hw/isa/piix4.c | 54 +++++++++++++++++++++-- hw/mips/gt64xxx_pci.c | 80 +++-------------------------------- hw/mips/malta.c | 7 +-- include/hw/mips/mips.h | 3 -- include/hw/southbridge/piix.h | 2 - 5 files changed, 59 insertions(+), 87 deletions(-)
Tested with [1]:
qemu-system-mipsel -M malta -kernel vmlinux-3.2.0-4-4kc-malta -hda \
debian_wheezy_mipsel_standard.qcow2 -append "root=/dev/sda1 console=tty0"
It was possible to log in as root and `poweroff` the machine.
[1] https://people.debian.org/~aurel32/qemu/mips/
v4:
As suggested by BALATON Zoltan, gt64120_register() is now resolved by
sysbus_create_simple() etc. This also fixes the nitpick pointed out by
Philippe.
v3:
The migration bug now gets fixed in gt64xxx_pci before any cleanup. As
suggested by PMM the patch is based on commit e735b55a8c11.
The code movement patch now moves the already fixed code. I might be a bit
too conservative here by removing Philippe's Reviewed-By tag.
As suggested by BALATON Zoltan, the redundant i8259[] attribute is now
resolved immediately after the code movement. As a side effect, it also
removes moved code which doesn't adhere to the coding style (local loop
variable).
To address BALATON Zoltan's comment and to reduce the number of required
Reviewed-By's, only piix4_set_irq() is modified to expect own DeviceState
paremeter. Up to v2, all remaining set_irq() functions were changed this
way.
The patch resolving piix4's singleton variable got split into two patches:
One which resolves the singleton variable and one which replaces magic
constants. The split patches should be more comprehensible.
Suggested by BALATON Zoltan, I took a chance to resolve gt64120_register(),
a method akin to the legacy init functions we're trying to get rid of.
v2:
isa/piix4: Fix PCI IRQ levels to be preserved in VMState
isa/piix4: Resolve redundant i8259[] attribute
Bernhard Beschow (7):
hw/mips/gt64xxx_pci: Fix PCI IRQ levels to be preserved during
migration
malta: Move PCI interrupt handling from gt64xxx_pci to piix4
hw/isa/piix4: Resolve redundant i8259[] attribute
hw/isa/piix4: Pass PIIX4State as opaque parameter for piix4_set_irq()
hw/isa/piix4: Resolve global instance variable
hw/isa/piix4: Replace some magic IRQ constants
hw/mips/gt64xxx_pci: Resolve gt64120_register()
hw/isa/piix4.c | 54 +++++++++++++++++++++--
hw/mips/gt64xxx_pci.c | 80 +++--------------------------------
hw/mips/malta.c | 7 +--
include/hw/mips/mips.h | 3 --
include/hw/southbridge/piix.h | 2 -
5 files changed, 59 insertions(+), 87 deletions(-)
--
2.35.1
Am 17. Februar 2022 10:19:17 UTC schrieb Bernhard Beschow <shentey@gmail.com>: >Tested with [1]: > > qemu-system-mipsel -M malta -kernel vmlinux-3.2.0-4-4kc-malta -hda \ > debian_wheezy_mipsel_standard.qcow2 -append "root=/dev/sda1 console=tty0" > >It was possible to log in as root and `poweroff` the machine. > >[1] https://people.debian.org/~aurel32/qemu/mips/ > > >v4: > As suggested by BALATON Zoltan, gt64120_register() is now resolved by > sysbus_create_simple() etc. This also fixes the nitpick pointed out by > Philippe. > >v3: > The migration bug now gets fixed in gt64xxx_pci before any cleanup. As > suggested by PMM the patch is based on commit e735b55a8c11. > The code movement patch now moves the already fixed code. I might be a bit > too conservative here by removing Philippe's Reviewed-By tag. > As suggested by BALATON Zoltan, the redundant i8259[] attribute is now > resolved immediately after the code movement. As a side effect, it also > removes moved code which doesn't adhere to the coding style (local loop > variable). > To address BALATON Zoltan's comment and to reduce the number of required > Reviewed-By's, only piix4_set_irq() is modified to expect own DeviceState > paremeter. Up to v2, all remaining set_irq() functions were changed this > way. > The patch resolving piix4's singleton variable got split into two patches: > One which resolves the singleton variable and one which replaces magic > constants. The split patches should be more comprehensible. > Suggested by BALATON Zoltan, I took a chance to resolve gt64120_register(), > a method akin to the legacy init functions we're trying to get rid of. > >v2: > isa/piix4: Fix PCI IRQ levels to be preserved in VMState > isa/piix4: Resolve redundant i8259[] attribute > >Bernhard Beschow (7): > hw/mips/gt64xxx_pci: Fix PCI IRQ levels to be preserved during > migration > malta: Move PCI interrupt handling from gt64xxx_pci to piix4 > hw/isa/piix4: Resolve redundant i8259[] attribute > hw/isa/piix4: Pass PIIX4State as opaque parameter for piix4_set_irq() > hw/isa/piix4: Resolve global instance variable > hw/isa/piix4: Replace some magic IRQ constants > hw/mips/gt64xxx_pci: Resolve gt64120_register() > > hw/isa/piix4.c | 54 +++++++++++++++++++++-- > hw/mips/gt64xxx_pci.c | 80 +++-------------------------------- > hw/mips/malta.c | 7 +-- > include/hw/mips/mips.h | 3 -- > include/hw/southbridge/piix.h | 2 - > 5 files changed, 59 insertions(+), 87 deletions(-) > @Phil: Would you also queue this series in your MIPS tree? Best regards, Bernhard
