Enable boot FIVR fault monitoring and logging
A boot FIVR fault will assert the CPU FIVR Fault GPIO and then
assert the thermtrip GPIO. This adds a check for if the CPU
FIVR Fault GPIO is asserted to determine whether to log a boot
FIVR fault or a CPU thermal trip.
Tested:
Pulled the CPU FIVR Fault GPIO high and moved the jumper to assert
a thermtrip. After the system shut down, confirmed that the event
was logged as a boot FIVR fault.
Change-Id: Ic4292a3fa9135c7367764f2b126937e33c5ad652
Signed-off-by: Jason M. Bills <jason.m.bills@linux.intel.com>
diff --git a/src/host_error_monitor.cpp b/src/host_error_monitor.cpp
index 36b6384..27acf5e 100644
--- a/src/host_error_monitor.cpp
+++ b/src/host_error_monitor.cpp
@@ -58,8 +58,10 @@
static boost::asio::posix::stream_descriptor err2Event(io);
static gpiod::line smiLine;
static boost::asio::posix::stream_descriptor smiEvent(io);
+static gpiod::line cpu1FIVRFaultLine;
static gpiod::line cpu1ThermtripLine;
static boost::asio::posix::stream_descriptor cpu1ThermtripEvent(io);
+static gpiod::line cpu2FIVRFaultLine;
static gpiod::line cpu2ThermtripLine;
static boost::asio::posix::stream_descriptor cpu2ThermtripEvent(io);
static gpiod::line cpu1VRHotLine;
@@ -131,6 +133,15 @@
"REDFISH_MESSAGE_ARGS=%s", "SMI Timeout", NULL);
}
+static void cpuBootFIVRFaultLog(const int cpuNum)
+{
+ std::string msg = "Boot FIVR Fault on CPU " + std::to_string(cpuNum);
+
+ sd_journal_send("MESSAGE=HostError: %s", msg.c_str(), "PRIORITY=%i",
+ LOG_INFO, "REDFISH_MESSAGE_ID=%s", "OpenBMC.0.1.CPUError",
+ "REDFISH_MESSAGE_ARGS=%s", msg.c_str(), NULL);
+}
+
static void cpuThermTripLog(const int cpuNum)
{
std::string msg = "CPU " + std::to_string(cpuNum) + " thermal trip";
@@ -276,6 +287,30 @@
return true;
}
+static bool requestGPIOInput(const std::string& name, gpiod::line& gpioLine)
+{
+ // Find the GPIO line
+ gpioLine = gpiod::find_line(name);
+ if (!gpioLine)
+ {
+ std::cerr << "Failed to find the " << name << " line.\n";
+ return false;
+ }
+
+ // Request GPIO input
+ try
+ {
+ gpioLine.request({__FUNCTION__, gpiod::line_request::DIRECTION_INPUT});
+ }
+ catch (std::exception&)
+ {
+ std::cerr << "Failed to request " << name << " input\n";
+ return false;
+ }
+
+ return true;
+}
+
static void startPowerCycle()
{
conn->async_method_call(
@@ -668,7 +703,14 @@
static void cpu1ThermtripAssertHandler()
{
- cpuThermTripLog(1);
+ if (cpu1FIVRFaultLine.get_value() == 0)
+ {
+ cpuBootFIVRFaultLog(1);
+ }
+ else
+ {
+ cpuThermTripLog(1);
+ }
}
static void cpu1ThermtripHandler()
@@ -699,7 +741,14 @@
static void cpu2ThermtripAssertHandler()
{
- cpuThermTripLog(2);
+ if (cpu2FIVRFaultLine.get_value() == 0)
+ {
+ cpuBootFIVRFaultLog(2);
+ }
+ else
+ {
+ cpuThermTripLog(2);
+ }
}
static void cpu2ThermtripHandler()
@@ -1387,6 +1436,13 @@
return -1;
}
+ // Request CPU1_FIVR_FAULT GPIO input
+ if (!host_error_monitor::requestGPIOInput(
+ "CPU1_FIVR_FAULT", host_error_monitor::cpu1FIVRFaultLine))
+ {
+ return -1;
+ }
+
// Request CPU1_THERMTRIP GPIO events
if (!host_error_monitor::requestGPIOEvents(
"CPU1_THERMTRIP", host_error_monitor::cpu1ThermtripHandler,
@@ -1396,6 +1452,13 @@
return -1;
}
+ // Request CPU2_FIVR_FAULT GPIO input
+ if (!host_error_monitor::requestGPIOInput(
+ "CPU2_FIVR_FAULT", host_error_monitor::cpu2FIVRFaultLine))
+ {
+ return -1;
+ }
+
// Request CPU2_THERMTRIP GPIO events
if (!host_error_monitor::requestGPIOEvents(
"CPU2_THERMTRIP", host_error_monitor::cpu2ThermtripHandler,