Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1332153 > unrolled thread
| Started by | Arnd Bergmann <arnd@arndb.de> |
|---|---|
| First post | 2016-02-11 16:40 +0100 |
| Last post | 2016-02-14 10:10 +0100 |
| Articles | 2 — 2 participants |
Back to article view | Back to linux.kernel
[PATCH] ARM: exynos: clarify KEYBOARD_SAMSUNG selection Arnd Bergmann <arnd@arndb.de> - 2016-02-11 16:40 +0100
Re: [PATCH] ARM: exynos: clarify KEYBOARD_SAMSUNG selection Krzysztof Kozlowski <k.kozlowski@samsung.com> - 2016-02-14 10:10 +0100
| From | Arnd Bergmann <arnd@arndb.de> |
|---|---|
| Date | 2016-02-11 16:40 +0100 |
| Subject | [PATCH] ARM: exynos: clarify KEYBOARD_SAMSUNG selection |
| Message-ID | <r11vz-2gb-5@gated-at.bofh.it> |
The samsung-keypad driver is implicitly selected by ARCH_EXYNOS4 (why?), but this fails if CONFIG_INPUT is a loadable module: drivers/input/built-in.o: In function `samsung_keypad_remove': drivers/input/keyboard/samsung-keypad.c:461: undefined reference to `input_unregister_device' drivers/input/built-in.o: In function `samsung_keypad_irq': drivers/input/keyboard/samsung-keypad.c:137: undefined reference to `input_event' drivers/input/built-in.o: In function `samsung_keypad_irq': include/linux/input.h:389: undefined reference to `input_event' drivers/input/built-in.o: In function `samsung_keypad_probe': drivers/input/keyboard/samsung-keypad.c:358: undefined reference to `devm_input_allocate_device' drivers/input/built-in.o:(.debug_addr+0x34): undefined reference to `input_set_capability' This changes the 'select' statement so we don't do it if CONFIG_INPUT=m. The problem does not happen on mainline kernels, as we don't normally build built-in input drivers when CONFIG_INPUT=m, but I am experimenting with a patch to change this, and the samsung keypad driver showed up as one example that was silently broken before. Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- arch/arm/mach-exynos/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm/mach-exynos/Kconfig b/arch/arm/mach-exynos/Kconfig index 8434a0f6334c..b63e64581c24 100644 --- a/arch/arm/mach-exynos/Kconfig +++ b/arch/arm/mach-exynos/Kconfig @@ -59,7 +59,7 @@ config ARCH_EXYNOS4 select CLKSRC_SAMSUNG_PWM if CPU_EXYNOS4210 select CPU_EXYNOS4210 select GIC_NON_BANKED - select KEYBOARD_SAMSUNG if INPUT_KEYBOARD + select KEYBOARD_SAMSUNG if INPUT=y && INPUT_KEYBOARD select MIGHT_HAVE_CACHE_L2X0 help Samsung EXYNOS4 (Cortex-A9) SoC based systems -- 2.7.0
[toc] | [next] | [standalone]
| From | Krzysztof Kozlowski <k.kozlowski@samsung.com> |
|---|---|
| Date | 2016-02-14 10:10 +0100 |
| Message-ID | <r20QO-FA-13@gated-at.bofh.it> |
| In reply to | #1332153 |
W dniu 12.02.2016 o 00:30, Arnd Bergmann pisze:
> The samsung-keypad driver is implicitly selected by ARCH_EXYNOS4 (why?),
> but this fails if CONFIG_INPUT is a loadable module:
How about removing the select entirely and adding it in defconfigs? It
was introduced by 49b999711ee7 ("ARM: EXYNOS: change HAVE_SAMSUNG_KEYPAD
to KEYBOARD_SAMSUNG") which looks like a mistake. The intention was to
indicate a HAVE, not to select a driver.
Moreover the Exynos3250 also has keypad but it is not selected.
Can you send a patch removing select and changing exynos+multi_v7
defconfigs?
Best regards,
Krzysztof
>
> drivers/input/built-in.o: In function `samsung_keypad_remove':
> drivers/input/keyboard/samsung-keypad.c:461: undefined reference to `input_unregister_device'
> drivers/input/built-in.o: In function `samsung_keypad_irq':
> drivers/input/keyboard/samsung-keypad.c:137: undefined reference to `input_event'
> drivers/input/built-in.o: In function `samsung_keypad_irq':
> include/linux/input.h:389: undefined reference to `input_event'
> drivers/input/built-in.o: In function `samsung_keypad_probe':
> drivers/input/keyboard/samsung-keypad.c:358: undefined reference to `devm_input_allocate_device'
> drivers/input/built-in.o:(.debug_addr+0x34): undefined reference to `input_set_capability'
>
> This changes the 'select' statement so we don't do it if CONFIG_INPUT=m.
> The problem does not happen on mainline kernels, as we don't normally
> build built-in input drivers when CONFIG_INPUT=m, but I am experimenting
> with a patch to change this, and the samsung keypad driver showed up
> as one example that was silently broken before.
>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> arch/arm/mach-exynos/Kconfig | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/arm/mach-exynos/Kconfig b/arch/arm/mach-exynos/Kconfig
> index 8434a0f6334c..b63e64581c24 100644
> --- a/arch/arm/mach-exynos/Kconfig
> +++ b/arch/arm/mach-exynos/Kconfig
> @@ -59,7 +59,7 @@ config ARCH_EXYNOS4
> select CLKSRC_SAMSUNG_PWM if CPU_EXYNOS4210
> select CPU_EXYNOS4210
> select GIC_NON_BANKED
> - select KEYBOARD_SAMSUNG if INPUT_KEYBOARD
> + select KEYBOARD_SAMSUNG if INPUT=y && INPUT_KEYBOARD
> select MIGHT_HAVE_CACHE_L2X0
> help
> Samsung EXYNOS4 (Cortex-A9) SoC based systems
>
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web