Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: infinite messages from acpitz0



Just as a follow-up, let me try to explain why you are only now seeing this behavior...

This entry from src/doc/CHANGES is responsible:

        sysmon_envsys(8): Enhancements to allow access to driver-internal
                limit values [pgoyette 20090710]

After this change, certain drivers (including acpi_tz) were updated to automatically monitor their limit values. You no longer need to make an entry in /etc/envsys.conf to enable monitoring.

For acpi_tz, if a zone has a Critical temperature setting, that value is used as the CritMax value for sysmon_envsys; otherwise, we will use the Hot temperature setting if present. For the WarnMax value, we use the lowest of the thresholds associated with Active Cooling levels. So on your machine, the WarnMax limit is 50C, and you get the event messages when the actual temp crosses that boundary.

Hope this helps.


On Wed, 23 Dec 2009, Paul Goyette wrote:

Based on the ATZ_F_VERBOSE output generated by the patch I supplied, it looks to me like we are behaving exactly as we should!

On Thu, 24 Dec 2009, Frank Wille wrote:

Ok. Here is the log:

(PG: Snipping only the acpitz0 stuff)

Dec 24 00:18:44 nx6310 /netbsd: acpitz0 at acpi0 (TZ0): active cooling
       level 0: 80.0C active cooling level 1: 70.0C active cooling
       level 2: 60.0C active cooling level 3: 50.0C
Dec 24 00:18:44 nx6310 /netbsd: acpitz0: zone temperature is now 53.0C
Dec 24 00:18:44 nx6310 /netbsd: acpitz0: active cooling level 3

This is correct action - the temp meets or exceeds the level-3 active cooling threshold. We have executed the tz's power-down routine for level 3 (I have no clue what actually gets turned off here, it is handled in the BIOS).

Dec 24 00:19:04 nx6310 /netbsd: acpitz0: zone temperature is now 53.0C
Dec 24 00:19:04 nx6310 /netbsd: acpitz0: warning over limit on 'temperature'

Dec 24 00:19:34 nx6310 /netbsd: acpitz0: zone temperature is now 45.0C
Dec 24 00:19:34 nx6310 /netbsd: acpitz0: normal state on 'temperature'
Dec 24 00:19:34 nx6310 /netbsd: acpitz0: no active cooling level

OK, our cooling efforts were successful. Whatever gets powered down by level-3 was enough to get the temperature to drop. We turn everything back on.

Dec 24 00:20:04 nx6310 /netbsd: acpitz0: zone temperature is now 50.0C
Dec 24 00:20:04 nx6310 /netbsd: acpitz0: active cooling level 3
Dec 24 00:20:04 nx6310 /netbsd: acpitz0: warning over limit on 'temperature'

Unfortunately, the temp has gone back up again, and has reached the threshold. So we report it, and we go back into thermal-reduction power off mode again.

Dec 24 00:20:34 nx6310 /netbsd: acpitz0: zone temperature is now 50.0C
Dec 24 00:21:04 nx6310 /netbsd: acpitz0: zone temperature is now 45.0C
Dec 24 00:21:04 nx6310 /netbsd: acpitz0: normal state on 'temperature'
Dec 24 00:21:04 nx6310 /netbsd: acpitz0: no active cooling level
Dec 24 00:21:34 nx6310 /netbsd: acpitz0: zone temperature is now 50.0C
Dec 24 00:21:34 nx6310 /netbsd: acpitz0: active cooling level 3
Dec 24 00:21:34 nx6310 /netbsd: acpitz0: warning over limit on 'temperature'
Dec 24 00:22:04 nx6310 /netbsd: acpitz0: zone temperature is now 50.0C
Dec 24 00:22:34 nx6310 /netbsd: acpitz0: zone temperature is now 45.0C
Dec 24 00:22:34 nx6310 /netbsd: acpitz0: normal state on 'temperature'
Dec 24 00:22:34 nx6310 /netbsd: acpitz0: no active cooling level
Dec 24 00:23:04 nx6310 /netbsd: acpitz0: zone temperature is now 50.0C
Dec 24 00:23:04 nx6310 /netbsd: acpitz0: active cooling level 3
Dec 24 00:23:04 nx6310 /netbsd: acpitz0: warning over limit on 'temperature'

Do you need more to get a picture, or is it enough? It's strange that
the zone0 temperature jumps between 45C and 50C.

Well, not necessarily that strange. As soon as the temp reaches the 50C threshold, we take action to manage the thermal load. This is the whole point behind "active cooling". The actions taken obviously work, since the temp quickly drops back down below the threshold.

The only thing strange I see is that the values are always EXACTLY 45C and 50C. Never anything in between. But that could simply be a result of the manufacturer's choice of sensor. (You said earlier that this problem goes away when you are running on battery power instead of on mains; what values do you get for acpitz0 under those conditions?)

I would suggest that perhaps you want to enable powerd (in rc.conf) and then tailor the entries in /etc/powerd/scripts to filter out the syslog messages from acpitz0. (Without enabling powerd, these messages are logged directly by the kernel with no chance to filter them.)

But anyway, thanks for having this problem - it did enable me to find a couple other issues (one with acpitz and another with sysmon). :)


-------------------------------------------------------------------------
|   Paul Goyette   | PGP DSS Key fingerprint: |  E-mail addresses:      |
| Customer Service | FA29 0E3B 35AF E8AE 6651 |  paul at whooppee.com   |
| Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette at juniper.net |
| Kernel Developer |                          | pgoyette at netbsd.org  |
-------------------------------------------------------------------------


-------------------------------------------------------------------------
|   Paul Goyette   | PGP DSS Key fingerprint: |  E-mail addresses:      |
| Customer Service | FA29 0E3B 35AF E8AE 6651 |  paul at whooppee.com   |
| Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette at juniper.net |
| Kernel Developer |                          | pgoyette at netbsd.org  |
-------------------------------------------------------------------------


Home | Main Index | Thread Index | Old Index