linux-next: build failure after merge of the voltage tree

DISCUSSION Randy Dunlap 2 participants 3 messages May 27, 2011 1 message not in the archive Multifunction devices (MFD)

lore

Share & tools

lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Randy Dunlap

author

2011-05-27 19:58 UTC #1 lore raw

Quoted 77 lines · On Fri, 27 May 2011 17:59:36 GMT Linux Kernel Mailing List wrote:

Gitweb:     http://git.kernel.org/linus/c01e36dde4632e0a7474fddf0716f1e54f01f13e
Commit:     c01e36dde4632e0a7474fddf0716f1e54f01f13e
Parent:     327e15af15248563c896c16adad6fc6e04bb5a4d
Author:     Jorge Eduardo Candelaria <[email protected]>
AuthorDate: Wed May 11 01:43:02 2011 -0500
Committer:  Liam Girdwood <[email protected]>
CommitDate: Fri May 27 10:49:09 2011 +0100

    linux-next: build failure after merge of the voltage tree
    
    On May 10, 2011, at 9:27 PM, Stephen Rothwell wrote:
    
    > Hi Jorge,
    >
    > On Tue, 10 May 2011 12:30:36 -0500 Jorge Eduardo Candelaria <[email protected]> wrote:
    >>
    >> On May 10, 2011, at 3:38 AM, Liam Girdwood wrote:
    >>
    >>> On Tue, 2011-05-10 at 12:44 +1000, Stephen Rothwell wrote:
    >>>> Hi Liam,
    >>>>
    >>>> After merging the voltage tree, today's linux-next build (x86_64
    >>>> allmodconfig) failed like this:
    >>>>
    >>>> ERROR: "tps65910_gpio_init" [drivers/mfd/tps65910.ko] undefined!
    >>>> ERROR: "tps65910_irq_init" [drivers/mfd/tps65910.ko] undefined!
    >>>> ERROR: "irq_modify_status" [drivers/mfd/tps65910-irq.ko] undefined!
    >>>> ERROR: "irq_set_chip_and_handler_name" [drivers/mfd/tps65910-irq.ko] undefined!
    >>>> ERROR: "handle_edge_irq" [drivers/mfd/tps65910-irq.ko] undefined!
    >>>>
    >>>> I have used the voltage tree from next-20110509 for today.
    >>>
    >>> Jorge, could you send a fix for this today.
    >>
    >> The following patch should solve this:
    >>
    >> From: Jorge Eduardo Candelaria <[email protected]>
    >> MFD: Fix TPS65910 build
    >>
    >> Support for tps65910 as a module is not available. The driver can
    >> only be compiled as built-in. OTOH, the regulator driver can still
    >> be built as module without breaking the compilation.
    >>
    >> Signed-off-by: Jorge Eduardo Candelaria <[email protected]>
    >
    > Today (even with the above patch included) I got these errors from the
    > x86_64 allmodconfig build:
    >
    > tps65910.c:(.text+0xf4140): undefined reference to `i2c_master_send'
    > drivers/built-in.o: In function `tps65910_i2c_read':
    > tps65910.c:(.text+0xf41d2): undefined reference to `i2c_transfer'
    > drivers/built-in.o: In function `tps65910_i2c_init':
    > tps65910.c:(.init.text+0xcb83): undefined reference to `i2c_register_driver'
    > drivers/built-in.o: In function `tps65910_i2c_exit':
    > tps65910.c:(.exit.text+0x6e0): undefined reference to `i2c_del_driver'
    >
    > I have used the voltage tree from next-20110509 again today.
    
    Following patch should fix the dependency problems. Please review:
    
    From: Jorge Eduardo Candelaria <[email protected]>
    [PATCH] MFD: TPS65910: Fix I2C dependency
    
    TPS65910 driver can only be compiled built-in, so the I2C driver
    should be as well.
    
    Signed-off-by: Jorge Eduardo Candelaria <[email protected]>
    Signed-off-by: Liam Girdwood <[email protected]>
---
 drivers/mfd/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

drivers/mfd/Kconfig +0 -0

index 05f882f..4990a49 100644

drivers/mfd/Kconfig config MFD_PM8XXX_IRQ

Quoted 9 lines

drivers/mfd/Kconfig +1 -2

@@ -721,7 +721,7 @@ config MFD_PM8XXX_IRQ
 
 config MFD_TPS65910
 	bool "TPS65910 Power Management chip"
-	depends on I2C && GPIOLIB
+	depends on I2C=y && GPIOLIB
 	select MFD_CORE
 	help
 	  if you say yes here you get support for the TPS65910 series of
--

