Is there any evidence or technical documentation to support the notion that System Integrity Protection can cause excessive use of the CPU by kernel_task in some situation?
Background
No battery. All Apple operating systems on an external hard disk drive, limited (by the MacBookPro8,2) to USB 2.0.
With Sierra, and with pre-release High Sierra, when not booted in safe mode I nearly always found that kernel_task hogged the CPU.
My usual workaround was safe mode.
After considering answers to the following questions:
- kernel_task consumes 500% CPU on AC power, sometimes (2012-11-24)
- How to disable the SpeedStep when using MacBook Pro without a battery? (2014-01-06)
- Kernel_task boosts CPU to 300%-500% (Yosemitte MBP 2011 without a battery) (2015-02-04)
- kernel_task hundreds of % CPU yet cpu frequency is being lowered (2015-04-01)
- Very high kernel_task CPU usage after updating to Yosemite 10.10.4 (2015-07-04)
- Macbook Pro with abnormal CPU usage and unrecognised battery. How to alternative SMC reset (2015-09-02)
– I aimed to move (set aside) the following file:
/System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/PlugIns/ACPI_SMC_PlatformPlugin.kext/Contents/Resources/MacBookPro8_2.plist
Prerequisite to the move:
- disable System Integrity Protection (SIP).
After using csrutil(1) in Recovery OS 17A264c to disable SIP, I booted High Sierra in normal mode …
… I was pleasantly susprised to find that with the .plist still in place:
- in normal mode without SIP, the CPU was no longer hogged.
Shut down, started, no hogging.
Started 10.12, no hogging, updated to 10.12.5, restarted, no hogging.
2017-06-10 around 16:50 the Mac unexpectedly stopped. When started, to 10.12.5, I noticed that SIP was re-enabled (not by me). No hogging.
Started Recovery OS 17A264c, disabled SIP, restarted to 10.12.5, no hogging …
