Felipe Balbi wrote:
Quoted 47 lines · Pali Roh?r
Quoted 46 lines · On Tuesday 17 September 2013 17:48:59 you wrote:
Quoted 29 lines · Pali Roh?r

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---

 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
(quoted diff) +0 -0

b/drivers/usb/musb/omap2430.c index f44e8b5..5c40252 100644 --- a/drivers/usb/musb/omap2430.c
+++ b/drivers/usb/musb/omap2430.c

drivers/usb/musb/omap2430.c static void · line 310

Quoted 10 lines
drivers/usb/musb/omap2430.c +3 -0
@@ -305,6 +305,9 @@ static void
 omap_musb_set_mailbox(struct omap2430_glue *glue)
 
 	default:
 		dev_dbg(dev, "ID float\n");
 
 	}
 
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);

let's add a wrapper for this:

static inline int usb_phy_notify(struct usb phy *x, unsigned val, void *v) {

Quoted 6 lines
Quoted 3 lines
	return atomic_notifier_call_chain(&x->notifier, val, v);

}

Where to add this wrapper? To omap2430.c? or some include file?

<linux/usb/phy.h>

Quoted 24 lines · On Tuesday 17 September 2013 17:49:17 Felipe Balbi wrote:
Quoted 9 lines · Pali Roh?r
Quoted 8 lines
(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index
8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev)
 
Quoted 8 lines

drivers/usb/phy/phy-twl4030-usb.c static int twl4030_usb_probe(struct platform_device *pdev) · line 708

Quoted 6 lines
drivers/usb/phy/phy-twl4030-usb.c +1 -0
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 
 		dev_warn(&pdev->dev, "could not create sysfs
 		file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);

BTW, this is a bugfix, send separately.

What to send separately?

This full patch 1/4 is bugfix. And I did not understand what you want. Maybe it could be easier for you to apply this small 3+2 lines patch how you need.

Quoted 5 lines
Quoted 2 lines

This hunk here (initializaing notifier head) is a separate bug fix and needs its own patch.

So will you do that? Or it is needed to resend this one line hunk again in new email again?

new patch, new email

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130917/c6cd20bc/attachment.sig>
#8 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 87 lines · On Wednesday 18 September 2013 03:49:42 Felipe Balbi wrote:
Quoted 86 lines · Pali Roh?r
Quoted 85 lines · On Tuesday 17 September 2013 18:08:35 Felipe Balbi wrote:
Quoted 48 lines · Pali Roh?r
Quoted 47 lines · On Tuesday 17 September 2013 17:48:59 you wrote:
Quoted 30 lines · Pali Roh?r

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---

 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
(quoted diff) +0 -0

b/drivers/usb/musb/omap2430.c index f44e8b5..5c40252 100644 --- a/drivers/usb/musb/omap2430.c
+++ b/drivers/usb/musb/omap2430.c

Review static void

Quoted 10 lines
(quoted diff) +3 -0
@@ -305,6 +305,9 @@ static void
 omap_musb_set_mailbox(struct omap2430_glue *glue)
 
 	default:
 		dev_dbg(dev, "ID float\n");
 
 	}
 
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);

let's add a wrapper for this:

static inline int usb_phy_notify(struct usb phy *x, unsigned val, void *v) {

Quoted 7 lines
Quoted 4 lines
	return atomic_notifier_call_chain(&x->notifier, val,
	v);

}

Where to add this wrapper? To omap2430.c? or some
include file?

<linux/usb/phy.h>

Quoted 25 lines · On Tuesday 17 September 2013 17:49:17 Felipe Balbi wrote:
Quoted 12 lines · Pali Roh?r
Quoted 11 lines
(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int
 twl4030_usb_probe(struct platform_device *pdev)
 
 	if (device_create_file(&pdev->dev,
 	&dev_attr_vbus))
 
Quoted 6 lines

drivers/usb/phy/phy-twl4030-usb.c static int

Quoted 4 lines
drivers/usb/phy/phy-twl4030-usb.c +1 -0
 		dev_warn(&pdev->dev, "could not create sysfs
 		file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);

BTW, this is a bugfix, send separately.

What to send separately?

This full patch 1/4 is bugfix. And I did not understand what you want. Maybe it could be easier for you to apply this small 3+2 lines patch how you need.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

This hunk here (initializaing notifier head) is a separate bug fix and needs its own patch.

So will you do that? Or it is needed to resend this one line hunk again in new email again?

new patch, new email

Guys, WHY ARE YOU SO STUPID AND ARROGANT?

Sorry but, need to copy full isolated patch/hunk from one mail to another is hassling. So what you want from me? Do all those non sense working only because yesterday you had bad day? Or what?

Everything needed with described information was in first mail. Also second hunk has full isolated "git diff" output, so it is for you really big problem to copy it? Or you did not see that patch?

I really did not understand what you wanted from me.

============================
==== BEGINNING OF PATCH ====
============================

This is bugfix and sending this patch separately from all other patches. This patch is visibly isolated from all others and could be readable also by disabled people. For other handicapped people I suggest to increase font size and other text settings in program which view this patch. For visually impaired people I suggest to use some text-to-speech software.

This is small 2 lines patch, diff starting after next visible break.

This patch initializing notifier head in tw4030 usb code which is missing. Initialization code is needed for using any atomic_notifier_* functions.

===========================
==== BEGINNING OF DIFF ====
===========================

drivers/usb/phy/phy-twl4030-usb.c +2 -0
index 8f78d2d..efe6155 100644
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct platform_device *pdev)
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 		dev_warn(&pdev->dev, "could not create sysfs file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
 	/* Our job is to use irqs and status from the power module
 	 * to keep the transceiver disabled when nothing's connected.
 	 *

======================
==== END OF PATCH ====
======================

PS: This is end of email and patch is above ^^^^

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/9c5970c0/attachment.sig>
#9 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 143 lines · Pali Roh?r
Quoted 87 lines · On Wednesday 18 September 2013 03:49:42 Felipe Balbi wrote:
Quoted 86 lines · Pali Roh?r
Quoted 85 lines · On Tuesday 17 September 2013 18:08:35 Felipe Balbi wrote:
Quoted 48 lines · Pali Roh?r
Quoted 47 lines · On Tuesday 17 September 2013 17:48:59 you wrote:
Quoted 30 lines · Pali Roh?r

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

---

 drivers/usb/musb/omap2430.c       |    3 +++
 drivers/usb/phy/phy-twl4030-usb.c |    2 ++
 2 files changed, 5 insertions(+)
(quoted diff) +0 -0

b/drivers/usb/musb/omap2430.c index f44e8b5..5c40252 100644 --- a/drivers/usb/musb/omap2430.c
+++ b/drivers/usb/musb/omap2430.c

Review static void

Quoted 10 lines
(quoted diff) +3 -0
@@ -305,6 +305,9 @@ static void
 omap_musb_set_mailbox(struct omap2430_glue *glue)
 
    default:
            dev_dbg(dev, "ID float\n");
 
    }
 
+
+   atomic_notifier_call_chain(&musb->xceiv->notifier,
+                   musb->xceiv->last_event, NULL);

let's add a wrapper for this:

static inline int usb_phy_notify(struct usb phy *x, unsigned val, void *v) {

Quoted 7 lines
Quoted 4 lines
      return atomic_notifier_call_chain(&x->notifier, val,
      v);

}

Where to add this wrapper? To omap2430.c? or some
include file?

<linux/usb/phy.h>

Quoted 25 lines · On Tuesday 17 September 2013 17:49:17 Felipe Balbi wrote:
Quoted 12 lines · Pali Roh?r
Quoted 11 lines
(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int
 twl4030_usb_probe(struct platform_device *pdev)
 
    if (device_create_file(&pdev->dev,
    &dev_attr_vbus))
 
Quoted 6 lines

drivers/usb/phy/phy-twl4030-usb.c static int twl4030_usb_probe(struct platform_device *pdev) · line 708

Quoted 4 lines
drivers/usb/phy/phy-twl4030-usb.c +1 -0
            dev_warn(&pdev->dev, "could not create sysfs
            file\n");
 
+   ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);

BTW, this is a bugfix, send separately.

What to send separately?

This full patch 1/4 is bugfix. And I did not understand what you want. Maybe it could be easier for you to apply this small 3+2 lines patch how you need.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

This hunk here (initializaing notifier head) is a separate bug fix and needs its own patch.

So will you do that? Or it is needed to resend this one line hunk again in new email again?

new patch, new email

Guys, WHY ARE YOU SO STUPID AND ARROGANT?

Sorry but, need to copy full isolated patch/hunk from one mail to another is hassling. So what you want from me? Do all those non sense working only because yesterday you had bad day? Or what?

Everything needed with described information was in first mail. Also second hunk has full isolated "git diff" output, so it is for you really big problem to copy it? Or you did not see that patch?

I really did not understand what you wanted from me.

============================
==== BEGINNING OF PATCH ====
============================

This is bugfix and sending this patch separately from all other patches. This patch is visibly isolated from all others and could be readable also by disabled people. For other handicapped people I suggest to increase font size and other text settings in program which view this patch. For visually impaired people I suggest to use some text-to-speech software.

This is small 2 lines patch, diff starting after next visible break.

This patch initializing notifier head in tw4030 usb code which is missing. Initialization code is needed for using any atomic_notifier_* functions.

===========================
==== BEGINNING OF DIFF ====
===========================

drivers/usb/phy/phy-twl4030-usb.c +2 -0
index 8f78d2d..efe6155 100644
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct platform_device *pdev)
        if (device_create_file(&pdev->dev, &dev_attr_vbus))
                dev_warn(&pdev->dev, "could not create sysfs file\n");
 
+       ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
        /* Our job is to use irqs and status from the power module
         * to keep the transceiver disabled when nothing's connected.
         *
Quoted 6 lines

======================
==== END OF PATCH ====
======================

PS: This is end of email and patch is above ^^^^

Hi Pali,

There is no need to be rude.

Felipe asked you to do the split since he believes that the notifier chain call for musb xceiv and the twl->phy notifier head init should be done in two separate patches.

I agree with him since is better to separate bugfixs from new features for different reasons. For example if is found that a feature has to be reverted, then this won't revert a bugfix causing a regression. Also, it is easier to review and the bugfix patch can be cherry-picked by stable kernels.

You said that to copy isolated patch/hunk from one mail to another is hassling so why do you want Felipe to do that? He like most most sub-system maintainers is very busy so if we want the kernel to scale we have to help them to make their live easier.

So, I don't understand from where your anger comes for just the need to split your patch in two and send correct patches with git send-email so Felipe can apply easily with git am mbox or whatever worfklow he has.

This is not the first time you contribute to the kernel so you know that the above patch s not only not in the correct format to be applied but also has a very offensive commit changelog that is not suitable for the kernel.

Take a deep breath, grab a mug of coffee and then just split your patches and repost as v2, I'm sure you spent more time writing that angry email that what it would have taken to do it properly ;-)

Quoted 3 lines

--
Pali Roh?r
pali.rohar at gmail.com

Best regards,
Javier

#10 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Hi!

Quoted 10 lines
Quoted 10 lines
Quoted 4 lines
Quoted 2 lines

So will you do that? Or it is needed to resend this one line hunk again in new email again?

new patch, new email

Guys, WHY ARE YOU SO STUPID AND ARROGANT?

Sorry but, need to copy full isolated patch/hunk from one mail to another is hassling. So what you want from me? Do all those non sense working only because yesterday you had bad day? Or what?

...

Quoted 8 lines

Hi Pali,

There is no need to be rude.

Felipe asked you to do the split since he believes that the notifier chain call for musb xceiv and the twl->phy notifier head init should be done in two separate patches.

Actually, there is need to be rude, because Felipe fails to act as maintainer. Instead of fixing bugs in his code, he bounces bugfix patches, points people to random READMEs and wastes everyones time.
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#11 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 18 lines · Pavel Machek

Hi!

Quoted 10 lines
Quoted 10 lines
Quoted 4 lines
Quoted 2 lines

So will you do that? Or it is needed to resend this one line hunk again in new email again?

new patch, new email

Guys, WHY ARE YOU SO STUPID AND ARROGANT?

Sorry but, need to copy full isolated patch/hunk from one mail to another is hassling. So what you want from me? Do all those non sense working only because yesterday you had bad day? Or what?

...

Quoted 5 lines

Hi Pali,

There is no need to be rude.

Quoted 9 lines
Quoted 3 lines

Felipe asked you to do the split since he believes that the notifier chain call for musb xceiv and the twl->phy notifier head init should be done in two separate patches.

Actually, there is need to be rude, because Felipe fails to act as maintainer. Instead of fixing bugs in his code, he bounces bugfix patches, points people to random READMEs and wastes everyones time.
Pavel

I don't know what are you talking about (if that happened in another thread then I need more context). Felipe is not bouncing any bugfix but just asked to split the patch in two since the patch was solving two separate issues so is way better to have it in two separate patches for the reasons I explained before.

So, as far as I can tell Felipe did exactly what I would expect from a maintainer. He took the time to review the patches sent to him and gave feedback. If the sender doesn't want to take his feedback into account and prefer to send pretty insulting emails instead that is his choice but I would say that is this not the greatest approach to get your code merged (to say the least).

Best regards,
Javier

#12 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Hi!

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 4 lines
Quoted 2 lines

So will you do that? Or it is needed to resend this one line hunk again in new email again?

new patch, new email

Guys, WHY ARE YOU SO STUPID AND ARROGANT?

Sorry but, need to copy full isolated patch/hunk from one mail to another is hassling. So what you want from me? Do all those non sense working only because yesterday you had bad day? Or what?

...

Quoted 6 lines
Quoted 3 lines

Actually, there is need to be rude, because Felipe fails to act as maintainer. Instead of fixing bugs in his code, he bounces bugfix patches, points people to random READMEs and wastes everyones time.

I don't know what are you talking about (if that happened in another thread then I need more context). Felipe is not bouncing any bugfix

Take a look here:

https://lkml.org/lkml/2013/9/17/286

I clearly state that patch can not be tested as required for "proper" submission, but offer patch anyway. I get irrelevant boilerplate on patch format.

Quoted 6 lines

but just asked to split the patch in two since the patch was solving two separate issues so is way better to have it in two separate patches for the reasons I explained before.

So, as far as I can tell Felipe did exactly what I would expect from a maintainer. He took the time to review the patches sent to him and

I'd expect maintainer to, well, maintain code. It means actually fixing bugs in his code, when he's pointed at them.

Quoted 4 lines

gave feedback. If the sender doesn't want to take his feedback into account and prefer to send pretty insulting emails instead that is his choice but I would say that is this not the greatest approach to get your code merged (to say the least).

Clearly not. But Pali found bug in code Felipe should maintain. Instead of "thank you for bug report, I applied this one line of your code to fix it", Pali got "new patch, new email" for his efforts. That is how you train dogs, not how you should treat kernel contributors.

Now, it is possible that Felipe just has problems with english, as he called me piece of wood in https://lkml.org/lkml/2013/9/17/476 , but he appears more arogant than usual over email.

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#13 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Hi!

Quoted 11 lines
Quoted 4 lines

gave feedback. If the sender doesn't want to take his feedback into account and prefer to send pretty insulting emails instead that is his choice but I would say that is this not the greatest approach to get your code merged (to say the least).

