You are not logged in.
Hey graysky, the link is down. Could you please reinstatement the link for another day or two? It looks like I will have to stay on 3.8 for some time.
Is there a pre-3.4-post-aspm/power-regression/-stable-kernel that will support modules like thinkpad_acpi, tp_smapi, bbswitch, etc. I tried linux-lts (3.0) which is too old and wouldn't let me compile these modules.
Offline
Has anyone noticed bigger power consumption (reduced battery time)/temperatures on laptops (preferably with i915 gpu) with kernel > 3.7.3 ? Is it only my idea?
Last edited by firewalker (2012-06-20 14:43:16)
Γίνε ρεαλιστής, μείνε ονειροπόλος ...
Offline
Offline
Yes, it seems to be related.
Γίνε ρεαλιστής, μείνε ονειροπόλος ...
Offline
There is an issue with 3.4 kernel. I have the same issue.
Back to 3.3.7 for me.
I will try a git-bisect on my free time.
Γίνε ρεαλιστής, μείνε ονειροπόλος ...
Offline
merging
There's no such thing as a stupid question, but there sure are a lot of inquisitive idiots !
Offline
There is an issue with 3.4 kernel. I have the same issue.
Back to 3.3.7 for me.
I will try a git-bisect on my free time.
Γίνε ρεαλιστής, μείνε ονειροπόλος ...
Offline
??? what's the point in posting the same thing over and over again?
There's no such thing as a stupid question, but there sure are a lot of inquisitive idiots !
Offline
Maybe related with this?
http://www.phoronix.com/scan.php?page=n … px=MTA3MzM
Did someone opened an upstream bug report?
Last edited by firewalker (2012-06-22 10:14:29)
Γίνε ρεαλιστής, μείνε ονειροπόλος ...
Offline
I can confirm this bug, and not restricted to Intel.
AMD Phenom II 965BE
Following a (as I thought) unrelated thread regarding load monitoring in general, this thread was brought to my attention. The only none 3.4 kernels in AUR are ice and fedora. ice is 3.2, and fedora is 3.3. I elected to install ice, booted and average load (queueing) has dropped by 33%, temps have dropped 9C compared with 3.4. I can also confirm that it's a 3.4 kernel issue, and not ck specifically (I personally prefer the ck kernels - may I have a 3.2/3.3 binary). I'll keep my eye on things with the 3.2 ice kernel for the next 24 hours, but honestly 24 minutes tells me enough.
Ryzen 5900X 12 core/24 thread - RTX 3090 FE 24 Gb, Asus B550-F Gaming MB, 128Gb Corsair DDR4, Fractal Design Define 7 XL, 5 HD (2 NvME PCI, 4SSD) + 1 x optical.
Linux user #545703
/ is the root of all problems.
Offline
@donniezazen I have a T420s which has very similar hardware compared to your machine (i7 2620, 8GB RAM, Intel 3000, NVIDIA Optimus 1GB). Which kernel do you have currently installed, what powersaving masures have you taken so far and how long does your battery last? (I assume your run a 6-Cell battery?)
Offline
I updated to 3.4.4-2-ARCH and I havn't noticed anything so far.
Athlon X2 240.
I use the powernow-k8 driver with the ondemand cpu governor to scale cpu freq.
It generally stays @ 800MHz and the temp is alright ~ 43 deg. cel.
At higher load(75-100%) when its clocked to the max(2.8G), it generally stays @65-85 deg. cel.
But thats normal for my system from what I've seen.
Last edited by debdj (2012-07-04 16:36:20)
Offline
@delexi Basically, bug started around late 3.3 or early 3.4. So, anything before that is good. I have not looked into 'how to compile kernel' so that I can get 3.3. I am instead using linux-ice which is 3.2. Optimus won't work with it. Other than that it's a solid kernel.
Here is my kernel line and rc.conf module and daemon line. I use no other modifications.
kernel /vmlinuz-linux-ice root=/dev/sda5 ro i915.i915_enable_rc6=1 i915.i915_enable_fbc=1 i915.i915_enable_fbc=1 i915.lvds_downclock=1 drm.vblankoffdelay=1MODULES=(thinkpad_acpi tp_smapi)DAEMONS=(hwclock syslog-ng dbus networkmanager netfs crond @adb @avahi-daemon @cupsd @bumblebeed @sshd @ntpd @alsa @acpid @sensors @hddtemp @thinkfan @laptop-mode lightdm)Offline
I updated to 3.4.4-2-ARCH and I havn't noticed anything so far.
Open a terminal and run:
uptimeHave a look at the three numbers at the end. If they are more than 1.0 then you may have the problem and just not know it. You would usually expect the first to be < 0.2 on an idle system.
Ryzen 5900X 12 core/24 thread - RTX 3090 FE 24 Gb, Asus B550-F Gaming MB, 128Gb Corsair DDR4, Fractal Design Define 7 XL, 5 HD (2 NvME PCI, 4SSD) + 1 x optical.
Linux user #545703
/ is the root of all problems.
Offline
@roken
You're right. I seem to have the problem too. A bit strange though.
When I have firefox and some terms open it's ~ 0.79 - 0.87. With completely idle it goes above 1.
EDIT: the man page of uptime and http://en.wikipedia.org/wiki/Load_%28computing%29 provides some insight on load avarage .
The above numbers were when my system was running @800MHz
With the same apps running it dropped to 0.37 - 0.48 (for the last minute) @2.8GHz
Still its a bit higher than usual.
Last edited by debdj (2012-07-05 13:38:40)
Offline
I had the same problem. I've find out on my laptop that problem occurs only on AC. After disconnecting AC temperature immediately goes down. I've also found that bfq scheduler was to blame. After switching to cfq problem disappears.
Last edited by b00rt00s (2012-07-31 01:15:21)
Offline
Offline
@donniezazen After longer observation It looks like the same happens on cfq and bfq. Also my problems where caused by linux-ck. On [core] kernel I don't see any difference between 3.4.7 kernel and 3.0.38. But after looking at the first post in the thread I see You are also using linux-ck. Maybe the patchset is the problem?
Offline
My laptop uses ~3W more with linux-ck over linux package. Powertop reports that audio codec is always active(100% usage) with -ck and shows estimated audio codec power usage ~2watts yet i can't force it to suspend. With linux audio codec is suspended when not used.
Offline