Turbostat: extend the configuration options

This commit adds a number of configuration options that overide
which features are enabled by the automated mechanisms.
These options should only be used to turn off unwanted features
or add (temporarily) support for a new CPU

diff --git a/src/collectd.conf.in b/src/collectd.conf.in
index 078f2dd90cd34147bb38c14663e92401da20c1f3..a5b82ec6944a4a871260df35162036be9ce1b7ea 100644 (file)
--- a/src/collectd.conf.in
+++ b/src/collectd.conf.in
@@ -1189,8 +1189,22 @@
 #</Plugin>
 
 #<Plugin turbostat>
-##     The following option should not be used if MSR_IA32_TEMPERATURE_TARGET is supported
-#      TCCActivationTemp 100
+##     None of the following option should be set manually
+##     This plugin automatically detect most optimal options
+##     Only set values here if:
+##     - The module ask you to
+##     - You want to disable the collection of some data
+##     - Your (intel) CPU is not supported (yet) by the module
+##     - The module generate a lot of errors 'MSR offset 0x... read failed'
+##     In the last two cases, please open a bug request
+#
+#      TCCActivationTemp "100"
+#      CoreCstates "392"
+#      PackageCstates "396"
+#      SystemManagementInterrupt true
+#      DigitalTemperatureSensor true
+#      PackageThermalManagement true
+#      RunningAveragePowerLimit "7"    
 #</Plugin>
 
 #<Plugin unixsock>

diff --git a/src/collectd.conf.pod b/src/collectd.conf.pod
index 642d7994fe1c947d76a9c5fd2def6f28fc8dc221..f5425042a2d2921df992d802fbd7a1201cb22275 100644 (file)
--- a/src/collectd.conf.pod
+++ b/src/collectd.conf.pod
@@ -6421,18 +6421,76 @@ Default: B<1978>
 
 =back
 
-=head2 Plugin C<Turbostat>
+=head2 Plugin C<turbostat>
 
 The I<Turbostat plugin> reads CPU frequency and C-state residency on modern
 Intel processors by using the new Model Specific Registers.
 
 =over 4
 
+=item B<CoreCstates> I<Bitmask(Integer)>
+
+Bitmask of the list of core C states supported by the processor.
+This option should only be used if the automated detection fails.
+Default value extracted from the cpu model and family.
+
+Currently supported C-states (by this plugin): 3, 6, 7
+
+Example: (1<<3)+(1<<6)+(1<<7) = 392 for all states
+
+=item B<PackageCstates> I<Bitmask(Integer)>
+
+Bitmask of the list of pacages C states supported by the processor.
+This option should only be used if the automated detection fails.
+Default value extracted from the cpu model and family.
+
+Currently supported C-states (by this plugin): 2, 3, 6, 7, 8, 9, 10
+
+Example: (1<<2)+(1<<3)+(1<<6)+(1<<7) = 396 for states 2, 3, 6 and 7
+
+=item B<SystemManagementInterrupt> I<true>|I<false>
+
+Boolean enabling the collection of the I/O System-Management Interrupt
+counter'. This option should only be used if the automated detection
+fails or if you want to disable this feature.
+
+=item B<DigitalTemperatureSensor> I<true>|I<false>
+
+Boolean enabling the collection of the temperature of each core.
+This option should only be used if the automated detectionfails or 
+if you want to disable this feature.
+
+=item B<DigitalTemperatureSensor> I<true>|I<false>
+
+Boolean enabling the collection of the temperature of each package.
+This option should only be used if the automated detectionfails or 
+if you want to disable this feature.
+
 =item B<TCCActivationTemp> I<Temperature>
 
 Thermal Control Circuit Activation Temperature of the installed
-CPU. This option should only be used if the automated detection
-fails. Default value extracted from B<MSR_IA32_TEMPERATURE_TARGET>
+CPU. This temperature is used when collecting the temperature of
+cores or packages. This option should only be used if the automated
+detection fails. Default value extracted from B<MSR_IA32_TEMPERATURE_TARGET>
+
+=item B<RunningAveragePowerLimit> I<Bitmask(Integer)>
+
+Bitmask of the list of elements to be thermally monitored. This option
+should only be used if the automated detection fails or if you want to
+disable some collections. The different bits of this bitmask accepted
+by this plugin are:
+
+=over 4
+
+=item 0 ('1'): Package
+
+=item 1 ('2'): DRAM
+
+=item 2 ('4'): Cores
+
+=item 3 ('8'): Embedded graphic device
+
+=back
 
 =back
 

diff --git a/src/turbostat.c b/src/turbostat.c
index ec8757e7a6040a8cc8c4d8eea48acbae15526e88..5ee45186ade3c2f7855886cdb183f9377e705107 100644 (file)
--- a/src/turbostat.c
+++ b/src/turbostat.c
@@ -81,17 +81,23 @@ static _Bool aperf_mperf_unstable;
  * Currently supported C-states (by this plugin): 3, 6, 7
  */
 static unsigned int do_core_cstate;
+static unsigned int config_core_cstate;
+static _Bool apply_config_core_cstate;
 
