linux-next: mfd tree build failure

linux-next: mfd tree build failure

[Stephen Rothwell]

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/power/pcf50633-charger.c: In function 'pcf50633_mbc_get_usb_online_status':
drivers/power/pcf50633-charger.c:147: error: 'struct pcf50633' has no member named 'mbc_pdev'

Caused by commit 4c2e3bacc050f0ebf884954bb007616c162564cd ("mfd: pcf50633
- fix error handling during probe") from the mfd tree interacting with
commit c329795052aa339850a45fab649ab97a36905136 ("pcf50633: Query charger
status directly") from the battery tree.

I have applied the following merge fixup patch and can carry it as
necessary.

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 23 Nov 2009 14:31:43 +1100
Subject: [PATCH] mfd: fix pcf50633-charger.c for fallout from merge

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 drivers/power/pcf50633-charger.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/power/pcf50633-charger.c b/drivers/power/pcf50633-charger.c
index d4a3e77..4e535bb 100644
--- a/drivers/power/pcf50633-charger.c
+++ b/drivers/power/pcf50633-charger.c
@@ -144,7 +144,8 @@ EXPORT_SYMBOL_GPL(pcf50633_mbc_get_status);
 
 int pcf50633_mbc_get_usb_online_status(struct pcf50633 *pcf)
 {
-	struct pcf50633_mbc *mbc  = platform_get_drvdata(pcf->mbc_pdev);
+	struct pcf50633_mbc *mbc =
+		platform_get_drvdata(pcf->pdevs[PCF50633_PDEV_ADC_IDX]);
 
 	if (!mbc)
 		return 0;
-- 
1.6.5.3

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

Hi Dmitry,

On Mon, Nov 23, 2009 at 02:35:03PM +1100, Stephen Rothwell wrote:
> Hi Samuel,
> 
> Today's linux-next build (x86_64 allmodconfig) failed like this:
> 
> drivers/power/pcf50633-charger.c: In function 'pcf50633_mbc_get_usb_online_status':
> drivers/power/pcf50633-charger.c:147: error: 'struct pcf50633' has no member named 'mbc_pdev'
> 
> Caused by commit 4c2e3bacc050f0ebf884954bb007616c162564cd ("mfd: pcf50633
> - fix error handling during probe") from the mfd tree interacting with
> commit c329795052aa339850a45fab649ab97a36905136 ("pcf50633: Query charger
> status directly") from the battery tree.
Could we split the above fix so that we get back to the:

	struct platform_device *rtc_pdev;
	struct platform_device *mbc_pdev;
	struct platform_device *adc_pdev;
	struct platform_device *input_pdev;
	struct platform_device *regulator_pdev[PCF50633_NUM_REGULATORS];

structure ? We can move to the index based structure after the next merge
window. That would make my life easier in terms of tree dependencies.
If you dont have time to do this change, I'll do it some time this week.

Thanks in advance.

Cheers,
Samuel.


> I have applied the following merge fixup patch and can carry it as
> necessary.
> 
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Mon, 23 Nov 2009 14:31:43 +1100
> Subject: [PATCH] mfd: fix pcf50633-charger.c for fallout from merge
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> ---
>  drivers/power/pcf50633-charger.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/power/pcf50633-charger.c b/drivers/power/pcf50633-charger.c
> index d4a3e77..4e535bb 100644
> --- a/drivers/power/pcf50633-charger.c
> +++ b/drivers/power/pcf50633-charger.c
> @@ -144,7 +144,8 @@ EXPORT_SYMBOL_GPL(pcf50633_mbc_get_status);
>  
>  int pcf50633_mbc_get_usb_online_status(struct pcf50633 *pcf)
>  {
> -	struct pcf50633_mbc *mbc  = platform_get_drvdata(pcf->mbc_pdev);
> +	struct pcf50633_mbc *mbc =
> +		platform_get_drvdata(pcf->pdevs[PCF50633_PDEV_ADC_IDX]);
>  
>  	if (!mbc)
>  		return 0;
> -- 
> 1.6.5.3
> 
> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/

-- 
Intel Open Source Technology Centre
http://oss.intel.com/

Re: linux-next: mfd tree build failure

[Dmitry Torokhov]

Hi Samuel,

On Mon, Nov 23, 2009 at 01:26:07PM +0100, Samuel Ortiz wrote:
> Hi Dmitry,
> 
> On Mon, Nov 23, 2009 at 02:35:03PM +1100, Stephen Rothwell wrote:
> > Hi Samuel,
> > 
> > Today's linux-next build (x86_64 allmodconfig) failed like this:
> > 
> > drivers/power/pcf50633-charger.c: In function 'pcf50633_mbc_get_usb_online_status':
> > drivers/power/pcf50633-charger.c:147: error: 'struct pcf50633' has no member named 'mbc_pdev'
> > 
> > Caused by commit 4c2e3bacc050f0ebf884954bb007616c162564cd ("mfd: pcf50633
> > - fix error handling during probe") from the mfd tree interacting with
> > commit c329795052aa339850a45fab649ab97a36905136 ("pcf50633: Query charger
> > status directly") from the battery tree.
> Could we split the above fix so that we get back to the:
> 
> 	struct platform_device *rtc_pdev;
> 	struct platform_device *mbc_pdev;
> 	struct platform_device *adc_pdev;
> 	struct platform_device *input_pdev;
> 	struct platform_device *regulator_pdev[PCF50633_NUM_REGULATORS];
> 
> structure ? We can move to the index based structure after the next merge
> window. That would make my life easier in terms of tree dependencies.
> If you dont have time to do this change, I'll do it some time this week.
> 

Please feel free to drop or reschedule the patch as you see fit. I don't
have any problem if it will go in later.

Thanks.

-- 
Dmitry

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

Hi Dmitry

On Mon, Nov 23, 2009 at 08:32:11AM -0800, Dmitry Torokhov wrote:
> > Could we split the above fix so that we get back to the:
> > 
> > 	struct platform_device *rtc_pdev;
> > 	struct platform_device *mbc_pdev;
> > 	struct platform_device *adc_pdev;
> > 	struct platform_device *input_pdev;
> > 	struct platform_device *regulator_pdev[PCF50633_NUM_REGULATORS];
> > 
> > structure ? We can move to the index based structure after the next merge
> > window. That would make my life easier in terms of tree dependencies.
> > If you dont have time to do this change, I'll do it some time this week.
> > 
> 
> Please feel free to drop or reschedule the patch as you see fit. I don't
> have any problem if it will go in later.
Yes, I'll requeue your patchset after the next merge window is closed.
Thanks for your understanding.

Cheers,
Samuel.

-- 
Intel Open Source Technology Centre
http://oss.intel.com/

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 271 bytes --]

Hi Samuel,

On Tue, 19 Jan 2010 11:59:27 +0100 Samuel Ortiz <sameo@linux.intel.com> wrote:
>
> Thanks for the report. It should be all fixed now.

Thanks.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

Hi Stephen,

On Tue, Jan 19, 2010 at 03:06:53PM +1100, Stephen Rothwell wrote:
> Hi Samuel,
> 
> Today's linux-next build (x86_64 allmodconfig) failed like this:
> 
> drivers/mfd/htc-i2cpld.c: In function 'htcpld_set_type':
> drivers/mfd/htc-i2cpld.c:106: error: 'irq_desc' undeclared (first use in this function)
> 
> Caused by commit b4714172cd2f1cd8694546169237c30c5b4fdf3b ("mfd: Add
> HTCPLD driver").
Thanks for the report. It should be all fixed now.

