On Tue, Feb 21, 2017 at 09:28:20AM +0800, Xunlei Pang wrote:
> Not kdump kernel starts dumping, just during nmi_shootdown_cpus(), if some
> MCE comes after crashing_cpu was set and we don't skip crashing_cpu, then
> the crashing cpu will enter mce handler and trigger the synchronization issue.
Ok, I went and looked at __crash_kexec() - so we do nmi_shootdown_cpus()
as part of machine_crash_shutdown(). The next thing we do is kexec the
new kernel in machine_kexec(kexec_crash_image). The picture is clearer
now.
--
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.
