The powerpc ppc6xx_defconfig builds failed on today's Linux next tag next-20240304.
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
Build log: --------- drivers/video/fbdev/chipsfb.c: In function 'chipsfb_pci_init': drivers/video/fbdev/chipsfb.c:401:31: error: invalid use of undefined type 'struct backlight_device' 401 | pmac_backlight->props.power = FB_BLANK_UNBLANK; | ^~
Steps to reproduce: # tuxmake --runtime podman --target-arch powerpc --toolchain gcc-13 --kconfig ppc6xx_defconfig - https://storage.tuxsuite.com/public/linaro/lkft/builds/2dDE297yLFrAr3gigIDy8...
Links: - https://qa-reports.linaro.org/lkft/linux-next-master/build/next-20240304/tes... - https://qa-reports.linaro.org/lkft/linux-next-master/build/next-20240304/tes... - https://qa-reports.linaro.org/lkft/linux-next-master/build/next-20240304/tes...
-- Linaro LKFT https://lkft.linaro.org
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
---
Not even compile tested! --- arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__
+#include <linux/backlight.h> #include <linux/fb.h> #include <linux/mutex.h>
On Mon, 04 Mar 2024, Jani Nikula jani.nikula@intel.com wrote:
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
Not even compile tested!
Naresh, please try this patch!
Michael, if this is fine by you, ack to merge via the drm subsystem along with the regressing commit?
BR, Jani.
arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__ +#include <linux/backlight.h> #include <linux/fb.h> #include <linux/mutex.h>
Hi Jani and Benjamin,
On Mon, 4 Mar 2024 at 15:31, Jani Nikula jani.nikula@intel.com wrote:
On Mon, 04 Mar 2024, Jani Nikula jani.nikula@intel.com wrote:
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
Not even compile tested!
Naresh, please try this patch!
Thanks for the proposed fix patch.
Steps to reproduce:
# tuxmake --runtime podman --target-arch powerpc --toolchain gcc-13 --kconfig ppc6xx_defconfig --kconfig-add CONFIG_PMAC_BACKLIGHT=y config debugkernel dtbs kernel modules xipkernel
# Applying patch set Applying: fbdev/chipsfb: Include <linux/backlight.h>
The reported build regression is fixed but build failed with below errors.
My two cents,
I should have copied the full build error log in the morning.
Few more build errors on powerpc builds, ------------------ drivers/macintosh/via-pmu-backlight.c: In function '__pmu_backlight_update_status': drivers/macintosh/via-pmu-backlight.c:74:21: error: implicit declaration of function 'backlight_get_brightness'; did you mean 'pmu_backlight_get_level_brightness'? [-Werror=implicit-function-declaration] 74 | int level = backlight_get_brightness(bd); | ^~~~~~~~~~~~~~~~~~~~~~~~ | pmu_backlight_get_level_brightness drivers/macintosh/via-pmu-backlight.c: At top level: drivers/macintosh/via-pmu-backlight.c:108:21: error: variable 'pmu_backlight_data' has initializer but incomplete type 108 | static const struct backlight_ops pmu_backlight_data = { | ^~~~~~~~~~~~~ drivers/macintosh/via-pmu-backlight.c:109:10: error: 'const struct backlight_ops' has no member named 'update_status' 109 | .update_status = pmu_backlight_update_status, | ^~~~~~~~~~~~~ drivers/macintosh/via-pmu-backlight.c:109:27: warning: excess elements in struct initializer 109 | .update_status = pmu_backlight_update_status, | ^~~~~~~~~~~~~~~~~~~~~~~~~~~ drivers/macintosh/via-pmu-backlight.c:109:27: note: (near initialization for 'pmu_backlight_data') drivers/macintosh/via-pmu-backlight.c: In function 'pmu_backlight_init': drivers/macintosh/via-pmu-backlight.c:136:37: error: storage size of 'props' isn't known 136 | struct backlight_properties props; | ^~~~~ drivers/macintosh/via-pmu-backlight.c:154:34: error: invalid application of 'sizeof' to incomplete type 'struct backlight_properties' 154 | memset(&props, 0, sizeof(struct backlight_properties)); | ^~~~~~ drivers/macintosh/via-pmu-backlight.c:155:22: error: 'BACKLIGHT_PLATFORM' undeclared (first use in this function) 155 | props.type = BACKLIGHT_PLATFORM; | ^~~~~~~~~~~~~~~~~~ drivers/macintosh/via-pmu-backlight.c:155:22: note: each undeclared identifier is reported only once for each function it appears in drivers/macintosh/via-pmu-backlight.c:157:14: error: implicit declaration of function 'backlight_device_register'; did you mean 'root_device_register'? [-Werror=implicit-function-declaration] 157 | bd = backlight_device_register(name, NULL, NULL, &pmu_backlight_data, | ^~~~~~~~~~~~~~~~~~~~~~~~~ | root_device_register drivers/macintosh/via-pmu-backlight.c:166:19: error: invalid use of undefined type 'struct backlight_device' 166 | level = bd->props.max_brightness; | ^~ drivers/macintosh/via-pmu-backlight.c:176:35: error: invalid use of undefined type 'struct backlight_device' 176 | bd->props.max_brightness / 15); | ^~ drivers/macintosh/via-pmu-backlight.c:179:11: error: invalid use of undefined type 'struct backlight_device' 179 | bd->props.brightness = level; | ^~ drivers/macintosh/via-pmu-backlight.c:180:11: error: invalid use of undefined type 'struct backlight_device' 180 | bd->props.power = FB_BLANK_UNBLANK; | ^~ drivers/macintosh/via-pmu-backlight.c:181:9: error: implicit declaration of function 'backlight_update_status'; did you mean 'pmu_backlight_update_status'? [-Werror=implicit-function-declaration] 181 | backlight_update_status(bd); | ^~~~~~~~~~~~~~~~~~~~~~~ | pmu_backlight_update_status drivers/macintosh/via-pmu-backlight.c:136:37: warning: unused variable 'props' [-Wunused-variable] 136 | struct backlight_properties props; | ^~~~~ drivers/macintosh/via-pmu-backlight.c: At top level: drivers/macintosh/via-pmu-backlight.c:108:35: error: storage size of 'pmu_backlight_data' isn't known 108 | static const struct backlight_ops pmu_backlight_data = { | ^~~~~~~~~~~~~~~~~~ drivers/macintosh/via-pmu-backlight.c:108:35: error: storage size of 'pmu_backlight_data' isn't known cc1: some warnings being treated as errors make[5]: *** [scripts/Makefile.build:244: drivers/macintosh/via-pmu-backlight.o] Error 1
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
Michael, if this is fine by you, ack to merge via the drm subsystem along with the regressing commit?
BR, Jani.
arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__
+#include <linux/backlight.h> #include <linux/fb.h> #include <linux/mutex.h>
-- Jani Nikula, Intel
-- Linaro LKFT https://lkft.linaro.org
Jani Nikula jani.nikula@intel.com writes:
On Mon, 04 Mar 2024, Jani Nikula jani.nikula@intel.com wrote:
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
Not even compile tested!
Naresh, please try this patch!
Michael, if this is fine by you, ack to merge via the drm subsystem along with the regressing commit?
I prefer Thomas' patch, but I'm fine with either to go in via drm.
cheers
On Tue, 05 Mar 2024, Michael Ellerman mpe@ellerman.id.au wrote:
I prefer Thomas' patch, but I'm fine with either to go in via drm.
Thanks!
Hi
Am 04.03.24 um 10:55 schrieb Jani Nikula:
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
Not even compile tested!
That's one of the cases that's hard to catch unless you get the config right.
arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__ +#include <linux/backlight.h>
Thanks, but I think this should go directly into chipsfb.c. I would have provided a patch already, if our mail server didn't have issues this morning. Let me try again.
