Message ID | 1476375902-11715-3-git-send-email-lvivier@redhat.com |
---|---|
State | New |
Headers | show |
On Thu, Oct 13, 2016 at 06:24:44PM +0200, Laurent Vivier wrote: > We have now the cpu_exec_realize() in realize, > so the init part must be in init. > > I've removed the cannot_destroy_with_object_finalize_yet field as > unsafe references have been moved to cpu_exec_realize(). > (tested with QOM command provided by commit 4c315c27 with > "athlon-x86_64-cpu") > > CC: Eduardo Habkost <ehabkost@redhat.com> > Signed-off-by: Laurent Vivier <lvivier@redhat.com> Resending the question I asked in my reply to v1: Instead of creating requiring each subclass to manually call cpu_exec_init()) on instance_init, why don't we move parts of cpu_exec_init()/cpu_exec_realize() code to cpu_common_initfn()? (TYPE_CPU's instance_init) (And if there's any code that needs to be run after the subclasses instance_init functions, we can just add a instance_post_init function to TYPE_CPU).
On 14/10/2016 15:33, Eduardo Habkost wrote: > On Thu, Oct 13, 2016 at 06:24:44PM +0200, Laurent Vivier wrote: >> We have now the cpu_exec_realize() in realize, >> so the init part must be in init. >> >> I've removed the cannot_destroy_with_object_finalize_yet field as >> unsafe references have been moved to cpu_exec_realize(). >> (tested with QOM command provided by commit 4c315c27 with >> "athlon-x86_64-cpu") >> >> CC: Eduardo Habkost <ehabkost@redhat.com> >> Signed-off-by: Laurent Vivier <lvivier@redhat.com> > > Resending the question I asked in my reply to v1: > > Instead of creating requiring each subclass to manually call > cpu_exec_init()) on instance_init, why don't we move parts of > cpu_exec_init()/cpu_exec_realize() code to cpu_common_initfn()? > (TYPE_CPU's instance_init) > > (And if there's any code that needs to be run after the > subclasses instance_init functions, we can just add a > instance_post_init function to TYPE_CPU). > It's done in PATCH 20/20. Is that what you want? Thanks, Laurent
On Fri, Oct 14, 2016 at 03:34:17PM +0200, Laurent Vivier wrote: > On 14/10/2016 15:33, Eduardo Habkost wrote: > > On Thu, Oct 13, 2016 at 06:24:44PM +0200, Laurent Vivier wrote: > >> We have now the cpu_exec_realize() in realize, > >> so the init part must be in init. > >> > >> I've removed the cannot_destroy_with_object_finalize_yet field as > >> unsafe references have been moved to cpu_exec_realize(). > >> (tested with QOM command provided by commit 4c315c27 with > >> "athlon-x86_64-cpu") > >> > >> CC: Eduardo Habkost <ehabkost@redhat.com> > >> Signed-off-by: Laurent Vivier <lvivier@redhat.com> > > > > Resending the question I asked in my reply to v1: > > > > Instead of creating requiring each subclass to manually call > > cpu_exec_init()) on instance_init, why don't we move parts of > > cpu_exec_init()/cpu_exec_realize() code to cpu_common_initfn()? > > (TYPE_CPU's instance_init) > > > > (And if there's any code that needs to be run after the > > subclasses instance_init functions, we can just add a > > instance_post_init function to TYPE_CPU). > > > > It's done in PATCH 20/20. > > Is that what you want? Yes (except that I would have inlined the the cpu_exec_init() code inside cpu_common_init()). I think I expected this to be done in a single step, that wouldn't require touching code for all architectures three times. Something like: 1) Move cpu->as, cpu->num_ases, cpu->thread_id, cpu->memory initialization, and "memory" property registration from cpu_exec_init() to cpu_common_init() (no architecture code touched). 2) (optional) Rename cpu_exec_init() to cpu_exec_realize() (only trivial changes in architecture code)
