| Message ID | 20240718133312.10324-5-philmd@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Reconstruct loongson ipi driver | expand |
On 18/7/24 15:32, Philippe Mathieu-Daudé wrote: > From: Bibo Mao <maobibo@loongson.cn> > > In preparation to extract common IPI code in few commits, > extract loongson_ipi_common_finalize(). > > Signed-off-by: Bibo Mao <maobibo@loongson.cn> > [PMD: Extracted from bigger commit, added commit description] > Co-Developed-by: Philippe Mathieu-Daudé <philmd@linaro.org> > Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> > Reviewed-by: Bibo Mao <maobibo@loongson.cn> > Tested-by: Bibo Mao <maobibo@loongson.cn> > Acked-by: Song Gao <gaosong@loongson.cn> > --- > hw/intc/loongson_ipi.c | 7 ++++++- > 1 file changed, 6 insertions(+), 1 deletion(-) > > diff --git a/hw/intc/loongson_ipi.c b/hw/intc/loongson_ipi.c > index d870af39c1..960d1e604f 100644 > --- a/hw/intc/loongson_ipi.c > +++ b/hw/intc/loongson_ipi.c > @@ -357,13 +357,18 @@ static void loongson_ipi_class_init(ObjectClass *klass, void *data) > dc->vmsd = &vmstate_loongson_ipi; > } > > -static void loongson_ipi_finalize(Object *obj) > +static void loongson_ipi_common_finalize(Object *obj) > { > LoongsonIPIState *s = LOONGSON_IPI(obj); > > g_free(s->cpu); Unfortunately the current code is bogus, @cpu is allocated in loongson_ipi_realize(), a DeviceRealize handler, so must be de-allocated in a DeviceUnrealize equivalent. Not at the instance finalize cleanup which is too late because we could leak after UNREALIZE -> REALIZE sequences. > } > > +static void loongson_ipi_finalize(Object *obj) > +{ > + loongson_ipi_common_finalize(obj); > +} > + > static const TypeInfo loongson_ipi_types[] = { > { > .name = TYPE_LOONGSON_IPI,
在 2024/7/23 下午4:41, Philippe Mathieu-Daudé 写道: > On 18/7/24 15:32, Philippe Mathieu-Daudé wrote: >> From: Bibo Mao <maobibo@loongson.cn> >> >> In preparation to extract common IPI code in few commits, >> extract loongson_ipi_common_finalize(). >> >> Signed-off-by: Bibo Mao <maobibo@loongson.cn> >> [PMD: Extracted from bigger commit, added commit description] >> Co-Developed-by: Philippe Mathieu-Daudé <philmd@linaro.org> >> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> >> Reviewed-by: Bibo Mao <maobibo@loongson.cn> >> Tested-by: Bibo Mao <maobibo@loongson.cn> >> Acked-by: Song Gao <gaosong@loongson.cn> >> --- >> hw/intc/loongson_ipi.c | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/hw/intc/loongson_ipi.c b/hw/intc/loongson_ipi.c >> index d870af39c1..960d1e604f 100644 >> --- a/hw/intc/loongson_ipi.c >> +++ b/hw/intc/loongson_ipi.c >> @@ -357,13 +357,18 @@ static void loongson_ipi_class_init(ObjectClass >> *klass, void *data) >> dc->vmsd = &vmstate_loongson_ipi; >> } >> -static void loongson_ipi_finalize(Object *obj) >> +static void loongson_ipi_common_finalize(Object *obj) >> { >> LoongsonIPIState *s = LOONGSON_IPI(obj); >> g_free(s->cpu); > > Unfortunately the current code is bogus, @cpu is allocated in > loongson_ipi_realize(), a DeviceRealize handler, so must be > de-allocated in a DeviceUnrealize equivalent. Not at the instance > finalize cleanup which is too late because we could leak after > UNREALIZE -> REALIZE sequences. > Yes, we should use DeviecUnrealize to free s->cpu. I will send a patch to correct it. Thanks. Song Gao >> } >> +static void loongson_ipi_finalize(Object *obj) >> +{ >> + loongson_ipi_common_finalize(obj); >> +} >> + >> static const TypeInfo loongson_ipi_types[] = { >> { >> .name = TYPE_LOONGSON_IPI,
On 23/7/24 12:00, gaosong wrote: > 在 2024/7/23 下午4:41, Philippe Mathieu-Daudé 写道: >> On 18/7/24 15:32, Philippe Mathieu-Daudé wrote: >>> From: Bibo Mao <maobibo@loongson.cn> >>> >>> In preparation to extract common IPI code in few commits, >>> extract loongson_ipi_common_finalize(). >>> >>> Signed-off-by: Bibo Mao <maobibo@loongson.cn> >>> [PMD: Extracted from bigger commit, added commit description] >>> Co-Developed-by: Philippe Mathieu-Daudé <philmd@linaro.org> >>> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> >>> Reviewed-by: Bibo Mao <maobibo@loongson.cn> >>> Tested-by: Bibo Mao <maobibo@loongson.cn> >>> Acked-by: Song Gao <gaosong@loongson.cn> >>> --- >>> hw/intc/loongson_ipi.c | 7 ++++++- >>> 1 file changed, 6 insertions(+), 1 deletion(-) >>> >>> diff --git a/hw/intc/loongson_ipi.c b/hw/intc/loongson_ipi.c >>> index d870af39c1..960d1e604f 100644 >>> --- a/hw/intc/loongson_ipi.c >>> +++ b/hw/intc/loongson_ipi.c >>> @@ -357,13 +357,18 @@ static void loongson_ipi_class_init(ObjectClass >>> *klass, void *data) >>> dc->vmsd = &vmstate_loongson_ipi; >>> } >>> -static void loongson_ipi_finalize(Object *obj) >>> +static void loongson_ipi_common_finalize(Object *obj) >>> { >>> LoongsonIPIState *s = LOONGSON_IPI(obj); >>> g_free(s->cpu); >> >> Unfortunately the current code is bogus, @cpu is allocated in >> loongson_ipi_realize(), a DeviceRealize handler, so must be >> de-allocated in a DeviceUnrealize equivalent. Not at the instance >> finalize cleanup which is too late because we could leak after >> UNREALIZE -> REALIZE sequences. >> > Yes, we should use DeviecUnrealize to free s->cpu. > > I will send a patch to correct it. No need, I already wrote it, currently testing various build configs before posting v6.
diff --git a/hw/intc/loongson_ipi.c b/hw/intc/loongson_ipi.c index d870af39c1..960d1e604f 100644 --- a/hw/intc/loongson_ipi.c +++ b/hw/intc/loongson_ipi.c @@ -357,13 +357,18 @@ static void loongson_ipi_class_init(ObjectClass *klass, void *data) dc->vmsd = &vmstate_loongson_ipi; } -static void loongson_ipi_finalize(Object *obj) +static void loongson_ipi_common_finalize(Object *obj) { LoongsonIPIState *s = LOONGSON_IPI(obj); g_free(s->cpu); } +static void loongson_ipi_finalize(Object *obj) +{ + loongson_ipi_common_finalize(obj); +} + static const TypeInfo loongson_ipi_types[] = { { .name = TYPE_LOONGSON_IPI,