Clearly not. But Pali found bug in code Felipe should maintain. Instead of "thank you for bug report, I applied this one line of your code to fix it", Pali got "new patch, new email" for his efforts. That is how you train dogs, not how you should treat kernel contributors.

Quoted 3 lines

Now, it is possible that Felipe just has problems with english, as he called me piece of wood in https://lkml.org/lkml/2013/9/17/476 , but he appears more arogant than usual over email.

Actually he called me "piece of wood with garbage in it". I guess I have right to be offended. I'm baby in the next email. Hmm.
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#14 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Hi,

Quoted 13 lines · Pavel Machek

Hi!

Quoted 11 lines
Quoted 4 lines

gave feedback. If the sender doesn't want to take his feedback into account and prefer to send pretty insulting emails instead that is his choice but I would say that is this not the greatest approach to get your code merged (to say the least).

Clearly not. But Pali found bug in code Felipe should maintain. Instead of "thank you for bug report, I applied this one line of your code to fix it", Pali got "new patch, new email" for his efforts. That is how you train dogs, not how you should treat kernel contributors.

Quoted 6 lines
Quoted 3 lines

Now, it is possible that Felipe just has problems with english, as he called me piece of wood in https://lkml.org/lkml/2013/9/17/476 , but he appears more arogant than usual over email.

Actually he called me "piece of wood with garbage in it". I guess I have right to be offended. I'm baby in the next email. Hmm.

what a baby. Grow up dude, just grow up. I said your commit log has garbage which shouldn't be there (there was a typo you're instead of your).

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/d419425d/attachment.sig>
#15 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

In reply to Pavel Machek
Quoted 20 lines · Pavel Machek

Hi!

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 4 lines
Quoted 2 lines

So will you do that? Or it is needed to resend this one line hunk again in new email again?

new patch, new email

Guys, WHY ARE YOU SO STUPID AND ARROGANT?

Sorry but, need to copy full isolated patch/hunk from one mail to another is hassling. So what you want from me? Do all those non sense working only because yesterday you had bad day? Or what?

...

Quoted 6 lines
Quoted 3 lines

Actually, there is need to be rude, because Felipe fails to act as maintainer. Instead of fixing bugs in his code, he bounces bugfix patches, points people to random READMEs and wastes everyones time.

I don't know what are you talking about (if that happened in another thread then I need more context). Felipe is not bouncing any bugfix

Quoted 8 lines

Take a look here:

https://lkml.org/lkml/2013/9/17/286

I clearly state that patch can not be tested as required for "proper" submission, but offer patch anyway. I get irrelevant boilerplate on patch format.

Felipe didn't complain about you not being to be able to test the patch (most of the times compile tested if enough) what he said was:

"Seriously though, read that file, you're commit log has garbage in it which shouldn't go to git history".

which is totally true, if you want to comment things that don't have to go to the backlog you can't comment between the --- after your s-o-b and before the first diff. That's were you should puts comments like "Hi! or Here's suggested patch. I don't have the hardware, so it is completely untested."

Quoted 15 lines
Quoted 6 lines

but just asked to split the patch in two since the patch was solving two separate issues so is way better to have it in two separate patches for the reasons I explained before.

So, as far as I can tell Felipe did exactly what I would expect from a maintainer. He took the time to review the patches sent to him and

I'd expect maintainer to, well, maintain code. It means actually fixing bugs in his code, when he's pointed at them.

Quoted 4 lines

gave feedback. If the sender doesn't want to take his feedback into account and prefer to send pretty insulting emails instead that is his choice but I would say that is this not the greatest approach to get your code merged (to say the least).

Quoted 6 lines

Clearly not. But Pali found bug in code Felipe should maintain. Instead of "thank you for bug report, I applied this one line of your code to fix it", Pali got "new patch, new email" for his efforts. That is how you train dogs, not how you should treat kernel contributors.

No, you misunderstood the role of the maintainers. Maintainers should be custodian of a part of the kernel but not responsible for every single line of code on their sub-systems. If a piece of code is buggy then the people using and that take care of that should be fixing and maintainers should review and suggest improvements to the patches. As long as a piece of code keep compiling then it is harmless even if is buggy and nobody cares about it. If it is so broken that it doesn't even compile then the maintainer can decide to just drop it since no one else seems to care about it.

If someone finds a bug on a piece of code is because that people care about that functionality. Maintainers are really busy people and can jump and fix any random bug that someone finds on a piece of code just because it is the subsystem they maintainer neither they have to blindly merge any crappy patch just because they don't have time (or interest) in fixing a reported bug on a piece of code.

Quoted 4 lines

Now, it is possible that Felipe just has problems with english, as he called me piece of wood in https://lkml.org/lkml/2013/9/17/476 , but he appears more arogant than usual over email.

Clearly he meant "your commit log has garbage" instead of you're, that's a typo.

Quoted 1 line

Pavel

But neither Felipe needs someone to defend him nor I have time to keep arguing with you.

Have nice day!
Javier

#16 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

In reply to Javier Martinez Canillas
Quoted 2 lines · On Wednesday 18 September 2013 15:57:13 Javier Martinez Canillas wrote:

to split the patch in two since the patch was solving two separate issues

My patch does not solving *two* issues. It is *one* regression and both parts of patch are needed for fixing it. Read commit message again. It does not make sense to split patch fixing kernel regression into more one line patches - or please clarify why.

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/e9dbd364/attachment.sig>
#17 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 9 lines · Pali Roh?r
Quoted 2 lines · On Wednesday 18 September 2013 15:57:13 Javier Martinez Canillas wrote:

to split the patch in two since the patch was solving two separate issues

My patch does not solving *two* issues. It is *one* regression and both parts of patch are needed for fixing it. Read commit message again. It does not make sense to split patch fixing kernel regression into more one line patches - or please clarify why.

issue 1) twl4030-usb.c doesn't initialize atomic notifier head issue 2) musb doesn't call notifier chain

do you need a drawing ?

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/09be804b/attachment.sig>
#18 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 14 lines · On Wednesday 18 September 2013 18:36:49 Felipe Balbi wrote:
Quoted 4 lines · Pali Roh?r

On Wednesday 18 September 2013 15:57:13 Javier Martinez Canillas

wrote:

Quoted 8 lines
Quoted 2 lines

to split the patch in two since the patch was solving two separate issues

My patch does not solving *two* issues. It is *one* regression and both parts of patch are needed for fixing it. Read commit message again. It does not make sense to split patch fixing kernel regression into more one line patches - or please clarify why.

Quoted 4 lines

issue 1) twl4030-usb.c doesn't initialize atomic notifier head issue 2) musb doesn't call notifier chain

do you need a drawing ?

Regression 1) power supply drivers not working since v3.5

Look at firsts emails.

Do you really have problem to do what *you* need with 1 line patch which I sent? Or what what you want?

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/7f96b2a0/attachment.sig>
#19 lore raw

Re: [PATCH 1/4] usb: musb: Call atomic_notifier_call_chain when status is changed

Quoted 20 lines · Pali Roh?r
Quoted 14 lines · On Wednesday 18 September 2013 18:36:49 Felipe Balbi wrote:
Quoted 4 lines · Pali Roh?r

On Wednesday 18 September 2013 15:57:13 Javier Martinez Canillas

wrote:

Quoted 8 lines
Quoted 2 lines

to split the patch in two since the patch was solving two separate issues

My patch does not solving *two* issues. It is *one* regression and both parts of patch are needed for fixing it. Read commit message again. It does not make sense to split patch fixing kernel regression into more one line patches - or please clarify why.

Quoted 4 lines

issue 1) twl4030-usb.c doesn't initialize atomic notifier head issue 2) musb doesn't call notifier chain

do you need a drawing ?

Quoted 7 lines

Regression 1) power supply drivers not working since v3.5

Look at firsts emails.

Do you really have problem to do what *you* need with 1 line patch which I sent? Or what what you want?

if you had already resent patches the way I requested, they'd already be applied. Instead you chose to troll the people who are trying to helping out.

You have two issues being fixed in the same patch and that's not acceptable.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/68a1dfef/attachment-0001.sig>
#20 lore raw

PATCH 1/2 usb: musb: Add missing ATOMIC_INIT_NOTIFIER_HEAD

&twl->phy.notifier is not initalized

drivers/usb/phy/phy-twl4030-usb.c +2 -0
index 8f78d2d..efe6155 100644
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct platform_device *pdev)
 	if (device_create_file(&pdev->dev, &dev_attr_vbus))
 		dev_warn(&pdev->dev, "could not create sysfs file\n");
 
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
 	/* Our job is to use irqs and status from the power module
 	 * to keep the transceiver disabled when nothing's connected.
 	 *

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/48b5af9e/attachment.sig>

#21 lore raw

Re: [PATCH usb 1/2] usb: musb: Add missing ATOMIC_INIT_NOTIFIER_HEAD

Quoted 18 lines · On Wednesday 18 September 2013 19:03:33 Pali Roh?r wrote:

&twl->phy.notifier is not initalized

(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev) if (device_create_file(&pdev->dev,

&dev_attr_vbus)) dev_warn(&pdev->dev, "could not create sysfs file\n");

drivers/usb/phy/phy-twl4030-usb.c +2 -0
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
 	/* Our job is to use irqs and status from the power module
 	 * to keep the transceiver disabled when nothing's
Quoted 1 line

connected. *

I sent above patch week ago. Did you already included it?

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130925/e77b7f6a/attachment.sig>
#22 lore raw

Re: [PATCH usb 1/2] usb: musb: Add missing ATOMIC_INIT_NOTIFIER_HEAD

Hi,

Quoted 15 lines · Pali Roh?r
Quoted 14 lines · On Wednesday 18 September 2013 19:03:33 Pali Roh?r wrote:

&twl->phy.notifier is not initalized

(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev) if (device_create_file(&pdev->dev,

&dev_attr_vbus)) dev_warn(&pdev->dev, "could not create sysfs file\n");

Quoted 7 lines
Quoted 5 lines
drivers/usb/phy/phy-twl4030-usb.c +2 -0
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
 	/* Our job is to use irqs and status from the power module
 	 * to keep the transceiver disabled when nothing's

connected. *

I sent above patch week ago. Did you already included it?

look at my testing branch.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130925/4582c849/attachment.sig>
#23 lore raw

Re: [PATCH usb 1/2] usb: musb: Add missing ATOMIC_INIT_NOTIFIER_HEAD

Quoted 26 lines · Felipe Balbi

Hi,

Quoted 15 lines · Pali Roh?r
Quoted 14 lines · On Wednesday 18 September 2013 19:03:33 Pali Roh?r wrote:

&twl->phy.notifier is not initalized

(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev) if (device_create_file(&pdev->dev,

&dev_attr_vbus)) dev_warn(&pdev->dev, "could not create sysfs file\n");

Quoted 7 lines
Quoted 5 lines
drivers/usb/phy/phy-twl4030-usb.c +2 -0
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
 	/* Our job is to use irqs and status from the power module
 	 * to keep the transceiver disabled when nothing's

connected. *

I sent above patch week ago. Did you already included it?

Quoted 1 line

look at my testing branch.

Felipe prefers you to go MAINTAINERS file for

git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git

entry. After few clicks, you'll find out your two patches in

https://git.kernel.org/cgit/linux/kernel/git/balbi/usb.git/log/?h=testing

. So yes, it seems that after 10+ flames Felipe was _not_ lazy to send, "Thanks for a patch" mail was too much work.

Apparently, we'll need Documentation/HowToBeGoodMaintainer file...

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#24 lore raw

Re: [PATCH usb 1/2] usb: musb: Add missing ATOMIC_INIT_NOTIFIER_HEAD

Hi,

Quoted 30 lines · Pavel Machek
Quoted 24 lines · Pali Roh?r
Quoted 15 lines · On Wednesday 18 September 2013 19:03:33 Pali Roh?r wrote:
Quoted 14 lines

&twl->phy.notifier is not initalized

(quoted diff) +0 -0

b/drivers/usb/phy/phy-twl4030-usb.c index 8f78d2d..efe6155 100644

drivers/usb/phy/phy-twl4030-usb.c +0 -0
@@ -705,6 +705,8 @@ static int twl4030_usb_probe(struct
 platform_device *pdev) if (device_create_file(&pdev->dev,

&dev_attr_vbus)) dev_warn(&pdev->dev, "could not create sysfs file\n");

Quoted 7 lines
Quoted 5 lines
drivers/usb/phy/phy-twl4030-usb.c +2 -0
+	ATOMIC_INIT_NOTIFIER_HEAD(&twl->phy.notifier);
+
 	/* Our job is to use irqs and status from the power module
 	 * to keep the transceiver disabled when nothing's

connected. *

I sent above patch week ago. Did you already included it?

Quoted 1 line

look at my testing branch.

Felipe prefers you to go MAINTAINERS file for

git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git

Quoted 8 lines

entry. After few clicks, you'll find out your two patches in

https://git.kernel.org/cgit/linux/kernel/git/balbi/usb.git/log/?h=testing

. So yes, it seems that after 10+ flames Felipe was _not_ lazy to send, "Thanks for a patch" mail was too much work.

Apparently, we'll need Documentation/HowToBeGoodMaintainer file...