Cheers,
Samuel.


> I have used the mfd tree from next-20100118 for today.
> 
> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/



-- 
Intel Open Source Technology Centre
http://oss.intel.com/

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 486 bytes --]

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/mfd/htc-i2cpld.c: In function 'htcpld_set_type':
drivers/mfd/htc-i2cpld.c:106: error: 'irq_desc' undeclared (first use in this function)

Caused by commit b4714172cd2f1cd8694546169237c30c5b4fdf3b ("mfd: Add
HTCPLD driver").

I have used the mfd tree from next-20100118 for today.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 475 bytes --]

Hi Sam,

On Mon, 11 Jan 2010 13:33:39 +0100 Samuel Ortiz <sameo@linux.intel.com> wrote:
>
> Actually, I fixed it. The problem was caused by the mfd and regulator drivers
> having the same name, i.e. max8925.ko. I renamed the regulator driver to
> max8925-regulator.ko, which also makes it more consistent with the rest of the
> regulator drivers.

Thanks.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

Hi Haojian,

On Mon, Jan 11, 2010 at 07:02:14AM -0500, Haojian Zhuang wrote:
> On Mon, Jan 11, 2010 at 5:51 AM, Mark Brown
> <broonie@opensource.wolfsonmicro.com> wrote:
> > On Mon, Jan 11, 2010 at 04:02:32PM +1100, Stephen Rothwell wrote:
> >> Hi Sam,
> >
> >> Today's linux-next build (x86_64_allmodconfig, gcc 4.4.0, binutils
> >> 2.19.1) failed like this:
> >
> >> ERROR: "max8925_set_bits" [drivers/video/backlight/max8925_bl.ko] undefined!
> >> ERROR: "max8925_reg_write" [drivers/video/backlight/max8925_bl.ko] undefined!
> >> ERROR: "max8925_reg_read" [drivers/video/backlight/max8925_bl.ko] undefined!
> >> ERROR: "max8925_set_bits" [drivers/regulator/max8925.ko] undefined!
> >> ERROR: "max8925_reg_read" [drivers/regulator/max8925.ko] undefined!
> >
> >> I am not sure what caused this, so I have used the version of the mfd
> >> tree from next-20100108 for today.
> >
> > CCing in Haojian who did the driver.  I've not looked but I'd guess that
> > the issue is missing EXPORT_SYMBOL{,_GPL}s on the symbols above.
> >
> 
> I'm sorry on inconvience. Now I attache the fix.
> 
> Samuel,
> Could you help to review and merge this patch?
Actually, I fixed it. The problem was caused by the mfd and regulator drivers
having the same name, i.e. max8925.ko. I renamed the regulator driver to
max8925-regulator.ko, which also makes it more consistent with the rest of the
regulator drivers.

Stephen, the fix is already pushed, you can resume pulling the latest mfd
bits. Thanks for letting us know about the failure.

Cheers,
Samuel.


> It's not caused by EXPORT_SYMBOL{,_GPL}. If I build in max8925 mfd
> driver, issue is resolved.
> Regulator subsystem could only be built in. And 8925 regulator driver
> calls i2c operations that is defined in mfd driver. If 8925 regulator
> is built it and mfd 8925 is built as module, it will cause linking
> error.
> 
> Thanks
> Haojian

> From cc08aae952302cbc4445f4db49fdfab5f9d779ed Mon Sep 17 00:00:00 2001
> From: Haojian Zhuang <haojian.zhuang@marvell.com>
> Date: Mon, 11 Jan 2010 14:38:39 -0500
> Subject: [PATCH] mfd: set max8925 as built in
> 
> Since regulator module is always built-in, MFD driver have to be built in.
> If MFD driver is built as module, linking error will occur.
> 
> Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
> ---
>  drivers/mfd/Kconfig |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig
> index 94e4d5e..b892c38 100644
> --- a/drivers/mfd/Kconfig
> +++ b/drivers/mfd/Kconfig
> @@ -186,7 +186,7 @@ config PMIC_ADP5520
>  
>  config MFD_MAX8925
>  	tristate "Maxim Semiconductor MAX8925 PMIC Support"
> -	depends on I2C
> +	depends on I2C=y
>  	select MFD_CORE
>  	help
>  	  Say yes here to support for Maxim Semiconductor MAX8925. This is
> -- 
> 1.5.6.5
> 


-- 
Intel Open Source Technology Centre
http://oss.intel.com/

Re: linux-next: mfd tree build failure

[Haojian Zhuang]

[-- Attachment #1: Type: text/plain, Size: 1379 bytes --]

On Mon, Jan 11, 2010 at 5:51 AM, Mark Brown
<broonie@opensource.wolfsonmicro.com> wrote:
> On Mon, Jan 11, 2010 at 04:02:32PM +1100, Stephen Rothwell wrote:
>> Hi Sam,
>
>> Today's linux-next build (x86_64_allmodconfig, gcc 4.4.0, binutils
>> 2.19.1) failed like this:
>
>> ERROR: "max8925_set_bits" [drivers/video/backlight/max8925_bl.ko] undefined!
>> ERROR: "max8925_reg_write" [drivers/video/backlight/max8925_bl.ko] undefined!
>> ERROR: "max8925_reg_read" [drivers/video/backlight/max8925_bl.ko] undefined!
>> ERROR: "max8925_set_bits" [drivers/regulator/max8925.ko] undefined!
>> ERROR: "max8925_reg_read" [drivers/regulator/max8925.ko] undefined!
>
>> I am not sure what caused this, so I have used the version of the mfd
>> tree from next-20100108 for today.
>
> CCing in Haojian who did the driver.  I've not looked but I'd guess that
> the issue is missing EXPORT_SYMBOL{,_GPL}s on the symbols above.
>

I'm sorry on inconvience. Now I attache the fix.

Samuel,
Could you help to review and merge this patch?

It's not caused by EXPORT_SYMBOL{,_GPL}. If I build in max8925 mfd
driver, issue is resolved.
Regulator subsystem could only be built in. And 8925 regulator driver
calls i2c operations that is defined in mfd driver. If 8925 regulator
is built it and mfd 8925 is built as module, it will cause linking
error.

Thanks
Haojian

[-- Attachment #2: 0001-mfd-set-max8925-as-built-in.patch --]
[-- Type: text/x-patch, Size: 883 bytes --]

From cc08aae952302cbc4445f4db49fdfab5f9d779ed Mon Sep 17 00:00:00 2001
From: Haojian Zhuang <haojian.zhuang@marvell.com>
Date: Mon, 11 Jan 2010 14:38:39 -0500
Subject: [PATCH] mfd: set max8925 as built in

Since regulator module is always built-in, MFD driver have to be built in.
If MFD driver is built as module, linking error will occur.

Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
---
 drivers/mfd/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig
index 94e4d5e..b892c38 100644
--- a/drivers/mfd/Kconfig
+++ b/drivers/mfd/Kconfig
@@ -186,7 +186,7 @@ config PMIC_ADP5520
 
 config MFD_MAX8925
 	tristate "Maxim Semiconductor MAX8925 PMIC Support"
-	depends on I2C
+	depends on I2C=y
 	select MFD_CORE
 	help
 	  Say yes here to support for Maxim Semiconductor MAX8925. This is
-- 
1.5.6.5

Re: linux-next: mfd tree build failure

[Mark Brown]

On Mon, Jan 11, 2010 at 04:02:32PM +1100, Stephen Rothwell wrote:
> Hi Sam,

> Today's linux-next build (x86_64_allmodconfig, gcc 4.4.0, binutils
> 2.19.1) failed like this:

> ERROR: "max8925_set_bits" [drivers/video/backlight/max8925_bl.ko] undefined!
> ERROR: "max8925_reg_write" [drivers/video/backlight/max8925_bl.ko] undefined!
> ERROR: "max8925_reg_read" [drivers/video/backlight/max8925_bl.ko] undefined!
> ERROR: "max8925_set_bits" [drivers/regulator/max8925.ko] undefined!
> ERROR: "max8925_reg_read" [drivers/regulator/max8925.ko] undefined!

> I am not sure what caused this, so I have used the version of the mfd
> tree from next-20100108 for today.

CCing in Haojian who did the driver.  I've not looked but I'd guess that
the issue is missing EXPORT_SYMBOL{,_GPL}s on the symbols above.

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 699 bytes --]

Hi Sam,

Today's linux-next build (x86_64_allmodconfig, gcc 4.4.0, binutils
2.19.1) failed like this:

ERROR: "max8925_set_bits" [drivers/video/backlight/max8925_bl.ko] undefined!
ERROR: "max8925_reg_write" [drivers/video/backlight/max8925_bl.ko] undefined!
ERROR: "max8925_reg_read" [drivers/video/backlight/max8925_bl.ko] undefined!
ERROR: "max8925_set_bits" [drivers/regulator/max8925.ko] undefined!
ERROR: "max8925_reg_read" [drivers/regulator/max8925.ko] undefined!

I am not sure what caused this, so I have used the version of the mfd
tree from next-20100108 for today.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Mark Brown]

On Tue, Nov 24, 2009 at 09:14:33PM +1100, Stephen Rothwell wrote:

> Sorry, I don't have the actual config (as I overwrite my build directory
> between each merge).  However, if I checkout the mfd tree (alone -
> nothing else in linux-next) and do "make ARCH=x86_64 allmodconfig", I get
> the following in my .config:

> CONFIG_I2C=m
> CONFIG_MFD_CORE=m
> CONFIG_MFD_WM831X=y

Oh, fail.  This is nothing to do with select, it's to do with bool.
Kbuild doesn't propagate the dependency from the boolean (which makes
sense for a lot of uses but not for this one).  The current "fix" used
by other drivers in the same situation is to depend on I2C=y which works
but means that we loose build coverage from allmodconfig.  I'll post a
patch for this shortly.

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 934 bytes --]

Hi Mark,

On Tue, 24 Nov 2009 09:57:16 +0000 Mark Brown <broonie@opensource.wolfsonmicro.com> wrote:
>
> There's something more seriously wrong than the select with your config
> here - the driver depends on (rather than selecting) I2C and you're also
> seeing undefined references to the I2C API.  The select on MFD_CORE
> should also be forcing that to be built in (it does normally).
> 
> Could you please post the actual configuration you're using?  I suspect
> some Kbuild bug has been triggered here...

Sorry, I don't have the actual config (as I overwrite my build directory
between each merge).  However, if I checkout the mfd tree (alone -
nothing else in linux-next) and do "make ARCH=x86_64 allmodconfig", I get
the following in my .config:

CONFIG_I2C=m
CONFIG_MFD_CORE=m
CONFIG_MFD_WM831X=y

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Mark Brown]

On Tue, Nov 24, 2009 at 04:37:53PM +1100, Stephen Rothwell wrote:

> drivers/built-in.o: In function `wm831x_i2c_remove':
> wm831x-core.c:(.text+0xd108b): undefined reference to `mfd_remove_devices'
> drivers/built-in.o: In function `wm831x_i2c_write_device':
> wm831x-core.c:(.text+0xd119c): undefined reference to `i2c_master_send'
> drivers/built-in.o: In function `wm831x_i2c_read_device':
> wm831x-core.c:(.text+0xd1208): undefined reference to `i2c_master_send'
> wm831x-core.c:(.text+0xd1237): undefined reference to `i2c_master_recv'

> Caused by commit 4f7231a3ff624ad8a0fdace424b10a8b262071cf ("mfd: Don't
> allow wm831x to be built as a module").  CONFIG_MFD_CORE is "m" and
> CONFIG_I2C is "m" while CONFIG_MFD_WM831X is "y".  "select" is a real
> problem :-(

There's something more seriously wrong than the select with your config
here - the driver depends on (rather than selecting) I2C and you're also
seeing undefined references to the I2C API.  The select on MFD_CORE
should also be forcing that to be built in (it does normally).

Could you please post the actual configuration you're using?  I suspect
some Kbuild bug has been triggered here...

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1711 bytes --]

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/built-in.o: In function `wm831x_i2c_remove':
wm831x-core.c:(.text+0xd108b): undefined reference to `mfd_remove_devices'
drivers/built-in.o: In function `wm831x_i2c_write_device':
wm831x-core.c:(.text+0xd119c): undefined reference to `i2c_master_send'
drivers/built-in.o: In function `wm831x_i2c_read_device':
wm831x-core.c:(.text+0xd1208): undefined reference to `i2c_master_send'
wm831x-core.c:(.text+0xd1237): undefined reference to `i2c_master_recv'
drivers/built-in.o: In function `wm831x_i2c_probe':
wm831x-core.c:(.text+0xd2099): undefined reference to `mfd_add_devices'
wm831x-core.c:(.text+0xd20cd): undefined reference to `mfd_add_devices'
wm831x-core.c:(.text+0xd2105): undefined reference to `mfd_add_devices'
wm831x-core.c:(.text+0xd2139): undefined reference to `mfd_add_devices'
wm831x-core.c:(.text+0xd21dd): undefined reference to `mfd_add_devices'
wm831x-core.c:(.text+0xd22c3): undefined reference to `mfd_remove_devices'
drivers/built-in.o: In function `wm831x_i2c_init':
wm831x-core.c:(.init.text+0xae6b): undefined reference to `i2c_register_driver'
drivers/built-in.o: In function `wm831x_i2c_exit':
wm831x-core.c:(.exit.text+0x522): undefined reference to `i2c_del_driver'

Caused by commit 4f7231a3ff624ad8a0fdace424b10a8b262071cf ("mfd: Don't
allow wm831x to be built as a module").  CONFIG_MFD_CORE is "m" and
CONFIG_I2C is "m" while CONFIG_MFD_WM831X is "y".  "select" is a real
problem :-(

I have used the version of the mfd tree from next-20091120 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Mark Brown]

On Mon, Nov 23, 2009 at 03:44:14PM +0100, Samuel Ortiz wrote:
> On Mon, Nov 23, 2009 at 01:11:13PM +0000, Mark Brown wrote:

> > I don't see getting them exported flying, especially not in the
> > timeframe we've got for the merge window.  Keeping them non-exported has
> > been a deliberate decision on the part of the genirq maintainers. 

> Out of curiosity, what's the reasoning behind thatdecision ?

