summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorAnton Blanchard <anton@samba.org>2011-11-30 00:23:16 +0000
committerBenjamin Herrenschmidt <benh@kernel.crashing.org>2011-12-08 14:02:24 +1100
commit549e88a134b3b393a4312e8d76628b9260eee57f (patch)
tree753fbc597df1b3771e75aa5fa75867c1a0672f45
parent90e8f57cf873188bc5cff445059ddeb72dc51d8c (diff)
downloadop-kernel-dev-549e88a134b3b393a4312e8d76628b9260eee57f.zip
op-kernel-dev-549e88a134b3b393a4312e8d76628b9260eee57f.tar.gz
powerpc/kdump: Delay before sending IPI on a system reset
If we enter the kdump code via system reset, wait a bit before sending the IPI to capture all secondary CPUs. Without it we race with the hypervisor that is issuing the system reset to each CPU. If the IPI gets there first the system reset oops output then shows the register state of the IPI handler which is not what we want. I took the opportunity to add defines for all the various delays we have. There's no need for cpu_relax when we are doing an mdelay, so remove them too. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
-rw-r--r--arch/powerpc/kernel/crash.c33
1 files changed, 26 insertions, 7 deletions
diff --git a/arch/powerpc/kernel/crash.c b/arch/powerpc/kernel/crash.c
index d446755..b942980 100644
--- a/arch/powerpc/kernel/crash.c
+++ b/arch/powerpc/kernel/crash.c
@@ -30,6 +30,20 @@
#include <asm/system.h>
#include <asm/setjmp.h>
+/*
+ * The primary CPU waits a while for all secondary CPUs to enter. This is to
+ * avoid sending an IPI if the secondary CPUs are entering
+ * crash_kexec_secondary on their own (eg via a system reset).
+ *
+ * The secondary timeout has to be longer than the primary. Both timeouts are
+ * in milliseconds.
+ */
+#define PRIMARY_TIMEOUT 500
+#define SECONDARY_TIMEOUT 1000
+
+#define IPI_TIMEOUT 10000
+#define REAL_MODE_TIMEOUT 10000
+
/* This keeps a track of which one is the crashing cpu. */
int crashing_cpu = -1;
static cpumask_t cpus_in_crash = CPU_MASK_NONE;
@@ -99,11 +113,9 @@ again:
* FIXME: Until we will have the way to stop other CPUs reliably,
* the crash CPU will send an IPI and wait for other CPUs to
* respond.
- * Delay of at least 10 seconds.
*/
- msecs = 10000;
+ msecs = IPI_TIMEOUT;
while ((cpumask_weight(&cpus_in_crash) < ncpus) && (--msecs > 0)) {
- cpu_relax();
mdelay(1);
}
@@ -163,11 +175,11 @@ again:
void crash_kexec_secondary(struct pt_regs *regs)
{
unsigned long flags;
- int msecs = 500;
+ int msecs = SECONDARY_TIMEOUT;
local_irq_save(flags);
- /* Wait 500ms for the primary crash CPU to signal its progress */
+ /* Wait for the primary crash CPU to signal its progress */
while (crashing_cpu < 0) {
if (--msecs < 0) {
/* No response, kdump image may not have been loaded */
@@ -176,7 +188,6 @@ void crash_kexec_secondary(struct pt_regs *regs)
}
mdelay(1);
- cpu_relax();
}
crash_ipi_callback(regs);
@@ -211,7 +222,7 @@ static void crash_kexec_wait_realmode(int cpu)
unsigned int msecs;
int i;
- msecs = 10000;
+ msecs = REAL_MODE_TIMEOUT;
for (i=0; i < nr_cpu_ids && msecs > 0; i++) {
if (i == cpu)
continue;
@@ -306,6 +317,14 @@ void default_machine_crash_shutdown(struct pt_regs *regs)
*/
crashing_cpu = smp_processor_id();
crash_save_cpu(regs, crashing_cpu);
+
+ /*
+ * If we came in via system reset, wait a while for the secondary
+ * CPUs to enter.
+ */
+ if (TRAP(regs) == 0x100)
+ mdelay(PRIMARY_TIMEOUT);
+
crash_kexec_prepare_cpus(crashing_cpu);
cpumask_set_cpu(crashing_cpu, &cpus_in_crash);
crash_kexec_wait_realmode(crashing_cpu);
OpenPOWER on IntegriCloud