binutils 2.35 marks .text.ftrace_trampoline sections WAX, which due to a shortcoming in the arm64 module loader causes STRICT_MODULE_RWX to reject all module loading. The result is a kernel that can load none of the built modules. The kernel patch seems to have already been merged, but we should wait until it lands in all keyworded arm64 kernels until we keyword binutils-2.35 for this arch. Other architectures are (from what I've read) are not affected See https://lkml.org/lkml/2020/9/7/923
> Assignee: bug-wranglers@gentoo.org → toolchain@gentoo.org Can you clarify what toolchain@ should do here?
Nothing to do here, this is just to keep track so we really don't accidentally keyword it until the patch lands in all kernels
Is there a bug against gentoo's kerne team to get patches in one way or another? We would set it as a blocker to this bug.
The bug has been acknowledged upstream and the patch will get backported to LTS kernels - should I still put up a bug on gentoo-sources anyways?
I think so. Otherwise how do we track the blocker action?
New enough kernel is stable on arm64 in all branches.