Unless you're actually implementing an interrupt controller those APIs
should never be called and people should be using flags on request_irq()
or similar.  Driver authors have been awfully fond of bypassing those
APIs, but this prevents them doing so.

None of this is really a practical issue unless you implement an IRQ
controller in a driver and that's very rare (or has been up until now).

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

On Mon, Nov 23, 2009 at 01:11:13PM +0000, Mark Brown wrote:
> On Mon, Nov 23, 2009 at 01:29:04PM +0100, Samuel Ortiz wrote:
> 
> > True. Mark, I think we should get those symbols exported, unless you have a
> > better idea ? Having wm831x as a boolean choice from Kconfig doesnt sound too
> > god to me.
> 
> I don't see getting them exported flying, especially not in the
> timeframe we've got for the merge window.  Keeping them non-exported has
> been a deliberate decision on the part of the genirq maintainers. 
Out of curiosity, what's the reasoning behind thatdecision ?

> I'll send a patch making it built in.
Thanks, patch applied.

Cheers,
Samuel.


> Given the typical applications for these parts building the driver in
> isn't the end of the world - there's a reasonable chance it'll be forced
> built in for other reasons due to device-specific configuration.

-- 
Intel Open Source Technology Centre
http://oss.intel.com/

Re: linux-next: mfd tree build failure

[Mark Brown]

On Mon, Nov 23, 2009 at 01:29:04PM +0100, Samuel Ortiz wrote:

> True. Mark, I think we should get those symbols exported, unless you have a
> better idea ? Having wm831x as a boolean choice from Kconfig doesnt sound too
> god to me.

I don't see getting them exported flying, especially not in the
timeframe we've got for the merge window.  Keeping them non-exported has
been a deliberate decision on the part of the genirq maintainers.  I'll
send a patch making it built in.

Given the typical applications for these parts building the driver in
isn't the end of the world - there's a reasonable chance it'll be forced
built in for other reasons due to device-specific configuration.

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

Hi Stephen,

On Mon, Nov 23, 2009 at 03:03:46PM +1100, Stephen Rothwell wrote:
> Hi Samuel,
> 
> On Mon, 23 Nov 2009 14:41:03 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> >
> > Today's linux-next build (x86_64 allmodconfig) failed like this:
> > 
> > ERROR: "irq_to_desc" [drivers/mfd/wm831x.ko] undefined!
> > ERROR: "handle_edge_irq" [drivers/mfd/wm831x.ko] undefined!
> > ERROR: "set_irq_chip_and_handler" [drivers/mfd/wm831x.ko] undefined!
> > ERROR: "set_irq_noprobe" [drivers/mfd/wm831x.ko] undefined!
> > 
> > Probably caused by commit 8fc686c7369173519817f0cfbb2df9a3671d5b8d ("mfd:
> > Move WM831x to generic IRQ").  Maybe a Kconfig dependency is missing?
> 
> It actually looks as though set_irq_chip_and_handler (at least) is not
> exported to modules.
True. Mark, I think we should get those symbols exported, unless you have a
better idea ? Having wm831x as a boolean choice from Kconfig doesnt sound too
god to me.

Cheers,
Samuel.


> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/



-- 
Intel Open Source Technology Centre
http://oss.intel.com/

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 786 bytes --]

Hi Samuel,

On Mon, 23 Nov 2009 14:41:03 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Today's linux-next build (x86_64 allmodconfig) failed like this:
> 
> ERROR: "irq_to_desc" [drivers/mfd/wm831x.ko] undefined!
> ERROR: "handle_edge_irq" [drivers/mfd/wm831x.ko] undefined!
> ERROR: "set_irq_chip_and_handler" [drivers/mfd/wm831x.ko] undefined!
> ERROR: "set_irq_noprobe" [drivers/mfd/wm831x.ko] undefined!
> 
> Probably caused by commit 8fc686c7369173519817f0cfbb2df9a3671d5b8d ("mfd:
> Move WM831x to generic IRQ").  Maybe a Kconfig dependency is missing?

It actually looks as though set_irq_chip_and_handler (at least) is not
exported to modules.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 644 bytes --]

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

ERROR: "irq_to_desc" [drivers/mfd/wm831x.ko] undefined!
ERROR: "handle_edge_irq" [drivers/mfd/wm831x.ko] undefined!
ERROR: "set_irq_chip_and_handler" [drivers/mfd/wm831x.ko] undefined!
ERROR: "set_irq_noprobe" [drivers/mfd/wm831x.ko] undefined!

Probably caused by commit 8fc686c7369173519817f0cfbb2df9a3671d5b8d ("mfd:
Move WM831x to generic IRQ").  Maybe a Kconfig dependency is missing?

I have used the mfd tree from next-20091120 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1467 bytes --]

Hi Samuel,

On Tue, 6 Oct 2009 13:02:02 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Today's linux-next build (x86_64 allmodconfig) failed like this:
> 
> drivers/built-in.o: In function `pm8607_write_device':
> 88pm8607.c:(.text+0xd02e9): undefined reference to `i2c_master_send'
> drivers/built-in.o: In function `pm8607_read_device':
> 88pm8607.c:(.text+0xd033b): undefined reference to `i2c_master_send'
> 88pm8607.c:(.text+0xd035c): undefined reference to `i2c_master_recv'
> drivers/built-in.o: In function `pm8607_init':
> 88pm8607.c:(.init.text+0xa9bd): undefined reference to `i2c_register_driver'
> drivers/built-in.o: In function `pm8607_probe':
> 88pm8607.c:(.devinit.text+0x4db6): undefined reference to `mfd_add_devices'
> drivers/built-in.o: In function `pm8607_exit':
> 88pm8607.c:(.exit.text+0x598): undefined reference to `i2c_del_driver'
> drivers/built-in.o: In function `pm8607_remove':
> 88pm8607.c:(.devexit.text+0x198): undefined reference to `mfd_remove_devices'
> 
> Obviously caused by commit 9ec6cd8f2235ddeee5b39068956f14238a12a617
> ("mfd: Add 88PM8607 driver").
> 
> This build has CONFIG_I2C=m and CONFIG_MFD_CORE=m.
> 
> I have used the version of the mfd tree form next-20091005 for today.

This build failure is still happening and I am still using the mfd tree
from next-20091005.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Haojian Zhuang]

> -----Original Message-----
> From: Stephen Rothwell [mailto:sfr@canb.auug.org.au]
> Sent: 2009年10月6日 11:54 AM
> To: Haojian Zhuang
> Cc: Samuel Ortiz; linux-next@vger.kernel.org; linux-kernel@vger.kernel.org; Mark Brown
> Subject: Re: linux-next: mfd tree build failure
>
> Hi Haojian,
>
> On Mon, 5 Oct 2009 19:35:05 -0700 Haojian Zhuang <hzhuang1@marvell.com> wrote:
>>
>> Could you send your .config file to me? The 88pm8607 patch is just for mfd-2.6 git tree. I didn't try it in linux-next git tree before. I can try it now. If I could have your .config file, it would be better.
>
> I am sorry, but I don't have the exact .config file any more (I have
> moved on in my creation of linux-next)  however, as I said, this was an
> allmodconfig build on x86_64.  I also noted that CONFIG_I2C=m and
> CONFIG_MFD_CORE=m, while the config entry that controls the building of
> this driver is a bool so must have been 'y'.  So from a builtin driver,
> you are referring to symbols that are only defined in modules.
>
> --
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/
>

Hi Stephen & Samuel,

Please ignore the original fix patch on 88pm8607 dependancy. I write a
new patch on it and paste it in below. Please help to review it.

