From: Lee Jones <hidden> Date: 2012-07-05 13:06:37
In this patch-set we provide the last batch of clean-ups related
to the Device Tree enablement of Snowball and other DB8500 based
chipsets. Most of these clean-ups are nice and easy
simplifications of code, or recommendations from particular
Mainline Maintainers.
.../devicetree/bindings/gpio/gpio-nmk.txt | 2 +-
arch/arm/boot/dts/db8500.dtsi | 18 +-
arch/arm/mach-ux500/board-mop500-regulators.c | 176 --------------------
arch/arm/mach-ux500/board-mop500-regulators.h | 2 -
arch/arm/mach-ux500/board-mop500.c | 20 +--
arch/arm/mach-ux500/cpu-db8500.c | 13 +-
drivers/input/misc/ab8500-ponkey.c | 2 +
drivers/pinctrl/pinctrl-nomadik.c | 4 +-
include/linux/mfd/abx500/ab8500.h | 3 -
9 files changed, 18 insertions(+), 222 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-05 13:06:41
If the Nomadik Pin Control driver is continuously bound and unbound
from the same device a number of times the kzalloc() will leak
memory. Replacing it with devm_kzalloc() will ensure memory is
freed automatically in the event of failure and unbinding.
Signed-off-by: Lee Jones <redacted>
---
drivers/pinctrl/pinctrl-nomadik.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
From: Lee Jones <hidden> Date: 2012-07-05 13:06:45
Any non-standard property should contain the vendor's identifier which
should be perpended onto the property name followed by a comma. This
aids in name-space collision prevention.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/boot/dts/db8500.dtsi | 18 +++++++++---------
1 file changed, 9 insertions(+), 9 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-05 13:06:49
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 2 ++
1 file changed, 2 insertions(+)
From: Lee Jones <hidden> Date: 2012-07-05 13:06:54
When pin control mapping tables are written the registered device
name is supplied for use in name-based searches within the pinctrl
driver. In the case of the DB8500 the string "pinctrl-db8500" is
used. However, when we register the driver with Device Tree, its
naming convention uses something that looks more like "pinctrl.2".
To work around the device naming inconsistencies between devices
registered via platform code and the ones registered by Device
Tree, we use AUXDATA to over-ride the Device Tree naming scheme.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/mach-ux500/board-mop500.c | 1 +
1 file changed, 1 insertion(+)
From: Lee Jones <hidden> Date: 2012-07-05 13:07:47
Here we remove some of the extra overhead we introduced to make
DT:ing the Snowball platform easier. Unfortunately we can't
remove everything. We are forced to leave cpufreq and DMA behind
until they have been successfully Device Tree enabled.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/mach-ux500/cpu-db8500.c | 10 ++--------
1 file changed, 2 insertions(+), 8 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-05 13:07:49
PMU registration is successfully completed by Device Tree now, so
there is no longer a requirement to register it from platform code.
This patch removes platform registration during a DT boot.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/mach-ux500/cpu-db8500.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-05 13:08:22
All Device Tree enablement for Snowball's Platform devices; LEDs, Keys,
Ethernet and all associated AB8500 Multi-Functional Devices are now
complete, so here we remove any trace of adding those devices when Device
Tree is enabled.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/mach-ux500/board-mop500.c | 17 -----------------
1 file changed, 17 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-05 13:08:56
There is no need to initialise the AB8500's regulator registers, as
most of this work is already carried out by framework features, so
we can safely remove all traces from platform code.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/mach-ux500/board-mop500-regulators.c | 176 -------------------------
arch/arm/mach-ux500/board-mop500-regulators.h | 2 -
arch/arm/mach-ux500/board-mop500.c | 2 -
include/linux/mfd/abx500/ab8500.h | 3 -
4 files changed, 183 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-05 13:10:00
Any non-standard property should contain the vendor's identifier which
should be perpended onto the property name followed by a comma. This
aids in name-space collision prevention. This patch ensures the
sleepmode property adheres to the rules.
Signed-off-by: Lee Jones <redacted>
---
.../devicetree/bindings/gpio/gpio-nmk.txt | 2 +-
drivers/pinctrl/pinctrl-nomadik.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
There is no need to initialise the AB8500's regulator registers, as
most of this work is already carried out by framework features, so
we can safely remove all traces from platform code.
Signed-off-by: Lee Jones <redacted>
It makes sense, but this patch is basically a revert of dfa3a824d
"mach-ux500: provide ab8500 init vector" from Bengt Jonsson, so
it would be nice to get an Ack from him.
Arnd
From: Lee Jones <hidden> Date: 2012-07-05 13:55:40
On 05/07/12 14:51, Arnd Bergmann wrote:
On Thursday 05 July 2012, Lee Jones wrote:
quoted
There is no need to initialise the AB8500's regulator registers, as
most of this work is already carried out by framework features, so
we can safely remove all traces from platform code.
Signed-off-by: Lee Jones <redacted>
It makes sense, but this patch is basically a revert of dfa3a824d
"mach-ux500: provide ab8500 init vector" from Bengt Jonsson, so
it would be nice to get an Ack from him.
Good idea.
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
There is no need to initialise the AB8500's regulator registers, as
most of this work is already carried out by framework features, so
we can safely remove all traces from platform code.
Signed-off-by: Lee Jones<redacted>
It makes sense, but this patch is basically a revert of dfa3a824d
"mach-ux500: provide ab8500 init vector" from Bengt Jonsson, so
it would be nice to get an Ack from him.
Bengt is on leave but I CC some other guys...
Yours,
Linus Walleij
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
There is no need to initialise the AB8500's regulator registers, as
most of this work is already carried out by framework features, so
we can safely remove all traces from platform code.
Signed-off-by: Lee Jones <redacted>
NACK, sorry.
These registers are used to set up sleep states for the regulators,
e.g that some regulators will turn off when the system go to sleep.
If you delete this code it's no longer possible to make the system
powerefficient, power consumption increases so it causes a
regression.
The proper thing to do is to take the AB8500 datasheet, go in and
read what these registers actually do, and provide the same
features through the framework, if just poking the registers
is deemed non-elegant (I see the point in that).
This is the AB8500 datasheet:
http://www.stericsson.com/developers/CD00291561_UM1031_AB8500_user_manual-rev5_CTDS_public.pdf
So please refactor this, don't just delete, that's destroying a lot
of useful stuff.
Yours,
Linus Walleij
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
If the Nomadik Pin Control driver is continuously bound and unbound
from the same device a number of times the kzalloc() will leak
memory. Replacing it with devm_kzalloc() will ensure memory is
freed automatically in the event of failure and unbinding.
Signed-off-by: Lee Jones <redacted>
A patch like this but more all-encompassing was already sent out
by me some days ago:
http://marc.info/?l=linux-kernel&m=134139303405730&w=2
Please test that one and see if it solves your problem.
Yours,
Linus Walleij
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
Any non-standard property should contain the vendor's identifier which
should be perpended onto the property name followed by a comma. This
aids in name-space collision prevention. This patch ensures the
sleepmode property adheres to the rules.
Signed-off-by: Lee Jones <redacted>
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
Any non-standard property should contain the vendor's identifier which
should be perpended onto the property name followed by a comma. This
aids in name-space collision prevention.
Signed-off-by: Lee Jones <redacted>
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
All Device Tree enablement for Snowball's Platform devices; LEDs, Keys,
Ethernet and all associated AB8500 Multi-Functional Devices are now
complete, so here we remove any trace of adding those devices when Device
Tree is enabled.
Signed-off-by: Lee Jones <redacted>
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
PMU registration is successfully completed by Device Tree now, so
there is no longer a requirement to register it from platform code.
This patch removes platform registration during a DT boot.
Signed-off-by: Lee Jones <redacted>
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
Acked-by: Linus Walleij <redacted>
Lee I think the in-kernel charging drivers in drivers/power/ab8500_* are
also using the IRQs, have you checked that these also do the right
thing? (I realize they may be hard to test, but atleast make an effort).
Please coordinate with Rajanikanth who is working on DT:ing
these drivers.
Yours,
Linus Walleij
From: Lee Jones <hidden> Date: 2012-07-06 07:45:43
On 06/07/12 07:55, Linus Walleij wrote:
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
quoted
There is no need to initialise the AB8500's regulator registers, as
most of this work is already carried out by framework features, so
we can safely remove all traces from platform code.
Signed-off-by: Lee Jones <redacted>
NACK, sorry.
