On Tue, Jul 30, 2019 at 05:17:56AM -0700, kernelci.org bot wrote:
next/master boot: 265 boots: 17 failed, 184 passed with 64 offline (next-20190730) Full Boot Summary: https://kernelci.org/boot/all/job/next/branch/master/kernel/next-20190730/ Full Build Summary: https://kernelci.org/build/next/branch/master/kernel/next-20190730/
For a while now all arm64 big endian clang built kernels have been failing, the kernel mounts the root filesystem but is unable to execute init due to an inability to understand the executable format:
08:55:25.999629 <6>[ 226.077194] Run /init as init process 08:55:31.066490 <4>[ 226.086518] request_module: kmod_concurrent_max (0) close to 0 (max_modprobes: 50), for module binfmt-464c, throttling... 08:55:31.085167 <4>[ 231.135458] request_module: modprobe binfmt-464c cannot be processed, kmod busy with 50 threads for more than 5 seconds now 08:55:35.745340 ShellCommand command timed out.: Sending # in case of corruption. Connection timeout 00:01:54, retry in 00:00:11 08:55:35.846536 # 08:55:35.849523 # 08:55:36.185339 <4>[ 231.154208] request_module: kmod_concurrent_max (0) close to 0 (max_modprobes: 50), for module binfmt-464c, throttling... 08:55:36.208673 <4>[ 236.255449] request_module: modprobe binfmt-464c cannot be processed, kmod busy with 50 threads for more than 5 seconds now 08:55:36.209013 <3>[ 236.269366] Failed to execute /init (error -8) 08:55:36.210161 <6>[ 236.285459] Run /sbin/init as init process 08:55:41.306737 <4>[ 236.294490] request_module: kmod_concurrent_max (0) close to 0 (max_modprobes: 50), for module binfmt-464c, throttling... 08:55:41.331547 <4>[ 241.375455] request_module: modprobe binfmt-464c cannot be processed, kmod busy with 50 threads for more than 5 seconds now 08:55:41.331837 <3>[ 241.389316] Starting init: /sbin/init exists but couldn't execute it (error -8)
(binfmt-464c is binfmt-misc, the fallback for unknown executable formats). The same kernel version built with GCC boots fine.
You can see a bunch of reports here (all the big endian failures):
https://kernelci.org/boot/all/job/next/branch/master/kernel/next-20190730/
It's possible that there's some infrastructure error that's causing the wrong ramdisk to be sent to the boards only for clang but I'd be a bit surprised.