Hello Richard, I've queued the uncontroversial bits of v1 for the next qom-cpu pull (FYI). This series attempts to introduce AlphaCPU subclasses in a less mechanical way and in light of efforts to coordinate the naming scheme across targets.
The -cpu ? support was only rebased for now; due to the type name changes this output has now changed and it does not include the aliases. @Eduardo: Note that I've used ObjectClass as return type to avoid unnecessary g_strdup(). Also note the use of a trivial TYPE() macro to enforce the type name scheme in the knowledge that string literals are being used, avoiding memory leaks. Regards, Andreas v1 -> v2: * Drop "2*" CPU types in favor of aliases * Adopt Eduardo's suggested naming scheme <name>-alpha-cpu * Move cpu_alpha_init() to cpu.c, turning alpha_translate_init() non-static * Split off alpha_cpu_realize() from cpu_alpha_init() * Drop illustratory alpha_cpu_reset() RFC for now Cc: Richard Henderson <r...@twiddle.net> Cc: Eduardo Habkost <ehabk...@redhat.com> Andreas Färber (5): target-alpha: Let cpu_alpha_init() return AlphaCPU alpha: Pass AlphaCPU array to Typhoon target-alpha: Avoid leaking the alarm timer over reset target-alpha: Turn CPU definitions into subclasses target-alpha: Add support for -cpu ? hw/alpha_dp264.c | 18 ++-- hw/alpha_sys.h | 2 +- hw/alpha_typhoon.c | 30 ++++--- target-alpha/cpu-qom.h | 3 + target-alpha/cpu.c | 219 ++++++++++++++++++++++++++++++++++++++++++++- target-alpha/cpu.h | 18 +++- target-alpha/sys_helper.c | 6 +- target-alpha/translate.c | 58 +----------- 8 Dateien geändert, 267 Zeilen hinzugefügt(+), 87 Zeilen entfernt(-) -- 1.7.10.4