These registers are used to set up sleep states for the regulators,
e.g that some regulators will turn off when the system go to sleep.
If you delete this code it's no longer possible to make the system
powerefficient, power consumption increases so it causes a
regression.
The proper thing to do is to take the AB8500 datasheet, go in and
read what these registers actually do, and provide the same
features through the framework, if just poking the registers
is deemed non-elegant (I see the point in that).
This is the AB8500 datasheet:
http://www.stericsson.com/developers/CD00291561_UM1031_AB8500_user_manual-rev5_CTDS_public.pdf
So please refactor this, don't just delete, that's destroying a lot
of useful stuff.
Sounds reasonable. I won't be doing the refactoring, as I'm moving onto
something else from Monday. Unless of course I find myself with lots of
spare time *chuckles*.
Mark,
Please take a note of this and also do not apply the regulator driver
counterpart of the patch-set (it was one of the ones you Reviewed-by in
the previous batch).
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
Here we remove some of the extra overhead we introduced to make
DT:ing the Snowball platform easier. Unfortunately we can't
remove everything. We are forced to leave cpufreq and DMA behind
until they have been successfully Device Tree enabled.
Signed-off-by: Lee Jones <redacted>
The patch is straight-forward but the commit message is confusing.
This just simplifies the code, the last two sentences about leaving
cpufreq and DMA behind you can put in your weekly report or
something :-)
Yours,
Linus Walleij
From: Lee Jones <hidden> Date: 2012-07-06 07:52:49
Sorry Dmitry, looks like I missed you off of the CC list on this one.
Here it is again:
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 2 ++
1 file changed, 2 insertions(+)
On Thu, Jul 5, 2012 at 3:06 PM, Lee Jones [off-list ref] wrote:
When pin control mapping tables are written the registered device
name is supplied for use in name-based searches within the pinctrl
driver. In the case of the DB8500 the string "pinctrl-db8500" is
used. However, when we register the driver with Device Tree, its
naming convention uses something that looks more like "pinctrl.2".
To work around the device naming inconsistencies between devices
registered via platform code and the ones registered by Device
Tree, we use AUXDATA to over-ride the Device Tree naming scheme.
Signed-off-by: Lee Jones <redacted>
Acked-by: Linus Walleij <redacted>
But I guess it won't work without the patch that Rob is herding.
So maybe that patch should go in with this one.
Yours,
Linus Walleij
From: Lee Jones <hidden> Date: 2012-07-06 09:30:26
Hi Linus,
Is this a NAK, or did you just miss it?
On 05/07/12 14:06, Lee Jones wrote:
quoted hunk
Here we remove some of the extra overhead we introduced to make
DT:ing the Snowball platform easier. Unfortunately we can't
remove everything. We are forced to leave cpufreq and DMA behind
until they have been successfully Device Tree enabled.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/mach-ux500/cpu-db8500.c | 10 ++--------
1 file changed, 2 insertions(+), 8 deletions(-)
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
From: Mark Brown <hidden> Date: 2012-07-06 11:23:25
On Fri, Jul 06, 2012 at 08:45:35AM +0100, Lee Jones wrote:
Please take a note of this and also do not apply the regulator
driver counterpart of the patch-set (it was one of the ones you
Reviewed-by in the previous batch).
From: Lee Jones <hidden> Date: 2012-07-06 11:52:45
On 06/07/12 12:23, Mark Brown wrote:
On Fri, Jul 06, 2012 at 08:45:35AM +0100, Lee Jones wrote:
quoted
Please take a note of this and also do not apply the regulator
driver counterpart of the patch-set (it was one of the ones you
Reviewed-by in the previous batch).
I've no idea what patches you're talking about, sorry.
Regarding patch:
[PATCH 08/15] regulator: Stop initialising AB8500's registers during bring-up
On 20/06/12 14:19, Lee Jones wrote:> On 20/06/12 14:08, Mark Brown wrote:
quoted
On Wed, Jun 20, 2012 at 01:56:44PM +0100, Lee Jones wrote:
quoted
There is no need to initialise the AB8500's regulator registers, as
most of this work is already carried out by framework features.
Does this have any dependency on the rest of the series?
Unfortunately yes.
It needs the "ARM: ux500:" stuff to go in first.
Then you added your Reviewed-by: until the ux500 stuff goes in.