Thanks
Haojian

>From 26174b67184759ac2fe33748eeb43d3738732e21 Mon Sep 17 00:00:00 2001
From: Haojian Zhuang <haojian.zhuang@marvell.com>
Date: Thu, 8 Oct 2009 09:07:05 -0400
Subject: [PATCH] mfd: fix the dependancy of 88PM8607

88PM8607 module is dependant on I2C. I2C module must be built-in.

Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
---
 drivers/mfd/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig
index fbe684b..00064e2 100644
--- a/drivers/mfd/Kconfig
+++ b/drivers/mfd/Kconfig
@@ -307,7 +307,7 @@ config EZX_PCAP

 config MFD_88PM8607
        bool "Support Marvell 88PM8607"
-       depends on I2C
+       depends on I2C=y
        select MFD_CORE
        help
          This supports for Marvell 88PM8607 Power Management IC. This includes
--
1.5.6.5

Re: linux-next: mfd tree build failure

[Haojian Zhuang]

> -----Original Message-----
> From: Stephen Rothwell [mailto:sfr@canb.auug.org.au]
> Sent: 2009年10月6日 11:54 AM
> To: Haojian Zhuang
> Cc: Samuel Ortiz; linux-next@vger.kernel.org; linux-kernel@vger.kernel.org; Mark Brown
> Subject: Re: linux-next: mfd tree build failure
>
> Hi Haojian,
>
> On Mon, 5 Oct 2009 19:35:05 -0700 Haojian Zhuang <hzhuang1@marvell.com> wrote:
>>
>> Could you send your .config file to me? The 88pm8607 patch is just for mfd-2.6 git tree. I didn't try it in linux-next git tree before. I can try it now. If I could have your .config file, it would be better.
>
> I am sorry, but I don't have the exact .config file any more (I have
> moved on in my creation of linux-next)  however, as I said, this was an
> allmodconfig build on x86_64.  I also noted that CONFIG_I2C=m and
> CONFIG_MFD_CORE=m, while the config entry that controls the building of
> this driver is a bool so must have been 'y'.  So from a builtin driver,
> you are referring to symbols that are only defined in modules.
>
> --
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/
>

Hi Stephen & Samuel,

Please ignore the original fix patch on 88pm8607 dependancy. I write a
new patch on it and paste it in below. Please help to review it.

Thanks
Haojian

From 26174b67184759ac2fe33748eeb43d3738732e21 Mon Sep 17 00:00:00 2001
From: Haojian Zhuang <haojian.zhuang@marvell.com>
Date: Thu, 8 Oct 2009 09:07:05 -0400
Subject: [PATCH] mfd: fix the dependancy of 88PM8607

88PM8607 module is dependant on I2C. I2C module must be built-in.

Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
---
 drivers/mfd/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig
index fbe684b..00064e2 100644
--- a/drivers/mfd/Kconfig
+++ b/drivers/mfd/Kconfig
@@ -307,7 +307,7 @@ config EZX_PCAP

 config MFD_88PM8607
        bool "Support Marvell 88PM8607"
-       depends on I2C
+       depends on I2C=y
        select MFD_CORE
        help
          This supports for Marvell 88PM8607 Power Management IC. This includes
--
1.5.6.5

RE: linux-next: mfd tree build failure

[Haojian Zhuang]

[-- Attachment #1: Type: text/plain, Size: 1260 bytes --]

Hi Stephen,

Thanks a lot for your help. I fix it now. Please help to review it. It's already passed the build test.

Best Regards
Haojian

-----Original Message-----
From: Stephen Rothwell [mailto:sfr@canb.auug.org.au] 
Sent: 2009年10月6日 11:54 AM
To: Haojian Zhuang
Cc: Samuel Ortiz; linux-next@vger.kernel.org; linux-kernel@vger.kernel.org; Mark Brown
Subject: Re: linux-next: mfd tree build failure

Hi Haojian,

On Mon, 5 Oct 2009 19:35:05 -0700 Haojian Zhuang <hzhuang1@marvell.com> wrote:
>
> Could you send your .config file to me? The 88pm8607 patch is just for mfd-2.6 git tree. I didn't try it in linux-next git tree before. I can try it now. If I could have your .config file, it would be better.

I am sorry, but I don't have the exact .config file any more (I have
moved on in my creation of linux-next)  however, as I said, this was an
allmodconfig build on x86_64.  I also noted that CONFIG_I2C=m and
CONFIG_MFD_CORE=m, while the config entry that controls the building of
this driver is a bool so must have been 'y'.  So from a builtin driver,
you are referring to symbols that are only defined in modules.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: 0002-mfd-fix-the-dependancy-issue-on-88PM8607.patch --]
[-- Type: application/octet-stream, Size: 959 bytes --]

From 543c564ba18189abbada3a16228a6193c5757213 Mon Sep 17 00:00:00 2001
From: Haojian Zhuang <haojian.zhuang@marvell.com>
Date: Tue, 6 Oct 2009 08:08:37 -0400
Subject: [PATCH] mfd: fix the dependancy issue on 88PM8607

88PM8607 depends on I2C and MFD_CORE. Since 88PM8607 is built-in
kernel, it also requires that I2C and MFD_CORE are built-in
kernel also.

Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
---
 drivers/mfd/Kconfig |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig
index fbe684b..6e9efcd 100644
--- a/drivers/mfd/Kconfig
+++ b/drivers/mfd/Kconfig
@@ -307,8 +307,7 @@ config EZX_PCAP
 
 config MFD_88PM8607
 	bool "Support Marvell 88PM8607"
-	depends on I2C
-	select MFD_CORE
+	depends on I2C=y && MFD_CORE=y
 	help
 	  This supports for Marvell 88PM8607 Power Management IC. This includes
 	  the I2C driver and the core APIs _only_, you have to select
-- 
1.5.6.5

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 841 bytes --]

Hi Haojian,

On Mon, 5 Oct 2009 19:35:05 -0700 Haojian Zhuang <hzhuang1@marvell.com> wrote:
>
> Could you send your .config file to me? The 88pm8607 patch is just for mfd-2.6 git tree. I didn't try it in linux-next git tree before. I can try it now. If I could have your .config file, it would be better.

I am sorry, but I don't have the exact .config file any more (I have
moved on in my creation of linux-next)  however, as I said, this was an
allmodconfig build on x86_64.  I also noted that CONFIG_I2C=m and
CONFIG_MFD_CORE=m, while the config entry that controls the building of
this driver is a bool so must have been 'y'.  So from a builtin driver,
you are referring to symbols that are only defined in modules.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

RE: linux-next: mfd tree build failure

[Haojian Zhuang]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="gb2312", Size: 1872 bytes --]

Hi Stephen,

Could you send your .config file to me? The 88pm8607 patch is just for mfd-2.6 git tree. I didn't try it in linux-next git tree before. I can try it now. If I could have your .config file, it would be better.

