Features/QDevCleanup: Difference between revisions

From QEMU
No edit summary
 
(2 intermediate revisions by one other user not shown)
Line 47: Line 47:
|-
|-
|  Refactor reset in qdev.c
|  Refactor reset in qdev.c
In Progress
Ready
|-
|-
|  Split monitor out of qdev.c
|  Split monitor out of qdev.c
In Progress
Ready
|-
|-
|  Split device_add out of qdev.c
|  Split device_add out of qdev.c
In Progress
Ready
|-
|-
|  Refactor hotplug in qdev
|  Refactor hotplug in qdev
In Progress
Ready
|-
|-
|  Get rid of no_user
|  Get rid of no_user
In Progress
Ready
|-
|-
|  Move sysbus out of qdev.c and use proper singleton
|  Move sysbus out of qdev.c and use proper singleton
|  Ready
|-
|  Make VMStateDescription have no code dependencies
|  Ready
|-
|  Add FIDL based auto-serialization
|  Ready
|-
|  Move vmstate_register out of qdev.c
|  In Progress
|  In Progress
|-
|-
Move vmstate_register out of qdev.c
Add a required name parameter to qdev
|  In Progress
|  In Progress
|-
|-
|  Build qdev.c in a library
|  Build qdev.c in a library
|  
| In Progress
|-
|-
|  Unit tests for qdev
|  Unit tests for qdev
Line 82: Line 91:
|
|
|-
|-
Refactor sysbus to not use target_phys_addr_t
Introduce hierarchical PIO/MMIO dispatch
|
|
|}
|}
Line 89: Line 98:


* http://repo.or.cz/w/qemu/aliguori.git features/qdev-cleanup
* http://repo.or.cz/w/qemu/aliguori.git features/qdev-cleanup
[[Category:Obsolete feature pages]]

Latest revision as of 14:45, 11 October 2016

Summary

Cleanup qdev's object model and move it into a shared library.

Owner

Detailed Summary

From the ML discussion:

All devices should have a DeviceState associated with them.  Otherwise, there's really no point in having qdev at all.

We have lots of devices today that don't have DeviceState's associated with them because the have a separate qdev representation with a reference to the non-DeviceState object.

We have non-DeviceState objects because otherwise we end up with an inheritance diamond.  We have this problem because we want to have relationships like: DeviceState <- SystemDeviceState <- ISADevice <- ISASerialDevice.

But ISASerialDevice is not the only type of serial device.  You can also have a SystemSerialDevice that's directly attached to the System bus.  That means you'd have to have:

SerialDevice -> ISASerialDevice -> SystemDeviceState -> DeviceState
                     -> SystemSerialDevice -> SystemDeviceState -> DeviceState

Which is a classic MI diamond.  The only way to resolve this modelling problem is to split out the common code and rely on a has-a relationship instead of an is-a.  That gives you:

ISASerialDevice->SystemDeviceState->DeviceState
SystemSerialDevice->SystemDeviceState->DeviceState

ISASerialDevice has-a SerialDevice
SystemSerialDevice has-a SerialDevice

And since we want SerialDevice inherit from a DeviceState (recall, all devices should have DeviceStates):

SerialDevice->DeviceState

No more MI diamond and all devices have DeviceStates.  Coincidentally, it matches more closely how hardware works..

Generally speaking, any time we have one device that needs to sit on multiple busses, we're going to have to model it in this fashion. 

TODO

Item Status
Refactor reset in qdev.c Ready
Split monitor out of qdev.c Ready
Split device_add out of qdev.c Ready
Refactor hotplug in qdev Ready
Get rid of no_user Ready
Move sysbus out of qdev.c and use proper singleton Ready
Make VMStateDescription have no code dependencies Ready
Add FIDL based auto-serialization Ready
Move vmstate_register out of qdev.c In Progress
Add a required name parameter to qdev In Progress
Build qdev.c in a library In Progress
Unit tests for qdev
Refactor serial to use proper modeling In Progress
Move serial (chip only) to libqdev
Unit tests for serial
Introduce hierarchical PIO/MMIO dispatch

Status