Bug 77751 - radeon: unable to set temperature warning levels
Summary: radeon: unable to set temperature warning levels
Status: NEW
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel) (show other bugs)
Hardware: All Linux
: P1 low
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-06-12 22:18 UTC by higuita
Modified: 2016-03-23 18:56 UTC (History)
1 user (show)

See Also:
Kernel Version: 3.15
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description higuita 2014-06-12 22:18:50 UTC
I have a old Radeon HD2600XT AGP card, but i think this should apply to all AMD/ATI cards.

Right now, when i use the "sensors" command i get this

radeon-pci-0100
Adapter: PCI adapter
temp-GPU:     +53.0°C  (crit = +120.0°C, hyst = +90.0°C)

But the values for critical and warning are too high for this card, specially as the fan is unable to run as fast as in the past. To me, right now, when the card reaches about 84ºC it start to lockup and force me to reboot to recover.

So i tried to configure the sensors.conf, but when running "sensors -s" i get:

Error: File /etc/sensors.d/local.conf, line 87: Failed to set value
Error: File /etc/sensors.d/local.conf, line 88: Failed to set value
radeon-pci-0100: At least one "set" statement failed

i have this configured:

chip "radeon-pci-*"
	label temp1 "temp-GPU"
	set temp1_crit 82
	set temp1_crit_hyst 79

Checking the /sys entry to try to manually set the values i see this:

# ls -l /sys/class/hwmon/hwmon1/
total 0
lrwxrwxrwx 1 root root    0 Jun 12 17:20 device -> ../../../0000:01:00.0
-r--r--r-- 1 root root 4096 Jun 12 22:14 name
drwxr-xr-x 2 root root    0 Jun 12 22:14 power
lrwxrwxrwx 1 root root    0 Jun 12 17:20 subsystem -> ../../../../../../class/hwmon
-r--r--r-- 1 root root 4096 Jun 12 22:14 temp1_crit
-r--r--r-- 1 root root 4096 Jun 12 22:14 temp1_crit_hyst
-r--r--r-- 1 root root 4096 Jun 12 22:14 temp1_input
-rw-r--r-- 1 root root 4096 Jun 12 22:14 uevent

So the temp1_crit and temp1_crit_hyst are readonly, that is why it fails to setup new values and we can only use hardcoded values.


it would be a good idea to allow one to change this value, and even better if one could automatically slow down the card when it reach the critical level, to avoid the crash.

Note You need to log in before you can comment on or make changes to this bug.