On 28/02/17 21:35, Mark Cave-Ayland wrote: > On 28/02/17 21:28, Thomas Huth wrote: > >>> Maybe worth mentioning that the prom-env test for sparc64 has just been >>> enabled a couple of commits ago (6b591ad613010f13). There was an issue >>> with that test for the sparc64 target last year, but IIRC it only >>> affected 32-bit hosts and has been fixed, so I think/hope this is a >>> different issue now... >> >> OK, I think this is definitively not related to the prom-env test, but >> to MTTCG (or something else in TCG): I can reproduce this problem >> reliably when I run QEMU in s390 mode with the Moon Buggy disk image >> from the QEMU advent calendar >> (http://www.qemu-advent-calendar.org/2014/download/s390-moon-buggy.tar.xz): >> >> qemu-system-s390x -nographic -kernel s390-bb.kernel \ >> -initrd s390-moon-buggy.initrd >> >> ... starts the kernel, but aborts very soon with: >> >> ERROR:qemu/translate-common.c:34:tcg_handle_interrupt: assertion failed: >> (qemu_mutex_iothread_locked()) >> Aborted (core dumped) > > FWIW I've spent some time running the "make check" regression tests > locally this evening and I can't seem to reproduce the failure here on > my normal qemu-system-ppc/qemu-system-sparc/qemu-system-sparc64 setup.
However I can reproduce it consistently booting my OpenBIOS test images under qemu-system-sparc: $ ./qemu-system-sparc -cdrom debian-40r4a-sparc-netinst.iso -boot d qemu-system-sparc: -cdrom /home/build/src/qemu/image/sparc32/debian-40r4a-sparc-netinst.iso: warning: bus=0,unit=2 is deprecated with this machine type ** ERROR:/home/build/src/qemu/git/qemu/translate-common.c:34:tcg_handle_interrupt: assertion failed: (qemu_mutex_iothread_locked()) Aborted ATB, Mark.
