Message ID | 20210422193902.2644064-9-f4bug@amsat.org |
---|---|
State | New |
Headers | show |
Series | cpu: Introduce SysemuCPUOps structure | expand |
On 4/22/21 12:38 PM, Philippe Mathieu-Daudé wrote: > See rationale in previous commit. Targets should use the vmsd field > of DeviceClass, not CPUClass. As migration is not important on the > avr/lm32/moxie targets, break the migration compatibility and set > the DeviceClass vmsd field. > > Signed-off-by: Philippe Mathieu-Daudé<f4bug@amsat.org> > --- > target/avr/cpu.c | 2 +- > target/lm32/cpu.c | 2 +- > target/moxie/cpu.c | 2 +- > 3 files changed, 3 insertions(+), 3 deletions(-) Do you not have to increment the vmstate versions? Though I can't see how the version numbers would even get compared, since we're looking for them in different locations? r~
On 4/23/21 3:16 AM, Richard Henderson wrote: > On 4/22/21 12:38 PM, Philippe Mathieu-Daudé wrote: >> See rationale in previous commit. Targets should use the vmsd field >> of DeviceClass, not CPUClass. As migration is not important on the >> avr/lm32/moxie targets, break the migration compatibility and set >> the DeviceClass vmsd field. >> >> Signed-off-by: Philippe Mathieu-Daudé<f4bug@amsat.org> >> --- >> target/avr/cpu.c | 2 +- >> target/lm32/cpu.c | 2 +- >> target/moxie/cpu.c | 2 +- >> 3 files changed, 3 insertions(+), 3 deletions(-) > > Do you not have to increment the vmstate versions? > > Though I can't see how the version numbers would even get compared, > since we're looking for them in different locations? Yes, I am not sure how that will work. Anyway, better safe than sorry!
diff --git a/target/avr/cpu.c b/target/avr/cpu.c index 37a8ebcc86f..3353bcb9fc7 100644 --- a/target/avr/cpu.c +++ b/target/avr/cpu.c @@ -213,7 +213,7 @@ static void avr_cpu_class_init(ObjectClass *oc, void *data) cc->set_pc = avr_cpu_set_pc; cc->memory_rw_debug = avr_cpu_memory_rw_debug; cc->get_phys_page_debug = avr_cpu_get_phys_page_debug; - cc->legacy_vmsd = &vms_avr_cpu; + dc->vmsd = &vms_avr_cpu; cc->disas_set_info = avr_cpu_disas_set_info; cc->gdb_read_register = avr_cpu_gdb_read_register; cc->gdb_write_register = avr_cpu_gdb_write_register; diff --git a/target/lm32/cpu.c b/target/lm32/cpu.c index c0388142528..e27c1a8bc97 100644 --- a/target/lm32/cpu.c +++ b/target/lm32/cpu.c @@ -241,7 +241,7 @@ static void lm32_cpu_class_init(ObjectClass *oc, void *data) cc->gdb_write_register = lm32_cpu_gdb_write_register; #ifndef CONFIG_USER_ONLY cc->get_phys_page_debug = lm32_cpu_get_phys_page_debug; - cc->legacy_vmsd = &vmstate_lm32_cpu; + dc->vmsd = &vmstate_lm32_cpu; #endif cc->gdb_num_core_regs = 32 + 7; cc->gdb_stop_before_watchpoint = true; diff --git a/target/moxie/cpu.c b/target/moxie/cpu.c index 3c81a088eaa..e5da3fa3f17 100644 --- a/target/moxie/cpu.c +++ b/target/moxie/cpu.c @@ -122,7 +122,7 @@ static void moxie_cpu_class_init(ObjectClass *oc, void *data) cc->set_pc = moxie_cpu_set_pc; #ifndef CONFIG_USER_ONLY cc->get_phys_page_debug = moxie_cpu_get_phys_page_debug; - cc->legacy_vmsd = &vmstate_moxie_cpu; + dc->vmsd = &vmstate_moxie_cpu; #endif cc->disas_set_info = moxie_cpu_disas_set_info; cc->tcg_ops = &moxie_tcg_ops;
See rationale in previous commit. Targets should use the vmsd field of DeviceClass, not CPUClass. As migration is not important on the avr/lm32/moxie targets, break the migration compatibility and set the DeviceClass vmsd field. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> --- target/avr/cpu.c | 2 +- target/lm32/cpu.c | 2 +- target/moxie/cpu.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-)