I only send my automated emails after properly testing them. They sit in 'testing' for as long as necessary. After all the test is done, they'll be moved to 'next' at which point you get your email notification and they reach linux-next for the next day.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131001/529e96bf/attachment-0001.sig>
#25 lore raw

PATCH 2/2 usb: musb: Call atomic_notifier_call_chain when status is changed

In reply to Felipe Balbi

More power supply drivers depends on vbus events and without it they not working. Power supply drivers using usb_register_notifier, so to deliver events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving vbus status and reporting bogus values to userspace and also to board platform data functions. Without proper data charger drivers trying to charge battery also when charger is disconnected or do not start charging when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with omap mailbox. This patch adding atomic_notifier_call_chain call from function omap_musb_set_mailbox.

drivers/usb/musb/omap2430.c +3 -1
index f44e8b5..5c40252 100644
@@ -305,6 +305,9 @@ static void omap_musb_set_mailbox(struct omap2430_glue *glue)
 	default:
 		dev_dbg(dev, "ID float\n");
 	}
+
+	atomic_notifier_call_chain(&musb->xceiv->notifier,
+			musb->xceiv->last_event, NULL);
 }
 
-------------- next part --------------

A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130918/1407c00d/attachment.sig>

#26 lore raw

PATCH 2/4 power: isp1704_charger: Fix driver to work with changes introduced in v3.5

In reply to Pali Rohár

* omap musb driver does not report USB_EVENT_ENUMERATED event anymore
* omap musb driver reporting USB_EVENT_VBUS when charger is connected
* read last event from phy->last_event (instead from ulpi register)
* do not call wall charger detection more times

---
 drivers/power/isp1704_charger.c |   91 +++++++++++++++++----------------------
 1 file changed, 40 insertions(+), 51 deletions(-)
drivers/power/isp1704_charger.c +40 -51
index fc04d19..f726cb6 100644
@@ -2,6 +2,7 @@
  * ISP1704 USB Charger Detection driver
  *
  * Copyright (C) 2010 Nokia Corporation
+ * Copyright (C) 2012 - 2013 Pali Roh?r <[email protected]>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -65,10 +66,6 @@ struct isp1704_charger {
 	unsigned		present:1;
 	unsigned		online:1;
 	unsigned		current_max;
-
-	/* temp storage variables */
-	unsigned long		event;
-	unsigned		max_power;
 };
 
 static inline int isp1704_read(struct isp1704_charger *isp, u32 reg)
@@ -231,56 +228,59 @@ static inline int isp1704_charger_detect(struct isp1704_charger *isp)
 	return ret;
 }
 
+static inline int isp1704_charger_detect_dcp(struct isp1704_charger *isp)
+{
+	if (isp1704_charger_detect(isp) &&
+		isp1704_charger_type(isp) == POWER_SUPPLY_TYPE_USB_DCP)
+		return true;
+	else
+		return false;
+}
+
 static void isp1704_charger_work(struct work_struct *data)
 {
-	int			detect;
-	unsigned long		event;
-	unsigned		power;
 	struct isp1704_charger	*isp =
 		container_of(data, struct isp1704_charger, work);
 	static DEFINE_MUTEX(lock);
 
-	event = isp->event;
-	power = isp->max_power;
-
 	mutex_lock(&lock);
 
-	if (event != USB_EVENT_NONE)
-		isp1704_charger_set_power(isp, 1);
-
-	switch (event) {
+	switch (isp->phy->last_event) {
 	case USB_EVENT_VBUS:
-		isp->online = true;
+		/* do not call wall charger detection more times */
+		if (!isp->present) {
+			isp->online = true;
+			isp->present = 1;
+			isp1704_charger_set_power(isp, 1);
+
+			/* detect wall charger */
+			if (isp1704_charger_detect_dcp(isp)) {
+				isp->psy.type = POWER_SUPPLY_TYPE_USB_DCP;
+				isp->current_max = 1800;
+			} else {
+				isp->psy.type = POWER_SUPPLY_TYPE_USB;
+				isp->current_max = 500;
+			}
 
-		/* detect charger */
-		detect = isp1704_charger_detect(isp);
-
-		if (detect) {
-			isp->present = detect;
-			isp->psy.type = isp1704_charger_type(isp);
+			/* enable data pullups */
+			if (isp->phy->otg->gadget)
+				usb_gadget_connect(isp->phy->otg->gadget);
 		}
 
-		switch (isp->psy.type) {
-		case POWER_SUPPLY_TYPE_USB_DCP:
-			isp->current_max = 1800;
-			break;
-		case POWER_SUPPLY_TYPE_USB_CDP:
+		if (isp->psy.type != POWER_SUPPLY_TYPE_USB_DCP) {
 			/*
 			 * Only 500mA here or high speed chirp
 			 * handshaking may break
 			 */
-			isp->current_max = 500;
-			/* FALLTHROUGH */
-		case POWER_SUPPLY_TYPE_USB:
-		default:
-			/* enable data pullups */
-			if (isp->phy->otg->gadget)
-				usb_gadget_connect(isp->phy->otg->gadget);
+			if (isp->current_max > 500)
+				isp->current_max = 500;
+
+			if (isp->current_max > 100)
+				isp->psy.type = POWER_SUPPLY_TYPE_USB_CDP;
 		}
 		break;
 	case USB_EVENT_NONE:
 		isp->online = false;
-		isp->current_max = 0;
 		isp->present = 0;
 		isp->current_max = 0;
 		isp->psy.type = POWER_SUPPLY_TYPE_USB;
@@ -298,12 +298,6 @@ static void isp1704_charger_work(struct work_struct *data)
 
 		isp1704_charger_set_power(isp, 0);
 		break;
-	case USB_EVENT_ENUMERATED:
-		if (isp->present)
-			isp->current_max = 1800;
-		else
-			isp->current_max = power;
-		break;
 	default:
 		goto out;
 	}
@@ -314,16 +308,11 @@ out:
 }
 
 static int isp1704_notifier_call(struct notifier_block *nb,
-		unsigned long event, void *power)
+		unsigned long val, void *v)
 {
 	struct isp1704_charger *isp =
 		container_of(nb, struct isp1704_charger, nb);
 
-	isp->event = event;
-
-	if (power)
-		isp->max_power = *((unsigned *)power);
-
 	schedule_work(&isp->work);
 
 	return NOTIFY_OK;
@@ -462,13 +451,13 @@ static int isp1704_charger_probe(struct platform_device *pdev)
 	if (isp->phy->otg->gadget)
 		usb_gadget_disconnect(isp->phy->otg->gadget);
 
+	if (isp->phy->last_event == USB_EVENT_NONE)
+		isp1704_charger_set_power(isp, 0);
+
 	/* Detect charger if VBUS is valid (the cable was already plugged). */
-	ret = isp1704_read(isp, ULPI_USB_INT_STS);
-	isp1704_charger_set_power(isp, 0);
-	if ((ret & ULPI_INT_VBUS_VALID) && !isp->phy->otg->default_a) {
-		isp->event = USB_EVENT_VBUS;
+	if (isp->phy->last_event == USB_EVENT_VBUS &&
+			!isp->phy->otg->default_a)
 		schedule_work(&isp->work);
-	}
 
 	return 0;
 fail2:
Signature · 1 line
1.7.10.4
#27 lore raw

Re: [PATCH 2/4] power: isp1704_charger: Fix driver to work with changes introduced in v3.5

Quoted 6 lines · Pali Roh?r

* omap musb driver does not report USB_EVENT_ENUMERATED event anymore
* omap musb driver reporting USB_EVENT_VBUS when charger is connected
* read last event from phy->last_event (instead from ulpi register)
* do not call wall charger detection more times

Applied, thanks a lot!

(Per others' review comments I am assuming that the second isp1704 patch will be reworked to support device-tree, so I am not applying it.)

Thanks,

Anton

#28 lore raw

PATCH 3/4 power: isp1704_charger: Add callback function set_current

In reply to Pali Rohár

This patch add callback function set_current to platform data. Driver will call this function when isp1704 change current and board provided this callback funtion in platform data.

This patch is needed for Nokia RX-51 to tell bq2415x charging chip about connected wallcharger events.

---
 drivers/power/isp1704_charger.c       |   16 ++++++++++++++++
 include/linux/power/isp1704_charger.h |    1 +
 2 files changed, 17 insertions(+)
drivers/power/isp1704_charger.c +16 -0
index f726cb6..4f20a1a 100644
@@ -91,6 +91,18 @@ static void isp1704_charger_set_power(struct isp1704_charger *isp, bool on)
 }
 
 /*
+ * Set charger current from isp1704 if a function for it
+ * has been provided with platform data.
+ */
+static void isp1704_charger_set_current(struct isp1704_charger *isp, int mA)
+{
+	struct isp1704_charger_data *board = isp->dev->platform_data;
+
+	if (board && board->set_current)
+		board->set_current(mA);
+}
+
+/*
  * Determine is the charging port DCP (dedicated charger) or CDP (Host/HUB
  * chargers).
  *
@@ -257,6 +269,7 @@ static void isp1704_charger_work(struct work_struct *data)
 			if (isp1704_charger_detect_dcp(isp)) {
 				isp->psy.type = POWER_SUPPLY_TYPE_USB_DCP;
 				isp->current_max = 1800;
+				isp1704_charger_set_current(isp, 1800);
 			} else {
 				isp->psy.type = POWER_SUPPLY_TYPE_USB;
 				isp->current_max = 500;
@@ -277,6 +290,8 @@ static void isp1704_charger_work(struct work_struct *data)
 
 			if (isp->current_max > 100)
 				isp->psy.type = POWER_SUPPLY_TYPE_USB_CDP;
+
+			isp1704_charger_set_current(isp, isp->current_max);
 		}
 		break;
 	case USB_EVENT_NONE:
@@ -297,6 +312,7 @@ static void isp1704_charger_work(struct work_struct *data)
 			usb_gadget_disconnect(isp->phy->otg->gadget);
 
 		isp1704_charger_set_power(isp, 0);
+		isp1704_charger_set_current(isp, 0);
 		break;
 	default:
 		goto out;
include/linux/power/isp1704_charger.h +1 -0
index 68096a6..d154b02 100644
@@ -24,6 +24,7 @@
 
 struct isp1704_charger_data {
 	void		(*set_power)(bool on);
+	void		(*set_current)(int mA);
 };
 
 #endif
Signature · 1 line
1.7.10.4
#29 lore raw

PATCH 4/4 RX-51: Add platform function and data for bq24150a charger

In reply to Pali Rohár

This patch will register bq24150a charger in RX-51 board data. Patch also adding platform function between isp1704 and bq2415x drivers for detecting charger type.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

---
 arch/arm/mach-omap2/board-rx51-peripherals.c |   56 +++++++++++++++++++++++++-
 1 file changed, 55 insertions(+), 1 deletion(-)
arch/arm/mach-omap2/board-rx51-peripherals.c +55 -1
index 9c2dd10..a993ffe 100644
@@ -25,6 +25,7 @@
 #include <linux/gpio_keys.h>
 #include <linux/mmc/host.h>
 #include <linux/power/isp1704_charger.h>
+#include <linux/power/bq2415x_charger.h>
 #include <linux/platform_data/spi-omap2-mcspi.h>
 #include <linux/platform_data/mtd-onenand-omap2.h>
 
@@ -270,6 +271,44 @@ static struct platform_device rx51_battery_device = {
 	.id	= -1,
 };
 
+static enum bq2415x_mode rx51_charger_mode = BQ2415X_MODE_OFF;
+static void *rx51_charger_hook_data;
+static void (*rx51_charger_hook)(enum bq2415x_mode mode, void *data);
+
+static int rx51_charger_set_hook(
+		void (*hook)(enum bq2415x_mode mode, void *data), void *data)
+{
+	rx51_charger_hook = hook;
+	rx51_charger_hook_data = data;
+	if (rx51_charger_hook)
+		rx51_charger_hook(rx51_charger_mode, rx51_charger_hook_data);
+	return 1;
+}
+
+static void rx51_charger_set_current(int mA)
+{
+	enum bq2415x_mode mode;
+
+	pr_info("RX-51: Charger current limit is %d mA\n", mA);
+
+	if (mA == 0)
+		mode = BQ2415X_MODE_OFF;
+	else if (mA < 500)
+		mode = BQ2415X_MODE_NONE;
+	else if (mA < 1800)
+		mode = BQ2415X_MODE_HOST_CHARGER;
+	else
+		mode = BQ2415X_MODE_DEDICATED_CHARGER;
+
+	if (rx51_charger_mode == mode)
+		return;
+
+	rx51_charger_mode = mode;
+
+	if (rx51_charger_hook)
+		rx51_charger_hook(rx51_charger_mode, rx51_charger_hook_data);
+}
+
 static void rx51_charger_set_power(bool on)
 {
 	gpio_set_value(RX51_USB_TRANSCEIVER_RST_GPIO, on);
@@ -277,6 +316,7 @@ static void rx51_charger_set_power(bool on)
 
 static struct isp1704_charger_data rx51_charger_data = {
 	.set_power	= rx51_charger_set_power,
+	.set_current	= rx51_charger_set_current,
 };
 
 static struct platform_device rx51_charger_device = {
@@ -1017,6 +1057,16 @@ static struct aic3x_pdata rx51_aic3x_data2 = {
 	.gpio_reset = 60,
 };
 
+static struct bq2415x_platform_data rx51_bq24150a_platform_data = {
+	.current_limit = 100,			/* mA */
+	.weak_battery_voltage = 3400,		/* mV */
+	.battery_regulation_voltage = 4200,	/* mV */
+	.charge_current = 650,			/* mA */
+	.termination_current = 100,		/* mA */
+	.resistor_sense = 68,			/* m ohm */
+	.set_mode_hook = &rx51_charger_set_hook,
+};
+
 static struct i2c_board_info __initdata rx51_peripherals_i2c_board_info_2[] = {
 	{
 		I2C_BOARD_INFO("tlv320aic3x", 0x18),
@@ -1044,7 +1094,11 @@ static struct i2c_board_info __initdata rx51_peripherals_i2c_board_info_2[] = {
 	{
 		I2C_BOARD_INFO("tpa6130a2", 0x60),
 		.platform_data = &rx51_tpa6130a2_data,
-	}
+	},
+	{
+		I2C_BOARD_INFO("bq24150a", 0x6b),
+		.platform_data = &rx51_bq24150a_platform_data,
+	},
 };
 
 static struct i2c_board_info __initdata rx51_peripherals_i2c_board_info_3[] = {
Signature · 1 line
1.7.10.4
#30 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Hi Pali,

Quoted 6 lines · Pali Roh?r

This patch will register bq24150a charger in RX-51 board data. Patch also adding platform function between isp1704 and bq2415x drivers for detecting charger type.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

cool :)

arch/arm/mach-omap2/board-rx51-peripherals.c

Quoted 3 lines

index 9c2dd10..a993ffe 100644

arch/arm/mach-omap2/board-rx51-peripherals.c +0 -0

AFAIK platform code for omap3 based boards is supposed to be removed in the near future [0]. Device Tree does not support the glue layer, so probably a small rx51 specific driver is needed.

