Richard Henderson <richard.hender...@linaro.org> writes: > On 4/27/23 14:16, Fabiano Rosas wrote: >> Richard Henderson <richard.hender...@linaro.org> writes: >> >>> On 4/26/23 19:00, Fabiano Rosas wrote: >>>> We're about to move the 32-bit CPUs under CONFIG_TCG, so adjust the >>>> query-cpu-model-expansion test to check against the cortex-a7, which >>>> is already under CONFIG_TCG. That allows the next patch to contain >>>> only code movement. >>>> >>>> While here add comments clarifying what we're testing. >>>> >>>> Signed-off-by: Fabiano Rosas<faro...@suse.de> >>>> Suggested-by: Philippe Mathieu-Daudé<phi...@linaro.org> >>>> --- >>>> tests/qtest/arm-cpu-features.c | 20 +++++++++++++++++--- >>>> 1 file changed, 17 insertions(+), 3 deletions(-) >>> >>> I don't see why you're changing the cpu model here. >>> Neither cpu will work, of course, but why change? >>> >> >> Because there's already a patch in master that puts the cortex-a7 under >> CONFIG_TCG, so I can have the whole if/else in this patch. >> >> If I keep the cortex-a15, this change needs to go into the next patch >> ("move cpu_tcg to tcg/cpu32.c") which moves the rest of the 32bit cpus, >> which was supposed to be only code movement. > > Well, I still think the change to a7 is wrong. > If the two patches need to be merged to break bisection, > then so be it -- just mention that fact in the commit message. >
I don't get why it would be wrong. The test just needs any cpu model that triggers the message at qmp_query_cpu_model_expansion.