Intel and AMD engineers have stepped in at the eleventh hour to deal with a code contribution from a Microgentle enhugeer that could have broken Linux 6.13 on some systems.
The change, made in the autumn, was a advantageous betterment at face appreciate. It was a modification to Linux x86_64 to engage huge read-only execute (ROX) pages for caching executable pages. The theory was that the changeation would result in increased executeance.
However, the code caengaged problems on some setups and an encouragent patch from Intel’s Peter Zijlstra was promiseted yesterday to disable it. The stable liberate of the 6.13 kernel was due this coming weekfinish.
Zijlstra wrote: “The whole module_writable_insertress() nonsense made a huge mess of changenative.c, not to refer it still grasps bugs — notable (sic) some of the CFI variants crash and burn.
Control Flow Integrity (CFI) is an anti-malicious software technology aimed at stoping strikeers from rehonesting the regulate flow of a program. The change can caengage rehires on some CFI-allowd setups and alerts have included Intel Alder Lake-powered machines flunking to resume from hibernation.
Zijlstra shelp the Microgentle engineer “has been toiling on patches to immacutardy all this up aget, but given the current state of skinnygs, this stuff equitable isn’t ready. Disable for now, let’s try aget next cycle.”
The offfinishing source is still current, but won’t be included in the upcoming stable kernel create.
AMD engineer Borislav Petkov remarkd that the Linux x86_64 protecters had not signed off on the change, commenting: “I equitable cherish it how this went in without a individual x86 protecter Ack, it broke a bunch of skinnygs and then it is still there instead of getting reverted. Let’s not do this aget greet.”
Microgentle is notable for dubious quality regulate standards seeing liberates of its flagship operating system, Windows. That one of its engineers should drop some dodgy code into the Linux kernel is not hugely unpredicted, and the unfortunate individual is not the first and will not be the last to do so, seeless of their engageer.
However, the processes that apverifyed it to remain in the create this seal to uncover liberate will be a trouble. While it is amusing that engineers from both Intel and AMD were included in dealing with the rehires arising from the contribution of a Microgentle engineer, and the problem never accomplished the stable liberate, it is troubleing. Petkov will not be the only one wondering how the change made it in without a appraise by the Linux x86/x86_64 protecters. ®