Thanks
Haojian

-----Original Message-----
From: Stephen Rothwell [mailto:sfr@canb.auug.org.au] 
Sent: 2009Äê10ÔÂ6ÈÕ 10:02 AM
To: Samuel Ortiz
Cc: linux-next@vger.kernel.org; linux-kernel@vger.kernel.org; Haojian Zhuang; Mark Brown
Subject: linux-next: mfd tree build failure

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/built-in.o: In function `pm8607_write_device':
88pm8607.c:(.text+0xd02e9): undefined reference to `i2c_master_send'
drivers/built-in.o: In function `pm8607_read_device':
88pm8607.c:(.text+0xd033b): undefined reference to `i2c_master_send'
88pm8607.c:(.text+0xd035c): undefined reference to `i2c_master_recv'
drivers/built-in.o: In function `pm8607_init':
88pm8607.c:(.init.text+0xa9bd): undefined reference to `i2c_register_driver'
drivers/built-in.o: In function `pm8607_probe':
88pm8607.c:(.devinit.text+0x4db6): undefined reference to `mfd_add_devices'
drivers/built-in.o: In function `pm8607_exit':
88pm8607.c:(.exit.text+0x598): undefined reference to `i2c_del_driver'
drivers/built-in.o: In function `pm8607_remove':
88pm8607.c:(.devexit.text+0x198): undefined reference to `mfd_remove_devices'

Obviously caused by commit 9ec6cd8f2235ddeee5b39068956f14238a12a617
("mfd: Add 88PM8607 driver").

This build has CONFIG_I2C=m and CONFIG_MFD_CORE=m.

I have used the version of the mfd tree form next-20091005 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/
ÿôèº{.nÇ+‰·Ÿ®‰­†+%ŠËÿ±éݶ\x17¥Šwÿº{.nÇ+‰·¥Š{±þG«éÿŠ{ayº\x1dʇڙë,j\a­¢f£¢·hšïêÿ‘êçz_è®\x03(­éšŽŠÝ¢j"ú\x1a¶^[m§ÿÿ¾\a«þG«éÿ¢¸?™¨è­Ú&£ø§~á¶iO•æ¬z·švØ^\x14\x04\x1a¶^[m§ÿÿÃ\fÿ¶ìÿ¢¸?–I¥

RE: linux-next: mfd tree build failure

[Haojian Zhuang]

Hi Stephen,

Could you send your .config file to me? The 88pm8607 patch is just for mfd-2.6 git tree. I didn't try it in linux-next git tree before. I can try it now. If I could have your .config file, it would be better.

Thanks
Haojian

-----Original Message-----
From: Stephen Rothwell [mailto:sfr@canb.auug.org.au] 
Sent: 2009年10月6日 10:02 AM
To: Samuel Ortiz
Cc: linux-next@vger.kernel.org; linux-kernel@vger.kernel.org; Haojian Zhuang; Mark Brown
Subject: linux-next: mfd tree build failure

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/built-in.o: In function `pm8607_write_device':
88pm8607.c:(.text+0xd02e9): undefined reference to `i2c_master_send'
drivers/built-in.o: In function `pm8607_read_device':
88pm8607.c:(.text+0xd033b): undefined reference to `i2c_master_send'
88pm8607.c:(.text+0xd035c): undefined reference to `i2c_master_recv'
drivers/built-in.o: In function `pm8607_init':
88pm8607.c:(.init.text+0xa9bd): undefined reference to `i2c_register_driver'
drivers/built-in.o: In function `pm8607_probe':
88pm8607.c:(.devinit.text+0x4db6): undefined reference to `mfd_add_devices'
drivers/built-in.o: In function `pm8607_exit':
88pm8607.c:(.exit.text+0x598): undefined reference to `i2c_del_driver'
drivers/built-in.o: In function `pm8607_remove':
88pm8607.c:(.devexit.text+0x198): undefined reference to `mfd_remove_devices'

Obviously caused by commit 9ec6cd8f2235ddeee5b39068956f14238a12a617
("mfd: Add 88PM8607 driver").

This build has CONFIG_I2C=m and CONFIG_MFD_CORE=m.

I have used the version of the mfd tree form next-20091005 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1240 bytes --]

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/built-in.o: In function `pm8607_write_device':
88pm8607.c:(.text+0xd02e9): undefined reference to `i2c_master_send'
drivers/built-in.o: In function `pm8607_read_device':
88pm8607.c:(.text+0xd033b): undefined reference to `i2c_master_send'
88pm8607.c:(.text+0xd035c): undefined reference to `i2c_master_recv'
drivers/built-in.o: In function `pm8607_init':
88pm8607.c:(.init.text+0xa9bd): undefined reference to `i2c_register_driver'
drivers/built-in.o: In function `pm8607_probe':
88pm8607.c:(.devinit.text+0x4db6): undefined reference to `mfd_add_devices'
drivers/built-in.o: In function `pm8607_exit':
88pm8607.c:(.exit.text+0x598): undefined reference to `i2c_del_driver'
drivers/built-in.o: In function `pm8607_remove':
88pm8607.c:(.devexit.text+0x198): undefined reference to `mfd_remove_devices'

Obviously caused by commit 9ec6cd8f2235ddeee5b39068956f14238a12a617
("mfd: Add 88PM8607 driver").

This build has CONFIG_I2C=m and CONFIG_MFD_CORE=m.

I have used the version of the mfd tree form next-20091005 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

On Mon, Jan 12, 2009 at 03:37:55PM +0530, Balaji Rao wrote:
> On Mon, Jan 12, 2009 at 02:03:18PM +1100, Stephen Rothwell wrote:
> > Hi Samuel,
> > 
> > Today's linux-next build (x86_64 allmodconfig) failed like this:
> > 
> > ERROR: "__set_irq_handler" [drivers/mfd/pcf50633-core.ko] undefined!
> > ERROR: "handle_level_irq" [drivers/mfd/pcf50633-core.ko] undefined!
> > 
> > Immediate cause is clearly commit
> > f52046b14b1e1a8a02ae48d0c69d39c5e204644f ("mfd: PCF50633 core driver")
> > which introduces this code.  The above functions don't appear to be
> > exported to modules.
> > 
> > I have dropped the mfd tree for today.
> 
> Oh, sorry about this. I can actually do without this. Here's the patch.
> Sameo, can you please apply it ?
Applied to both my for-next and for-linus branches.
If Linus doesnt apply the patch, I'll resend a pull request tomorrow.

Cheers,
Samuel.
 
>     pcf50633: Remove references to non-exported functions
>     
>     Remove references to set_irq_type and handle_level_irq which are not
>     exported to modules.
>     
>     Signed-off-by: Balaji Rao <balajirrao@openmoko.org>
> 
> diff --git a/drivers/mfd/pcf50633-core.c b/drivers/mfd/pcf50633-core.c
> index 24508e2..ea9488e 100644
> --- a/drivers/mfd/pcf50633-core.c
> +++ b/drivers/mfd/pcf50633-core.c
> @@ -626,7 +626,6 @@ static int __devinit pcf50633_probe(struct i2c_client *client,
>  	}
>  
>  	if (client->irq) {
> -		set_irq_handler(client->irq, handle_level_irq);
>  		ret = request_irq(client->irq, pcf50633_irq,
>  				IRQF_TRIGGER_LOW, "pcf50633", pcf);
>  

-- 
Intel Open Source Technology Centre
http://oss.intel.com/

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

Hi Linus,

On Mon, 12 Jan 2009 15:37:55 +0530 Balaji Rao <balajirrao@openmoko.org> wrote:
>
> On Mon, Jan 12, 2009 at 02:03:18PM +1100, Stephen Rothwell wrote:
> > 
> > Today's linux-next build (x86_64 allmodconfig) failed like this:
> > 
> > ERROR: "__set_irq_handler" [drivers/mfd/pcf50633-core.ko] undefined!
> > ERROR: "handle_level_irq" [drivers/mfd/pcf50633-core.ko] undefined!
> > 
> > Immediate cause is clearly commit
> > f52046b14b1e1a8a02ae48d0c69d39c5e204644f ("mfd: PCF50633 core driver")
> > which introduces this code.  The above functions don't appear to be
> > exported to modules.
> > 
> > I have dropped the mfd tree for today.
> 
> Oh, sorry about this. I can actually do without this. Here's the patch.
> Sameo, can you please apply it ?

