On 11/14/19 12:50 PM, Oliver O'Halloran wrote:
On Thu, Nov 14, 2019 at 6:19 PM Madhavan Srinivasan
<ma...@linux.vnet.ibm.com> wrote:
When a root user or a user with CAP_SYS_ADMIN
privilege use trace_imc performance monitoring
unit events, to monitor application or KVM threads,
may result in a checkstop (System crash). Reason
being frequent switch of the "trace/accumulation"
mode of In-Memory Collection hardware.
This patch disables trace_imc pmu unit, but will
be re-enabled at a later stage with a fix patchset.
---
arch/powerpc/platforms/powernv/opal-imc.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/platforms/powernv/opal-imc.c
b/arch/powerpc/platforms/powernv/opal-imc.c
index e04b20625cb9..5790f078771f 100644
--- a/arch/powerpc/platforms/powernv/opal-imc.c
+++ b/arch/powerpc/platforms/powernv/opal-imc.c
@@ -285,7 +285,12 @@ static int opal_imc_counters_probe(struct platform_device
*pdev)
domain = IMC_DOMAIN_THREAD;
break;
case IMC_TYPE_TRACE:
- domain = IMC_DOMAIN_TRACE;
+ /* Using trace_imc events to monitor
+ * application or KVM thread performances
+ * may result in a checkstop (system crash).
+ * So disabling it for now.
+ */
+ domain = -1;
break;
default:
pr_warn("IMC Unknown Device type \n");
--
2.21.0
Does this need a Fixes: tag?
I was thinking of adding this commit as a fixes tag for fix patchset.
But if thats
not right, i can add the fixes tag along with a request to send to
Stable and post a v2
Maddy
