Am 08.04.2015 um 17:22 schrieb Peter Maydell: > On 8 April 2015 at 16:19, Andreas Färber <afaer...@suse.de> wrote: >> Hello, >> >> With rcutorture fixed, I am still seeing the following failure in -rc2. >> It is only for /x86_64/acpi/..., not for /i386/acpi/...; only on ppc >> host, not on ppc64 or ppc64le (so not endianness), nor on i586 (so not >> 32-bitness alone). >> >> Alex was suggesting it may be a ppc TCG problem, but I'm not seeing >> anything suspicious in this QTest since v2.2.0: >> >> 3a9c86d tests: bios-tables-test: add support for testing bridges >> 71f4be2 bios-tables-test: split piix4 and q35 tests >> b8e665e qtests: Specify image format explicitly > > This is basically the only test in "make check" that tries to > actually execute any TCG code, so it's a bit of a canary for > general TCG issues. If you try just running a random guest on > this QEMU (eg qemu-system-i386 and see if it can at least execute > the BIOS properly) does it work or crash?
I've resorted to setting up a ppc chroot on ppc64, and I can reproduce that qemu-system-i386 shows SeaBIOS okay, but qemu-system-x86_64 immediately aborts with the quoted error. Regards, Andreas -- SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Felix Imendörffer, Jane Smithard, Jennifer Guild, Dilip Upmanyu, Graham Norton; HRB 21284 (AG Nürnberg)