Hi,

Current mainline build fails when GPIO is not enabled:

warning: (MFD_TPS65910) selects GPIO_TPS65910 which has unmet direct dependencies (GPIOLIB && MFD_TPS65910)

which causes:

In file included from drivers/gpio/tps65910-gpio.c:21:
include/linux/mfd/tps65910.h:774: error: field 'gpio' has incomplete type drivers/gpio/tps65910-gpio.c: In function 'tps65910_gpio_get':
drivers/gpio/tps65910-gpio.c:25: warning: type defaults to 'int' in declaration of '__mptr' drivers/gpio/tps65910-gpio.c:25: warning: initialization from incompatible pointer type drivers/gpio/tps65910-gpio.c: In function 'tps65910_gpio_set':
drivers/gpio/tps65910-gpio.c:39: warning: type defaults to 'int' in declaration of '__mptr' drivers/gpio/tps65910-gpio.c:39: warning: initialization from incompatible pointer type drivers/gpio/tps65910-gpio.c: In function 'tps65910_gpio_output':
drivers/gpio/tps65910-gpio.c:52: warning: type defaults to 'int' in declaration of '__mptr' drivers/gpio/tps65910-gpio.c:52: warning: initialization from incompatible pointer type drivers/gpio/tps65910-gpio.c: In function 'tps65910_gpio_input':
drivers/gpio/tps65910-gpio.c:63: warning: type defaults to 'int' in declaration of '__mptr' drivers/gpio/tps65910-gpio.c:63: warning: initialization from incompatible pointer type drivers/gpio/tps65910-gpio.c: In function 'tps65910_gpio_init':
drivers/gpio/tps65910-gpio.c:96: error: implicit declaration of function 'gpiochip_add' make[3]: *** [drivers/gpio/tps65910-gpio.o] Error 1
make[2]: *** [drivers/gpio] Error 2
make[1]: *** [drivers] Error 2

---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***

Liam Girdwood

2011-05-27 21:13 UTC #2 lore raw

Quoted 109 lines · Randy Dunlap

Quoted 77 lines · On Fri, 27 May 2011 17:59:36 GMT Linux Kernel Mailing List wrote:

Gitweb:     http://git.kernel.org/linus/c01e36dde4632e0a7474fddf0716f1e54f01f13e
Commit:     c01e36dde4632e0a7474fddf0716f1e54f01f13e
Parent:     327e15af15248563c896c16adad6fc6e04bb5a4d
Author:     Jorge Eduardo Candelaria <[email protected]>
AuthorDate: Wed May 11 01:43:02 2011 -0500
Committer:  Liam Girdwood <[email protected]>
CommitDate: Fri May 27 10:49:09 2011 +0100

    linux-next: build failure after merge of the voltage tree
    
    On May 10, 2011, at 9:27 PM, Stephen Rothwell wrote:
    
    > Hi Jorge,
    >
    > On Tue, 10 May 2011 12:30:36 -0500 Jorge Eduardo Candelaria <[email protected]> wrote:
    >>
    >> On May 10, 2011, at 3:38 AM, Liam Girdwood wrote:
    >>
    >>> On Tue, 2011-05-10 at 12:44 +1000, Stephen Rothwell wrote:
    >>>> Hi Liam,
    >>>>
    >>>> After merging the voltage tree, today's linux-next build (x86_64
    >>>> allmodconfig) failed like this:
    >>>>
    >>>> ERROR: "tps65910_gpio_init" [drivers/mfd/tps65910.ko] undefined!
    >>>> ERROR: "tps65910_irq_init" [drivers/mfd/tps65910.ko] undefined!
    >>>> ERROR: "irq_modify_status" [drivers/mfd/tps65910-irq.ko] undefined!
    >>>> ERROR: "irq_set_chip_and_handler_name" [drivers/mfd/tps65910-irq.ko] undefined!
    >>>> ERROR: "handle_edge_irq" [drivers/mfd/tps65910-irq.ko] undefined!
    >>>>
    >>>> I have used the voltage tree from next-20110509 for today.
    >>>
    >>> Jorge, could you send a fix for this today.
    >>
    >> The following patch should solve this:
    >>
    >> From: Jorge Eduardo Candelaria <[email protected]>
    >> MFD: Fix TPS65910 build
    >>
    >> Support for tps65910 as a module is not available. The driver can
    >> only be compiled as built-in. OTOH, the regulator driver can still
    >> be built as module without breaking the compilation.
    >>
    >> Signed-off-by: Jorge Eduardo Candelaria <[email protected]>
    >
    > Today (even with the above patch included) I got these errors from the
    > x86_64 allmodconfig build:
    >
    > tps65910.c:(.text+0xf4140): undefined reference to `i2c_master_send'
    > drivers/built-in.o: In function `tps65910_i2c_read':
    > tps65910.c:(.text+0xf41d2): undefined reference to `i2c_transfer'
    > drivers/built-in.o: In function `tps65910_i2c_init':
    > tps65910.c:(.init.text+0xcb83): undefined reference to `i2c_register_driver'
    > drivers/built-in.o: In function `tps65910_i2c_exit':
    > tps65910.c:(.exit.text+0x6e0): undefined reference to `i2c_del_driver'
    >
    > I have used the voltage tree from next-20110509 again today.
    
