Commit 0cced40e authored by Hidetoshi Seto's avatar Hidetoshi Seto Committed by Tony Luck
Browse files

[IA64] kdump: Short path to freeze CPUs 



Setting monarch_cpu = -1 to let slaves frozen might not work, because
there might be slaves being late, not entered the rendezvous yet.
Such slaves might be caught in while (monarch_cpu == -1) loop.

Use kdump_in_progress instead of monarch_cpus to break INIT rendezvous
and let all slaves enter DIE_INIT_SLAVE_LEAVE smoothly.

And monarch no longer need to manage rendezvous if once kdump_in_progress
is set, catch the monarch in DIE_INIT_MONARCH_ENTER then.

Signed-off-by: default avatarHidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Haren Myneni <hbabu@us.ibm.com>
Cc: kexec@lists.infradead.org
Acked-by: default avatarFenghua Yu <fenghua.yu@intel.com>
Signed-off-by: default avatarTony Luck <tony.luck@intel.com>
parent 5959906e

arch/ia64/kernel/crash.c

+6 −9
Original line number Diff line number Diff line
@@ -127,14 +127,13 @@ machine_crash_shutdown(struct pt_regs *pt) 
	 * If an INIT is asserted here:

	 * - All receivers might be slaves, since some of cpus could already

	 *   be frozen and INIT might be masked on monarch.  In this case,

	 *   all slaves will park in while (monarch_cpu == -1) loop before

	 *   DIE_INIT_SLAVE_ENTER that for waiting monarch enters.

	 *	=> TBD: freeze all slaves

	 *   all slaves will be frozen soon since kdump_in_progress will let

	 *   them into DIE_INIT_SLAVE_LEAVE.

	 * - One might be a monarch, but INIT rendezvous will fail since

	 *   at least this cpu already have INIT masked so it never join

	 *   to the rendezvous.  In this case, all slaves and monarch will

	 *   be frozen after timeout of the INIT rendezvous.

	 *	=> TBD: freeze them without waiting timeout

	 *   be frozen soon with no wait since the INIT rendezvous is skipped

	 *   by kdump_in_progress.

	 */

	kdump_smp_send_stop();

	/* not all cpu response to IPI, send INIT to freeze them */
@@ -187,6 +186,7 @@ kdump_init_notifier(struct notifier_block *self, unsigned long val, void *data) 
				break;

			/* fall through */

		case DIE_INIT_SLAVE_LEAVE:

		case DIE_INIT_MONARCH_ENTER:

		case DIE_MCA_RENDZVOUS_LEAVE:

			unw_init_running(kdump_cpu_freeze, NULL);

			break;
@@ -217,7 +217,6 @@ kdump_init_notifier(struct notifier_block *self, unsigned long val, void *data) 
		if (kdump_on_init && (nd->sos->rv_rc != 1)) {

			if (atomic_inc_return(&kdump_in_progress) != 1)

				kdump_freeze_monarch = 1;

			*(nd->monarch_cpu) = -1;

		}

		break;

	case DIE_INIT_MONARCH_LEAVE:
@@ -228,10 +227,8 @@ kdump_init_notifier(struct notifier_block *self, unsigned long val, void *data) 
	case DIE_MCA_MONARCH_LEAVE:

		/* *(nd->data) indicate if MCA is recoverable */

		if (kdump_on_fatal_mca && !(*(nd->data))) {

			if (atomic_inc_return(&kdump_in_progress) == 1) {

				*(nd->monarch_cpu) = -1;

			if (atomic_inc_return(&kdump_in_progress) == 1)

				machine_kdump_on_init();

			}

			/* We got fatal MCA while kdump!? No way!! */

		}

		break;

arch/ia64/kernel/mca.c

+13 −2
Original line number Diff line number Diff line
@@ -1682,14 +1682,25 @@ ia64_init_handler(struct pt_regs *regs, struct switch_stack *sw, 


	if (!sos->monarch) {

		ia64_mc_info.imi_rendez_checkin[cpu] = IA64_MCA_RENDEZ_CHECKIN_INIT;



#ifdef CONFIG_KEXEC

		while (monarch_cpu == -1 && !atomic_read(&kdump_in_progress))

			udelay(1000);

#else

		while (monarch_cpu == -1)

			cpu_relax();	/* spin until monarch enters */

#endif



		NOTIFY_INIT(DIE_INIT_SLAVE_ENTER, regs, (long)&nd, 1);

		NOTIFY_INIT(DIE_INIT_SLAVE_PROCESS, regs, (long)&nd, 1);



#ifdef CONFIG_KEXEC

		while (monarch_cpu != -1 && !atomic_read(&kdump_in_progress))

			udelay(1000);

#else

		while (monarch_cpu != -1)

			cpu_relax();	/* spin until monarch leaves */

#endif



		NOTIFY_INIT(DIE_INIT_SLAVE_LEAVE, regs, (long)&nd, 1);

CRA Git | Maintained and supported by SUSTech CRA and CCSE