Re: [PATCH 08/10] build: Add SPICE_CFLAGS and SPICE_LIBS to relevant files

[Christophe de Dinechin <dinechin@redhat.com>]

On 2020-06-30 at 01:08 CEST, Gerd Hoffmann wrote...
>   Hi,
>
>>  obj-$(CONFIG_PC) += pc.o pc_sysfw.o
>> +pc.o-cflags += $(SPICE_CFLAGS)
>
> Hmm, looks strange.  Why does pc.c need spice?

It includes ui/qemu-spice.h, and I did not check why.
Turns out this is not needed. So I'll remove it.

>
>> +qmp-cmds.o-cflags += $(SPICE_CFLAGS)
>> +hmp-cmds.o-cflags += $(SPICE_CFLAGS)
>
> spice monitor commands need this I guess?

Yes.

>
>> +misc.o-cflags += $(SPICE_CFLAGS)
>
> Why this?

qemu_using_spice and qemu_spice_migrate_info

>
>> +vl.o-cflags := $(GPROF_CFLAGS) $(SDL_CFLAGS) $(SPICE_CFLAGS)
>
> spice init probably

Yes.
.
>
>> -stub-obj-y += vmgenid.o
>>  stub-obj-y += sysbus.o
>>  stub-obj-y += tpm.o
>>  stub-obj-y += trace-control.o
>> +stub-obj-y += vmgenid.o
>
> Huh?

I sent it separately as a trivial patch. Wrong alphabetical order, and where
that change was placed was causing a conflict on each rebase with a spice.c
stub I had at some point "at the right spot" ;-)

>
>> -spice-app.mo-cflags := $(GIO_CFLAGS)
>> -spice-app.mo-libs := $(GIO_LIBS)
>> +spice-app.mo-cflags := $(GIO_CFLAGS) $(SPICE_CFLAGS)
>> +spice-app.mo-libs := $(GIO_LIBS) $(SPICE_LIBS)
>
> Good.
>
>> --- a/util/module.c
>> +++ b/util/module.c
>> @@ -22,11 +22,11 @@
>>  #ifdef CONFIG_MODULE_UPGRADES
>>  #include "qemu-version.h"
>>  #endif
>> -#ifdef CONFIG_TRACE_RECORDER
>>  #include "trace/recorder.h"
>> -#endif
>>
>>
>> +RECORDER(modules, 16, "QEMU load modules");
>> +
>>  typedef struct ModuleEntry
>>  {
>>      void (*init)(void);
>> @@ -85,6 +85,15 @@ void register_dso_module_init(void (*fn)(void), module_init_type type)
>>  {
>>      ModuleEntry *e;
>>
>> +#ifdef CONFIG_TRACE_RECORDER
>> +    static const char *name[] = {
>> +        "MIGRATION", "BLOCK", "OPTS", "QOM",
>> +        "TRACE", "XEN_BACKEND", "LIBQOS", "FUZZ_TARGET",
>> +        "MAX"
>> +    };
>> +#endif
>> +    record(modules, "Register DSO module init %p type %u %+s",
>> +           fn, type, name[type]);
>>      init_lists();
>>
>>      e = g_malloc0(sizeof(*e));
>
> Unrelated change.
>
> (the recorder stuff should probably integrate with qemu trace support,
> so you can record any trace point qemu has, but that'll be another patch
> series ...)

I sent it separately, and fixed the leftover patch.

>
> take care,
>   Gerd


--
Cheers,
Christophe de Dinechin (IRC c3d)