On 5 January 2015 at 16:31, alvise rigo <a.r...@virtualopensystems.com> wrote: > It's just because in patch 1/4 of this series we use > find_machine_info(machine->cpu_model), which could be a NULL pointer.
Well, don't do that, because in that case you'll do the wrong thing if we take the default value of the cpu model. But I don't think patch 1 is the right approach anyway. (Also, if you have a series where patch B requires a change or fix that is put in patch A, then A should come before B in the series.) thanks -- PMM