[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PATCH 00/22] Indirection Cleanup
From: |
Juan Quintela |
Subject: |
[Qemu-devel] [PATCH 00/22] Indirection Cleanup |
Date: |
Mon, 24 Aug 2009 13:03:21 +0200 |
This patch series clean up "half" converted qemu drivers that had changed from:
struct FOOState
to
typedef PCIFOOState {
PCIDevice dev;
FOOState foo;
} PCIFOOState;
It just moves PCIDevice to be the 1st field of FOOState.
Once there, other cleanups were done:
a - pci_dev pointer from FOOState to PCIFOOState is removed, jsut use s->dev
The field is leave only in the drivers that also emulate isa.
b- Once there, transformo
PCIFOOState *s = (PCIFOOState *)pci_dev
to
PCIFOOState *s = DO_UPCAST(PCIFOOState, dev, pci_dev)
where pci_dev is a PCI_DEVICE.
c- again, once there, remove all the casts from void * (they are not
needed since '89)
PCIFOOState *s = (PCIFOOState *)opaque;
to
PCIFOOState *s = opaque;
d- Start of vga.c cleanup. It is not trivial, as just now
VGAState == VGACommonState, functions need to be changed to use the right
value.
ToDo:
- pcnet: It needs a different approach, because it can be both a PCIDevice
or a SysBus device.
- vga: It have to separate the common part for the not common part, problems now
is that VGAState is used for both the common state and the standard vga.
To make things more interesting, different bits are "inherited" from vga
in different devices:
- cirrus_vga
- vmware_vga
- blizzard (vga that is not pci)
- vga common state has a pci_dev pointer, that is only needed by std vga,
as cirrus_vga stores it (with this patches) otherplace, blizzard is not pci
.... you get the idea
- I guess some other driver is missing, but my fast grep didn't found it.
Later, Juan.
Juan Quintela (22):
eepro100: convert casts to DO_UPCAST()
eepro100: cast a void * makes no sense
eepro100: Remove unused indirection of PCIDevice
es1370: Remove unused indirection of PCIES1370State and ES1370State
ne2000: remove casts from void *
ne2000: pci_dev has this very value with the right type
ne2000: Remove unneeded double indirection of PCINE2000State
ne2000: change pci_dev to is_pci
pci: remove casts from void *
rtl8139: Remove unneeded double indirection of PCIRTL8139State
rtl8139: remove pointless cast from void *
lsi53c895a: remove pointless cast from void *
lsi53c895a: use DO_UPCAST to cast from PCIDevice
lsi53c895a: rename PCIDevice field from pci_dev to dev (consistence)
lsi53c895a: LSIState is a PCIDevice is a DeviceHost
usb-ohci: Remove unneeded double indirection of OHCIPCIState
cirrus_vga: Remove unneeded double indirection of PCICirrusVGAState
cirrus_vga: remove pointless cast from void *
cirrus_vga: change use of pci_dev for is_pci
Introduce vga_common_reset() to be able to typcheck vga_reset()
vga: Rename vga_state -> vga
Everything outside of vga.c should use VGACommonState
hw/cirrus_vga.c | 59 ++++++++++++++++++++++---------------------------
hw/eepro100.c | 65 ++++++++++++++++++++++--------------------------------
hw/es1370.c | 31 ++++++++-----------------
hw/lsi53c895a.c | 65 ++++++++++++++++++++++++++++---------------------------
hw/ne2000.c | 41 ++++++++++++++--------------------
hw/pci.c | 8 +++---
hw/rtl8139.c | 42 ++++++++++++----------------------
hw/usb-ohci.c | 34 +++++++++++-----------------
hw/vga.c | 26 +++++++++++++--------
hw/vga_int.h | 12 ++++------
hw/vmware_vga.c | 4 +-
11 files changed, 170 insertions(+), 217 deletions(-)
- [Qemu-devel] [PATCH 00/22] Indirection Cleanup,
Juan Quintela <=
- [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense, Juan Quintela, 2009/08/24
- Re: [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense, Stefan Weil, 2009/08/24
- Re: [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense, Markus Armbruster, 2009/08/24
- [Qemu-devel] Coding style, C++ compatible code (was Re: [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense), Stefan Weil, 2009/08/26
- Re: [Qemu-devel] Coding style, C++ compatible code (was Re: [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense), Gerd Hoffmann, 2009/08/26
- Re: [Qemu-devel] Coding style, C++ compatible code (was Re: [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense), Jamie Lokier, 2009/08/26
- Re: [Qemu-devel] Coding style, C++ compatible code (was Re: [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense), malc, 2009/08/26
- Re: [Qemu-devel] Coding style, C++ compatible code (was Re: [Qemu-devel] [PATCH 02/22] eepro100: cast a void * makes no sense), Jamie Lokier, 2009/08/26