However, the ux500 patch has been NACKed, so please don't take the aforementioned patch in either.
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
From: Mark Brown <hidden> Date: 2012-07-06 12:34:45
On Fri, Jul 06, 2012 at 12:52:35PM +0100, Lee Jones wrote:
Then you added your Reviewed-by: until the ux500 stuff goes in.
However, the ux500 patch has been NACKed, so please don't take the aforementioned patch in either.
No problem, if I punted on it like that I'll have deleted the mail and
so wouldn't have applied it anyway unless you resent it.
-------------- 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/20120706/3c448997/attachment.sig>
Hi Lee,
On Fri, Jul 06, 2012 at 08:52:43AM +0100, Lee Jones wrote:
quoted hunk
Sorry Dmitry, looks like I missed you off of the CC list on this one.
Here it is again:
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 2 ++
1 file changed, 2 insertions(+)
I do not think this is right - you are storing irq_dbr in irq_dbf. Also,
I think you want to do this earlier, like this:
- ponkey->irq_dbf = irq_dbf;
- ponkey->irq_dbr = irq_dbr;
+ ponkey->irq_dbf = ab8500_irq_get_virq(ab8500, irq_dbf);
+ ponkey->irq_dbr = ab8500_irq_get_virq(ab8500, irq_dbr);
Thanks.
--
Dmitry
From: Lee Jones <hidden> Date: 2012-07-06 17:51:21
So I am. Great spot Dmitry.
I'll fix and resubmit next week.
Sent from my mobile Linux device.
On Jul 6, 2012 6:40 PM, "Dmitry Torokhov" [off-list ref] wrote:
Hi Lee,
On Fri, Jul 06, 2012 at 08:52:43AM +0100, Lee Jones wrote:
quoted
Sorry Dmitry, looks like I missed you off of the CC list on this one.
Here it is again:
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 2 ++
1 file changed, 2 insertions(+)
I do not think this is right - you are storing irq_dbr in irq_dbf. Also,
I think you want to do this earlier, like this:
- ponkey->irq_dbf = irq_dbf;
- ponkey->irq_dbr = irq_dbr;
+ ponkey->irq_dbf = ab8500_irq_get_virq(ab8500, irq_dbf);
+ ponkey->irq_dbr = ab8500_irq_get_virq(ab8500, irq_dbr);
Thanks.
--
Dmitry
From: Lee Jones <hidden> Date: 2012-07-09 18:50:54
Better?
From 9ef8a52c00dedbd63b6eecc086f8bd688d2cb66a Mon Sep 17 00:00:00 2001
From: Lee Jones <redacted>
Date: Wed, 27 Jun 2012 08:42:08 +0100
Subject: [PATCH 1/1] ARM: ux500: Clean-up temporary Device Tree enablement
helpers
Here we remove some of the extra overhead we introduced to make
DT:ing the Snowball platform easier.
Signed-off-by: Lee Jones <redacted>
---
arch/arm/mach-ux500/cpu-db8500.c | 10 ++--------
1 file changed, 2 insertions(+), 8 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-09 18:53:08
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 2 ++
1 file changed, 2 insertions(+)
On Mon, Jul 09, 2012 at 08:53:01PM +0200, Lee Jones wrote:
quoted hunk
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 2 ++
1 file changed, 2 insertions(+)
@@ -82,6 +82,7 @@ static int __devinit ab8500_ponkey_probe(struct platform_device *pdev)input_set_capability(input,EV_KEY,KEY_POWER);+ponkey->irq_dbf=ab8500_irq_get_virq(ab8500,ponkey->irq_dbf);
*sigh*
Can we please do the assignment earlier? Look, we have a few lines
above:
ponkey->idev = input;
ponkey->ab8500 = ab8500;
ponkey->irq_dbf = irq_dbf;
ponkey->irq_dbr = irq_dbr;
Why do you keep these assignments and then override them with
translations? Why don't you do
ponkey->idev = input;
ponkey->ab8500 = ab8500;
ponkey->irq_dbf = ab8500_irq_get_virq(ab8500, irq_dbf);
ponkey->irq_dbr = ab8500_irq_get_virq(ab8500, irq_dbr);
to begin with?
Thanks.
