| Message ID | 1359304210-5073-1-git-send-email-afaerber@suse.de |
|---|---|
| State | New |
| Headers | show |
On 27 January 2013 16:30, Andreas Färber <afaerber@suse.de> wrote: > In the initial conversion of CPU models to QOM types, model names were > mapped 1:1 to type names. As a side effect this gained us a type "any", > which is now a device. > > To avoid "-device any" silliness and to pave the way for compiling > multiple targets into one executable, adopt a <name>-<arch>-cpu scheme. > This leads to names like arm926-arm-cpu but is easiest to handle. > > No functional changes for -cpu arguments or -cpu ? output. > > Suggested-by: Eduardo Habkost <ehabkost@redhat.com> > Signed-off-by: Andreas Färber <afaerber@suse.de> Acked-by: Peter Maydell <peter.maydell@linaro.org> I assume you're going to submit this via the qom subtree. -- PMM
Am 29.01.2013 12:03, schrieb Peter Maydell: > On 27 January 2013 16:30, Andreas Färber <afaerber@suse.de> wrote: >> In the initial conversion of CPU models to QOM types, model names were >> mapped 1:1 to type names. As a side effect this gained us a type "any", >> which is now a device. >> >> To avoid "-device any" silliness and to pave the way for compiling >> multiple targets into one executable, adopt a <name>-<arch>-cpu scheme. >> This leads to names like arm926-arm-cpu but is easiest to handle. >> >> No functional changes for -cpu arguments or -cpu ? output. >> >> Suggested-by: Eduardo Habkost <ehabkost@redhat.com> >> Signed-off-by: Andreas Färber <afaerber@suse.de> > > Acked-by: Peter Maydell <peter.maydell@linaro.org> I just re-reviewed this: env->cpu_model_str is set in cpu_arm_init(), so this patch is good to go (unlike unicore32). > I assume you're going to submit this via the qom subtree. It applies to master now and is an arm-internal leaf patch. Since you have another target-arm.next patch that Blue didn't pick up so far, feel free to send a pull for both. But I'm fine taking it through my tree if you prefer. Thanks, Andreas
On 29 January 2013 12:00, Andreas Färber <afaerber@suse.de> wrote: > Am 29.01.2013 12:03, schrieb Peter Maydell: >> I assume you're going to submit this via the qom subtree. > > It applies to master now and is an arm-internal leaf patch. Since you > have another target-arm.next patch that Blue didn't pick up so far, feel > free to send a pull for both. But I'm fine taking it through my tree if > you prefer. OK, I put it in my target-arm pullreq. -- PMM
diff --git a/target-arm/cpu.c b/target-arm/cpu.c index d1a4c82..1c6a628 100644 --- a/target-arm/cpu.c +++ b/target-arm/cpu.c @@ -204,12 +204,15 @@ void arm_cpu_realize(ARMCPU *cpu) static ObjectClass *arm_cpu_class_by_name(const char *cpu_model) { ObjectClass *oc; + char *typename; if (!cpu_model) { return NULL; } - oc = object_class_by_name(cpu_model); + typename = g_strdup_printf("%s-" TYPE_ARM_CPU, cpu_model); + oc = object_class_by_name(typename); + g_free(typename); if (!oc || !object_class_dynamic_cast(oc, TYPE_ARM_CPU) || object_class_is_abstract(oc)) { return NULL; @@ -789,14 +792,15 @@ static void arm_cpu_class_init(ObjectClass *oc, void *data) static void cpu_register(const ARMCPUInfo *info) { TypeInfo type_info = { - .name = info->name, .parent = TYPE_ARM_CPU, .instance_size = sizeof(ARMCPU), .instance_init = info->initfn, .class_size = sizeof(ARMCPUClass), }; + type_info.name = g_strdup_printf("%s-" TYPE_ARM_CPU, info->name); type_register(&type_info); + g_free((void *)type_info.name); } static const TypeInfo arm_cpu_type_info = { diff --git a/target-arm/helper.c b/target-arm/helper.c index 7a10fdd..eb7b291 100644 --- a/target-arm/helper.c +++ b/target-arm/helper.c @@ -1303,9 +1303,9 @@ static gint arm_cpu_list_compare(gconstpointer a, gconstpointer b) name_a = object_class_get_name(class_a); name_b = object_class_get_name(class_b); - if (strcmp(name_a, "any") == 0) { + if (strcmp(name_a, "any-" TYPE_ARM_CPU) == 0) { return 1; - } else if (strcmp(name_b, "any") == 0) { + } else if (strcmp(name_b, "any-" TYPE_ARM_CPU) == 0) { return -1; } else { return strcmp(name_a, name_b); @@ -1316,9 +1316,14 @@ static void arm_cpu_list_entry(gpointer data, gpointer user_data) { ObjectClass *oc = data; CPUListState *s = user_data; + const char *typename; + char *name; + typename = object_class_get_name(oc); + name = g_strndup(typename, strlen(typename) - strlen("-" TYPE_ARM_CPU)); (*s->cpu_fprintf)(s->file, " %s\n", - object_class_get_name(oc)); + name); + g_free(name); } void arm_cpu_list(FILE *f, fprintf_function cpu_fprintf)
In the initial conversion of CPU models to QOM types, model names were mapped 1:1 to type names. As a side effect this gained us a type "any", which is now a device. To avoid "-device any" silliness and to pave the way for compiling multiple targets into one executable, adopt a <name>-<arch>-cpu scheme. This leads to names like arm926-arm-cpu but is easiest to handle. No functional changes for -cpu arguments or -cpu ? output. Suggested-by: Eduardo Habkost <ehabkost@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de> --- target-arm/cpu.c | 8 ++++++-- target-arm/helper.c | 11 ++++++++--- 2 Dateien geändert, 14 Zeilen hinzugefügt(+), 5 Zeilen entfernt(-)