Bug#838206: linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle
Control: tag -1 moreinfo On Sun, 18 Sep 2016 19:08:32 +0200 Alessio wrote: > 2016-09-18 15:44 GMT+02:00 Ben Hutchings : > > Control: tag -1 moreinfo > > > > On Sun, 2016-09-18 at 14:46 +0200, Alessio Gaeta wrote: > >> Package: src:linux > >> Version: 4.7.2-1+s1 > >> Severity: important > >> > >> Hello, > >> > >> I'm facing a power issue with the Linux 4.7 kernel. > >> > >> Every time the system goes idle (the configured delay passes and the > >> display > >> goes black), the frequency of all cores get maxed, even if all cores are > >> idle. > >> > >> Evident symptoms are the fan spinning loudly and the errors logged in > >> journal > >> stating the thermal threshold being surpassed and the CPU throttled. > >> > >> In detail, the frequency of all cores raises from ~800 MHz to 2.6 GHz for > >> all 8 > >> cores of my i7. I can reproduce the behavior 100% of times, by connecting > >> via > >> ssh to the affected system and launching powertop. > >> > >> The issue does not exists in Linux 4.6 (to wich I will stick for now), so I > >> guess the problem resides in kernel, not in other packages. > >> > >> You have all system information here below (gathered by repotbug). Ask any > >> further info if needed. Thank you! > > [...] > > > > Please test with bbswitch and acpi_call removed. > > Hello, > > tested again: > > 1. systemctl stop bumblebeed.service > 2. modprobe -r bbswitch > 3. modprobe -r acpi_call > 4. systemctl stop vboxweb.service > 5. modprobe -r vboxpci vboxnetflt vboxnetadp vboxdrv > > (removed all out-of-tree modules, just to be sure). > > Same result: on display idle the CPU frequency got maxed on all cores. > > Tell me if you need more info. OK, so something (other than those modules) is keeping the CPUs busy. Please install linux-perf and use perf (as root) to show what it's busy with: perf record -a sleep 10 # creates 'perf.data' in current dir perf report > perf_report.txt # will show some warnings; don't worry about them Ben. -- Ben Hutchings No political challenge can be met by shopping. - George Monbiot signature.asc Description: This is a digitally signed message part
Processed: Re: Bug#838206: linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle
Processing control commands: > tag -1 moreinfo Bug #838206 [src:linux] linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle Added tag(s) moreinfo. -- 838206: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=838206 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems
Bug#838206: linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle
2016-09-18 15:44 GMT+02:00 Ben Hutchings : > Control: tag -1 moreinfo > > On Sun, 2016-09-18 at 14:46 +0200, Alessio Gaeta wrote: >> Package: src:linux >> Version: 4.7.2-1+s1 >> Severity: important >> >> Hello, >> >> I'm facing a power issue with the Linux 4.7 kernel. >> >> Every time the system goes idle (the configured delay passes and the display >> goes black), the frequency of all cores get maxed, even if all cores are >> idle. >> >> Evident symptoms are the fan spinning loudly and the errors logged in journal >> stating the thermal threshold being surpassed and the CPU throttled. >> >> In detail, the frequency of all cores raises from ~800 MHz to 2.6 GHz for >> all 8 >> cores of my i7. I can reproduce the behavior 100% of times, by connecting via >> ssh to the affected system and launching powertop. >> >> The issue does not exists in Linux 4.6 (to wich I will stick for now), so I >> guess the problem resides in kernel, not in other packages. >> >> You have all system information here below (gathered by repotbug). Ask any >> further info if needed. Thank you! > [...] > > Please test with bbswitch and acpi_call removed. Hello, tested again: 1. systemctl stop bumblebeed.service 2. modprobe -r bbswitch 3. modprobe -r acpi_call 4. systemctl stop vboxweb.service 5. modprobe -r vboxpci vboxnetflt vboxnetadp vboxdrv (removed all out-of-tree modules, just to be sure). Same result: on display idle the CPU frequency got maxed on all cores. Tell me if you need more info.
