Skip to contentSkip to conversation

linux-next: build failure after merge of the drivers-x86 tree

DISCUSSION Stephen Rothwell 3 participants 4 messages Feb 6, 2025 Linux-next tree

lore
Share & tools
lore
#0 lore raw

Hi all,

After merging the drivers-x86 tree, today's linux-next build (x86_64_allmodconfig) failed like this:

In file included from include/linux/kobject.h:20,
from include/linux/energy_model.h:7,
from include/linux/device.h:16,
from include/linux/acpi.h:14,
from drivers/platform/x86/samsung-galaxybook.c:14:
drivers/platform/x86/samsung-galaxybook.c: In function 'galaxybook_fw_attr_init':
drivers/platform/x86/samsung-galaxybook.c:1014:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1014 | sysfs_attr_init(&fw_attr.display_name);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~

drivers/platform/x86/samsung-galaxybook.c:1020:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1020 | sysfs_attr_init(&fw_attr.current_value);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~

Caused by commit

f97634611408 ("platform/x86: samsung-galaxybook: Add samsung-galaxybook driver")

I guess this was never built with CONFIG_DEBUG_LOCK_ALLOC set.

I have used the drivers-x86 tree from next-20250205 for today.

Signature · 2 lines
Cheers,
Stephen Rothwell
#1 lore raw
Quoted 24 lines · Stephen Rothwell

Hi all,

After merging the drivers-x86 tree, today's linux-next build (x86_64_allmodconfig) failed like this:

In file included from include/linux/kobject.h:20,
from include/linux/energy_model.h:7,
from include/linux/device.h:16,
from include/linux/acpi.h:14,
from drivers/platform/x86/samsung-galaxybook.c:14:
drivers/platform/x86/samsung-galaxybook.c: In function 'galaxybook_fw_attr_init':
drivers/platform/x86/samsung-galaxybook.c:1014:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1014 | sysfs_attr_init(&fw_attr.display_name);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~

drivers/platform/x86/samsung-galaxybook.c:1020:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1020 | sysfs_attr_init(&fw_attr.current_value);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~
Quoted 7 lines

Caused by commit

f97634611408 ("platform/x86: samsung-galaxybook: Add samsung-galaxybook driver")

I guess this was never built with CONFIG_DEBUG_LOCK_ALLOC set.

I have used the drivers-x86 tree from next-20250205 for today.

Apparently it wasn't.

However, I've an LKP success report for f97634611408 (prior to pushing it to for-next, I always wait for LKP).

Why LKP didn't catch it despite claiming it built with x86_64_allyesconfig (successfully)?? Did LKP not build the tree??

I've pulled the commit from for-next until the problem is resolved to not keep breaking builds. Joshua, could you please take a look at it.

Signature · 1 line
 i.
#2 lore raw

Den tors 6 feb. 2025 kl 11:27 skrev Ilpo Järvinen
<[email protected]>:

Quoted 40 lines · Stephen Rothwell
Quoted 25 lines

Hi all,

After merging the drivers-x86 tree, today's linux-next build (x86_64_allmodconfig) failed like this:

In file included from include/linux/kobject.h:20,
from include/linux/energy_model.h:7,
from include/linux/device.h:16,
from include/linux/acpi.h:14,
from drivers/platform/x86/samsung-galaxybook.c:14:
drivers/platform/x86/samsung-galaxybook.c: In function 'galaxybook_fw_attr_init':
drivers/platform/x86/samsung-galaxybook.c:1014:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1014 | sysfs_attr_init(&fw_attr.display_name);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~

drivers/platform/x86/samsung-galaxybook.c:1020:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1020 | sysfs_attr_init(&fw_attr.current_value);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~
Quoted 7 lines

Caused by commit

f97634611408 ("platform/x86: samsung-galaxybook: Add samsung-galaxybook driver")

I guess this was never built with CONFIG_DEBUG_LOCK_ALLOC set.

I have used the drivers-x86 tree from next-20250205 for today.

Apparently it wasn't.

However, I've an LKP success report for f97634611408 (prior to pushing it to for-next, I always wait for LKP).

Quoted 9 lines

Why LKP didn't catch it despite claiming it built with x86_64_allyesconfig (successfully)?? Did LKP not build the tree??

I've pulled the commit from for-next until the problem is resolved to not keep breaking builds. Joshua, could you please take a look at it.

--
i.

Hi Ilpo and all,
Yes, great that there are tests in place :) Sorry this was one of the later changes and I can't remember now when exactly I did the full testing vs not as per the kernel documentation (as there were several iterations to the patch for this driver) but I believe I have identified what should be fixed and can send it shortly.

Ilpo would you prefer a patch on top of the existing patch (i.e. just a diff of these 2 lines) or would you rather that I create a v11 of the original patch and send the whole thing again?

Thanks again!

Best regards,
Joshua

#3 lore raw
Quoted 59 lines · Joshua Grisham

Den tors 6 feb. 2025 kl 11:27 skrev Ilpo Järvinen
<[email protected]>:

Quoted 40 lines · Stephen Rothwell
Quoted 25 lines

Hi all,

After merging the drivers-x86 tree, today's linux-next build (x86_64_allmodconfig) failed like this:

In file included from include/linux/kobject.h:20,
from include/linux/energy_model.h:7,
from include/linux/device.h:16,
from include/linux/acpi.h:14,
from drivers/platform/x86/samsung-galaxybook.c:14:
drivers/platform/x86/samsung-galaxybook.c: In function 'galaxybook_fw_attr_init':
drivers/platform/x86/samsung-galaxybook.c:1014:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1014 | sysfs_attr_init(&fw_attr.display_name);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~

drivers/platform/x86/samsung-galaxybook.c:1020:33: error: 'fw_attr' is a pointer; did you mean to use '->'?
1020 | sysfs_attr_init(&fw_attr.current_value);
| ^
include/linux/sysfs.h:55:10: note: in definition of macro 'sysfs_attr_init'

   55 |         (attr)->key = &__key;                           \
      |          ^~~~
Quoted 7 lines

Caused by commit

f97634611408 ("platform/x86: samsung-galaxybook: Add samsung-galaxybook driver")

I guess this was never built with CONFIG_DEBUG_LOCK_ALLOC set.

I have used the drivers-x86 tree from next-20250205 for today.

Apparently it wasn't.

However, I've an LKP success report for f97634611408 (prior to pushing it to for-next, I always wait for LKP).

Quoted 9 lines

Why LKP didn't catch it despite claiming it built with x86_64_allyesconfig (successfully)?? Did LKP not build the tree??

I've pulled the commit from for-next until the problem is resolved to not keep breaking builds. Joshua, could you please take a look at it.

--
i.

Hi Ilpo and all,
Yes, great that there are tests in place :) Sorry this was one of the later changes and I can't remember now when exactly I did the full testing vs not as per the kernel documentation (as there were several iterations to the patch for this driver) but I believe I have identified what should be fixed and can send it shortly.

Quoted 5 lines

Ilpo would you prefer a patch on top of the existing patch (i.e. just a diff of these 2 lines) or would you rather that I create a v11 of the original patch and send the whole thing again?

Thanks again!

I've kept the change in review-ilpo-next branch. An incremental patch is fine (it's easier to see its correctness on a glance too).

Signature · 1 line
 i.