This is now in the upstream kernel tree ...

---------------------------------------------------------------------------
From: Balaji Rao <balajirrao@openmoko.org>

    pcf50633: Remove references to non-exported functions
    
    Remove references to set_irq_type and handle_level_irq which are not
    exported to modules.
    
    Signed-off-by: Balaji Rao <balajirrao@openmoko.org>

diff --git a/drivers/mfd/pcf50633-core.c b/drivers/mfd/pcf50633-core.c
index 24508e2..ea9488e 100644
--- a/drivers/mfd/pcf50633-core.c
+++ b/drivers/mfd/pcf50633-core.c
@@ -626,7 +626,6 @@ static int __devinit pcf50633_probe(struct i2c_client *client,
 	}
 
 	if (client->irq) {
-		set_irq_handler(client->irq, handle_level_irq);
 		ret = request_irq(client->irq, pcf50633_irq,
 				IRQF_TRIGGER_LOW, "pcf50633", pcf);
 

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

Re: linux-next: mfd tree build failure

[Balaji Rao]

On Mon, Jan 12, 2009 at 02:03:18PM +1100, Stephen Rothwell wrote:
> Hi Samuel,
> 
> Today's linux-next build (x86_64 allmodconfig) failed like this:
> 
> ERROR: "__set_irq_handler" [drivers/mfd/pcf50633-core.ko] undefined!
> ERROR: "handle_level_irq" [drivers/mfd/pcf50633-core.ko] undefined!
> 
> Immediate cause is clearly commit
> f52046b14b1e1a8a02ae48d0c69d39c5e204644f ("mfd: PCF50633 core driver")
> which introduces this code.  The above functions don't appear to be
> exported to modules.
> 
> I have dropped the mfd tree for today.

Oh, sorry about this. I can actually do without this. Here's the patch.
Sameo, can you please apply it ?

    pcf50633: Remove references to non-exported functions
    
    Remove references to set_irq_type and handle_level_irq which are not
    exported to modules.
    
    Signed-off-by: Balaji Rao <balajirrao@openmoko.org>

diff --git a/drivers/mfd/pcf50633-core.c b/drivers/mfd/pcf50633-core.c
index 24508e2..ea9488e 100644
--- a/drivers/mfd/pcf50633-core.c
+++ b/drivers/mfd/pcf50633-core.c
@@ -626,7 +626,6 @@ static int __devinit pcf50633_probe(struct i2c_client *client,
 	}
 
 	if (client->irq) {
-		set_irq_handler(client->irq, handle_level_irq);
 		ret = request_irq(client->irq, pcf50633_irq,
 				IRQF_TRIGGER_LOW, "pcf50633", pcf);
 

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 570 bytes --]

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

ERROR: "__set_irq_handler" [drivers/mfd/pcf50633-core.ko] undefined!
ERROR: "handle_level_irq" [drivers/mfd/pcf50633-core.ko] undefined!

Immediate cause is clearly commit
f52046b14b1e1a8a02ae48d0c69d39c5e204644f ("mfd: PCF50633 core driver")
which introduces this code.  The above functions don't appear to be
exported to modules.

I have dropped the mfd tree for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

On Tue, Aug 05, 2008 at 07:46:18PM +0100, ian wrote:
> On Tue, 2008-08-05 at 19:40 +0200, Samuel Ortiz wrote:
> > 
> > All TMIO devices depend on ARM, and the MMC and NAND drivers now
> > depend on a new symbol (MFD_TMIO) selected when choosing one of the
> > TMIO devices.
> 
> Sounds good.
> 
> Whens the merge date for mainline btw ?
Later this week, hopefully.

Cheers,
Samuel.
 

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 667 bytes --]

HI Samuel,

On Tue, 5 Aug 2008 19:40:48 +0200 Samuel Ortiz <sameo@openedhand.com> wrote:
>
> On Wed, Aug 06, 2008 at 02:16:14AM +1000, Stephen Rothwell wrote:
> > On Tue, 5 Aug 2008 18:08:42 +0200 Samuel Ortiz <sameo@openedhand.com> wrote:
> > >
> > > :-) I'll take care of it, yep.
> > 
> > Thanks, I look forward to it.
> It should be fixed by now.
> All TMIO devices depend on ARM, and the MMC and NAND drivers now depend
> on a new symbol (MFD_TMIO) selected when choosing one of the TMIO devices.

OK, thanks.  What I fetched looks good.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mfd tree build failure

[ian]

On Tue, 2008-08-05 at 19:05 +0300, Adrian Bunk wrote:
> 
> The "Multifunction device drivers" menu depends on HAS_IOMEM.

Actually, this is wrong, it shouldnt.

Re: linux-next: mfd tree build failure

[ian]

On Tue, 2008-08-05 at 19:40 +0200, Samuel Ortiz wrote:
> 
> All TMIO devices depend on ARM, and the MMC and NAND drivers now
> depend on a new symbol (MFD_TMIO) selected when choosing one of the
> TMIO devices.

Sounds good.

Whens the merge date for mainline btw ?

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

Hi Stephen,

On Wed, Aug 06, 2008 at 02:16:14AM +1000, Stephen Rothwell wrote:
> On Tue, 5 Aug 2008 18:08:42 +0200 Samuel Ortiz <sameo@openedhand.com> wrote:
> >
> > :-) I'll take care of it, yep.
> 
> Thanks, I look forward to it.
It should be fixed by now.
All TMIO devices depend on ARM, and the MMC and NAND drivers now depend
on a new symbol (MFD_TMIO) selected when choosing one of the TMIO devices.

Cheers,
Samuel.

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 271 bytes --]

Hi Samuel,

On Tue, 5 Aug 2008 18:08:42 +0200 Samuel Ortiz <sameo@openedhand.com> wrote:
>
> :-) I'll take care of it, yep.

Thanks, I look forward to it.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

On Wed, Aug 06, 2008 at 02:05:40AM +1000, Stephen Rothwell wrote:
> Hi Ian,
> 
> On Tue, 05 Aug 2008 16:39:54 +0100 ian <spyro@f2s.com> wrote:
> >
> > I'd have expected the other TMIO drivers to fail too.
> > 
> > get_irq_data is defined in linux/irq.h only if 
> > 
> > CONFIG_GENERIC_HARDIRQs is set and you arent on S390 (so why doesnt S390
> > fail?)
> > 
> > I dont see this as a big problem. As far as Im concerned, all the TMIO
> > drivers can depend on CONFIG_ARM.
> > 
> > It may not be a bad idea to add CONFIG_MFD_TMIO_SUBDEVICE so that TMIO
> > sub device drivers can only be selected when TMIO cores are selected.
> > This would neatly fix the earlier problem, rather than making al lthe
> > subdevices CONFIG_ARM.
> 
> It'd be real nice if someone did a patch to that effect :-)
:-) I'll take care of it, yep.