[0] https://lkml.org/lkml/2013/8/12/70

-- Sebastian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130909/f57a5e5d/attachment-0001.sig>

#31 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

In reply to Pali Rohár
Quoted 81 lines · On Sunday 08 September 2013 10:50:39 Pali Roh?r wrote:

This patch will register bq24150a charger in RX-51 board data. Patch also adding platform function between isp1704 and bq2415x drivers for detecting charger type.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

---
 arch/arm/mach-omap2/board-rx51-peripherals.c |   56

+++++++++++++++++++++++++- 1 file changed, 55 insertions(+), 1 deletion(-)

(quoted diff) +0 -0

b/arch/arm/mach-omap2/board-rx51-peripherals.c index 9c2dd10..a993ffe 100644

arch/arm/mach-omap2/board-rx51-peripherals.c +9 -0
@@ -25,6 +25,7 @@
 #include <linux/gpio_keys.h>
 #include <linux/mmc/host.h>
 #include <linux/power/isp1704_charger.h>
+#include <linux/power/bq2415x_charger.h>
 #include <linux/platform_data/spi-omap2-mcspi.h>
 #include <linux/platform_data/mtd-onenand-omap2.h>
 
@@ -270,6 +271,44 @@ static struct platform_device
 rx51_battery_device = { .id	= -1,
 };
 
+static enum bq2415x_mode rx51_charger_mode =
 BQ2415X_MODE_OFF; +static void *rx51_charger_hook_data;
+static void (*rx51_charger_hook)(enum bq2415x_mode mode, void
 *data); +
+static int rx51_charger_set_hook(
+		void (*hook)(enum bq2415x_mode mode, void *data), void
 *data) +{
+	rx51_charger_hook = hook;
+	rx51_charger_hook_data = data;
+	if (rx51_charger_hook)
+		rx51_charger_hook(rx51_charger_mode,

rx51_charger_hook_data); + return 1;

arch/arm/mach-omap2/board-rx51-peripherals.c +24 -0
+}
+
+static void rx51_charger_set_current(int mA)
+{
+	enum bq2415x_mode mode;
+
+	pr_info("RX-51: Charger current limit is %d mA\n", mA);
+
+	if (mA == 0)
+		mode = BQ2415X_MODE_OFF;
+	else if (mA < 500)
+		mode = BQ2415X_MODE_NONE;
+	else if (mA < 1800)
+		mode = BQ2415X_MODE_HOST_CHARGER;
+	else
+		mode = BQ2415X_MODE_DEDICATED_CHARGER;
+
+	if (rx51_charger_mode == mode)
+		return;
+
+	rx51_charger_mode = mode;
+
+	if (rx51_charger_hook)
+		rx51_charger_hook(rx51_charger_mode,

rx51_charger_hook_data); +}

arch/arm/mach-omap2/board-rx51-peripherals.c +2 -0
+
 static void rx51_charger_set_power(bool on)
 {
 	gpio_set_value(RX51_USB_TRANSCEIVER_RST_GPIO, on);
@@ -277,6 +316,7 @@ static void rx51_charger_set_power(bool
 on)
 
 static struct isp1704_charger_data rx51_charger_data = {
 	.set_power	= rx51_charger_set_power,
+	.set_current	= rx51_charger_set_current,
 };
 

static struct platform_device rx51_charger_device = {

arch/arm/mach-omap2/board-rx51-peripherals.c static struct aic3x_pdata · line 1061

Quoted 5 lines
arch/arm/mach-omap2/board-rx51-peripherals.c +1 -0
@@ -1017,6 +1057,16 @@ static struct aic3x_pdata
 rx51_aic3x_data2 = { .gpio_reset = 60,
 };
 
+static struct bq2415x_platform_data
 rx51_bq24150a_platform_data = { +	.current_limit = 100,			

/*

Quoted 25 lines

mA */

arch/arm/mach-omap2/board-rx51-peripherals.c +8 -0
+	.weak_battery_voltage = 3400,		/* mV */
+	.battery_regulation_voltage = 4200,	/* mV */
+	.charge_current = 650,			/* mA */
+	.termination_current = 100,		/* mA */
+	.resistor_sense = 68,			/* m ohm */
+	.set_mode_hook = &rx51_charger_set_hook,
+};
+
 static struct i2c_board_info __initdata

rx51_peripherals_i2c_board_info_2[] = { {

arch/arm/mach-omap2/board-rx51-peripherals.c +5 -1
 		I2C_BOARD_INFO("tlv320aic3x", 0x18),
@@ -1044,7 +1094,11 @@ static struct i2c_board_info __initdata
 rx51_peripherals_i2c_board_info_2[] = { {
 		I2C_BOARD_INFO("tpa6130a2", 0x60),
 		.platform_data = &rx51_tpa6130a2_data,
-	}
+	},
+	{
+		I2C_BOARD_INFO("bq24150a", 0x6b),
+		.platform_data = &rx51_bq24150a_platform_data,
+	},
 };
 
 static struct i2c_board_info __initdata
Quoted 1 line

rx51_peripherals_i2c_board_info_3[] = {

Tony, can you look and review this board patch?

I think that this patch series it the most important for Nokia N900, because it finally bringing charging support. And without charging battery it very hard to use phone which has power supply only from battery.

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130920/f25e367a/attachment.sig>
maintainer
#32 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Quoted 7 lines · Pali Roh?r
Quoted 6 lines · On Sunday 08 September 2013 10:50:39 Pali Roh?r wrote:

This patch will register bq24150a charger in RX-51 board data. Patch also adding platform function between isp1704 and bq2415x drivers for detecting charger type.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

...

Quoted 7 lines

arch/arm/mach-omap2/board-rx51-peripherals.c static void rx51_charger_set_power(bool · line 321

Quoted 7 lines
arch/arm/mach-omap2/board-rx51-peripherals.c +1 -0
@@ -277,6 +316,7 @@ static void rx51_charger_set_power(bool
 on)
 
 static struct isp1704_charger_data rx51_charger_data = {
 	.set_power	= rx51_charger_set_power,
+	.set_current	= rx51_charger_set_current,
 };

We want to get rid of the platform data callbacks here, there no longer any need to keep these under arch/arm.

Quoted 1 line

Tony, can you look and review this board patch?

Yes, looks like this can all be done in the driver nowadays. You can use drivers/reset for the set_power. Or if it's really controlling the regulator, then the regulator framework. The info can be passed in a .dts file for those.

The .set_current you can do in the driver based on the compatible flag.

Quoted 4 lines

I think that this patch series it the most important for Nokia N900, because it finally bringing charging support. And without charging battery it very hard to use phone which has power supply only from battery.

Right, let's get this driver updated to use the device tree based init and that way this file is no longer needed. I would like to $ grep -i grandmom ~/.phonebook | call too :)

I forgot how this charger is wired up, but maybe take a look at commit d7bf353f (bq24190_charger: Add support for TI BQ24190 Battery Charger) for the DT parts.

Regards,

Tony

#33 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Hello,

Quoted 17 lines · On Monday 23 September 2013 20:03:00 you wrote:
Quoted 8 lines · Pali Roh?r
Quoted 7 lines · On Sunday 08 September 2013 10:50:39 Pali Roh?r wrote:

This patch will register bq24150a charger in RX-51 board data. Patch also adding platform function between isp1704 and bq2415x drivers for detecting charger type.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

...

Quoted 5 lines
Quoted 5 lines
arch/arm/mach-omap2/board-rx51-peripherals.c +0 -0
@@ -277,6 +316,7 @@ static void
 rx51_charger_set_power(bool on)
 
 static struct isp1704_charger_data rx51_charger_data = {
 
Quoted 9 lines
Quoted 5 lines
Quoted 5 lines
arch/arm/mach-omap2/board-rx51-peripherals.c +1 -0
 	.set_power	= rx51_charger_set_power,
 
+	.set_current	= rx51_charger_set_current,

};

We want to get rid of the platform data callbacks here, there no longer any need to keep these under arch/arm.

Where to put rx51 board specified functions?
It cannot go to DT, because DT does not support C/ASM code.

Quoted 10 lines
Quoted 1 line

Tony, can you look and review this board patch?

Yes, looks like this can all be done in the driver nowadays. You can use drivers/reset for the set_power. Or if it's really controlling the regulator, then the regulator framework. The info can be passed in a .dts file for those.

The .set_current you can do in the driver based on the compatible flag.

It is not as simple as it looks. This is reason why I submited this patch long time after I submited bq2415x driver.

Problem is that for rx51 is needed specific function which connect to two drivers (bq2415x and isp1704) plus it call specific rx51 board functions.

Something which cannot be in DT (unless DT support C/ASM code).

Quoted 9 lines
Quoted 4 lines

I think that this patch series it the most important for Nokia N900, because it finally bringing charging support. And without charging battery it very hard to use phone which has power supply only from battery.

Right, let's get this driver updated to use the device tree based init and that way this file is no longer needed. I would like to $ grep -i grandmom ~/.phonebook | call too :)

Patches for modem support are being prepared for upstreaming :-) so after that it is up to you to create "call" script as you want

Quoted 7 lines

I forgot how this charger is wired up, but maybe take a look at commit d7bf353f (bq24190_charger: Add support for TI BQ24190 Battery Charger) for the DT parts.

Regards,

Tony

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130923/2395c3e1/attachment.sig>
#34 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Quoted 8 lines · Pali Roh?r

It is not as simple as it looks. This is reason why I submited this patch long time after I submited bq2415x driver.

Problem is that for rx51 is needed specific function which connect to two drivers (bq2415x and isp1704) plus it call specific rx51 board functions.

Something which cannot be in DT (unless DT support C/ASM code).

mh could isp1704 driver expose the data via the regulator framework? Then the bq2415x chip can just use the regulator framework. This should make converting to DT easy (by giving the bq2415x chip the isp1704 as regulator) and uses existing standard interfaces.

Quoted 2 lines

Patches for modem support are being prepared for upstreaming :-) so after that it is up to you to create "call" script as you want

I'm on it. RFC round will be sent out this week. It seems hwmod data is already finished, since i didn't get feedback for that patch :)

-- Sebastian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130923/853a6857/attachment.sig>

#35 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Quoted 9 lines · On Monday 23 September 2013 22:00:09 Sebastian Reichel wrote:
Quoted 8 lines · Pali Roh?r

It is not as simple as it looks. This is reason why I submited this patch long time after I submited bq2415x driver.

Problem is that for rx51 is needed specific function which connect to two drivers (bq2415x and isp1704) plus it call specific rx51 board functions.

Quoted 5 lines
Quoted 2 lines

Something which cannot be in DT (unless DT support C/ASM code).

mh could isp1704 driver expose the data via the regulator framework?

No, isp1704 is power supply driver and export data via power supply (sysfs) interface. It is not regulator but charger driver.

Quoted 14 lines

Then the bq2415x chip can just use the regulator framework. This should make converting to DT easy (by giving the bq2415x chip the isp1704 as regulator) and uses existing standard interfaces.

Quoted 3 lines

Patches for modem support are being prepared for upstreaming :-) so after that it is up to you to create "call" script as you want

I'm on it. RFC round will be sent out this week. It seems hwmod data is already finished, since i didn't get feedback for that patch :)

-- Sebastian

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130923/07adf26c/attachment.sig>
#36 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Quoted 10 lines · Pali Roh?r
Quoted 9 lines · On Monday 23 September 2013 22:00:09 Sebastian Reichel wrote:
Quoted 8 lines · Pali Roh?r

It is not as simple as it looks. This is reason why I submited this patch long time after I submited bq2415x driver.

Problem is that for rx51 is needed specific function which connect to two drivers (bq2415x and isp1704) plus it call specific rx51 board functions.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Something which cannot be in DT (unless DT support C/ASM code).

mh could isp1704 driver expose the data via the regulator framework?

No, isp1704 is power supply driver and export data via power supply (sysfs) interface. It is not regulator but charger driver.

well it does not charge the battery directly, but just provides a power line with 5 Volt and a specified amount of current to the system, doesn't it?

Quoted 1 line

From my POV this is a candiate for the regulator framework:

https://www.kernel.org/doc/Documentation/power/regulator/overview.txt

-- Sebastian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130923/458b2b62/attachment.sig>

maintainer
#37 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Quoted 20 lines · Sebastian Reichel
Quoted 10 lines · Pali Roh?r
Quoted 9 lines · On Monday 23 September 2013 22:00:09 Sebastian Reichel wrote:
Quoted 8 lines · Pali Roh?r

It is not as simple as it looks. This is reason why I submited this patch long time after I submited bq2415x driver.

Problem is that for rx51 is needed specific function which connect to two drivers (bq2415x and isp1704) plus it call specific rx51 board functions.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Something which cannot be in DT (unless DT support C/ASM code).

mh could isp1704 driver expose the data via the regulator framework?

No, isp1704 is power supply driver and export data via power supply (sysfs) interface. It is not regulator but charger driver.

Quoted 7 lines

well it does not charge the battery directly, but just provides a power line with 5 Volt and a specified amount of current to the system, doesn't it?

From my POV this is a candiate for the regulator framework:

https://www.kernel.org/doc/Documentation/power/regulator/overview.txt

Yes I think we should be able handle that rail with the regulator framework.

Regards,

Tony

#38 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

In reply to Sebastian Reichel

Hi!

Quoted 6 lines
Quoted 2 lines

No, isp1704 is power supply driver and export data via power supply (sysfs) interface. It is not regulator but charger driver.

well it does not charge the battery directly, but just provides a power line with 5 Volt and a specified amount of current to the system, doesn't it?

I don't want to talk about isp1704, but...

LiION charger is just something that provides 4.2V power line with C/10 current limitation...

(IOW line between regulators and chargers is very very thin, if it even exists).

Pavel

(who did use laboratory power supply as emergency battery charger)

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#39 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

In reply to Sebastian Reichel
Quoted 17 lines · On Monday 23 September 2013 22:47:15 Sebastian Reichel wrote:
Quoted 16 lines · Pali Roh?r
Quoted 14 lines · On Monday 23 September 2013 22:00:09 Sebastian Reichel wrote:
Quoted 10 lines · Pali Roh?r

It is not as simple as it looks. This is reason why I submited this patch long time after I submited bq2415x driver.

Problem is that for rx51 is needed specific function which connect to two drivers (bq2415x and isp1704) plus it call specific rx51 board functions.

Something which cannot be in DT (unless DT support C/ASM code).

mh could isp1704 driver expose the data via the regulator framework?

Quoted 8 lines
Quoted 3 lines

No, isp1704 is power supply driver and export data via power supply (sysfs) interface. It is not regulator but charger driver.

well it does not charge the battery directly, but just provides a power line with 5 Volt and a specified amount of current to the system, doesn't it?

No, isp1704 driver is doing fastcharger detection (and then export charger type via sysfs power supply) based on musb usb events.

Real charging (enabling/disabling and setting properties) is done by bq24150a chip which has own power driver bq2415x_charger.

As already wrote this is not simple and this is reason why I sent board data and functions only now...

Quoted 6 lines

From my POV this is a candiate for the regulator framework:

https://www.kernel.org/doc/Documentation/power/regulator/overv
iew.txt

-- Sebastian

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130924/23254e9e/attachment.sig>
#40 lore raw

Re: [PATCH 4/4] RX-51: Add platform function and data for bq24150a charger

Hi,

Quoted 9 lines · Pali Roh?r

No, isp1704 driver is doing fastcharger detection (and then export charger type via sysfs power supply) based on musb usb events.

Real charging (enabling/disabling and setting properties) is done by bq24150a chip which has own power driver bq2415x_charger.

As already wrote this is not simple and this is reason why I sent board data and functions only now...

Yes, I'm aware of this. Technically the isp1704 does not provide the 5 volt line, but for the system as a whole that fact does not really matter.

The isp1704 can provide its functionality via the regulator API as a simple regulator device. It provides information about the power "it can supply".

The bq24150a on the other hand can just use the regulator via the consumer interface.

The regulator framework provides capabilities for events:
"Regulators can notify consumers of external events"

-- Sebastian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130924/649da08c/attachment.sig>

#41 lore raw
In reply to Pali Rohár
Quoted 11 lines · On Sunday 08 September 2013 10:50:35 Pali Roh?r wrote:

This patch series finally bringing support for charging battery on Nokia N900 (RX-51) without any proprietary Nokia bits in userspace.

Pali Roh?r (4):
usb: musb: Call atomic_notifier_call_chain when status is
changed power: isp1704_charger: Fix driver to work with changes introduced in v3.5
power: isp1704_charger: Add callback function set_current
RX-51: Add platform function and data for bq24150a charger

Quoted 6 lines
 arch/arm/mach-omap2/board-rx51-peripherals.c |   56
+++++++++++++- drivers/power/isp1704_charger.c              |
 107 ++++++++++++++------------ drivers/usb/musb/omap2430.c  
               |    3 + drivers/usb/phy/phy-twl4030-usb.c    
       |    2 + include/linux/power/isp1704_charger.h       

| 1 + 5 files changed, 117 insertions(+), 52 deletions(-)

Hello, can you review this patch series?

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130914/5e97a21a/attachment.sig>
#42 lore raw
In reply to Pali Rohár
Quoted 11 lines · On Sunday 08 September 2013 10:50:35 Pali Roh?r wrote:

This patch series finally bringing support for charging battery on Nokia N900 (RX-51) without any proprietary Nokia bits in userspace.

