Laszlo Ersek <ler...@redhat.com> writes: > On 04/26/18 11:18, Laszlo Ersek wrote: >> On 04/26/18 08:26, Markus Armbruster wrote: >>> Laszlo Ersek <ler...@redhat.com> writes: >>> >>> [...] >>>> In brief, I think extending configure / the build system would only help >>>> with the less painful part of this (the scalar mapping), and so it's not >>>> worth doing. >>> >>> We're going to leave deprecated query-cpus alone (see Eric's review of >>> the previous patch). >> >> Yes. >> >>> Does that change matters? >> >> It does. >> >>> PS: Instead of configure changes, #TARGET_BASE_ARCH might do. >> >> Does that already exist as a macro? Hmm... After grepping my build dir, >> it doesn't seem to, but maybe I can change that. Because, >> TARGET_BASE_ARCH is exactly what we need. > > Acerbic discovery of the day: @CpuInfoArch has "x86", while configure > produces: > > TARGET_NAME TARGET_BASE_ARCH > i386 i386 > x86_64 i386 > > Note how "i386" does not match "x86".
Review fail. Just three weeks ago, we could still have fixed query-cpus-fast... > I guess I'll have to keep the sysemu_target_to_cpuinfo_arch() function. Assuming this is the only offender, you could also s/i386/x86/ before you pass it to qapi_enum_parse(). Pick what you hate less.