--
Dmitry
From: Lee Jones <hidden> Date: 2012-07-09 20:34:43
Why do you keep these assignments and then override them with
translations? Why don't you do
ponkey->idev = input;
ponkey->ab8500 = ab8500;
ponkey->irq_dbf = ab8500_irq_get_virq(ab8500, irq_dbf);
ponkey->irq_dbr = ab8500_irq_get_virq(ab8500, irq_dbr);
to begin with?
Ah, I see what you mean now.
Sorry, just glossed over your first email.
From: Lee Jones <hidden> Date: 2012-07-09 20:35:54
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
On Mon, Jul 9, 2012 at 8:50 PM, Lee Jones [off-list ref] wrote:
Better?
Yep!
From 9ef8a52c00dedbd63b6eecc086f8bd688d2cb66a Mon Sep 17 00:00:00 2001
From: Lee Jones <redacted>
Date: Wed, 27 Jun 2012 08:42:08 +0100
Subject: [PATCH 1/1] ARM: ux500: Clean-up temporary Device Tree enablement
helpers
Here we remove some of the extra overhead we introduced to make
DT:ing the Snowball platform easier.
Signed-off-by: Lee Jones <redacted>
On Mon, Jul 9, 2012 at 10:35 PM, Lee Jones [off-list ref] wrote:
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
From: Lee Jones <hidden> Date: 2012-07-09 20:42:43
On 09/07/12 22:37, Linus Walleij wrote:
On Mon, Jul 9, 2012 at 10:35 PM, Lee Jones [off-list ref] wrote:
quoted
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
Acked-by: Linus Walleij <redacted>
I've just sent a new one following Dmitry's suggestion.
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
On Mon, Jul 09, 2012 at 10:35:48PM +0200, Lee Jones wrote:
quoted hunk
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
From: Lee Jones <hidden> Date: 2012-07-09 20:56:55
On 09/07/12 22:46, Dmitry Torokhov wrote:
On Mon, Jul 09, 2012 at 10:35:48PM +0200, Lee Jones wrote:
quoted
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
@@ -74,8 +74,8 @@ static int __devinit ab8500_ponkey_probe(struct platform_device *pdev)ponkey->idev=input;ponkey->ab8500=ab8500;-ponkey->irq_dbf=irq_dbf;-ponkey->irq_dbr=irq_dbr;+ponkey->irq_dbf=ab8500_irq_get_virq(ab8500,ponkey->irq_dbf);+ponkey->irq_dbr=ab8500_irq_get_virq(ab8500,ponkey->irq_dbr);
^^^^^^^^^^^^^^^
These are zeroes now. Man, you are killing me...
I've been traveling all day and my brain is dead.
Hopefully you have a sense of humor. :)
Give me two mins.
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
From: Lee Jones <hidden> Date: 2012-07-09 20:59:49
Sorry for the cock-up(s).
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
On Mon, Jul 9, 2012 at 10:59 PM, Lee Jones [off-list ref] wrote:
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
Now since I was fooled by the last patch, thinking you had tested it before
submitting (obviously not) - please send a test log of some
cat /dev/input/event* for this one before I ACK it this time.
Yours,
Linus Walleij
From: Lee Jones <hidden> Date: 2012-07-09 21:21:41
On 09/07/12 23:03, Linus Walleij wrote:
On Mon, Jul 9, 2012 at 10:59 PM, Lee Jones [off-list ref] wrote:
quoted
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
Now since I was fooled by the last patch, thinking you had tested it before
submitting (obviously not) - please send a test log of some
cat /dev/input/event* for this one before I ACK it this time.
I did test it, but must have only looked at the one IRQ incrementing.
I can provide a log tomorrow if it makes you feel better. :)
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
On Mon, Jul 09, 2012 at 10:59:43PM +0200, Lee Jones wrote:
Sorry for the cock-up(s).
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
Acked-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
I think it would be best if this got merged through whatever tree that
was used to introduce ab8500_irq_get_virq(); otherwise it will have to
wait till I resync with mainline sometime in the middle of the next
release cycle.
Thanks.
From: Lee Jones <hidden> Date: 2012-07-10 06:22:40
On 10/07/12 08:17, Dmitry Torokhov wrote:
On Mon, Jul 09, 2012 at 10:59:43PM +0200, Lee Jones wrote:
quoted
Sorry for the cock-up(s).