Pali Roh?r (4):
usb: musb: Call atomic_notifier_call_chain when status is
changed power: isp1704_charger: Fix driver to work with changes introduced in v3.5
power: isp1704_charger: Add callback function set_current
RX-51: Add platform function and data for bq24150a charger

Quoted 6 lines
 arch/arm/mach-omap2/board-rx51-peripherals.c |   56
+++++++++++++- drivers/power/isp1704_charger.c              |
 107 ++++++++++++++------------ drivers/usb/musb/omap2430.c  
               |    3 + drivers/usb/phy/phy-twl4030-usb.c    
       |    2 + include/linux/power/isp1704_charger.h       

| 1 + 5 files changed, 117 insertions(+), 52 deletions(-)

Anton, can you review this patch series, specially power supply code (changes to isp1704_charger.c driver)?

Signature · 9 lines
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130921/ac95098d/attachment.sig>
#43 lore raw

v2 cover Add support for charging battery in Nokia RX-51

In reply to Pali Rohár

This patch series finally bringing support for charging battery on Nokia N900 (RX-51) without any proprietary Nokia bits in userspace.

I removed hook function from bq2415x_charger driver and replaced it by Jenny TC power_supply notifier code. This simplify usage of bq2415x driver and also allow adding DT support in future because driver does not need board specific function anymore (only static data).

I tested this series with linux 3.12-rc5 on Nokia N900.

Pali Rohár (3):
power_supply: Add power_supply notifier
bq2415x_charger: Use power_supply notifier for automode
RX-51: Add platform data for bq24150a charger

 arch/arm/mach-omap2/board-rx51-peripherals.c |   17 +++++-
 drivers/power/bq2415x_charger.c              |   77 ++++++++++++++++++++------
 drivers/power/power_supply_core.c            |   17 ++++++
 include/linux/power/bq2415x_charger.h        |   48 ++--------------
 include/linux/power_supply.h                 |   10 ++++
 5 files changed, 108 insertions(+), 61 deletions(-)
Signature · 1 line
1.7.9.5
#44 lore raw

PATCH v2 1/3 power_supply: Add power_supply notifier

This patch adds a notifier chain to the power_supply.
This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties. One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

---
 drivers/power/power_supply_core.c |   17 +++++++++++++++++
 include/linux/power_supply.h      |   10 ++++++++++
 2 files changed, 27 insertions(+)
drivers/power/power_supply_core.c +17 -0
index 00e6672..08bce22 100644
@@ -24,6 +24,9 @@
 struct class *power_supply_class;
 EXPORT_SYMBOL_GPL(power_supply_class);
 
+ATOMIC_NOTIFIER_HEAD(power_supply_notifier);
+EXPORT_SYMBOL_GPL(power_supply_notifier);
+
 static struct device_type power_supply_dev_type;
 
 static bool __power_supply_is_supplied_by(struct power_supply *supplier,
@@ -80,6 +83,8 @@ static void power_supply_changed_work(struct work_struct *work)
 		class_for_each_device(power_supply_class, NULL, psy,
 				      __power_supply_changed_work);
 		power_supply_update_leds(psy);
+		atomic_notifier_call_chain(&power_supply_notifier,
+				PSY_EVENT_PROP_CHANGED, psy);
 		kobject_uevent(&psy->dev->kobj, KOBJ_CHANGE);
 		spin_lock_irqsave(&psy->changed_lock, flags);
 	}
@@ -347,6 +352,18 @@ static void power_supply_dev_release(struct device *dev)
 	kfree(dev);
 }
 
+int power_supply_reg_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_register(&power_supply_notifier, nb);
+}
+EXPORT_SYMBOL_GPL(power_supply_reg_notifier);
+
+void power_supply_unreg_notifier(struct notifier_block *nb)
+{
+	atomic_notifier_chain_unregister(&power_supply_notifier, nb);
+}
+EXPORT_SYMBOL_GPL(power_supply_unreg_notifier);
+
 #ifdef CONFIG_THERMAL
 static int power_supply_read_temp(struct thermal_zone_device *tzd,
 		unsigned long *temp)
include/linux/power_supply.h +10 -0
index 5c26006..c6f52c0 100644
@@ -158,6 +158,13 @@ enum power_supply_type {
 	POWER_SUPPLY_TYPE_USB_ACA,	/* Accessory Charger Adapters */
 };
 
+enum power_supply_notifier_events {
+	PSY_EVENT_NONE,
+	PSY_EVENT_PROP_CHANGED,
+	PSY_EVENT_BATTERY,
+	PSY_EVENT_CABLE,
+};
+
 union power_supply_propval {
 	int intval;
 	const char *strval;
@@ -235,6 +242,9 @@ struct power_supply_info {
 	int use_for_apm;
 };
 
+extern struct atomic_notifier_head    power_supply_notifier;
+extern int power_supply_reg_notifier(struct notifier_block *nb);
+extern void power_supply_unreg_notifier(struct notifier_block *nb);
 extern struct power_supply *power_supply_get_by_name(const char *name);
 extern void power_supply_changed(struct power_supply *psy);
 extern int power_supply_am_i_supplied(struct power_supply *psy);
Signature · 1 line
1.7.9.5
Reviewed-by on v2 1/3
#45 lore raw

Re: [PATCH v2 1/3] power_supply: Add power_supply notifier

Quoted 10 lines · Pali Rohár

This patch adds a notifier chain to the power_supply.
This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties. One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

Quoted 2 lines
Quoted 1 line

+extern struct atomic_notifier_head power_supply_notifier;

You may want to delete some spaces here.

Thanks,
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#46 lore raw

Re: [PATCH v2 1/3] power_supply: Add power_supply notifier

Quoted 18 lines · On Tuesday 19 November 2013 14:10:06 Pavel Machek wrote:
Quoted 12 lines · Pali Rohár

This patch adds a notifier chain to the power_supply. This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties. One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

Quoted 2 lines
Quoted 7 lines
Quoted 2 lines

+extern struct atomic_notifier_head
power_supply_notifier;

You may want to delete some spaces here.

Thanks,
Pavel

Ok, note that this patch I copied from:
https://lkml.org/lkml/2013/9/23/181

Signature · 2 lines
Pali Rohár
[email protected]
Tested-by on v2 1/3
#47 lore raw

Re: [PATCH v2 1/3] power_supply: Add power_supply notifier

In reply to Pali Rohár

Hi,

Quoted 11 lines · Pali Rohár

This patch adds a notifier chain to the power_supply.
This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties. One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

[...]

-- Sebastian

Acked-by on v2 1/3
#48 lore raw

RE: [PATCH v2 1/3] power_supply: Add power_supply notifier

In reply to Pali Rohár
Quoted 12 lines

Subject: [PATCH v2 1/3] power_supply: Add power_supply notifier

This patch adds a notifier chain to the power_supply.
This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties.
One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

Quoted 3 lines

---

#49 lore raw

Re: [PATCH v2 1/3] power_supply: Add power_supply notifier

Quoted 12 lines · Jenny
Quoted 12 lines

Subject: [PATCH v2 1/3] power_supply: Add power_supply notifier

This patch adds a notifier chain to the power_supply.
This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties.
One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

Quoted 5 lines
Quoted 3 lines

---

Applied, thanks a lot!

Anton

#50 lore raw

Re: [PATCH v2 1/3] power_supply: Add power_supply notifier

In reply to Pali Rohár
Quoted 12 lines · Pali Rohár

This patch adds a notifier chain to the power_supply.
This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties. One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

...

Review

Quoted 2 lines
(quoted diff) +2 -0
+enum power_supply_notifier_events {
+	PSY_EVENT_NONE,

This one is not needed.

Review

Quoted 4 lines
(quoted diff) +4 -0
+	PSY_EVENT_PROP_CHANGED,
+	PSY_EVENT_BATTERY,
+	PSY_EVENT_CABLE,
+};

The only event that is currently used in your patch series is EVENT_PROP_CHANGED... So, I applied the patch with the following changes:

include/linux/power_supply.h +2 -4
index c6f52c0..0c2a260 100644
@@ -15,6 +15,7 @@
 #include <linux/init.h>
 #include <linux/slab.h>
 #include <linux/device.h>
+#include <linux/notifier.h>
 #include <linux/err.h>
 #include <linux/power_supply.h>
 #include <linux/thermal.h>
@@ -159,10 +159,7 @@ enum power_supply_type {
 };
 
 enum power_supply_notifier_events {
-	PSY_EVENT_NONE,
 	PSY_EVENT_PROP_CHANGED,
-	PSY_EVENT_BATTERY,
-	PSY_EVENT_CABLE,
 };
 
 union power_supply_propval {
@@ -242,7 +239,7 @@ struct power_supply_info {
 	int use_for_apm;
 };
 
-extern struct atomic_notifier_head    power_supply_notifier;
+extern struct atomic_notifier_head power_supply_notifier;
 extern int power_supply_reg_notifier(struct notifier_block *nb);
 extern void power_supply_unreg_notifier(struct notifier_block *nb);
 extern struct power_supply *power_supply_get_by_name(const char *name);

--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#51 lore raw

RE: [PATCH v2 1/3] power_supply: Add power_supply notifier

Anton,

I don't see this patch in Linux tree. Any update on this would be helpful

Thanks
Jenny

Quoted 68 lines

-----Original Message-----
From: Anton Vorontsov [mailto:[email protected]] On Behalf Of Anton Vorontsov
Sent: Monday, December 02, 2013 3:54 AM
To: Pali Rohár
Cc: David Woodhouse; Tony Lindgren; Russell King; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Tc, Jenny
Subject: Re: [PATCH v2 1/3] power_supply: Add power_supply notifier

Quoted 12 lines · Pali Rohár

This patch adds a notifier chain to the power_supply.
This notifier helps drivers in other subsystem to listen to changes in power supply subsystem. This would help to take some actions in those drivers on changing the power supply properties.
One such scenario is to increase/decrease system performance based on the battery capacity/voltage. Another scenario is to adjust the h/w peak current detection voltage/current thresholds based on battery voltage/capacity. The notifier helps drivers to listen to changes in power_suppy susbystem without polling the power_supply properties

...

Review

Quoted 2 lines
(quoted diff) +2 -0
+enum power_supply_notifier_events {
+	PSY_EVENT_NONE,

This one is not needed.

Review

Quoted 4 lines
(quoted diff) +4 -0
+	PSY_EVENT_PROP_CHANGED,
+	PSY_EVENT_BATTERY,
+	PSY_EVENT_CABLE,
+};

The only event that is currently used in your patch series is EVENT_PROP_CHANGED... So, I applied the patch with the following changes:

include/linux/power_supply.h index renamed from include/linux/power_supply.h +0 -0

c6f52c0..0c2a260 100644

include/linux/power_supply.h +2 -4
@@ -15,6 +15,7 @@
 #include <linux/init.h>
 #include <linux/slab.h>
 #include <linux/device.h>
+#include <linux/notifier.h>
 #include <linux/err.h>
 #include <linux/power_supply.h>
 #include <linux/thermal.h>
@@ -159,10 +159,7 @@ enum power_supply_type {  };
 
 enum power_supply_notifier_events {
-	PSY_EVENT_NONE,
 	PSY_EVENT_PROP_CHANGED,
-	PSY_EVENT_BATTERY,
-	PSY_EVENT_CABLE,
 };
 
 union power_supply_propval {
@@ -242,7 +239,7 @@ struct power_supply_info {
 	int use_for_apm;
 };
 
-extern struct atomic_notifier_head    power_supply_notifier;
+extern struct atomic_notifier_head power_supply_notifier;
 extern int power_supply_reg_notifier(struct notifier_block *nb);  extern void

power_supply_unreg_notifier(struct notifier_block *nb); extern struct power_supply *power_supply_get_by_name(const char *name);

#53 lore raw

PATCH v2 2/3 bq2415x_charger: Use power_supply notifier for automode

In reply to Pali Rohár

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

---
 drivers/power/bq2415x_charger.c       |   77 +++++++++++++++++++++++++--------
 include/linux/power/bq2415x_charger.h |   48 +++-----------------
 2 files changed, 65 insertions(+), 60 deletions(-)
drivers/power/bq2415x_charger.c +60 -17
index 0727f92..d89583d 100644
@@ -1,7 +1,7 @@
 /*
  * bq2415x charger driver
  *
- * Copyright (C) 2011-2012  Pali Rohár <[email protected]>
+ * Copyright (C) 2011-2013  Pali Rohár <[email protected]>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -170,6 +170,7 @@ struct bq2415x_device {
 	struct bq2415x_platform_data init_data;
 	struct power_supply charger;
 	struct delayed_work work;
+	struct notifier_block nb;
 	enum bq2415x_mode reported_mode;/* mode reported by hook function */
 	enum bq2415x_mode mode;		/* current configured mode */
 	enum bq2415x_chip chip;
@@ -791,24 +792,53 @@ static int bq2415x_set_mode(struct bq2415x_device *bq, enum bq2415x_mode mode)
 
 }
 
-/* hook function called by other driver which set reported mode */
-static void bq2415x_hook_function(enum bq2415x_mode mode, void *data)
+static int bq2415x_notifier_call(struct notifier_block *nb,
+		unsigned long val, void *v)
 {
-	struct bq2415x_device *bq = data;
+	struct bq2415x_device *bq =
+		container_of(nb, struct bq2415x_device, nb);
+	struct power_supply *psy = v;
+	enum bq2415x_mode mode;
+	union power_supply_propval prop;
+	int ret;
+	int mA;
 
-	if (!bq)
-		return;
+	if (val != PSY_EVENT_PROP_CHANGED)
+		return NOTIFY_OK;
+
+	if (strcmp(psy->name, bq->init_data.notify_device) != 0)
+		return NOTIFY_OK;
+
+	dev_dbg(bq->dev, "notifier call was called\n");
+
+	ret = psy->get_property(psy, POWER_SUPPLY_PROP_CURRENT_MAX, &prop);
+	if (ret != 0)
+		return NOTIFY_OK;
+
+	mA = prop.intval;
+
+	if (mA == 0)
+		mode = BQ2415X_MODE_OFF;
+	else if (mA < 500)
+		mode = BQ2415X_MODE_NONE;
+	else if (mA < 1800)
+		mode = BQ2415X_MODE_HOST_CHARGER;
+	else
+		mode = BQ2415X_MODE_DEDICATED_CHARGER;
+
+	if (bq->reported_mode == mode)
+		return NOTIFY_OK;
 
-	dev_dbg(bq->dev, "hook function was called\n");
 	bq->reported_mode = mode;
 
 	/* if automode is not enabled do not tell about reported_mode */
 	if (bq->automode < 1)
-		return;
+		return NOTIFY_OK;
 
 	sysfs_notify(&bq->charger.dev->kobj, NULL, "reported_mode");
 	bq2415x_set_mode(bq, bq->reported_mode);
 
+	return NOTIFY_OK;
 }
 
 /**** timer functions ****/
@@ -1508,6 +1538,7 @@ static int bq2415x_probe(struct i2c_client *client,
 	int num;
 	char *name;
 	struct bq2415x_device *bq;
+	struct power_supply *psy;
 
 	if (!client->dev.platform_data) {
 		dev_err(&client->dev, "platform data not set\n");
@@ -1569,16 +1600,27 @@ static int bq2415x_probe(struct i2c_client *client,
 		goto error_4;
 	}
 
