Messages in this thread | | | Date | Mon, 14 Jan 2008 15:58:54 +0100 | From | Ingo Molnar <> | Subject | Re: [PATCH x86] [15/16] Force __cpuinit on for CONFIG_PM without HOTPLUG_CPU |
| |
* Sam Ravnborg <sam@ravnborg.org> wrote:
> > ok, i've dropped this patch from x86.git for now: > > > > Subject: x86: force __cpuinit on for CONFIG_PM without HOTPLUG_CPU > > From: Andi Kleen <ak@suse.de> > > > > but longer-term, shouldnt these annotations be automated? We'll see > > a constant stream of them, all around the clock as people regularly > > get it wrong (because it's not intuitive). > > Would be great to have them automated - just dunno how to do it. Do > you see a feasible way to do it?
a good starting point would be to make the warnings a lot more self-explanatory. Right now it's often non-obvious trying to figure out how the dependencies are structured and which one should be changed to get rid of the bug.
> Short term modpost etc. will be enhanced to be less dependent on the > actual configuration when performing the checks. It should not matter > if CONFIG_HOTPLUG_CPU is enabled or not when we check the __cpuint > annotations but this is how we see it today so far too many faults > slip through.
a good way i see is to:
- improve the debug output so that it's obvious at a glance what the problem is.
- quickly reach a close-to-100%-perfect stage, brute-force. Drop __init* annotations en masse if they are not perfectly layered. Whoever reintroduces them will then have to do it perfectly.
- then turn these into hard failures (which they _are_ in a significant percentage of the situations).
once it's a hard build failure, people will fix them quickly and automated tests will pick them up as well.
Ingo
| |