Laying To Rest That Odd Linux Kernel Regression
Former Red Hat employee Dave Jones has provided some closure to that Linux 3.18 kernel bug that was initially viewed as a "worrisome regression" and turned out to be very difficult to track with no official fix within the mainline Linux kernel.
The bug wasn't fixed for Linux 3.18 final but various other bugs / potentially bad code was cleaned-up in the process of tracking down and isolating this lock-up issue that Dave Jones first reported on one of his systems. The bug went unresolved and at the end of December is when Dave Jones left Red Hat and had to return his hardware -- including the affected system.
The last update I wrote on this issue was that That Peculiar Linux 3.18 Kernel Bug Might Be Closed Soon. Dave wrote a final update on the bug today to offer some closure but its exact cause remains undetermined.
It appears that the issue was related to the system's High Precision Event Timer (HPET) being scribbled, but what was causing this on the recent kernel code is unanswered. Dave views the possible explanations as either his Trinity testing program doing something very odd to write over the HPET or that his system was affected by a hardware bug. The likely explanation is the latter that he was affected by some weird hardware issue that just began to show itself on recent kernels.
Dave's final comments can be read via this blog post. In a separate post he also laid out yesterday how his involvement with various open-source projects will be impacted now that he's let Red Hat: his upstream testing won't continue at the same level as in the past, he'll keep running Coverity scans of the kernel, he will no longer be maintaining his Fedora packages, and other Fedora developers have picked up his Fedora kernel work.