Messages in this thread | | | Date | Fri, 2 Mar 2012 16:11:56 -0500 | From | Don Zickus <> | Subject | Re: [PATCH 1/2] x86, reboot: revert stop_other_cpus to using IRQ with NMI fallback |
| |
On Fri, Mar 02, 2012 at 03:41:57PM -0500, Seiji Aguchi wrote: > > > + /* if the REBOOT_VECTOR didn't work, try with the NMI */ > > + if ((num_online_cpus() > 1) && (!smp_no_nmi_ipi)) { > > + if (register_nmi_handler(NMI_LOCAL, smp_stop_nmi_callback, > > + NMI_FLAG_FIRST, "smp_stop")) > > register_nmi_handler() doesn't work if kernel panics in nmi context because kzalloc() is called in register_nmi_handler() and > it may sleep. > > register_nmi_handler() should be replaced with notifier_chain like kdump does.
I don't think kdump uses the notifier_chain.
Cheers, Don
| |