Latest Linux kernel release candidate was a sticky mess
4.14 rc2 tarball trapped the unwary with missed patch
Linus Torvalds is not noted as having the most even of tempers, but after a weekend spent scuba diving a glitch in the latest Linux kernel release candidate saw the Linux overlord merely label the mess "nasty".
The release cycle was following its usual cadence when Torvalds announced Linux 4.14 release candidate 2, just after 5:00PM on Sunday, September 24th.
It was a mostly unremarkable release, Torvalds told the Linux Kernel Mailing List: driver updates, fixes to most key architectures (x86, MIPS, System 390, PA-RISC, PowerPC and ARM) and the like.
“Nothing stands out, although hopefully we've gotten over all the x86 ASID issues,” he wrote, “Knock wood” [ASID is the Address Space Identifier - Ed].
“Go test it”, he concluded. And the Linux faithful did, but quickly, as Miss Clavel of Madeline fame might put it, found “something's not quite right”.
Randy Dunlap reported the problem first, noting the tarball “gives me a file that does not apply cleanly to v4.13”.
Linus agreed, flicking the question to Konstantin Ryabitsev, who solved it and took the hit for it by posting the following:
“Ick, looks like the cache truncation bug is still not fixed in the upstream cgit and the EPEL update blew away my locally fixed package (which is my fault, as I should have versionlocked it). The fixed version is now in place and truncated caches are purged. It should generate valid patches now.”
Which, after all, is what testing is for. And Linus' strongest language wasn't of the error, but at the notion that anybody would still be running patches the hard way rather than installing Git.
“I'm surprised that people still even use those nasty patches and tar-balls”, he wrote. ®