Processed: Re: Bug#838206: linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle
Processing control commands: > tag -1 moreinfo Bug #838206 [src:linux] linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle Added tag(s) moreinfo. -- 838206: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=838206 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems
Bug#838206: linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle
Control: tag -1 moreinfo On Sun, 2016-09-18 at 14:46 +0200, Alessio Gaeta wrote: > Package: src:linux > Version: 4.7.2-1+s1 > Severity: important > > Hello, > > I'm facing a power issue with the Linux 4.7 kernel. > > Every time the system goes idle (the configured delay passes and the display > goes black), the frequency of all cores get maxed, even if all cores are idle. > > Evident symptoms are the fan spinning loudly and the errors logged in journal > stating the thermal threshold being surpassed and the CPU throttled. > > In detail, the frequency of all cores raises from ~800 MHz to 2.6 GHz for all > 8 > cores of my i7. I can reproduce the behavior 100% of times, by connecting via > ssh to the affected system and launching powertop. > > The issue does not exists in Linux 4.6 (to wich I will stick for now), so I > guess the problem resides in kernel, not in other packages. > > You have all system information here below (gathered by repotbug). Ask any > further info if needed. Thank you! [...] Please test with bbswitch and acpi_call removed. Ben. -- Ben Hutchings Klipstein's 4th Law of Prototyping and Production: A fail-safe circuit will destroy others. signature.asc Description: This is a digitally signed message part
Bug#838206: linux-image-4.7.0-1-amd64: [Linux Kernel 4.7] Power issue: Intel i7 CPU frequency get maxed on display idle
Package: src:linux Version: 4.7.2-1+s1 Severity: important Hello, I'm facing a power issue with the Linux 4.7 kernel. Every time the system goes idle (the configured delay passes and the display goes black), the frequency of all cores get maxed, even if all cores are idle. Evident symptoms are the fan spinning loudly and the errors logged in journal stating the thermal threshold being surpassed and the CPU throttled. In detail, the frequency of all cores raises from ~800 MHz to 2.6 GHz for all 8 cores of my i7. I can reproduce the behavior 100% of times, by connecting via ssh to the affected system and launching powertop. The issue does not exists in Linux 4.6 (to wich I will stick for now), so I guess the problem resides in kernel, not in other packages. You have all system information here below (gathered by repotbug). Ask any further info if needed. Thank you! -- Alessio Gaeta -- Package-specific info: ** Version: Linux version 4.7.0-1-amd64 (debian-kernel@lists.debian.org) (gcc version 5.4.1 20160803 (Debian 5.4.1-1) ) #1 SMP Debian 4.7.2-1 (2016-08-28) ** Command line: BOOT_IMAGE=/boot/vmlinuz-4.7.0-1-amd64 root=UUID=325f7845-cb20-4f45-9328-6a9ce9133f5a ro acpi_backlight=vendor quiet ** Tainted: OE (12288) * Out-of-tree module has been loaded. * Unsigned module has been loaded (currently expected). ** Kernel log: [4.649062] iTCO_wdt: Intel TCO WatchDog Timer Driver v1.11 [4.649104] iTCO_wdt: Found a Cougar Point TCO device (Version=2, TCOBASE=0x0460) [4.649285] iTCO_wdt: initialized. heartbeat=30 sec (nowayout=0) [4.650851] asus_wmi: ASUS WMI generic driver loaded [4.652548] asus_wmi: Initialization: 0x1 [4.652753] asus_wmi: SFUN value: 0x1a0af7 [4.653693] input: Asus WMI hotkeys as /devices/platform/asus-nb-wmi/input/input17 [4.662844] asus_wmi: Number of fans: 0 [4.669120] acpi_call: module verification failed: signature and/or required key missing - tainting kernel [4.677496] Non-volatile memory driver v1.3 [4.718296] ath: phy0: Enable LNA combining [4.720944] ath: phy0: ASPM enabled: 0x42 [4.720953] ath: EEPROM regdomain: 0x60 [4.720956] ath: EEPROM indicates we should expect a direct regpair map [4.720961] ath: Country alpha2 being used: 00 [4.720964] ath: Regpair used: 0x60 [4.725655] ieee80211 phy0: Selected rate control algorithm 'minstrel_ht' [4.725958] ieee80211 phy0: Atheros AR9285 Rev:2 mem=0xc900010c, irq=17 [4.750243] [drm] Memory usable by graphics device = 2048M [4.750248] [drm] Replacing VGA console driver [4.750890] Console: switching to colour dummy device 80x25 [4.756320] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [4.756323] [drm] Driver supports precise vblank timestamp query. [4.758406] vgaarb: device changed decodes: PCI::00:02.0,olddecodes=io+mem,decodes=none:owns=io+mem [4.758974] intel_rapl: Found RAPL domain package [4.758977] intel_rapl: Found RAPL domain core [4.758979] intel_rapl: Found RAPL domain uncore [4.778466] ACPI: Video Device [GFX0] (multi-head: yes rom: yes post: no) [4.778897] input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:4c/LNXVIDEO:00/input/input18 [4.779487] ACPI: Video Device [GFX0] (multi-head: yes rom: no post: no) [4.779909] input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/LNXVIDEO:01/input/input19 [4.784877] EXT4-fs (sda2): mounted filesystem with ordered data mode. Opts: (null) [4.892652] snd_hda_intel :00:1b.0: bound :00:02.0 (ops i915_audio_component_bind_ops [i915]) [4.892659] [drm] Initialized i915 1.6.0 20160425 for :00:02.0 on minor 0 [4.893710] fbcon: inteldrmfb (fb0) is primary device [4.923758] random: nonblocking pool is initialized [4.982716] EXT4-fs (sdb5): mounted filesystem with ordered data mode. Opts: (null) [5.099244] Console: switching to colour frame buffer device 170x48 [5.101549] i915 :00:02.0: fb0: inteldrmfb frame buffer device [5.128538] input: HDA Intel PCH Mic as /devices/pci:00/:00:1b.0/sound/card0/input15 [5.128596] input: HDA Intel PCH Headphone as /devices/pci:00/:00:1b.0/sound/card0/input16 [5.128647] input: HDA Intel PCH HDMI/DP,pcm=3 as /devices/pci:00/:00:1b.0/sound/card0/input20 [5.309850] bbswitch: version 0.8 [5.309860] bbswitch: Found integrated VGA device :00:02.0: \_SB_.PCI0.GFX0 [5.309866] bbswitch: Found discrete VGA device :01:00.0: \_SB_.PCI0.PEGR.GFX0 [5.309881] ACPI Warning: \_SB.PCI0.PEGR.GFX0._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20160422/nsarguments-95) [5.310017] bbswitch: detected an Optimus _DSM function [5.310031] pci :01:00.0: enabling device ( -> 0003) [5.310090] bbswitch: Succesfully loaded. Discrete card :01:00.0 is on [5.310788] bbswitch: disabling discrete graphics [5.310798] ACPI Warning: \_SB