RE: [PATCH v4 2/5] x86/mce: Add per-bank CMCI storm mitigation

[Luck, Tony]

> > ld: vmlinux.o: in function `machine_check_poll':
> > /home/boris/kernel/2nd/linux/arch/x86/kernel/cpu/mce/core.c:683: undefined reference to `track_cmci_storm'
> > make[1]: *** [scripts/Makefile.vmlinux:35: vmlinux] Error 1
> > make: *** [Makefile:1249: vmlinux] Error 2
> > make: *** Waiting for unfinished jobs....
> >
>
> Ah, this is with CONFIG_MCE_INTEL=n and everything =y. Is there an automated
> way to test every config, not just random, combination in a subsystem?
>
> I'll try to add something like this to my flow. It seems allnoconfig,
> defconfig, etc. aren't enough. And it's too easy to overlook during code review.

I'm a bit surprised that lkp didn't complain. It used to do a zillion builds with combinations
of CONFIG options that were relevant to the patch series.

I'll spin a v5 with an inline stub function to fix this.

Boris: Have you seen anything else that needs fixing?

-Tony