linux-next-2023-0815: s390/block/dasd: build error

linux-next-2023-0815: s390/block/dasd: build error

[Randy Dunlap]

[-- Attachment #1: Type: text/plain, Size: 298 bytes --]

I have spent some time on this but I don't see where the problem is.

ERROR: modpost: "bdev_mark_dead" [drivers/s390/block/dasd_mod.ko] undefined!

CONFIG_BLOCK=y, bdev.o is built and contains the missing symbol.

Full randconfig file is attached.

Hopefully I'm just overlooking something.

~Randy

[-- Attachment #2: config-r5177.gz --]
[-- Type: application/gzip, Size: 19900 bytes --]

Re: linux-next-2023-0815: s390/block/dasd: build error

[Sven Schnelle]

Randy Dunlap <rdunlap@infradead.org> writes:

> I have spent some time on this but I don't see where the problem is.
>
> ERROR: modpost: "bdev_mark_dead" [drivers/s390/block/dasd_mod.ko] undefined!
>
> CONFIG_BLOCK=y, bdev.o is built and contains the missing symbol.
>
> Full randconfig file is attached.
>
> Hopefully I'm just overlooking something.

The EXPORT_SYMBOL_GPL is surrounded by #ifdef CONFIG_DASD, but i think
it should be '#ifdef CONFIG_DASD_MODULE'. This was introduced by

381f678306ce ("block: consolidate __invalidate_device and fsync_bdev")

There was already a thread about this:

https://www.spinics.net/lists/linux-btrfs/msg138633.html

regards,
Sven

Re: linux-next-2023-0815: s390/block/dasd: build error

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 852 bytes --]

Hi Randy,

On Tue, 15 Aug 2023 22:34:19 -0700 Randy Dunlap <rdunlap@infradead.org> wrote:
>
> I have spent some time on this but I don't see where the problem is.
> 
> ERROR: modpost: "bdev_mark_dead" [drivers/s390/block/dasd_mod.ko] undefined!
> 
> CONFIG_BLOCK=y, bdev.o is built and contains the missing symbol.
> 
> Full randconfig file is attached.
> 
> Hopefully I'm just overlooking something.

In your config, CONFIG_DASD=m which means "#ifdef CONFIG_DASD" will
fail since CONFIG_DASD_MODULE is defined instead ... so maybe it should
be "#if IS_ENABLED(CONFIG_DASD)" in block/bdev.c.  This also fails an
s390 allmodconfig build (obviously).

So this is caused by commit

  381f678306ce ("block: consolidate __invalidate_device and fsync_bdev")

from the vfs-brauner branch in linux-next.
-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next-2023-0815: s390/block/dasd: build error

[Christian Brauner]

On Wed, Aug 16, 2023 at 05:14:39PM +1000, Stephen Rothwell wrote:
> Hi Randy,
> 
> On Tue, 15 Aug 2023 22:34:19 -0700 Randy Dunlap <rdunlap@infradead.org> wrote:
> >
> > I have spent some time on this but I don't see where the problem is.
> > 
> > ERROR: modpost: "bdev_mark_dead" [drivers/s390/block/dasd_mod.ko] undefined!
> > 
> > CONFIG_BLOCK=y, bdev.o is built and contains the missing symbol.
> > 
> > Full randconfig file is attached.
> > 
> > Hopefully I'm just overlooking something.
> 
> In your config, CONFIG_DASD=m which means "#ifdef CONFIG_DASD" will
> fail since CONFIG_DASD_MODULE is defined instead ... so maybe it should
> be "#if IS_ENABLED(CONFIG_DASD)" in block/bdev.c.  This also fails an
> s390 allmodconfig build (obviously).
> 
> So this is caused by commit
> 
>   381f678306ce ("block: consolidate __invalidate_device and fsync_bdev")
> 
> from the vfs-brauner branch in linux-next.

Thanks for reporting this everyone. Fixed in-tree.