    Following patch should fix the dependency problems. Please review:
    
    From: Jorge Eduardo Candelaria <[email protected]>
    [PATCH] MFD: TPS65910: Fix I2C dependency
    
    TPS65910 driver can only be compiled built-in, so the I2C driver
    should be as well.
    
    Signed-off-by: Jorge Eduardo Candelaria <[email protected]>
    Signed-off-by: Liam Girdwood <[email protected]>
---
 drivers/mfd/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

drivers/mfd/Kconfig +0 -0

index 05f882f..4990a49 100644

drivers/mfd/Kconfig config MFD_PM8XXX_IRQ

Quoted 9 lines

drivers/mfd/Kconfig +1 -2

@@ -721,7 +721,7 @@ config MFD_PM8XXX_IRQ
 
 config MFD_TPS65910
 	bool "TPS65910 Power Management chip"
-	depends on I2C && GPIOLIB
+	depends on I2C=y && GPIOLIB
 	select MFD_CORE
 	help
 	  if you say yes here you get support for the TPS65910 series of
--

Hi,

Current mainline build fails when GPIO is not enabled:

warning: (MFD_TPS65910) selects GPIO_TPS65910 which has unmet direct dependencies (GPIOLIB && MFD_TPS65910)

which causes:

Quoted 10 lines

drivers/gpio/tps65910-gpio.c: In function 'tps65910_gpio_input':
drivers/gpio/tps65910-gpio.c:63: warning: type defaults to 'int' in declaration of '__mptr' drivers/gpio/tps65910-gpio.c:63: warning: initialization from incompatible pointer type drivers/gpio/tps65910-gpio.c: In function 'tps65910_gpio_init':
drivers/gpio/tps65910-gpio.c:96: error: implicit declaration of function 'gpiochip_add' make[3]: *** [drivers/gpio/tps65910-gpio.o] Error 1
make[2]: *** [drivers/gpio] Error 2
make[1]: *** [drivers] Error 2

Hi,

gah, looks this was broken by moving the tps65910 gpio directory move.

This should fix it though:-

From: Liam Girdwood <[email protected]>
Date: Fri, 27 May 2011 22:06:52 +0100
Subject: [PATCH] mfd: tps65910 - fix build breakage caused by tps65910 gpio directory move.

Signed-off-by Liam Girdwood <[email protected]>

---
 drivers/mfd/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

drivers/mfd/Kconfig +1 -1

index b6c2677..0f09c05 100644
@@ -721,7 +721,7 @@ config MFD_PM8XXX_IRQ
 
 config MFD_TPS65910
 	bool "TPS65910 Power Management chip"
-	depends on I2C=y
+	depends on I2C=y && GPIOLIB
 	select MFD_CORE
 	select GPIO_TPS65910
 	help

Signature · 1 line

1.7.4.1

Randy Dunlap

author Acked-by

2011-05-27 21:22 UTC #3 lore raw

Quoted 5 lines

Hi,
gah, looks this was broken by moving the tps65910 gpio directory move.
This should fix it though:-

It does. Thanks.

Acked-by Randy Dunlap <[email protected]>

Quoted 22 lines

From: Liam Girdwood <[email protected]>
Date: Fri, 27 May 2011 22:06:52 +0100
Subject: [PATCH] mfd: tps65910 - fix build breakage caused by tps65910 gpio directory move.
Signed-off-by Liam Girdwood <[email protected]>
---
 drivers/mfd/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/mfd/Kconfig +1 -1
index b6c2677..0f09c05 100644
@@ -721,7 +721,7 @@ config MFD_PM8XXX_IRQ
 
 config MFD_TPS65910
 	bool "TPS65910 Power Management chip"
-	depends on I2C=y
+	depends on I2C=y && GPIOLIB
 	select MFD_CORE
 	select GPIO_TPS65910
 	help

Signature · 2 lines

~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***