Cheers,
Samuel.

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 816 bytes --]

Hi Ian,

On Tue, 05 Aug 2008 16:39:54 +0100 ian <spyro@f2s.com> wrote:
>
> I'd have expected the other TMIO drivers to fail too.
> 
> get_irq_data is defined in linux/irq.h only if 
> 
> CONFIG_GENERIC_HARDIRQs is set and you arent on S390 (so why doesnt S390
> fail?)
> 
> I dont see this as a big problem. As far as Im concerned, all the TMIO
> drivers can depend on CONFIG_ARM.
> 
> It may not be a bad idea to add CONFIG_MFD_TMIO_SUBDEVICE so that TMIO
> sub device drivers can only be selected when TMIO cores are selected.
> This would neatly fix the earlier problem, rather than making al lthe
> subdevices CONFIG_ARM.

It'd be real nice if someone did a patch to that effect :-)

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mfd tree build failure

[Adrian Bunk]

On Tue, Aug 05, 2008 at 04:39:54PM +0100, ian wrote:
> On Wed, 2008-08-06 at 00:08 +1000, Stephen Rothwell wrote:
> > Hi Samuel,
> > 
> > Today's linux-next build (sparc allmodconfig) failed like this:
> > 
> > drivers/mfd/t7l66xb.c:167: error: implicit declaration of function
> > `get_irq_data'
> 
> Hi.
> 
> I'd have expected the other TMIO drivers to fail too.
> 
> get_irq_data is defined in linux/irq.h only if 
> 
> CONFIG_GENERIC_HARDIRQs is set and you arent on S390 (so why doesnt S390
> fail?)
>...

The "Multifunction device drivers" menu depends on HAS_IOMEM.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: linux-next: mfd tree build failure

[ian]

On Wed, 2008-08-06 at 00:08 +1000, Stephen Rothwell wrote:
> Hi Samuel,
> 
> Today's linux-next build (sparc allmodconfig) failed like this:
> 
> drivers/mfd/t7l66xb.c:167: error: implicit declaration of function
> `get_irq_data'

Hi.

I'd have expected the other TMIO drivers to fail too.

get_irq_data is defined in linux/irq.h only if 

CONFIG_GENERIC_HARDIRQs is set and you arent on S390 (so why doesnt S390
fail?)

I dont see this as a big problem. As far as Im concerned, all the TMIO
drivers can depend on CONFIG_ARM.

It may not be a bad idea to add CONFIG_MFD_TMIO_SUBDEVICE so that TMIO
sub device drivers can only be selected when TMIO cores are selected.
This would neatly fix the earlier problem, rather than making al lthe
subdevices CONFIG_ARM.

Re: linux-next: mfd tree build failure

[ian]

On Tue, 2008-08-05 at 15:09 +0200, Samuel Ortiz wrote:
> 
> > Should I redo the two drivers patchsets, or would you prefer a patch
> > fixing this after the fact?
> No worries, I'll fix the Kconfigs later today.

Awesome, thanks!

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1264 bytes --]

Hi Samuel,

Today's linux-next build (sparc allmodconfig) failed like this:

drivers/mfd/t7l66xb.c:167: error: implicit declaration of function `get_irq_data'
drivers/mfd/t7l66xb.c:177: error: implicit declaration of function `generic_handle_irq'
drivers/mfd/t7l66xb.c:182: error: implicit declaration of function `get_irq_chip_data'
drivers/mfd/t7l66xb.c:224: error: implicit declaration of function `set_irq_chip'
drivers/mfd/t7l66xb.c:225: error: implicit declaration of function `set_irq_chip_data'
drivers/mfd/t7l66xb.c:226: error: implicit declaration of function `set_irq_handler'
drivers/mfd/t7l66xb.c:226: error: `handle_level_irq' undeclared (first use in this function)
drivers/mfd/t7l66xb.c:226: error: (Each undeclared identifier is reported only once
drivers/mfd/t7l66xb.c:226: error: for each function it appears in.)
drivers/mfd/t7l66xb.c:232: error: implicit declaration of function `set_irq_type'
drivers/mfd/t7l66xb.c:233: error: implicit declaration of function `set_irq_data'
drivers/mfd/t7l66xb.c:234: error: implicit declaration of function `set_irq_chained_handler'

This has actually been happening since Aug 1.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 270 bytes --]

Hi Samuel, Ian,

On Tue, 5 Aug 2008 15:09:00 +0200 Samuel Ortiz <sameo@openedhand.com> wrote:
>
> No worries, I'll fix the Kconfigs later today.

Thanks.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mfd tree build failure

[Samuel Ortiz]

Hi Ian,

On Tue, Aug 05, 2008 at 02:03:14PM +0100, ian wrote:
> On Tue, 2008-08-05 at 15:09 +1000, Stephen Rothwell wrote:
> > 
> > I reverted commit 4d6da2de1b09228989efb67a1f9e6d2ced075540 ("mfd:
> > driver for the TMIO NAND controller") until it has tighter controls in
> > its Kconfig patch. i.e. I have no idea what arches this should be
> > built on.
> 
> Since there is no way to know wether an arch has readsw() and friends at
> config time, I'd propose that bot the TMIO NAND and MMC drivers be built
> only on ARM for now. If anyone every uses them on another platform, they
> can add the support.
> 
> Should I redo the two drivers patchsets, or would you prefer a patch
> fixing this after the fact?
No worries, I'll fix the Kconfigs later today.

Cheers,
Samuel.

Re: linux-next: mfd tree build failure

[ian]

On Tue, 2008-08-05 at 15:09 +1000, Stephen Rothwell wrote:
> 
> I reverted commit 4d6da2de1b09228989efb67a1f9e6d2ced075540 ("mfd:
> driver for the TMIO NAND controller") until it has tighter controls in
> its Kconfig patch. i.e. I have no idea what arches this should be
> built on.

Since there is no way to know wether an arch has readsw() and friends at
config time, I'd propose that bot the TMIO NAND and MMC drivers be built
only on ARM for now. If anyone every uses them on another platform, they
can add the support.

Should I redo the two drivers patchsets, or would you prefer a patch
fixing this after the fact?

linux-next: mfd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 763 bytes --]

Hi Samuel,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/mtd/nand/tmio_nand.c: In function 'tmio_nand_write_buf':
drivers/mtd/nand/tmio_nand.c:251: error: implicit declaration of function 'writesw'
drivers/mtd/nand/tmio_nand.c: In function 'tmio_nand_read_buf':
drivers/mtd/nand/tmio_nand.c:258: error: implicit declaration of function 'readsw'

Most architectures don't implement readsw/writesw.

I reverted commit 4d6da2de1b09228989efb67a1f9e6d2ced075540 ("mfd: driver
for the TMIO NAND controller") until it has tighter controls in its
Kconfig patch. i.e. I have no idea what arches this should be built on.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]