On Mon, Feb 13, 2017 at 11:33 AM, Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
>
> We never retrigger any of the special vectors, just the device interrupts
> can get that treatment and there only the few edge interrupts connected to
> the IO-APIC are relevant.
Don't we walk up to the local apic when we walk the parent chain? I
didn't check how the parent chain thing is connected..
Anyway, I'm wondering if the WARN_ON might give some ideas (and I just
noticed that I made the WARN_ON() be WARN_ON_ONCE() in the patch I
sent out, because I'm so used to trying to avoid constant streams of
warnings, but in this case we migth want them - particularly if
Gabriel then tries it with a broken kernel and we get a "it halts
after _this_ one" situation.
> Boris has access to a machine with a similar CPU. It shows the symptom of
> boot hang, but only after Boris used Gabriels config file. The normal SuSe
> config just boots fine. Again, nothing stands out there which immediately
> rings a bell. :(
Oh, I must have missed that email.
Boris - have you tried to narrow down the exact config isse with the
ktest config bisection script? I haven't used it myself, but I'm
adding StevenR to see if he has suggestions.
Linus
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html