-	if (bq->init_data.set_mode_hook) {
-		if (bq->init_data.set_mode_hook(
-				bq2415x_hook_function, bq)) {
-			bq->automode = 1;
+	if (bq->init_data.notify_device) {
+		bq->nb.notifier_call = bq2415x_notifier_call;
+		ret = power_supply_reg_notifier(&bq->nb);
+		if (ret) {
+			dev_err(bq->dev, "failed to reg notifier: %d\n", ret);
+			goto error_5;
+		}
+		psy = power_supply_get_by_name(bq->init_data.notify_device);
+		if (psy) {
+			/* Query for initial reported_mode and set it */
+			bq2415x_notifier_call(&bq->nb,
+					PSY_EVENT_PROP_CHANGED, psy);
 			bq2415x_set_mode(bq, bq->reported_mode);
-			dev_info(bq->dev, "automode enabled\n");
 		} else {
-			bq->automode = -1;
-			dev_info(bq->dev, "automode failed\n");
+			dev_info(bq->dev, "notifier power supply device (%s) "
+				"for automode is not registred yet... "
+				"automode will not work without that device\n",
+				bq->init_data.notify_device);
 		}
+		bq->automode = 1;
+		dev_info(bq->dev, "automode enabled\n");
 	} else {
 		bq->automode = -1;
 		dev_info(bq->dev, "automode not supported\n");
@@ -1590,6 +1632,7 @@ static int bq2415x_probe(struct i2c_client *client,
 	dev_info(bq->dev, "driver registered\n");
 	return 0;
 
+error_5:
 error_4:
 	bq2415x_sysfs_exit(bq);
 error_3:
@@ -1610,8 +1653,8 @@ static int bq2415x_remove(struct i2c_client *client)
 {
 	struct bq2415x_device *bq = i2c_get_clientdata(client);
 
-	if (bq->init_data.set_mode_hook)
-		bq->init_data.set_mode_hook(NULL, NULL);
+	if (bq->init_data.notify_device)
+		power_supply_unreg_notifier(&bq->nb);
 
 	bq2415x_sysfs_exit(bq);
 	bq2415x_power_supply_exit(bq);
include/linux/power/bq2415x_charger.h +5 -43

66 lines not on this page · raw message

Signature · 1 line
1.7.9.5
Reviewed-by on v2 2/3
#54 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 6 lines · Pali Rohár

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#55 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 9 lines · Pavel Machek
Quoted 6 lines · Pali Rohár

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

Applied, thanks!

Anton

#56 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

In reply to Pali Rohár

Hi,

Quoted 4 lines · Pali Rohár

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

I'm wondering if the watchdog thread should check some values, like e.g. battery temperature. It should stop charging the battery if some critical battery temperature threshold is reached.

-- Sebastian

#57 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 10 lines · On Sunday 24 November 2013 18:00:00 Sebastian Reichel wrote:

Hi,

Quoted 6 lines · Pali Rohár

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

Quoted 6 lines

I'm wondering if the watchdog thread should check some values, like e.g. battery temperature. It should stop charging the battery if some critical battery temperature threshold is reached.

-- Sebastian

For checking battery temperature is needed third driver (rx51_battery.ko). And I think this should not be implemented in driver itself but in charger manager framework...

Currently on Maemo 5 this is handled in userspace (with open source dsme daemon).

Signature · 2 lines
Pali Rohár
[email protected]
#58 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 7 lines · Pali Rohár
Quoted 4 lines

I'm wondering if the watchdog thread should check some values, like e.g. battery temperature. It should stop charging the battery if some critical battery temperature threshold is reached.

For checking battery temperature is needed third driver (rx51_battery.ko).

I know, but that doesn't mean one can skip the check.

Quoted 2 lines

And I think this should not be implemented in driver itself but in charger manager framework...

There is a reason, that the HW has a watchdog. I think the safety relevant stuff should be done in the watchdog thread. This ensures, that charging depends on reading the safety relevant sensors.

Of course the watchdog stuff could be moved into the charger manager framework...

Quoted 2 lines

Currently on Maemo 5 this is handled in userspace (with open source dsme daemon).

I assume it currently also takes care of the bq2415x watchdog? That means if the daemon dies charging will stop, because the watchdog does no longer trigger.

When your patch is applied you have introduced a safety issue. When the daemon dies charging will continue and temperature is no longer checked.

-- Sebastian

#59 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 8 lines · On Sunday 24 November 2013 20:26:09 Sebastian Reichel wrote:
Quoted 2 lines · Pali Rohár

Currently on Maemo 5 this is handled in userspace (with open source dsme daemon).

I assume it currently also takes care of the bq2415x watchdog? That means if the daemon dies charging will stop, because the watchdog does no longer trigger.

Quoted 5 lines

When your patch is applied you have introduced a safety issue. When the daemon dies charging will continue and temperature is no longer checked.

-- Sebastian

No dsme checking battery temperature and does not handle bq24510 timer (previously this was in closed bme daemon which functionality is now in kernel drivers). But dsme daemon also kicking tlw4030 watchdog, so when daemon dies after 30s tlw4030 reboot device.

But right you can implement correctly this in userspace (e.g. when daemon not running/crashed, you can restart daemon or reboot system or disable charing, whatever...) and you do not need to have it in kernel...

Signature · 2 lines
Pali Rohár
[email protected]
#60 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 13 lines · Pali Rohár
Quoted 12 lines · On Sunday 24 November 2013 20:26:09 Sebastian Reichel wrote:
Quoted 2 lines · Pali Rohár

Currently on Maemo 5 this is handled in userspace (with open source dsme daemon).

I assume it currently also takes care of the bq2415x watchdog? That means if the daemon dies charging will stop, because the watchdog does no longer trigger.

When your patch is applied you have introduced a safety issue. When the daemon dies charging will continue and temperature is no longer checked.

Quoted 5 lines
Quoted 1 line

-- Sebastian

No dsme checking battery temperature and does not handle bq24510 timer (previously this was in closed bme daemon which functionality is now in kernel drivers).

Ok, so then the temperature checks are done by bme.

Quoted 2 lines

But dsme daemon also kicking tlw4030 watchdog, so when daemon dies after 30s tlw4030 reboot device.

Yes, but that does not matter if it does not take care of the charging. If the dsme daemon is not started for some reason phone will not be rebootet and phone charges.

Quoted 4 lines

But right you can implement correctly this in userspace (e.g. when daemon not running/crashed, you can restart daemon or reboot system or disable charing, whatever...) and you do not need to have it in kernel...

That does not change the fact, that the kernel openes a safety issue in default configuration. To be on the safe side charging must be disabled until some safety checking daemon enables it.

-- Sebastian

#61 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

In reply to Pali Rohár

Hi!

Quoted 21 lines · Pali Rohár
Quoted 8 lines · On Sunday 24 November 2013 20:26:09 Sebastian Reichel wrote:
Quoted 2 lines · Pali Rohár

Currently on Maemo 5 this is handled in userspace (with open source dsme daemon).

I assume it currently also takes care of the bq2415x watchdog? That means if the daemon dies charging will stop, because the watchdog does no longer trigger.

Quoted 5 lines

When your patch is applied you have introduced a safety issue. When the daemon dies charging will continue and temperature is no longer checked.

-- Sebastian

No dsme checking battery temperature and does not handle bq24510 timer (previously this was in closed bme daemon which functionality is now in kernel drivers). But dsme daemon also kicking tlw4030 watchdog, so when daemon dies after 30s tlw4030 reboot device.

Quoted 4 lines

But right you can implement correctly this in userspace (e.g. when daemon not running/crashed, you can restart daemon or reboot system or disable charing, whatever...) and you do not need to have it in kernel...

Charging handling really should be in kernel.

1st) you want charging to work with init=/bin/bash, and you want it with regular (not n900-specific) debian/arm.

2st) normally hardware, firmware or kernel does the charging. (PCs, Sharp Zaurus). Yes, it pulls us farther away from Maemo...

Where can I get dsme sources?
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#62 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 2 lines · On Monday 25 November 2013 15:01:27 Pavel Machek wrote:

Where can I get dsme sources?
Pavel

dsme daemon: https://gitorious.org/community-ssu/dsme
dsme thermal plugin: https://gitorious.org/rx51-bme-replacement/dsme-thermalobject-surface

dsme thermal plugin is responsible for checking temperature of battery and send it to dsme damon. dsme daemon itself checking temperature limits and do something (e.g reboot device)

in dsme thermal plugin git history you can find both old version (which asking temperature from proprietary nokia bme daemon) and new version which reading temperature from power supply kernel modules.

Signature · 2 lines
Pali Rohár
[email protected]
#63 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 10 lines · Pali Rohár
Quoted 2 lines · On Monday 25 November 2013 15:01:27 Pavel Machek wrote:

Where can I get dsme sources?
Pavel

dsme daemon: https://gitorious.org/community-ssu/dsme
dsme thermal plugin: https://gitorious.org/rx51-bme-replacement/dsme-thermalobject-surface

dsme thermal plugin is responsible for checking temperature of battery and send it to dsme damon. dsme daemon itself checking temperature limits and do something (e.g reboot device)

Thanks for links!

AFAICT.... dsme daemon does some kind of system management, not directly controlling charging, right? It does monitor system temperature.

Now.. Imagine phone left in car in charger (on sun). Likely temperature will reach high values, kernel is charging, dsme will reboot the system, but kernel will start charging again, dsme will reboot again, ...

And it is not only high temperatures that are problem for li-ion charging; battery should not be fast charged below 5C and should not be charged below 0C. (Again, both are likely to happen if you leave your phone in car).

AFAICT, we should simply disable charging below 5C or above 45C.

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#64 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 16 lines · On Monday 25 November 2013 16:18:39 Pavel Machek wrote:
Quoted 10 lines · Pali Rohár
Quoted 3 lines · On Monday 25 November 2013 15:01:27 Pavel Machek wrote:

Where can I get dsme sources?

Pavel

dsme daemon: https://gitorious.org/community-ssu/dsme
dsme thermal plugin:
https://gitorious.org/rx51-bme-replacement/dsme-thermalobje
ct-surface

Quoted 4 lines

dsme thermal plugin is responsible for checking temperature of battery and send it to dsme damon. dsme daemon itself checking temperature limits and do something (e.g reboot device)

Quoted 6 lines

Thanks for links!

AFAICT.... dsme daemon does some kind of system management, not directly controlling charging, right? It does monitor system temperature.

Yes.

Quoted 10 lines

Now.. Imagine phone left in car in charger (on sun). Likely temperature will reach high values, kernel is charging, dsme will reboot the system, but kernel will start charging again, dsme will reboot again, ...

And it is not only high temperatures that are problem for li-ion charging; battery should not be fast charged below 5C and should not be charged below 0C. (Again, both are likely to happen if you leave your phone in car).

Quoted 4 lines

AFAICT, we should simply disable charging below 5C or above 45C.

Pavel

Similar patch (as this) can be added to check for temperature from specified power supply driver (in timer reset thread). This can be easy and will fix this problem.

Signature · 2 lines
Pali Rohár
[email protected]
#65 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

In reply to Pali Rohár

Hi

Quoted 53 lines · Pali Rohár

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

---
 drivers/power/bq2415x_charger.c       |   77 +++++++++++++++++++++++++--------
 include/linux/power/bq2415x_charger.h |   48 +++-----------------
 2 files changed, 65 insertions(+), 60 deletions(-)
drivers/power/bq2415x_charger.c +12 -6
index 0727f92..d89583d 100644
@@ -1,7 +1,7 @@
 /*
  * bq2415x charger driver
  *
- * Copyright (C) 2011-2012  Pali Rohár <[email protected]>
+ * Copyright (C) 2011-2013  Pali Rohár <[email protected]>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -170,6 +170,7 @@ struct bq2415x_device {
        struct bq2415x_platform_data init_data;
        struct power_supply charger;
        struct delayed_work work;
+       struct notifier_block nb;
        enum bq2415x_mode reported_mode;/* mode reported by hook function */
        enum bq2415x_mode mode;         /* current configured mode */
        enum bq2415x_chip chip;
@@ -791,24 +792,53 @@ static int bq2415x_set_mode(struct bq2415x_device *bq, enum bq2415x_mode mode)
 
 }
 
-/* hook function called by other driver which set reported mode */
-static void bq2415x_hook_function(enum bq2415x_mode mode, void *data)
+static int bq2415x_notifier_call(struct notifier_block *nb,
+               unsigned long val, void *v)
 {
-       struct bq2415x_device *bq = data;
+       struct bq2415x_device *bq =
+               container_of(nb, struct bq2415x_device, nb);
+       struct power_supply *psy = v;
+       enum bq2415x_mode mode;
+       union power_supply_propval prop;
+       int ret;
+       int mA;
 
-       if (!bq)
-               return;
+       if (val != PSY_EVENT_PROP_CHANGED)

drivers/power/bq2415x_charger.c static int bq2415x_set_mode(struct bq2415x_device *bq, enum bq2415x_mode mode) · line 816

Quoted 10 lines
drivers/power/bq2415x_charger.c +10 -0
+               return NOTIFY_OK;
+
+       if (strcmp(psy->name, bq->init_data.notify_device) != 0)
+               return NOTIFY_OK;
+
+       dev_dbg(bq->dev, "notifier call was called\n");
+
+       ret = psy->get_property(psy, POWER_SUPPLY_PROP_CURRENT_MAX, &prop);
+       if (ret != 0)
+               return NOTIFY_OK;

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

Michael

Quoted 168 lines
drivers/power/bq2415x_charger.c +38 -11
+
+       mA = prop.intval;
+
+       if (mA == 0)
+               mode = BQ2415X_MODE_OFF;
+       else if (mA < 500)
+               mode = BQ2415X_MODE_NONE;
+       else if (mA < 1800)
+               mode = BQ2415X_MODE_HOST_CHARGER;
+       else
+               mode = BQ2415X_MODE_DEDICATED_CHARGER;
+
+       if (bq->reported_mode == mode)
+               return NOTIFY_OK;
 
-       dev_dbg(bq->dev, "hook function was called\n");
        bq->reported_mode = mode;
 
        /* if automode is not enabled do not tell about reported_mode */
        if (bq->automode < 1)
-               return;
+               return NOTIFY_OK;
 
        sysfs_notify(&bq->charger.dev->kobj, NULL, "reported_mode");
        bq2415x_set_mode(bq, bq->reported_mode);
 
+       return NOTIFY_OK;
 }
 
 /**** timer functions ****/
@@ -1508,6 +1538,7 @@ static int bq2415x_probe(struct i2c_client *client,
        int num;
        char *name;
        struct bq2415x_device *bq;
+       struct power_supply *psy;
 
        if (!client->dev.platform_data) {
                dev_err(&client->dev, "platform data not set\n");
@@ -1569,16 +1600,27 @@ static int bq2415x_probe(struct i2c_client *client,
                goto error_4;
        }
 