Best regards Thomas
#include <linux/fb.h> #include <linux/mutex.h>
Am 04.03.24 um 11:32 schrieb Thomas Zimmermann: [...]
arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__ +#include <linux/backlight.h>
Thanks, but I think this should go directly into chipsfb.c. I would have provided a patch already, if our mail server didn't have issues this morning. Let me try again.
Posted at
https://lore.kernel.org/dri-devel/20240304103820.16708-1-tzimmermann@suse.de...
Best regards Thomas
#include <linux/fb.h> #include <linux/mutex.h>
Le 04/03/2024 à 11:32, Thomas Zimmermann a écrit :
Hi
Am 04.03.24 um 10:55 schrieb Jani Nikula:
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
Not even compile tested!
That's one of the cases that's hard to catch unless you get the config right.
arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__ +#include <linux/backlight.h>
Thanks, but I think this should go directly into chipsfb.c. I would have provided a patch already, if our mail server didn't have issues this morning. Let me try again.
asm/backlight.h needs it for struct backlight_device
At least if you don't want to include linux/backlight.h in asm/backlight.h, then you need a forward declaration of struct backlight_device;
Best regards Thomas
#include <linux/fb.h> #include <linux/mutex.h>
Christophe Leroy christophe.leroy@csgroup.eu writes:
Le 04/03/2024 à 11:32, Thomas Zimmermann a écrit :
Hi
Am 04.03.24 um 10:55 schrieb Jani Nikula:
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
Not even compile tested!
That's one of the cases that's hard to catch unless you get the config right.
arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__ +#include <linux/backlight.h>
Thanks, but I think this should go directly into chipsfb.c. I would have provided a patch already, if our mail server didn't have issues this morning. Let me try again.
asm/backlight.h needs it for struct backlight_device
At least if you don't want to include linux/backlight.h in asm/backlight.h, then you need a forward declaration of struct backlight_device;
It's preferable for asm headers not to include linux headers (to avoid loops), so a forward declaration would be better IMHO.
cheers
Hi
Am 05.03.24 um 01:36 schrieb Michael Ellerman:
Christophe Leroy christophe.leroy@csgroup.eu writes:
Le 04/03/2024 à 11:32, Thomas Zimmermann a écrit :
Hi
Am 04.03.24 um 10:55 schrieb Jani Nikula:
Removal of the backlight include from fb.h uncovered an implicit dependency in powerpc asm/backlight.h. Add the explicit include.
Reported-by: Naresh Kamboju naresh.kamboju@linaro.org Closes: https://lore.kernel.org/r/CA+G9fYsAk5TbqqxFC2W4oHLGA0CbTHMxbeq8QayFXTU75Yiue... Fixes: 11b4eedfc87d ("fbdev: Do not include <linux/backlight.h> in header") Cc: Thomas Zimmermann tzimmermann@suse.de Cc: Helge Deller deller@gmx.de Cc: linux-fbdev@vger.kernel.org Signed-off-by: Jani Nikula jani.nikula@intel.com
Not even compile tested!
That's one of the cases that's hard to catch unless you get the config right.
arch/powerpc/include/asm/backlight.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/include/asm/backlight.h b/arch/powerpc/include/asm/backlight.h index 1b5eab62ed04..275d5bb9aa04 100644 --- a/arch/powerpc/include/asm/backlight.h +++ b/arch/powerpc/include/asm/backlight.h @@ -10,6 +10,7 @@ #define __ASM_POWERPC_BACKLIGHT_H #ifdef __KERNEL__ +#include <linux/backlight.h>
Thanks, but I think this should go directly into chipsfb.c. I would have provided a patch already, if our mail server didn't have issues this morning. Let me try again.
asm/backlight.h needs it for struct backlight_device
At least if you don't want to include linux/backlight.h in asm/backlight.h, then you need a forward declaration of struct backlight_device;
It's preferable for asm headers not to include linux headers (to avoid loops), so a forward declaration would be better IMHO.
The asm header shouldn't include <linux/fb.h> either. I'll send an update to my patch to clean up everything.
Best regards Thomas
cheers