Messages in this thread | | | Date | Mon, 17 Nov 2008 18:35:27 +0100 | From | Ingo Molnar <> | Subject | Re: [PATCH] x86: fix wakeup_cpu with numaq/es7000 v2 |
| |
* Cyrill Gorcunov <gorcunov@gmail.com> wrote:
> [Ingo Molnar - Mon, Nov 17, 2008 at 05:52:24PM +0100] > | > | * Yinghai Lu <yinghai@kernel.org> wrote: > | > | > Impact: fix wakeup path with numaq and es7000 > | > | applied to tip/x86/quirks, thanks Yinghai! > | > | A couple of comments: > | > | > +static inline void inquire_remote_apic(int apicid) > | > +{ > | > + if (apic_verbosity >= APIC_DEBUG) > | > + __inquire_remote_apic(apicid); > | > +} > | > > Btw, Ingo, Yinghai, > > it's a bit weird but I found that we use apic_verbosity > here to _do_ something (ie touching circuit) but on any > other cases apic_verbosity is just a additional logging > filter so -- what the initial apic_verbosity purpose was? > If it's just a logging filter (to which I'm more biased) > this case is plain wrong. But if apic_verbosity is not > like this -- then it's fine. Since I was not initial author > of the code -- I'm asking you :)
the reason why it's dependent on apic_verbosity is that the inquire_remote_apic() function is a debug function which we only call if we fail to start up a secondary CPU.
Perhaps renaming it to debug_print_remote_apic() would have made that more obvious.
Ingo
| |