-       if (bq->init_data.set_mode_hook) {
-               if (bq->init_data.set_mode_hook(
-                               bq2415x_hook_function, bq)) {
-                       bq->automode = 1;
+       if (bq->init_data.notify_device) {
+               bq->nb.notifier_call = bq2415x_notifier_call;
+               ret = power_supply_reg_notifier(&bq->nb);
+               if (ret) {
+                       dev_err(bq->dev, "failed to reg notifier: %d\n", ret);
+                       goto error_5;
+               }
+               psy = power_supply_get_by_name(bq->init_data.notify_device);
+               if (psy) {
+                       /* Query for initial reported_mode and set it */
+                       bq2415x_notifier_call(&bq->nb,
+                                       PSY_EVENT_PROP_CHANGED, psy);
                        bq2415x_set_mode(bq, bq->reported_mode);
-                       dev_info(bq->dev, "automode enabled\n");
                } else {
-                       bq->automode = -1;
-                       dev_info(bq->dev, "automode failed\n");
+                       dev_info(bq->dev, "notifier power supply device (%s) "
+                               "for automode is not registred yet... "
+                               "automode will not work without that device\n",
+                               bq->init_data.notify_device);
                }
+               bq->automode = 1;
+               dev_info(bq->dev, "automode enabled\n");
        } else {
                bq->automode = -1;
                dev_info(bq->dev, "automode not supported\n");
@@ -1590,6 +1632,7 @@ static int bq2415x_probe(struct i2c_client *client,
        dev_info(bq->dev, "driver registered\n");
        return 0;
 
+error_5:
 error_4:
        bq2415x_sysfs_exit(bq);
 error_3:
@@ -1610,8 +1653,8 @@ static int bq2415x_remove(struct i2c_client *client)
 {
        struct bq2415x_device *bq = i2c_get_clientdata(client);
 
-       if (bq->init_data.set_mode_hook)
-               bq->init_data.set_mode_hook(NULL, NULL);
+       if (bq->init_data.notify_device)
+               power_supply_unreg_notifier(&bq->nb);
 
        bq2415x_sysfs_exit(bq);
        bq2415x_power_supply_exit(bq);
include/linux/power/bq2415x_charger.h +5 -44
index 8dcc0f4..50762af 100644
@@ -1,7 +1,7 @@
 /*
  * bq2415x charger driver
  *
- * Copyright (C) 2011-2012  Pali Rohár <[email protected]>
+ * Copyright (C) 2011-2013  Pali Rohár <[email protected]>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -31,46 +31,9 @@
  * termination current. It it is less or equal to zero, configuring charge
  * and termination current will not be possible.
  *
- * Function set_mode_hook is needed for automode (setting correct current
- * limit when charger is connected/disconnected or setting boost mode).
- * When is NULL, automode function is disabled. When is not NULL, it must
- * have this prototype:
- *
- *    int (*set_mode_hook)(
- *      void (*hook)(enum bq2415x_mode mode, void *data),
- *      void *data)
- *
- * hook is hook function (see below) and data is pointer to driver private
- * data
- *
- * bq2415x driver will call it as:
- *
- *    platform_data->set_mode_hook(bq2415x_hook_function, bq2415x_device);
- *
- * Board/platform function set_mode_hook return non zero value when hook
- * function was successful registered. Platform code should call that hook
- * function (which get from pointer, with data) every time when charger
- * was connected/disconnected or require to enable boost mode. bq2415x
- * driver then will set correct current limit, enable/disable charger or
- * boost mode.
- *
- * Hook function has this prototype:
- *
- *    void hook(enum bq2415x_mode mode, void *data);
- *
- * mode is bq2415x mode (charger or boost)
- * data is pointer to driver private data (which get from
- * set_charger_type_hook)
- *
- * When bq driver is being unloaded, it call function:
- *
- *    platform_data->set_mode_hook(NULL, NULL);
- *
- * (hook function and driver private data are NULL)
- *
- * After that board/platform code must not call driver hook function! It
- * is possible that pointer to hook function will not be valid and calling
- * will cause undefined result.
+ * For automode support is needed to provide name of power supply device
+ * in value notify_device. Device driver must immediately report property
+ * POWER_SUPPLY_PROP_CURRENT_MAX when current changed.
  */
 
 /* Supported modes with maximal current limit */
@@ -89,8 +52,7 @@ struct bq2415x_platform_data {
        int charge_current;             /* mA */
        int termination_current;        /* mA */
        int resistor_sense;             /* m ohm */
-       int (*set_mode_hook)(void (*hook)(enum bq2415x_mode mode, void *data),
-                            void *data);
+       const char *notify_device;      /* name */
 };
 
 #endif
--

1.7.9.5

Quoted 4 lines

--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#66 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 3 lines · On Sunday 24 November 2013 18:18:03 Michael Trimarchi wrote:

Hi

On Tue, Nov 19, 2013 at 11:18 AM, Pali Rohár

Quoted 16 lines · <[email protected]> wrote:
Quoted 16 lines

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

---

 drivers/power/bq2415x_charger.c       |   77
 +++++++++++++++++++++++++--------
 include/linux/power/bq2415x_charger.h |   48

+++----------------- 2 files changed, 65 insertions(+), 60 deletions(-)

...

Quoted 30 lines
Quoted 21 lines
drivers/power/bq2415x_charger.c +11 -3
-       struct bq2415x_device *bq = data;
+       struct bq2415x_device *bq =
+               container_of(nb, struct bq2415x_device, nb);
+       struct power_supply *psy = v;
+       enum bq2415x_mode mode;
+       union power_supply_propval prop;
+       int ret;
+       int mA;
 
-       if (!bq)
-               return;
+       if (val != PSY_EVENT_PROP_CHANGED)
+               return NOTIFY_OK;
+
+       if (strcmp(psy->name, bq->init_data.notify_device)

!= 0) + return NOTIFY_OK;

drivers/power/bq2415x_charger.c +4 -0
+
+       dev_dbg(bq->dev, "notifier call was called\n");
+
+       ret = psy->get_property(psy,
Quoted 3 lines

POWER_SUPPLY_PROP_CURRENT_MAX, &prop); + if (ret !=
0)
+ return NOTIFY_OK;

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

Quoted 2 lines

Michael

I do not see reason why I cannot use it. When isp1704 driver send PSY_EVENT_PROP_CHANGED then property POWER_SUPPLY_PROP_CURRENT_MAX is already updated and can be read by get_property function.

Signature · 2 lines
Pali Rohár
[email protected]
#67 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Hi

Quoted 52 lines · Pali Rohár
Quoted 3 lines · On Sunday 24 November 2013 18:18:03 Michael Trimarchi wrote:

Hi

On Tue, Nov 19, 2013 at 11:18 AM, Pali Rohár

Quoted 16 lines · <[email protected]> wrote:
Quoted 16 lines

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

---

 drivers/power/bq2415x_charger.c       |   77
 +++++++++++++++++++++++++--------
 include/linux/power/bq2415x_charger.h |   48

+++----------------- 2 files changed, 65 insertions(+), 60 deletions(-)

...

Quoted 30 lines
Quoted 21 lines
drivers/power/bq2415x_charger.c +11 -3
-       struct bq2415x_device *bq = data;
+       struct bq2415x_device *bq =
+               container_of(nb, struct bq2415x_device, nb);
+       struct power_supply *psy = v;
+       enum bq2415x_mode mode;
+       union power_supply_propval prop;
+       int ret;
+       int mA;
 
-       if (!bq)
-               return;
+       if (val != PSY_EVENT_PROP_CHANGED)
+               return NOTIFY_OK;
+
+       if (strcmp(psy->name, bq->init_data.notify_device)

!= 0) + return NOTIFY_OK;

drivers/power/bq2415x_charger.c +4 -0
+
+       dev_dbg(bq->dev, "notifier call was called\n");
+
+       ret = psy->get_property(psy,
Quoted 3 lines

POWER_SUPPLY_PROP_CURRENT_MAX, &prop); + if (ret !=
0)
+ return NOTIFY_OK;

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

Quoted 8 lines
Quoted 2 lines

Michael

I do not see reason why I cannot use it. When isp1704 driver send PSY_EVENT_PROP_CHANGED then property POWER_SUPPLY_PROP_CURRENT_MAX is already updated and can be read by get_property function.

https://lists.ubuntu.com/archives/kernel-team/2013-January/025206.html

I found and equivalent scenario that I was trying to said

Michael

Quoted 3 lines

--
Pali Rohár
[email protected]

--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#68 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

On Mon, Nov 25, 2013 at 08:16:34PM +0100, Michael Trimarchi wrote:
...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

...

Quoted 3 lines

https://lists.ubuntu.com/archives/kernel-team/2013-January/025206.html

I found and equivalent scenario that I was trying to said

That's a good question, actually. Even though in Pali's case the notifier is atomic (so that we are pretty confident that the object won't be unregistered), there is still a possiblity of a dead lock, for example. So notifiers should be careful to not hold any locks, because the other driver might call get_property(), which might acquire locks.

Thanks,

Anton

#69 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Hi Anton

Quoted 5 lines · Anton Vorontsov

On Mon, Nov 25, 2013 at 08:16:34PM +0100, Michael Trimarchi wrote:
...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

Quoted 8 lines

...

Quoted 3 lines

https://lists.ubuntu.com/archives/kernel-team/2013-January/025206.html

I found and equivalent scenario that I was trying to said

That's a good question, actually. Even though in Pali's case the notifier is atomic (so that we are pretty confident that the object won't be unregistered), there is still a possiblity of a dead lock, for example. So

So if the get_property is a sleeping function it will be a deadlock. Right?

Michael

Quoted 6 lines

notifiers should be careful to not hold any locks, because the other driver might call get_property(), which might acquire locks.

Thanks,

Anton

#70 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 9 lines · Michael Trimarchi
Quoted 8 lines · Anton Vorontsov

On Mon, Nov 25, 2013 at 08:16:34PM +0100, Michael Trimarchi wrote:
...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

...

Quoted 2 lines

https://lists.ubuntu.com/archives/kernel-team/2013-January/025206.html

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

I found and equivalent scenario that I was trying to said

That's a good question, actually. Even though in Pali's case the notifier is atomic (so that we are pretty confident that the object won't be unregistered), there is still a possiblity of a dead lock, for example. So

So if the get_property is a sleeping function it will be a deadlock. Right?

All kind of bad things might happen, yes. But before that I would expect a bunch of warnings from might_sleep() stuff.

I would recommend to test the patches using preempt/smp kernels + various DEBUG_ kernel options set.

Thanks,

Anton

#71 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Hi

Quoted 16 lines · Anton Vorontsov
Quoted 15 lines · Michael Trimarchi
Quoted 13 lines · Anton Vorontsov

On Mon, Nov 25, 2013 at 08:16:34PM +0100, Michael Trimarchi wrote:
...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

...

Quoted 3 lines

https://lists.ubuntu.com/archives/kernel-team/2013-January/025206.html

I found and equivalent scenario that I was trying to said

That's a good question, actually. Even though in Pali's case the notifier is atomic (so that we are pretty confident that the object won't be unregistered), there is still a possiblity of a dead lock, for example. So

Quoted 8 lines
Quoted 1 line

So if the get_property is a sleeping function it will be a deadlock. Right?

All kind of bad things might happen, yes. But before that I would expect a bunch of warnings from might_sleep() stuff.

I would recommend to test the patches using preempt/smp kernels + various DEBUG_ kernel options set.

Is it more simple to make it not atomic and use a mutex in get_property?

Michael

Quoted 3 lines

Thanks,

Anton

#72 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 19 lines · Michael Trimarchi
Quoted 18 lines · Anton Vorontsov
Quoted 16 lines · Michael Trimarchi
Quoted 5 lines · Anton Vorontsov

On Mon, Nov 25, 2013 at 08:16:34PM +0100, Michael Trimarchi wrote:
...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

Quoted 8 lines

...

Quoted 3 lines

https://lists.ubuntu.com/archives/kernel-team/2013-January/025206.html

I found and equivalent scenario that I was trying to said

That's a good question, actually. Even though in Pali's case the notifier is atomic (so that we are pretty confident that the object won't be unregistered), there is still a possiblity of a dead lock, for example. So

So if the get_property is a sleeping function it will be a deadlock. Right?

Quoted 8 lines
Quoted 6 lines

All kind of bad things might happen, yes. But before that I would expect a bunch of warnings from might_sleep() stuff.

I would recommend to test the patches using preempt/smp kernels + various DEBUG_ kernel options set.

Is it more simple to make it not atomic and use a mutex in get_property?

I just built a kernel with CONFIG_DEBUG_ATOMIC_SLEEP to test another driver and got the following output for bq2415x:

[    7.667449] Workqueue: events power_supply_changed_work
[    7.673034] [<c0015c28>] (unwind_backtrace+0x0/0xe0) from [<c0011e1c>] (show_stack+0x10/0x14)
[    7.682098] [<c0011e1c>] (show_stack+0x10/0x14) from [<c052cdd0>] (dump_stack+0x78/0xac)
[    7.690704] [<c052cdd0>] (dump_stack+0x78/0xac) from [<c052a044>] (__schedule_bug+0x48/0x60)
[    7.699645] [<c052a044>] (__schedule_bug+0x48/0x60) from [<c053071c>] (__schedule+0x74/0x638)
[    7.708618] [<c053071c>] (__schedule+0x74/0x638) from [<c05301fc>] (schedule_timeout+0x1dc/0x24c)
[    7.718017] [<c05301fc>] (schedule_timeout+0x1dc/0x24c) from [<c05316ec>] (wait_for_common+0x138/0x17c)
[    7.727966] [<c05316ec>] (wait_for_common+0x138/0x17c) from [<c0362a70>] (omap_i2c_xfer+0x340/0x4a0)
[    7.737640] [<c0362a70>] (omap_i2c_xfer+0x340/0x4a0) from [<c035d928>] (__i2c_transfer+0x40/0x74)
[    7.747039] [<c035d928>] (__i2c_transfer+0x40/0x74) from [<c035e22c>] (i2c_transfer+0x6c/0x90)
[    7.756195] [<c035e22c>] (i2c_transfer+0x6c/0x90) from [<c037ad24>] (bq2415x_i2c_write+0x48/0x78)
[    7.765563] [<c037ad24>] (bq2415x_i2c_write+0x48/0x78) from [<c037ae60>] (bq2415x_set_weak_battery_voltage+0x4c/0x50)
[    7.776824] [<c037ae60>] (bq2415x_set_weak_battery_voltage+0x4c/0x50) from [<c037bce8>] (bq2415x_set_mode+0xdc/0x14c)
[    7.788085] [<c037bce8>] (bq2415x_set_mode+0xdc/0x14c) from [<c037bfb8>] (bq2415x_notifier_call+0xa8/0xb4)
[    7.798309] [<c037bfb8>] (bq2415x_notifier_call+0xa8/0xb4) from [<c005f228>] (notifier_call_chain+0x38/0x68)
[    7.808715] [<c005f228>] (notifier_call_chain+0x38/0x68) from [<c005f284>] (__atomic_notifier_call_chain+0x2c/0x3c)
[    7.819732] [<c005f284>] (__atomic_notifier_call_chain+0x2c/0x3c) from [<c005f2a8>] (atomic_notifier_call_chain+0x14/0x18)
[    7.831420] [<c005f2a8>] (atomic_notifier_call_chain+0x14/0x18) from [<c0378078>] (power_supply_changed_work+0x6c/0xb8)
[    7.842864] [<c0378078>] (power_supply_changed_work+0x6c/0xb8) from [<c00556c0>] (process_one_work+0x248/0x440)
[    7.853546] [<c00556c0>] (process_one_work+0x248/0x440) from [<c0055d6c>] (worker_thread+0x208/0x350)
[    7.863372] [<c0055d6c>] (worker_thread+0x208/0x350) from [<c005b0ac>] (kthread+0xc8/0xdc)
[    7.872131] [<c005b0ac>] (kthread+0xc8/0xdc) from [<c000e138>] (ret_from_fork+0x14/0x3c)

-- Sebastian

#73 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Hi

Quoted 55 lines · Sebastian Reichel
Quoted 19 lines · Michael Trimarchi
Quoted 18 lines · Anton Vorontsov
Quoted 16 lines · Michael Trimarchi
Quoted 5 lines · Anton Vorontsov

On Mon, Nov 25, 2013 at 08:16:34PM +0100, Michael Trimarchi wrote:
...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

So you can read this value without any type of synchronization with the power_supply_core
and sysfs implementation?

