Add power_envelope & soc_power sensors
#112
Conversation
trindenau
commented
Dec 23, 2024
•
trindenau
commented
lanserv/mellanox-bf/mlx-bf.sdrs
Outdated
| endsdr | ||
|
|
||
|
|
||
| #ddr_temp |
There was a problem hiding this comment.
Again, extra space and the comment should be #power_envelope. The comment can also be omitted.
|
Please add more description in the commit message |
trindenau
force-pushed
the
addNewSensors
branch
2 times, most recently
from
404f4f9 to
77c26c4
Compare
trindenau
force-pushed
the
addNewSensors
branch
from
77c26c4 to
adef23b
Compare
|
@trindenau - please update commit message |
| # Get SOC power info # | ||
| ################################### | ||
| #If the file doesn’t exists try to load the module | ||
| SOC_POWER_PATH="/sys/kernel/debug/mlxbf-ptm/monitors/status/total_power" |
There was a problem hiding this comment.
@trindenau - is this path defined in Yochai arch doc? if not we need to make sure he define the path
There was a problem hiding this comment.
This is the right path according to the arch doc
lanserv/mellanox-bf/set_emu_param.sh
Outdated
| #If the file doesn’t exists try to load the module | ||
| SOC_POWER_PATH="/sys/kernel/debug/mlxbf-ptm/monitors/status/total_power" | ||
| if [ ! -f "$SOC_POWER_PATH" ]; then | ||
| modprobe mlxbf-ptm |
There was a problem hiding this comment.
@trindenau - arch state the we need to load the driver? what happens in secure Linux where we can't load it or driver does not exist?
There was a problem hiding this comment.
No it doesn't I added it.
Do you want to get read of that or find a way to add it to secure linux?
| soc_power=$(cat "$SOC_POWER_PATH") | ||
| # Remove all the number after the decimal point – it can cause issues in the ipmb | ||
| if ! [[ "$soc_power" =~ ^-?[0-9]+(\.[0-9]+)?$ ]]; then | ||
| echo "Error: soc_power is not a valid number" |
lanserv/mellanox-bf/set_emu_param.sh
Outdated
| else | ||
| soc_power=$(cat "$SOC_POWER_PATH") | ||
| # Remove all the number after the decimal point – it can cause issues in the ipmb | ||
| if ! [[ "$soc_power" =~ ^-?[0-9]+(\.[0-9]+)?$ ]]; then |
There was a problem hiding this comment.
- power can be negative? do we have such requirement?
- the power reading is in decimal? not hex?
There was a problem hiding this comment.
The power is in decimal. and I got read of the minus
| ################################### | ||
| # Get power envelope info # | ||
| ################################### | ||
| POWER_ENVELOPE_PATH="/sys/kernel/debug/mlxbf-ptm/monitors/status/power_envelope" |
There was a problem hiding this comment.
@trindenau - same as above, this need to be defined in the arch
There was a problem hiding this comment.
This is the right path according to the arch doc
lanserv/mellanox-bf/set_emu_param.sh
Outdated
| POWER_ENVELOPE_PATH="/sys/kernel/debug/mlxbf-ptm/monitors/status/power_envelope" | ||
| if [ ! -f "$POWER_ENVELOPE_PATH" ]; then | ||
| #the module loaded in the soc_power routine | ||
| echo "Error: power_envelope file still not found after loading module" |
lanserv/mellanox-bf/set_emu_param.sh
Outdated
| ################################### | ||
| POWER_ENVELOPE_PATH="/sys/kernel/debug/mlxbf-ptm/monitors/status/power_envelope" | ||
| if [ ! -f "$POWER_ENVELOPE_PATH" ]; then | ||
| #the module loaded in the soc_power routine |
There was a problem hiding this comment.
@trindenau - this comment not clear? what does it means?
There was a problem hiding this comment.
I got read of that
lanserv/mellanox-bf/set_emu_param.sh
Outdated
| remove_sensor "soc_power" | ||
| else | ||
| soc_power=$(cat "$SOC_POWER_PATH") | ||
| # Remove all the number after the decimal point – it can cause issues in the ipmb |
There was a problem hiding this comment.
@trindenau - this is only a check it doesn't remove anything?
lanserv/mellanox-bf/set_emu_param.sh
Outdated
| remove_sensor "power_envelope" | ||
| else | ||
| power_envelope=$(cat "$POWER_ENVELOPE_PATH") | ||
| # Remove all the number after the decimal point – it can cause issues in the ipmb |
trindenau
force-pushed
the
addNewSensors
branch
5 times, most recently
from
2c3ac58 to
b0d700d
Compare
Add `power_envelope` and `soc_power` sensors.
Both will derive their values from `dbugfs`.
testd by: ipmitool -I ipmb sensors:
soc_power | 22.000 | Watts | ok | na | 5.000 | na | na | na | na
power_envelope | 65.000 | Watts | ok | na | na | 10.000 | 150.000 | na |na
Or by running the TestRedFishSensorSchema test.
Fixes jira https://redmine.mellanox.com/issues/4016386.
trindenau
force-pushed
the
addNewSensors
branch
from
b0d700d to
297b205
Compare
