Here Is Why The Linux 4.2 Kernel Is Messing Up On Some Ubuntu Systems
Written by Michael Larabel in Software on 13 July 2015. Page 2 of 2. 11 Comments

Finally this morning, the modified Phoronix Test Suite and Phoromatic had an answer: CONFIG_OF and friends. For some reason, the Ubuntu vanilla kernel packages have started enabling CONFIG_OF and its various sub-options by default for their x86/x86_64 kernel builds on Linux 4.2... This option is for OpenFirmware. While OpenFirmware was used by the original OLPC-XO, this option really isn't a big deal for x86 systems but rather ARM/SPARC/PowerPC. I have yet to see anywhere with reasoning why Ubuntu developers made this Kconfig change, but I can only imagine that it's due to trying to get a closer, more-universal kernel config across their supported architectures. Anyhow, when manually trying a kernel build with CONFIG_OF enabled, indeed on some 64-bit Intel/AMD systems it does cause issues at boot time. On other x86 systems, however, there weren't problems even with it enabled. There were also some false positives and strangeness around the Ubuntu mainline kernel packaging also enabling CONFIG_CPUFREQ_DT as built into their amd64 kernel even though this DeviceTree CPUFreq driver shouldn't really be needed/relevant as far as I know on x86 platforms.

This testing just finished recently, but will be exploring more this afternoon... So at the end of the day there's some new odd kernel configuration options enabled for the Ubuntu mainline kernel builds that they did while moving to Linux 4.2 Git; hopefully the Ubuntu developers will disable CONFIG_OF on x86 systems and some other bloating options. While CONFIG_OF may not be needed for x86, there does seem to be some issues in the kernel code itself given the odd behavior across different systems regardless of whether it makes sense enabling this option for unneeded systems. That's the information I know for now, stay tuned for updates.

For those interested in the changes to the Phoronix Test Suite / Phoromatic for improving the kernel testing, fallbacks on failed kernels, and the other modifications, those aren't being put into the open-source repository at this point. Prior to doing so a lot of the code would need to be cleaned-up, a proper plug-in architecture for the Phoromatic Server implemented in full for handling these more niche features, and cleaning up of various hacks when just trying to get the tests done this weekend -- it's basically just in a proof-of-concept state and could be made much better and enterprise-ready with enough support. For now it will be limited to just Phoronix Test Suite enterprise clients as future work items -- unless there's any sponsors interested for finishing up these capabilities for the open-source, public code. Organizations in need of commercial support, etc, for the Phoronix Test Suite and Phoromatic and OpenBenchmarking.org can contact us. If you're just an end-user and appreciate the work being done with our daily kernel benchmarking at LinuxBenchmarking.com and other efforts around automated Linux benchmarking, please consider subscribing to Phoronix Premium or making a PayPal tip or even Bitcoin.

If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal tips are also graciously accepted. Thanks for your support.


Related Articles
About The Author
Author picture

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter or contacted via MichaelLarabel.com.

Trending Linux News