Commit 311f594d authored by Jay Lan's avatar Jay Lan Committed by Tony Luck

[IA64] kdump on INIT needs multi-nodes sync-up (v.2)

The current implementation of kdump on INIT events would enter
kdump processing on DIE_INIT_MONARCH_ENTER and DIE_INIT_SLAVE_ENTER
events. Thus, the monarch cpu would go ahead and boot up the kdump

On SN shub2 systems, this out-of-sync situation causes some slave
cpus on different nodes to enter POD.

This patch moves kdump entry points to DIE_INIT_MONARCH_LEAVE and
DIE_INIT_SLAVE_LEAVE. It also sets kdump_in_progress variable in
the DIE_INIT_MONARCH_PROCESS event to not dump all active stack
traces to the console in the case of kdump.

I have tested this patch on an SN machine and a HP RX2600.
Signed-off-by: default avatarJay Lan <jlan@sgi.com>
Acked-by: default avatarZou Nan hai <nanhai.zou@intel.com>
Signed-off-by: default avatarTony Luck <tony.luck@intel.com>
parent ae67e498
...@@ -156,23 +156,29 @@ kdump_init_notifier(struct notifier_block *self, unsigned long val, void *data) ...@@ -156,23 +156,29 @@ kdump_init_notifier(struct notifier_block *self, unsigned long val, void *data)
if (!kdump_on_init) if (!kdump_on_init)
return NOTIFY_DONE; return NOTIFY_DONE;
if (val != DIE_INIT_MONARCH_ENTER && if (val != DIE_INIT_MONARCH_LEAVE &&
val != DIE_INIT_SLAVE_ENTER && val != DIE_INIT_SLAVE_LEAVE &&
val != DIE_INIT_MONARCH_PROCESS &&
val != DIE_MCA_RENDZVOUS_LEAVE && val != DIE_MCA_RENDZVOUS_LEAVE &&
val != DIE_MCA_MONARCH_LEAVE) val != DIE_MCA_MONARCH_LEAVE)
return NOTIFY_DONE; return NOTIFY_DONE;
nd = (struct ia64_mca_notify_die *)args->err; nd = (struct ia64_mca_notify_die *)args->err;
/* Reason code 1 means machine check rendezvous*/ /* Reason code 1 means machine check rendezvous*/
if ((val == DIE_INIT_MONARCH_ENTER || val == DIE_INIT_SLAVE_ENTER) && if ((val == DIE_INIT_MONARCH_LEAVE || val == DIE_INIT_SLAVE_LEAVE
nd->sos->rv_rc == 1) || val == DIE_INIT_MONARCH_PROCESS) && nd->sos->rv_rc == 1)
return NOTIFY_DONE; return NOTIFY_DONE;
switch (val) { switch (val) {
case DIE_INIT_MONARCH_ENTER: case DIE_INIT_MONARCH_PROCESS:
atomic_set(&kdump_in_progress, 1);
*(nd->monarch_cpu) = -1;
break;
case DIE_INIT_MONARCH_LEAVE:
machine_kdump_on_init(); machine_kdump_on_init();
break; break;
case DIE_INIT_SLAVE_ENTER: case DIE_INIT_SLAVE_LEAVE:
if (atomic_read(&kdump_in_progress))
unw_init_running(kdump_cpu_freeze, NULL); unw_init_running(kdump_cpu_freeze, NULL);
break; break;
case DIE_MCA_RENDZVOUS_LEAVE: case DIE_MCA_RENDZVOUS_LEAVE:
...@@ -215,8 +221,10 @@ static ctl_table sys_table[] = { ...@@ -215,8 +221,10 @@ static ctl_table sys_table[] = {
static int static int
machine_crash_setup(void) machine_crash_setup(void)
{ {
/* be notified before default_monarch_init_process */
static struct notifier_block kdump_init_notifier_nb = { static struct notifier_block kdump_init_notifier_nb = {
.notifier_call = kdump_init_notifier, .notifier_call = kdump_init_notifier,
.priority = 1,
}; };
int ret; int ret;
if((ret = register_die_notifier(&kdump_init_notifier_nb)) != 0) if((ret = register_die_notifier(&kdump_init_notifier_nb)) != 0)
......
...@@ -1477,6 +1477,10 @@ default_monarch_init_process(struct notifier_block *self, unsigned long val, voi ...@@ -1477,6 +1477,10 @@ default_monarch_init_process(struct notifier_block *self, unsigned long val, voi
struct task_struct *g, *t; struct task_struct *g, *t;
if (val != DIE_INIT_MONARCH_PROCESS) if (val != DIE_INIT_MONARCH_PROCESS)
return NOTIFY_DONE; return NOTIFY_DONE;
#ifdef CONFIG_KEXEC
if (atomic_read(&kdump_in_progress))
return NOTIFY_DONE;
#endif
/* /*
* FIXME: mlogbuf will brim over with INIT stack dumps. * FIXME: mlogbuf will brim over with INIT stack dumps.
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment