On 06/04/2014 07:11 PM, Vivek Goyal wrote: > On Wed, Jun 04, 2014 at 01:58:40AM +0530, Srivatsa S. Bhat wrote: >> On 05/28/2014 07:01 PM, Vivek Goyal wrote: >>> On Tue, May 27, 2014 at 04:25:34PM +0530, Srivatsa S. Bhat wrote: >>>> If we try to perform a kexec when the machine is in ST (Single-Threaded) mode >>>> (ppc64_cpu --smt=off), the kexec operation doesn't succeed properly, and we >>>> get the following messages during boot: >>>> [...] >>>> diff --git a/kernel/kexec.c b/kernel/kexec.c >>>> index c8380ad..28c5706 100644 >>>> --- a/kernel/kexec.c >>>> +++ b/kernel/kexec.c >>>> @@ -1683,6 +1683,14 @@ int kernel_kexec(void) >>>> kexec_in_progress = true; >>>> kernel_restart_prepare(NULL); >>>> migrate_to_reboot_cpu(); >>>> + >>>> + /* >>>> + * migrate_to_reboot_cpu() disables CPU hotplug assuming that >>>> + * no further code needs to use CPU hotplug (which is true in >>>> + * the reboot case). However, the kexec path depends on using >>>> + * CPU hotplug again; so re-enable it here. >>>> + */ >>>> + cpu_hotplug_enable(); >>>> printk(KERN_EMERG "Starting new kernel\n"); >>>> machine_shutdown(); >>> >>> After migrate_to_reboot_cpu(), we are calling machine_shutdown() which >>> calls disable_nonboot_cpus() and which in turn calls _cpu_down(). >>> >> >> Hmm? I see only 'arm' calling disable_nonboot_cpus() from machine_shutdown(). >> None of the other architectures call it. Is that a leftover in arm? > > You are right. I did not notice that only arm is doing that. Looks like > it is calling into some platform code, I am not sure what exactly arm > does for disabling cpu. > > x86 code calls stop_other_cpus() in machine_shutdown() which sends > REBOOT_VECTOR to other cpus and calls stop_this_cpu() which in turn > does. > > for (;;) > halt(); > > IIUC, upon receipt of certain interrupts cpu will come out of halt state. > Not sure how safe it is from kexec point of view as we will be replacing > original kernel that means if cpu comes out of halt state it might be > running some random code. > > Eric/hpa might know better the context here and what safeguards us on x86. > > So one should not make cpu spin on some code as kexec will change that > code. It should be some other platform specific mechanism which brings > cpu in to hlt like state. So that way arm seems to be doing right thing. > > I am not sure what powerpc does to stop cpus. > powerpc shepherds all CPUs to a safe state, by making them run kexec_smp_down(), and eventually those CPUs end up calling kexec_wait() in assembly. Regards, Srivatsa S. Bhat
