Re: 2.4.19pre8aa2

[Eyal Lebedinsky]

Keith Owens wrote:
> 
> On Sun, 05 May 2002 09:34:25 +1000,
> Eyal Lebedinsky <[e-mail]> wrote:
> >I agree with this. However, since this driver cannot be built, and
> >since the latest modutils will exit badly for unresolved, I strongly
> >believe that the comx driver should not be offered (disable it in
> >the Config.in) until it is fixed - 2.4 being a stable kernel.
> >
> >I had to wrap 'depmod' with a script to ignore failures in order
> >to get through a full build (which includes the kernel plus a few
> >extra modules like NVdriver, dc395, lm_sensors).
> 
> You should not have to wrap depmod.  Standard 2.4.* modutils returns 0
> for unresolved errors, for compatibility with previous behaviour.  You
> have to do depmod -u to get a non-zero return code.  The -u flag was
> added in modutils 2.4.7 and defaults to off.  In modutils 2.5 it will
> default to on, that will break code that relies on the existing
> behaviour.  I will not change modutils default behaviour in 2.4.
> 
> So why is your version of depmod breaking?  Either you are specifying
> -u or your distribution has hacked depmod to default -u to on.  Check
> depmod.c for variable flag_unresolved_error, it should default to 0.

You are right, it is not the unresolved that caused it but the non
ELF objects in there (it used not to care before):

# /sbin/depmod-2.4.15 -ae ; echo $?
depmod: /lib/modules/2.4.19-pre8-aa2/ksyms is not an ELF file
depmod: /lib/modules/2.4.19-pre8-aa2/soundconf is not an ELF file
1

--
Eyal Lebedinsky ([e-mail]) <http://samba.org/eyal/>