On 14/10/2016 15:48, Eduardo Habkost wrote: > On Fri, Oct 14, 2016 at 03:34:17PM +0200, Laurent Vivier wrote: >> On 14/10/2016 15:33, Eduardo Habkost wrote: >>> On Thu, Oct 13, 2016 at 06:24:44PM +0200, Laurent Vivier wrote: >>>> We have now the cpu_exec_realize() in realize, >>>> so the init part must be in init. >>>> >>>> I've removed the cannot_destroy_with_object_finalize_yet field as >>>> unsafe references have been moved to cpu_exec_realize(). >>>> (tested with QOM command provided by commit 4c315c27 with >>>> "athlon-x86_64-cpu") >>>> >>>> CC: Eduardo Habkost <ehabkost@redhat.com> >>>> Signed-off-by: Laurent Vivier <lvivier@redhat.com> >>> >>> Resending the question I asked in my reply to v1: >>> >>> Instead of creating requiring each subclass to manually call >>> cpu_exec_init()) on instance_init, why don't we move parts of >>> cpu_exec_init()/cpu_exec_realize() code to cpu_common_initfn()? >>> (TYPE_CPU's instance_init) >>> >>> (And if there's any code that needs to be run after the >>> subclasses instance_init functions, we can just add a >>> instance_post_init function to TYPE_CPU). >>> >> >> It's done in PATCH 20/20. >> >> Is that what you want? > > Yes (except that I would have inlined the the cpu_exec_init() > code inside cpu_common_init()). > > I think I expected this to be done in a single step, that > wouldn't require touching code for all architectures three times. > Something like: Well, I've added several steps to help to review and break nothing. > 1) Move cpu->as, cpu->num_ases, cpu->thread_id, cpu->memory > initialization, and "memory" property registration from > cpu_exec_init() to cpu_common_init() (no architecture code > touched). system_memory (for the "memory" property) is declared as static in exec.c, so we can't move it to cpu_common_init(). Laurent
On Fri, Oct 14, 2016 at 04:12:20PM +0200, Laurent Vivier wrote: > > > On 14/10/2016 15:48, Eduardo Habkost wrote: > > On Fri, Oct 14, 2016 at 03:34:17PM +0200, Laurent Vivier wrote: > >> On 14/10/2016 15:33, Eduardo Habkost wrote: > >>> On Thu, Oct 13, 2016 at 06:24:44PM +0200, Laurent Vivier wrote: > >>>> We have now the cpu_exec_realize() in realize, > >>>> so the init part must be in init. > >>>> > >>>> I've removed the cannot_destroy_with_object_finalize_yet field as > >>>> unsafe references have been moved to cpu_exec_realize(). > >>>> (tested with QOM command provided by commit 4c315c27 with > >>>> "athlon-x86_64-cpu") > >>>> > >>>> CC: Eduardo Habkost <ehabkost@redhat.com> > >>>> Signed-off-by: Laurent Vivier <lvivier@redhat.com> > >>> > >>> Resending the question I asked in my reply to v1: > >>> > >>> Instead of creating requiring each subclass to manually call > >>> cpu_exec_init()) on instance_init, why don't we move parts of > >>> cpu_exec_init()/cpu_exec_realize() code to cpu_common_initfn()? > >>> (TYPE_CPU's instance_init) > >>> > >>> (And if there's any code that needs to be run after the > >>> subclasses instance_init functions, we can just add a > >>> instance_post_init function to TYPE_CPU). > >>> > >> > >> It's done in PATCH 20/20. > >> > >> Is that what you want? > > > > Yes (except that I would have inlined the the cpu_exec_init() > > code inside cpu_common_init()). > > > > I think I expected this to be done in a single step, that > > wouldn't require touching code for all architectures three times. > > Something like: > > Well, I've added several steps to help to review and break nothing. To me, it made review harder. Architecture maintainers have to review 3 different patches that touch their code, instead of a single one. Anyway, I would still give my Reviewed-by line for the i386 part, if you believe it's better to do it this way. > > > 1) Move cpu->as, cpu->num_ases, cpu->thread_id, cpu->memory > > initialization, and "memory" property registration from > > cpu_exec_init() to cpu_common_init() (no architecture code > > touched). > > system_memory (for the "memory" property) is declared as static in > exec.c, so we can't move it to cpu_common_init(). In this case we can't inline it, that's true. But it still doesn't require touching each architecture 3 times (we could just create a cpu_exec_instance_init() function in exec.c and call it from cpu_common_init()). (Or we could move system_memory to MachineState, but that could be done in a follow-up patch).