Quoted 8 lines

...

Quoted 3 lines

https://lists.ubuntu.com/archives/kernel-team/2013-January/025206.html

I found and equivalent scenario that I was trying to said

That's a good question, actually. Even though in Pali's case the notifier is atomic (so that we are pretty confident that the object won't be unregistered), there is still a possiblity of a dead lock, for example. So

So if the get_property is a sleeping function it will be a deadlock. Right?

Quoted 8 lines
Quoted 6 lines

All kind of bad things might happen, yes. But before that I would expect a bunch of warnings from might_sleep() stuff.

I would recommend to test the patches using preempt/smp kernels + various DEBUG_ kernel options set.

Is it more simple to make it not atomic and use a mutex in get_property?

I just built a kernel with CONFIG_DEBUG_ATOMIC_SLEEP to test another driver and got the following output for bq2415x:

[    7.667449] Workqueue: events power_supply_changed_work
[    7.673034] [<c0015c28>] (unwind_backtrace+0x0/0xe0) from [<c0011e1c>] (show_stack+0x10/0x14)
[    7.682098] [<c0011e1c>] (show_stack+0x10/0x14) from [<c052cdd0>] (dump_stack+0x78/0xac)
[    7.690704] [<c052cdd0>] (dump_stack+0x78/0xac) from [<c052a044>] (__schedule_bug+0x48/0x60)
[    7.699645] [<c052a044>] (__schedule_bug+0x48/0x60) from [<c053071c>] (__schedule+0x74/0x638)
[    7.708618] [<c053071c>] (__schedule+0x74/0x638) from [<c05301fc>] (schedule_timeout+0x1dc/0x24c)
[    7.718017] [<c05301fc>] (schedule_timeout+0x1dc/0x24c) from [<c05316ec>] (wait_for_common+0x138/0x17c)
[    7.727966] [<c05316ec>] (wait_for_common+0x138/0x17c) from [<c0362a70>] (omap_i2c_xfer+0x340/0x4a0)
[    7.737640] [<c0362a70>] (omap_i2c_xfer+0x340/0x4a0) from [<c035d928>] (__i2c_transfer+0x40/0x74)
[    7.747039] [<c035d928>] (__i2c_transfer+0x40/0x74) from [<c035e22c>] (i2c_transfer+0x6c/0x90)
[    7.756195] [<c035e22c>] (i2c_transfer+0x6c/0x90) from [<c037ad24>] (bq2415x_i2c_write+0x48/0x78)
[    7.765563] [<c037ad24>] (bq2415x_i2c_write+0x48/0x78) from [<c037ae60>] (bq2415x_set_weak_battery_voltage+0x4c/0x50)
[    7.776824] [<c037ae60>] (bq2415x_set_weak_battery_voltage+0x4c/0x50) from [<c037bce8>] (bq2415x_set_mode+0xdc/0x14c)
[    7.788085] [<c037bce8>] (bq2415x_set_mode+0xdc/0x14c) from [<c037bfb8>] (bq2415x_notifier_call+0xa8/0xb4)
[    7.798309] [<c037bfb8>] (bq2415x_notifier_call+0xa8/0xb4) from [<c005f228>] (notifier_call_chain+0x38/0x68)
[    7.808715] [<c005f228>] (notifier_call_chain+0x38/0x68) from [<c005f284>] (__atomic_notifier_call_chain+0x2c/0x3c)
[    7.819732] [<c005f284>] (__atomic_notifier_call_chain+0x2c/0x3c) from [<c005f2a8>] (atomic_notifier_call_chain+0x14/0x18)
[    7.831420] [<c005f2a8>] (atomic_notifier_call_chain+0x14/0x18) from [<c0378078>] (power_supply_changed_work+0x6c/0xb8)
[    7.842864] [<c0378078>] (power_supply_changed_work+0x6c/0xb8) from [<c00556c0>] (process_one_work+0x248/0x440)
[    7.853546] [<c00556c0>] (process_one_work+0x248/0x440) from [<c0055d6c>] (worker_thread+0x208/0x350)
[    7.863372] [<c0055d6c>] (worker_thread+0x208/0x350) from [<c005b0ac>] (kthread+0xc8/0xdc)
[    7.872131] [<c005b0ac>] (kthread+0xc8/0xdc) from [<c000e138>] (ret_from_fork+0x14/0x3c)
Quoted 1 line

-- Sebastian

I have already give my opinion about this problem

Michael

#74 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

In reply to Sebastian Reichel

And another one using that evil mail-followup-to header:

Mail-Followup-To: Pali Rohár <[email protected]>,
Anton Vorontsov <[email protected]>,
Michael Trimarchi <[email protected]>,
David Woodhouse <[email protected]>,
Tony Lindgren <[email protected]>,
Russell King <[email protected]>, [email protected], Linux OMAP Mailing List <[email protected]>, [email protected], [email protected], [email protected]

which results in all the above being moved into the To: header by unsuspecting recipients who reply to your message.

Signature · 3 lines
FTTC broadband for 0.8mile line: 5.8Mbps down 500kbps up.  Estimation
in database were 13.1 to 19Mbit for a good line, about 7.5+ for a bad.
Estimate before purchase was "up to 13.2Mbit".
#75 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

Quoted 11 lines · Russell King - ARM Linux

And another one using that evil mail-followup-to header:

Mail-Followup-To: Pali Rohár <[email protected]>,
Anton Vorontsov <[email protected]>,
Michael Trimarchi <[email protected]>,
David Woodhouse <[email protected]>,
Tony Lindgren <[email protected]>,
Russell King <[email protected]>, [email protected], Linux OMAP Mailing List <[email protected]>, [email protected], [email protected], [email protected]

Quoted 2 lines

which results in all the above being moved into the To: header by unsuspecting recipients who reply to your message.

Thanks for the hint. Should be fixed now.

-- Sebastian

#76 lore raw

Re: [PATCH v2 2/3] bq2415x_charger: Use power_supply notifier for automode

In reply to Pali Rohár
Quoted 7 lines · On Tuesday 19 November 2013 11:18:04 Pali Rohár wrote:

This patch removing set_mode_hook function from board data and replacing it with new string variable of notifier power supply device. After this change it is possible to add DT support because driver does not need specific board function anymore. Only static data and name of power supply device is required.

Quoted 8 lines
---
 drivers/power/bq2415x_charger.c       |   77

+++++++++++++++++++++++++--------
include/linux/power/bq2415x_charger.h | 48
+++----------------- 2 files changed, 65 insertions(+), 60 deletions(-)

Hi Anton, is there any problem with this patch?

Signature · 2 lines
Pali Rohár
[email protected]
#77 lore raw

PATCH v2 3/3 RX-51: Add platform data for bq24150a charger

In reply to Pali Rohár

This patch will register bq24150a charger in RX-51 board data.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

---
 arch/arm/mach-omap2/board-rx51-peripherals.c |   17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)
arch/arm/mach-omap2/board-rx51-peripherals.c +16 -1
index 70fcaff..95e6638 100644
@@ -25,6 +25,7 @@
 #include <linux/gpio_keys.h>
 #include <linux/mmc/host.h>
 #include <linux/power/isp1704_charger.h>
+#include <linux/power/bq2415x_charger.h>
 #include <linux/platform_data/spi-omap2-mcspi.h>
 #include <linux/platform_data/mtd-onenand-omap2.h>
 #include <linux/hsi/hsi.h>
@@ -1275,6 +1276,16 @@ static struct aic3x_pdata rx51_aic3x_data2 = {
 	.gpio_reset = 60,
 };
 
+static struct bq2415x_platform_data rx51_bq24150a_platform_data = {
+	.current_limit = 100,			/* mA */
+	.weak_battery_voltage = 3400,		/* mV */
+	.battery_regulation_voltage = 4200,	/* mV */
+	.charge_current = 650,			/* mA */
+	.termination_current = 100,		/* mA */
+	.resistor_sense = 68,			/* m ohm */
+	.notify_device = "isp1704",
+};
+
 static struct i2c_board_info __initdata rx51_peripherals_i2c_board_info_2[] = {
 	{
 		I2C_BOARD_INFO("tlv320aic3x", 0x18),
@@ -1302,7 +1313,11 @@ static struct i2c_board_info __initdata rx51_peripherals_i2c_board_info_2[] = {
 	{
 		I2C_BOARD_INFO("tpa6130a2", 0x60),
 		.platform_data = &rx51_tpa6130a2_data,
-	}
+	},
+	{
+		I2C_BOARD_INFO("bq24150a", 0x6b),
+		.platform_data = &rx51_bq24150a_platform_data,
+	},
 };
 
 static struct i2c_board_info __initdata rx51_peripherals_i2c_board_info_3[] = {
Signature · 1 line
1.7.9.5
Reviewed-by on v2 3/3
#78 lore raw

Re: [PATCH v2 3/3] RX-51: Add platform data for bq24150a charger

Quoted 6 lines · Pali Rohár

This patch will register bq24150a charger in RX-51 board data.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

Pavel
BTW, current code says:

That seems to limit charging voltage to 3.9V, with comment saying 4.11V... which means battery will not be charged fully. does it make sense to raise limit to 4.11V here?

Aha, and actually will not 4.2V requested by
rx51_bq24150a_platform_data trigger -EINVAL?

/* set battery regulation voltage in mV */ static int bq2415x_set_battery_regulation_voltage(struct bq2415x_device *bq,
int mV)

{
        int val = (mV/10 - 350) / 2;

        if (val < 0)
                val = 0;
        else if (val > 94) /* FIXME: Max is 94 or 122 ? Set max value
                ? */
                return -EINVAL;

        return bq2415x_i2c_write_mask(bq, BQ2415X_REG_VOLTAGE, val,
                        BQ2415X_MASK_VO, BQ2415X_SHIFT_VO);
}
Quoted 19 lines
arch/arm/mach-omap2/board-rx51-peripherals.c +10 -0
@@ -1275,6 +1276,16 @@ static struct aic3x_pdata rx51_aic3x_data2 = {
 	.gpio_reset = 60,
 };
 
+static struct bq2415x_platform_data rx51_bq24150a_platform_data = {
+	.current_limit = 100,			/* mA */
+	.weak_battery_voltage = 3400,		/* mV */
+	.battery_regulation_voltage = 4200,	/* mV */
+	.charge_current = 650,			/* mA */
+	.termination_current = 100,		/* mA */
+	.resistor_sense = 68,			/* m ohm */
+	.notify_device = "isp1704",
+};
+
 static struct i2c_board_info __initdata rx51_peripherals_i2c_board_info_2[] = {
 	{
 		I2C_BOARD_INFO("tlv320aic3x", 0x18),
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#79 lore raw

Re: [PATCH v2 3/3] RX-51: Add platform data for bq24150a charger

Quoted 36 lines · On Tuesday 19 November 2013 14:32:13 Pavel Machek wrote:
Quoted 8 lines · Pali Rohár

This patch will register bq24150a charger in RX-51 board data.

So finally charging battery on Nokia N900 (RX-51) working automatically without any proprietary Nokia bits in userspace.

Pavel
BTW, current code says:

That seems to limit charging voltage to 3.9V, with comment saying 4.11V... which means battery will not be charged fully. does it make sense to raise limit to 4.11V here?

Aha, and actually will not 4.2V requested by
rx51_bq24150a_platform_data trigger -EINVAL?

/* set battery regulation voltage in mV */ static int bq2415x_set_battery_regulation_voltage(struct bq2415x_device *bq,
int mV)

{
        int val = (mV/10 - 350) / 2;

        if (val < 0)
                val = 0;
        else if (val > 94) /* FIXME: Max is 94 or 122 ? Set
max value ? */
                return -EINVAL;

        return bq2415x_i2c_write_mask(bq, BQ2415X_REG_VOLTAGE,
Quoted 2 lines

val, BQ2415X_MASK_VO, BQ2415X_SHIFT_VO); }

Above bq2415x_set_battery_regulation_voltage function is updated in 3.13, see commit 3ed5cd79b82e9a055e0df3275eeb471ea0f1e2d7.

Signature · 2 lines
Pali Rohár
[email protected]
#80 lore raw
In reply to Pali Rohár

Hi Pali,

Quoted 2 lines · Pali Rohár

This patch series finally bringing support for charging battery on Nokia N900 (RX-51) without any proprietary Nokia bits in userspace.

YAY! :)

Quoted 6 lines

I removed hook function from bq2415x_charger driver and replaced it by Jenny TC power_supply notifier code. This simplify usage of bq2415x driver and also allow adding DT support in future because driver does not need board specific function anymore (only static data).

I tested this series with linux 3.12-rc5 on Nokia N900.

cool :)

Do you want me to take this up and add DT support?

-- Sebastian

maintainer
#81 lore raw
Quoted 13 lines · Sebastian Reichel

Hi Pali,

Quoted 2 lines · Pali Rohár

This patch series finally bringing support for charging battery on Nokia N900 (RX-51) without any proprietary Nokia bits in userspace.

YAY! :)

Quoted 5 lines

I removed hook function from bq2415x_charger driver and replaced it by Jenny TC power_supply notifier code. This simplify usage of bq2415x driver and also allow adding DT support in future because driver does not need board specific function anymore (only static data).

Quoted 5 lines
Quoted 1 line

I tested this series with linux 3.12-rc5 on Nokia N900.

cool :)

Do you want me to take this up and add DT support?

Great news, yeah let's add this with DT support instead as we're planning to switch over to DT only booting anyways. The values can be probably set directly in the charger driver based on the compatible flag.

Note that we can also still add the I2C devices as needed from the pdata-quirks.c, but in this case there should not be any need for it :)

Regards,

Tony

#82 lore raw
In reply to Sebastian Reichel
Quoted 15 lines · On Tuesday 19 November 2013 15:44:09 Sebastian Reichel wrote:

Hi Pali,

Quoted 3 lines · Pali Rohár

This patch series finally bringing support for charging battery on Nokia N900 (RX-51) without any proprietary Nokia bits in userspace.

YAY! :)

Quoted 6 lines

I removed hook function from bq2415x_charger driver and replaced it by Jenny TC power_supply notifier code. This simplify usage of bq2415x driver and also allow adding DT support in future because driver does not need board specific function anymore (only static data).

Quoted 7 lines
Quoted 1 line

I tested this series with linux 3.12-rc5 on Nokia N900.

cool :)

Do you want me to take this up and add DT support?

-- Sebastian

Hello, if you have time, can you add DT support code?

Signature · 2 lines
Pali Rohár
[email protected]