powerpc/watchdog: Fix marking of stuck CPUs

When the SMP detector finds other CPUs stuck, it iterates over
them and marks them as stuck. This pulls them out of the pending
mask and allows the detector to continue with remaining good
CPUs (if nmi_watchdog=panic is not enabled).

The code to dothat was buggy because when setting a CPU stuck,
if the pending mask became empty, it resets it to keep the
watchdog running. However the iterator will continue to run
over the new pending mask and mark remaining good CPUs sas stuck.

Fix this by doing it with cpumask bitwise operations.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

1 files changed, 9 insertions, 6 deletions

diff --git a/arch/powerpc/kernel/watchdog.c b/arch/powerpc/kernel/watchdog.c
index 7d16dafa1bb6..12e90ae712fe 100644
--- a/arch/powerpc/kernel/watchdog.c
+++ b/arch/powerpc/kernel/watchdog.c
@@ -101,10 +101,10 @@ static void wd_lockup_ipi(struct pt_regs *regs)
                 nmi_panic(regs, "Hard LOCKUP");
 }
 
-static void set_cpu_stuck(int cpu, u64 tb)
+static void set_cpumask_stuck(const struct cpumask *cpumask, u64 tb)
 {
-        cpumask_set_cpu(cpu, &wd_smp_cpus_stuck);
-        cpumask_clear_cpu(cpu, &wd_smp_cpus_pending);
+        cpumask_or(&wd_smp_cpus_stuck, &wd_smp_cpus_stuck, cpumask);
+        cpumask_andnot(&wd_smp_cpus_pending, &wd_smp_cpus_pending, cpumask);
         if (cpumask_empty(&wd_smp_cpus_pending)) {
                 wd_smp_last_reset_tb = tb;
                 cpumask_andnot(&wd_smp_cpus_pending,
@@ -112,6 +112,10 @@ static void set_cpu_stuck(int cpu, u64 tb)
                                 &wd_smp_cpus_stuck);
         }
 }
+static void set_cpu_stuck(int cpu, u64 tb)
+{
+        set_cpumask_stuck(cpumask_of(cpu), tb);
+}
 
 static void watchdog_smp_panic(int cpu, u64 tb)
 {
@@ -140,9 +144,8 @@ static void watchdog_smp_panic(int cpu, u64 tb)
         }
         smp_flush_nmi_ipi(1000000);
 
-        /* Take the stuck CPU out of the watch group */
-        for_each_cpu(c, &wd_smp_cpus_pending)
-                set_cpu_stuck(c, tb);
+        /* Take the stuck CPUs out of the watch group */
+        set_cpumask_stuck(&wd_smp_cpus_pending, tb);
 
         wd_smp_unlock(&flags);