Re: [PATCH 18/28] kbuild: fail kernel compilation in case of unresolved module symbols

From: Sam Ravnborg
Date: Sun Sep 24 2006 - 18:42:57 EST

On Sun, Sep 24, 2006 at 11:36:43PM +0100, Al Viro wrote:
> On Mon, Sep 25, 2006 at 12:35:34AM +0200, Sam Ravnborg wrote:
> > On what architectures do you see lots of these warnings - maybe fixable?
> > Otherwise I could do something like this:
> Try allmodconfig someday...

My cross-compile setup severely broke due to me changing dev-machine.
Will restore it when I get my hand on a proper dev machine again - current
box consumes too much power for normal use (almost 300W) and is not much
quicker than my older amd64 box :-(

> aviro@icy:/usr/src/cross-kernel/volatile/work$ grep -l WARNING.*undefined ../logs/*/X4c
> ../logs/alpha-SMP/X4c
> ../logs/alpha/X4c
> ../logs/arm/X4c
> ../logs/armv/X4c
> ../logs/chestnut/X4c
> ../logs/frv/X4c
> ../logs/ia64/X4c
> ../logs/m32r/X4c
> ../logs/m68k/X4c
> ../logs/ppc/X4c
> ../logs/ppc44x/X4c
> ../logs/ppc64/X4c
> ../logs/s390/X4c
> ../logs/s390x/X4c
> ../logs/sparc32/X4c
> ../logs/sparc64/X4c
> ../logs/sun3/X4c
> ../logs/sun4/X4c
> ../logs/uml-i386/X4c
> That's out of 25 targets. The only variants that do _NOT_ trigger are
> amd64, amd64-UP, i386 and uml-amd64.

How many of these ougth to be fixed then?
> I more or less agree with rationale behind making that default, but I'd
> very much appreciate a way to override that. For now I've just made the
> -w line unconditional, but the fewer infrastructure patches I've to carry...
OK. Will include it in next round of kbuild updates.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at