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Signed-off-by: Lee Jones <redacted>
Acked-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
I think it would be best if this got merged through whatever tree that
was used to introduce ab8500_irq_get_virq(); otherwise it will have to
wait till I resync with mainline sometime in the middle of the next
release cycle.
Thanks.
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
On Mon, Jul 9, 2012 at 11:21 PM, Lee Jones [off-list ref] wrote:
On 09/07/12 23:03, Linus Walleij wrote:
quoted
Now since I was fooled by the last patch, thinking you had tested it
before
submitting (obviously not) - please send a test log of some
cat /dev/input/event* for this one before I ACK it this time.
(Hm, sorry for grumpy mode...)
I did test it, but must have only looked at the one IRQ incrementing.
That's not enough, cat /dev/input/event/* whatever event node is used
by the ponkey, press it and verify you get some garbage (=events)
in the console.
Yours,
Linus Walleij
From: Mark Brown <hidden> Date: 2012-07-12 16:37:57
On Tue, Jul 10, 2012 at 11:08:39PM +0200, Linus Walleij wrote:
On Mon, Jul 9, 2012 at 11:21 PM, Lee Jones [off-list ref] wrote:
quoted
I did test it, but must have only looked at the one IRQ incrementing.
That's not enough, cat /dev/input/event/* whatever event node is used
by the ponkey, press it and verify you get some garbage (=events)
in the console.
There's the evtest debug program which is very handy for testing input
API stuff (Google should throw up copies, it's just a .c file).
From: Lee Jones <hidden> Date: 2012-07-13 13:43:45
On 10/07/12 22:08, Linus Walleij wrote:
On Mon, Jul 9, 2012 at 11:21 PM, Lee Jones [off-list ref] wrote:
quoted
On 09/07/12 23:03, Linus Walleij wrote:
quoted
quoted
Now since I was fooled by the last patch, thinking you had tested it
before
submitting (obviously not) - please send a test log of some
cat /dev/input/event* for this one before I ACK it this time.
(Hm, sorry for grumpy mode...)
quoted
I did test it, but must have only looked at the one IRQ incrementing.
That's not enough, cat /dev/input/event/* whatever event node is used
by the ponkey, press it and verify you get some garbage (=events)
in the console.
--
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
On Fri, Jul 13, 2012 at 3:43 PM, Lee Jones [off-list ref] wrote:
On 10/07/12 22:08, Linus Walleij wrote:
quoted
That's not enough, cat /dev/input/event/* whatever event node is used
by the ponkey, press it and verify you get some garbage (=events)
in the console.
Yes, garbage seen, works fine.
OK!
Acked-by: on whatever Dmitry Acked.
Yours,
Linus Walleij
From: Lee Jones <hidden> Date: 2012-07-16 07:35:45
I think it would be best if this got merged through whatever tree that
was used to introduce ab8500_irq_get_virq(); otherwise it will have to
wait till I resync with mainline sometime in the middle of the next
release cycle.
I agree. Sam that's you, patch below. :)
From: Lee Jones <redacted>
Date: Tue, 26 Jun 2012 10:46:50 +0100
Subject: [PATCH 1/1] Input: ab8500-ponkey: Create AB8500 domain IRQ mapping
Before we can use any domain allocated IRQ, we need to first create a
map between the Hardware IRQ (hwirq) and the Linux Virtual IRQ (virq).
We do this with a helper function provided by the AB8500 IRQ domain
controller called ab8500_irq_get_virq(). We need to do this for both
IRQs which the Power-On-Key driver uses; one for button press, the other
for button depress.
Acked-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Acked-by: Linus Walleij <redacted>
Signed-off-by: Lee Jones <redacted>
---
drivers/input/misc/ab8500-ponkey.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
From: Samuel Ortiz <hidden> Date: 2012-07-16 13:57:31
Hi Lee,
On Mon, Jul 16, 2012 at 08:35:37AM +0100, Lee Jones wrote:
quoted
I think it would be best if this got merged through whatever tree that
was used to introduce ab8500_irq_get_virq(); otherwise it will have to
wait till I resync with mainline sometime in the middle of the next
release cycle.
I agree. Sam that's you, patch below. :)
Applied, thanks.
Cheers,
Samuel.
--
Intel Open Source Technology Centre
http://oss.intel.com/