Has anybody attempted to use kexec with QEMU and -machine vexpress-a9? I'm getting this:
-------------------------- root@vexpress-qemu:~# kexec --load --command-line="$(cat /proc/cmdline)" --dtb=/boot/vexpress-v2p-ca9.dtb /boot/zImage-4.1.42-yocto-standard --debug Try gzip decompression. Try LZMA decompression. lzma_decompress_file: read on /boot/zImage-4.1.42-yocto-standard of 65536 bytes failed kernel: 0xb6a34008 kernel_size: 0x3561a0 MEMORY RANGES 0000000060000000-0000000067ffffff (0) zImage header: 0x016f2818 0x00000000 0x003561a0 zImage size 0x3561a0, file size 0x3561a0 kexec_load: entry = 0x60008000 flags = 0x280000 nr_segments = 2 segment[0].buf = 0xb6a34008 segment[0].bufsz = 0x3561a0 segment[0].mem = 0x60008000 segment[0].memsz = 0x357000 segment[1].buf = 0x56840 segment[1].bufsz = 0x37fa segment[1].mem = 0x610b8000 segment[1].memsz = 0x4000 root@vexpress-qemu:~# kexec --exec --debug [ 105.499955] kexec: Starting new kernel [ 105.505925] Disabling non-boot CPUs ... [ 105.506919] Bye! -------------------------- And it hangs there forever. I've experimented with some different address ranges, but this didn't help and they look quite similar to what U-Boot uses. This works if I try the same thing on the Beaglebone-hardware, so the kexec mechanism itself must be in decent shape. -- Kristian