On Fri, 14 Oct 2016 10:48:58 -0300 Eduardo Habkost <ehabkost@redhat.com> wrote: > On Fri, Oct 14, 2016 at 03:34:17PM +0200, Laurent Vivier wrote: > > On 14/10/2016 15:33, Eduardo Habkost wrote: > > > On Thu, Oct 13, 2016 at 06:24:44PM +0200, Laurent Vivier wrote: > > >> We have now the cpu_exec_realize() in realize, > > >> so the init part must be in init. > > >> > > >> I've removed the cannot_destroy_with_object_finalize_yet field as > > >> unsafe references have been moved to cpu_exec_realize(). > > >> (tested with QOM command provided by commit 4c315c27 with > > >> "athlon-x86_64-cpu") > > >> > > >> CC: Eduardo Habkost <ehabkost@redhat.com> > > >> Signed-off-by: Laurent Vivier <lvivier@redhat.com> > > > > > > Resending the question I asked in my reply to v1: > > > > > > Instead of creating requiring each subclass to manually call > > > cpu_exec_init()) on instance_init, why don't we move parts of > > > cpu_exec_init()/cpu_exec_realize() code to cpu_common_initfn()? > > > (TYPE_CPU's instance_init) > > > > > > (And if there's any code that needs to be run after the > > > subclasses instance_init functions, we can just add a > > > instance_post_init function to TYPE_CPU). > > > > > > > It's done in PATCH 20/20. > > > > Is that what you want? > > Yes (except that I would have inlined the the cpu_exec_init() > code inside cpu_common_init()). > > I think I expected this to be done in a single step, that > wouldn't require touching code for all architectures three times. > Something like: > > 1) Move cpu->as, cpu->num_ases, cpu->thread_id, cpu->memory > initialization, and "memory" property registration from > cpu_exec_init() to cpu_common_init() (no architecture code > touched). > 2) (optional) Rename cpu_exec_init() to cpu_exec_realize() (only > trivial changes in architecture code) I'd do all of it in 1 step - split cpu_exec_init on init/realize parts (comment in cpu_exec_init says that qom/cpu.c can't be used for "memory" property) - call cpu_exec_init() from cpu_common_init() - s/cpu_exec_init/cpu_exec_realize/ in target-* Follow up patches 2) 1 patch, could move parts of split cpu_exec_init() to cpu_common_init() if that makes sense. 3) 1 patch, could move cpu_exec_realize() into per target *_realizefn() it would be not small patch but still trivial 4) 1 patch, do similar (#3) thing for unrealize Perhaps #3,4 could be done in a more generic way in qom/cpu.c but I don't have a good idea how to do it.
On Fri, Oct 14, 2016 at 04:33:07PM +0200, Igor Mammedov wrote: > On Fri, 14 Oct 2016 10:48:58 -0300 > Eduardo Habkost <ehabkost@redhat.com> wrote: > > > On Fri, Oct 14, 2016 at 03:34:17PM +0200, Laurent Vivier wrote: > > > On 14/10/2016 15:33, Eduardo Habkost wrote: > > > > On Thu, Oct 13, 2016 at 06:24:44PM +0200, Laurent Vivier wrote: > > > >> We have now the cpu_exec_realize() in realize, > > > >> so the init part must be in init. > > > >> > > > >> I've removed the cannot_destroy_with_object_finalize_yet field as > > > >> unsafe references have been moved to cpu_exec_realize(). > > > >> (tested with QOM command provided by commit 4c315c27 with > > > >> "athlon-x86_64-cpu") > > > >> > > > >> CC: Eduardo Habkost <ehabkost@redhat.com> > > > >> Signed-off-by: Laurent Vivier <lvivier@redhat.com> > > > > > > > > Resending the question I asked in my reply to v1: > > > > > > > > Instead of creating requiring each subclass to manually call > > > > cpu_exec_init()) on instance_init, why don't we move parts of > > > > cpu_exec_init()/cpu_exec_realize() code to cpu_common_initfn()? > > > > (TYPE_CPU's instance_init) > > > > > > > > (And if there's any code that needs to be run after the > > > > subclasses instance_init functions, we can just add a > > > > instance_post_init function to TYPE_CPU). > > > > > > > > > > It's done in PATCH 20/20. > > > > > > Is that what you want? > > > > Yes (except that I would have inlined the the cpu_exec_init() > > code inside cpu_common_init()). > > > > I think I expected this to be done in a single step, that > > wouldn't require touching code for all architectures three times. > > Something like: > > > > 1) Move cpu->as, cpu->num_ases, cpu->thread_id, cpu->memory > > initialization, and "memory" property registration from > > cpu_exec_init() to cpu_common_init() (no architecture code > > touched). > > 2) (optional) Rename cpu_exec_init() to cpu_exec_realize() (only > > trivial changes in architecture code) > I'd do all of it in 1 step > - split cpu_exec_init on init/realize parts > (comment in cpu_exec_init says that qom/cpu.c can't be used for > "memory" property) > - call cpu_exec_init() from cpu_common_init() > - s/cpu_exec_init/cpu_exec_realize/ in target-* Personally, I don't mind if it is done in 1 or 2 steps. I would just like to avoid changing architecture code 3 times. In addition to make review easier, it would make the decision to merge it easier for the maintainer who's going to do it (no need to wait for Acked-bys/Reviewed-bys from multiple architecture maintainers).
diff --git a/target-i386/cpu.c b/target-i386/cpu.c index b977130..4741dd6 100644 --- a/target-i386/cpu.c +++ b/target-i386/cpu.c @@ -3158,8 +3158,11 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp) cpu->phys_bits = 32; } } - cpu_exec_init(cs); - cpu_exec_realize(cs, &error_abort); + cpu_exec_realize(cs, &local_err); + if (local_err != NULL) { + error_propagate(errp, local_err); + return; + } if (tcg_enabled()) { tcg_x86_init(); @@ -3365,6 +3368,7 @@ static void x86_cpu_initfn(Object *obj) FeatureWord w; cs->env_ptr = env; + cpu_exec_init(cs); object_property_add(obj, "family", "int", x86_cpuid_version_get_family, @@ -3538,11 +3542,6 @@ static void x86_cpu_common_class_init(ObjectClass *oc, void *data) cc->cpu_exec_exit = x86_cpu_exec_exit; dc->cannot_instantiate_with_device_add_yet = false; - /* - * Reason: x86_cpu_initfn() calls cpu_exec_init(), which saves the - * object in cpus -> dangling pointer after final object_unref(). - */ - dc->cannot_destroy_with_object_finalize_yet = true; } static const TypeInfo x86_cpu_type_info = {
We have now the cpu_exec_realize() in realize, so the init part must be in init. I've removed the cannot_destroy_with_object_finalize_yet field as unsafe references have been moved to cpu_exec_realize(). (tested with QOM command provided by commit 4c315c27 with "athlon-x86_64-cpu") CC: Eduardo Habkost <ehabkost@redhat.com> Signed-off-by: Laurent Vivier <lvivier@redhat.com> --- target-i386/cpu.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-)