 /*
  * Bitmask of the list of pacages C states supported by the processor.
  * Currently supported C-states (by this plugin): 2, 3, 6, 7, 8, 9, 10
  */
 static unsigned int do_pkg_cstate;
+static unsigned int config_pkg_cstate;
+static _Bool apply_config_pkg_cstate;
 
 /*
  * Boolean indicating if the processor supports 'I/O System-Management Interrupt counter'
  */
 static _Bool do_smi;
+static _Bool config_smi;
+static _Bool apply_config_smi;
 
 /*
  * Boolean indicating if the processor supports 'Digital temperature sensor'
@@ -102,6 +108,8 @@ static _Bool do_smi;
  *  - Temperatures above the tcc_activation_temp are not recorded
  */
 static _Bool do_dts;
+static _Bool config_dts;
+static _Bool apply_config_dts;
 
 /*
  * Boolean indicating if the processor supports 'Package thermal management'
@@ -112,6 +120,8 @@ static _Bool do_dts;
  *  - Temperatures above the tcc_activation_temp are not recorded
  */
 static _Bool do_ptm;
+static _Bool config_ptm;
+static _Bool apply_config_ptm;
 
 /*
  * Thermal Control Circuit Activation Temperature as configured by the user.
@@ -121,6 +131,8 @@ static _Bool do_ptm;
 static unsigned int tcc_activation_temp;
 
 static unsigned int do_rapl;
+static unsigned int config_rapl;
+static _Bool apply_config_rapl;
 static double rapl_energy_units;
 
 #define RAPL_PKG               (1 << 0)
@@ -218,7 +230,13 @@ static cdtime_t time_even, time_odd, time_delta;
 
 static const char *config_keys[] =
 {
+       "CoreCstates",
+       "PackageCstates",
+       "SystemManagementInterrupt",
+       "DigitalTemperatureSensor",
+       "PackageThermalManagement",
        "TCCActivationTemp",
+       "RunningAveragePowerLimit",
 };
 static const int config_keys_num = STATIC_ARRAY_SIZE (config_keys);
 
@@ -981,6 +999,20 @@ probe_cpu()
                return -1;
        }
 
+       /* Override detected values with configuration */
+       if (apply_config_core_cstate)
+               do_core_cstate = config_core_cstate;
+       if (apply_config_pkg_cstate)
+               do_pkg_cstate = config_pkg_cstate;
+       if (apply_config_smi)
+               do_smi = config_smi;
+       if (apply_config_dts)
+               do_dts = config_dts;
+       if (apply_config_ptm)
+               do_ptm = config_ptm;
+       if (apply_config_rapl)
+               do_rapl = config_rapl;
+
        if (do_rapl) {
                unsigned long long msr;
                if (get_msr(0, MSR_RAPL_POWER_UNIT, &msr))
@@ -1537,7 +1569,43 @@ turbostat_config(const char *key, const char *value)
        long unsigned int tmp_val;
        char *end;
 
-       if (strcasecmp("TCCActivationTemp", key) == 0) {
+       if (strcasecmp("CoreCstates", key) == 0) {
+               tmp_val = strtoul(value, &end, 0);
+               if (*end != '\0' || tmp_val > UINT_MAX) {
+                       ERROR("Turbostat plugin: Invalid CoreCstates '%s'",
+                             value);
+                       return -1;
+               }
+               config_core_cstate = (unsigned int) tmp_val;
+               apply_config_core_cstate = 1;
+       } else if (strcasecmp("PackageCstates", key) == 0) {
+               tmp_val = strtoul(value, &end, 0);
+               if (*end != '\0' || tmp_val > UINT_MAX) {
+                       ERROR("Turbostat plugin: Invalid PackageCstates '%s'",
+                             value);
+                       return -1;
+               }
+               config_pkg_cstate = (unsigned int) tmp_val;
+               apply_config_pkg_cstate = 1;
+       } else if (strcasecmp("SystemManagementInterrupt", key) == 0) {
+               config_smi = IS_TRUE(value);
+               apply_config_smi = 1;
+       } else if (strcasecmp("DigitalTemperatureSensor", key) == 0) {
+               config_dts = IS_TRUE(value);
+               apply_config_dts = 1;
+       } else if (strcasecmp("PackageThermalManagement", key) == 0) {
+               config_ptm = IS_TRUE(value);
+               apply_config_ptm = 1;
+       } else if (strcasecmp("RunningAveragePowerLimit", key) == 0) {
+               tmp_val = strtoul(value, &end, 0);
+               if (*end != '\0' || tmp_val > UINT_MAX) {
+                       ERROR("Turbostat plugin: Invalid RunningAveragePowerLimit '%s'",
+                             value);
+                       return -1;
+               }
+               config_rapl = (unsigned int) tmp_val;
+               apply_config_rapl = 1;
+       } else if (strcasecmp("TCCActivationTemp", key) == 0) {
                tmp_val = strtoul(value, &end, 0);
                if (*end != '\0' || tmp_val > UINT_MAX) {
                        ERROR("Turbostat plugin: Invalid TCCActivationTemp '%s'",