Re: CONFIG_NLS=m resulting in undefined reference to utf16s_to_utf8scausing build failure

[Len Brown]

> > drivers/built-in.o: In function `dsm_label_utf16s_to_utf8s':
> > /usr/src/Linux-Next/linux-next/drivers/pci/pci-label.c:195: undefined reference to `utf16s_to_utf8s'
> > make: *** [.tmp_vmlinux1] Error 1
> > 
> > The drivers/pci/pci-label.c is compiled as part of vmlinux as defined by
> > the makefile drivers/pci/Makefile. 
> > 
> > #
> > # ACPI Related PCI FW Functions
> > # ACPI _DSM provided firmware instance and string name
> > #
> > obj-$(CONFIG_ACPI)    += pci-acpi.o pci-label.o
> > 
> > # SMBIOS provided firmware instance and labels
> > obj-$(CONFIG_DMI)    += pci-label.o
> > 
> > This could be prevented by one of the following methods -
> > 
> > * forcing CONFIG_NLS=y if CONFIG_PCI=y 
> > * move `utf16s_to_utf8s' out of nls_base.c to a new file to be always
> > built into vmlinux
> > 
> > Please correct me if i missed something here and let me know what would be the right
> > way to fix this build failure.
> 
> So there's no other ACPI code that depends on this function?  I'd hate
> to make PCI depend on NLS, so assuming ACPI doesn't need it (in which
> case the select or depend should be there), I guess it should be pulled
> into a common file that will always be included.

No, nothing in the ACPI code uses utf16s_to_utf8s.

I don't know anything about NLS, but it appears to be
something that file systems (befs, cifs, fat, hfs, isofs, jfs, etc)
use to to support native languages.

BTW. I don't understand why pci-label.o appears twice above.
It used to be just the 2nd one, that depends on CONFIG_DMI.

cheers,
Len Brown, Intel Open Source Technology Center


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/