On 08/14/2017 05:47 PM, Thomas Huth wrote: > On 14.08.2017 17:33, Cornelia Huck wrote: >> On Mon, 14 Aug 2017 17:34:15 +0300 >> "Michael S. Tsirkin" <m...@redhat.com> wrote: >> >>> Speed up tests on host systems with kvm support. >>> In particular, this fixes tests with --disable-tcg. > [...] >>> diff --git a/tests/pnv-xscom-test.c b/tests/pnv-xscom-test.c >>> index 5adc3fd..7be396f 100644 >>> --- a/tests/pnv-xscom-test.c >>> +++ b/tests/pnv-xscom-test.c >>> @@ -84,7 +84,7 @@ static void test_cfam_id(const void *data) >>> char *args; >>> const PnvChip *chip = data; >>> >>> - args = g_strdup_printf("-M powernv,accel=tcg -cpu %s", >>> chip->cpu_model); >>> + args = g_strdup_printf("-M powernv,accel=kvm:tcg -cpu %s", >>> chip->cpu_model); >>> >>> qtest_start(args); >>> test_xscom_cfam_id(chip); >>> @@ -112,7 +112,7 @@ static void test_core(const void *data) >>> char *args; >>> const PnvChip *chip = data; >>> >>> - args = g_strdup_printf("-M powernv,accel=tcg -cpu %s", >>> chip->cpu_model); >>> + args = g_strdup_printf("-M powernv,accel=kvm:tcg -cpu %s", >>> chip->cpu_model); >>> >>> qtest_start(args); >>> test_xscom_core(chip); >> >> These seem to test a specified cpu model explicitly, so I'm not sure >> how well this will work with kvm. > > AFAIK the powernv machine does not work with KVM yet (Cédric, can you > confirm?), so you better drop the changes to this file.
No. the powernv machine does not work with KVM. C. >>> diff --git a/tests/prom-env-test.c b/tests/prom-env-test.c >>> index eac207b..0a10f06 100644 >>> --- a/tests/prom-env-test.c >>> +++ b/tests/prom-env-test.c >>> @@ -50,7 +50,7 @@ static void test_machine(const void *machine) >>> /* The pseries firmware boots much faster without the default devices >>> */ >>> extra_args = strcmp(machine, "pseries") == 0 ? "-nodefaults" : ""; >>> >>> - args = g_strdup_printf("-M %s,accel=tcg %s -prom-env >>> 'use-nvramrc?=true' " >>> + args = g_strdup_printf("-M %s,accel=kvm:tcg %s -prom-env >>> 'use-nvramrc?=true' " >>> "-prom-env 'nvramrc=%x %x l!' ", >>> (const char *)machine, extra_args, MAGIC, >>> ADDRESS); >> >> The header explicitly states that this is supposed to test tcg mode. > > This change should be OK here. But it is currently not necessary yet: > The test only runs on SPARC and PPC machines, and both do not support > the --disable-tcg switch yet. > > Thomas >