[PATCH v2] powerpc/pseries: Enable hardlockup watchdog for PowerVM partitions

PowerVM will not arbitrarily oversubscribe or stop guests, page out the
guest kernel text to a NFS volume connected by carrier pigeon to abacus
based storage, etc., as a KVM host might. So PowerVM guests are not
likely to be killed by the hard lockup watchdog in normal operation,
even with shared processor LPARs which still get a minimum allotment of
CPU time.

Enable the hard lockup detector by default on !KVM guests, which we will
assume is PowerVM. It has been useful in finding problems on bare metal
kernels.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
v2: Fix 64e build by including kvm_guest.h

 arch/powerpc/kernel/setup_64.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index e42b85e4f1aa..428058dc5114 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c

@@ -33,6 +33,7 @@
 #include <linux/pgtable.h>
 
 #include <asm/debugfs.h>
+#include <asm/kvm_guest.h>
 #include <asm/io.h>
 #include <asm/kdump.h>
 #include <asm/prom.h>

@@ -939,16 +940,20 @@ u64 hw_nmi_get_sample_period(int watchdog_thresh)
  * disable it by default. Book3S has a soft-nmi hardlockup detector based
  * on the decrementer interrupt, so it does not suffer from this problem.
  *
- * It is likely to get false positives in VM guests, so disable it there
- * by default too.
+ * It is likely to get false positives in KVM guests, so disable it there
+ * by default too. PowerVM will not stop or arbitrarily oversubscribe
+ * CPUs, but give a minimum regular allotment even with SPLPAR, so enable
+ * the detector for non-KVM guests, assume PowerVM.
  */
 static int __init disable_hardlockup_detector(void)
 {
 #ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF
 	hardlockup_detector_disable();
 #else
-	if (firmware_has_feature(FW_FEATURE_LPAR))
-		hardlockup_detector_disable();
+	if (firmware_has_feature(FW_FEATURE_LPAR)) {
+		if (is_kvm_guest())
+			hardlockup_detector_disable();
+	}
 #endif
 
 	return 0;

-- 
2.23.0

On Wed, 23 Jun 2021 12:15:28 +1000, Nicholas Piggin wrote:

PowerVM will not arbitrarily oversubscribe or stop guests, page out the
guest kernel text to a NFS volume connected by carrier pigeon to abacus
based storage, etc., as a KVM host might. So PowerVM guests are not
likely to be killed by the hard lockup watchdog in normal operation,
even with shared processor LPARs which still get a minimum allotment of
CPU time.

[...]

Applied to powerpc/next.

[1/1] powerpc/pseries: Enable hardlockup watchdog for PowerVM partitions
      https://git.kernel.org/powerpc/c/633c8e9800f3884a26b2af59be8ce27696ad6ebf

cheers