Skip to contentSkip to conversation

hw/core/or-irq: Mark the device with cannot_instantiate_with_device_add_yet

PATCH Superseded Thomas Huth 4 participants 4 messages Jan 27, 2017

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hw/core/or-irq: Mark the device with cannot_instantiate_with_device_add_yet

The "or-irq" device is just used internally. It's strange to see this device showing up in the "-device ?" help text. Let's mark it with cannot_instantiate_with_device_add_yet to hide it from the users.

---
 hw/core/or-irq.c | 3 +++
 1 file changed, 3 insertions(+)
hw/core/or-irq.c +3 -0
index 1ac090d..2487fa3 100644
@@ -89,6 +89,9 @@ static void or_irq_class_init(ObjectClass *klass, void *data)
     dc->props = or_irq_properties;
     dc->realize = or_irq_realize;
     dc->vmsd = &vmstate_or_irq;
+
+    /* It's just used internally, and should not be exposed to users */
+    dc->cannot_instantiate_with_device_add_yet = true;
 }
 
 static const TypeInfo or_irq_type_info = {
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 3 lines · Thomas Huth

The "or-irq" device is just used internally. It's strange to see this device showing up in the "-device ?" help text. Let's mark it with cannot_instantiate_with_device_add_yet to hide it from the users.

I agree that it is strange to be showing up to users, but is this really the best way to do this?

I thought we were trying to get rid of cannot_instantiate_with_device_add_yet. Would it maybe be better to have an internal filed that sets this device as internal?

Thanks,

Alistair

Quoted 24 lines
---
 hw/core/or-irq.c | 3 +++
 1 file changed, 3 insertions(+)
hw/core/or-irq.c +3 -1
index 1ac090d..2487fa3 100644
@@ -89,6 +89,9 @@ static void or_irq_class_init(ObjectClass *klass, void *data)
     dc->props = or_irq_properties;
     dc->realize = or_irq_realize;
     dc->vmsd = &vmstate_or_irq;
+
+    /* It's just used internally, and should not be exposed to users */
+    dc->cannot_instantiate_with_device_add_yet = true;
 }
 
 static const TypeInfo or_irq_type_info = {
--

1.8.3.1

#2 lore raw
Quoted 11 lines · Alistair Francis
Quoted 3 lines · Thomas Huth

The "or-irq" device is just used internally. It's strange to see this device showing up in the "-device ?" help text. Let's mark it with cannot_instantiate_with_device_add_yet to hide it from the users.

I agree that it is strange to be showing up to users, but is this really the best way to do this?

I thought we were trying to get rid of cannot_instantiate_with_device_add_yet. Would it maybe be better to have an internal filed that sets this device as internal?

Yes, I agree; the reason for cannot_instantiate_with_device_add_yet being such a long and ugly name is so it's easy to grep for things we in theory might want to fix some day. "We'd rather this didn't show up in help" is a different question.

(Like a lot of devices, it's pretty useless via -device, though, because we have no way to wire up qemu_irq lines on the command line.)

thanks
-- PMM

Reviewed-by
#3 lore raw
Quoted 14 lines · Peter Maydell
Quoted 11 lines · Alistair Francis
Quoted 3 lines · Thomas Huth

The "or-irq" device is just used internally. It's strange to see this device showing up in the "-device ?" help text. Let's mark it with cannot_instantiate_with_device_add_yet to hide it from the users.

I agree that it is strange to be showing up to users, but is this really the best way to do this?

I thought we were trying to get rid of cannot_instantiate_with_device_add_yet. Would it maybe be better to have an internal filed that sets this device as internal?

Quoted 8 lines

Yes, I agree; the reason for cannot_instantiate_with_device_add_yet being such a long and ugly name is so it's easy to grep for things we in theory might want to fix some day. "We'd rather this didn't show up in help" is a different question.

(Like a lot of devices, it's pretty useless via -device, though, because we have no way to wire up qemu_irq lines on the command line.)

And that's precisely why it needs cannot_instantiate_with_device_add_yet set. Like any device that needs code to wire it up.

The commit message should state that. There's no need to argue whether we want to hide it from users.

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
---
 hw/core/or-irq.c | 3 +++
 1 file changed, 3 insertions(+)
Quoted 10 lines
Quoted 10 lines

hw/core/or-irq.c static void or_irq_class_init(ObjectClass *klass, void *data) · line 93

Quoted 10 lines
hw/core/or-irq.c +2 -0
index 1ac090d..2487fa3 100644
@@ -89,6 +89,9 @@ static void or_irq_class_init(ObjectClass *klass, void *data)
     dc->props = or_irq_properties;
     dc->realize = or_irq_realize;
     dc->vmsd = &vmstate_or_irq;
+
+    /* It's just used internally, and should not be exposed to users */

Make that

         /* Reason: needs to be wired up, e.g. by stm32f205_soc_realize() */
Quoted 4 lines
Quoted 4 lines

hw/core/or-irq.c static void or_irq_class_init(ObjectClass *klass, void *data) · line 97

Quoted 4 lines
hw/core/or-irq.c +1 -0
+    dc->cannot_instantiate_with_device_add_yet = true;
 }
 
 static const TypeInfo or_irq_type_info = {

With that and a rephrased commit message