Am 17. Februar 2022 10:19:17 UTC schrieb Bernhard Beschow <shentey@gmail.com>: >Tested with [1]: > > qemu-system-mipsel -M malta -kernel vmlinux-3.2.0-4-4kc-malta -hda \ > debian_wheezy_mipsel_standard.qcow2 -append "root=/dev/sda1 console=tty0" > >It was possible to log in as root and `poweroff` the machine. > >[1] https://people.debian.org/~aurel32/qemu/mips/ > > >v4: > As suggested by BALATON Zoltan, gt64120_register() is now resolved by > sysbus_create_simple() etc. This also fixes the nitpick pointed out by > Philippe. > >v3: > The migration bug now gets fixed in gt64xxx_pci before any cleanup. As > suggested by PMM the patch is based on commit e735b55a8c11. > The code movement patch now moves the already fixed code. I might be a bit > too conservative here by removing Philippe's Reviewed-By tag. > As suggested by BALATON Zoltan, the redundant i8259[] attribute is now > resolved immediately after the code movement. As a side effect, it also > removes moved code which doesn't adhere to the coding style (local loop > variable). > To address BALATON Zoltan's comment and to reduce the number of required > Reviewed-By's, only piix4_set_irq() is modified to expect own DeviceState > paremeter. Up to v2, all remaining set_irq() functions were changed this > way. > The patch resolving piix4's singleton variable got split into two patches: > One which resolves the singleton variable and one which replaces magic > constants. The split patches should be more comprehensible. > Suggested by BALATON Zoltan, I took a chance to resolve gt64120_register(), > a method akin to the legacy init functions we're trying to get rid of. > >v2: > isa/piix4: Fix PCI IRQ levels to be preserved in VMState > isa/piix4: Resolve redundant i8259[] attribute > >Bernhard Beschow (7): > hw/mips/gt64xxx_pci: Fix PCI IRQ levels to be preserved during > migration > malta: Move PCI interrupt handling from gt64xxx_pci to piix4 > hw/isa/piix4: Resolve redundant i8259[] attribute > hw/isa/piix4: Pass PIIX4State as opaque parameter for piix4_set_irq() > hw/isa/piix4: Resolve global instance variable > hw/isa/piix4: Replace some magic IRQ constants > hw/mips/gt64xxx_pci: Resolve gt64120_register() > > hw/isa/piix4.c | 54 +++++++++++++++++++++-- > hw/mips/gt64xxx_pci.c | 80 +++-------------------------------- > hw/mips/malta.c | 7 +-- > include/hw/mips/mips.h | 3 -- > include/hw/southbridge/piix.h | 2 - > 5 files changed, 59 insertions(+), 87 deletions(-) > Hi, all patches have at least one Reviewed-by tag. What's missing for merging? Best regards, Bernhard
On 25/2/22 21:36, Bernhard Beschow wrote: > Am 17. Februar 2022 10:19:17 UTC schrieb Bernhard Beschow <shentey@gmail.com>: >> Tested with [1]: >> >> qemu-system-mipsel -M malta -kernel vmlinux-3.2.0-4-4kc-malta -hda \ >> debian_wheezy_mipsel_standard.qcow2 -append "root=/dev/sda1 console=tty0" >> >> It was possible to log in as root and `poweroff` the machine. >> >> [1] https://people.debian.org/~aurel32/qemu/mips/ >> >> >> v4: >> As suggested by BALATON Zoltan, gt64120_register() is now resolved by >> sysbus_create_simple() etc. This also fixes the nitpick pointed out by >> Philippe. >> >> v3: >> The migration bug now gets fixed in gt64xxx_pci before any cleanup. As >> suggested by PMM the patch is based on commit e735b55a8c11. >> The code movement patch now moves the already fixed code. I might be a bit >> too conservative here by removing Philippe's Reviewed-By tag. >> As suggested by BALATON Zoltan, the redundant i8259[] attribute is now >> resolved immediately after the code movement. As a side effect, it also >> removes moved code which doesn't adhere to the coding style (local loop >> variable). >> To address BALATON Zoltan's comment and to reduce the number of required >> Reviewed-By's, only piix4_set_irq() is modified to expect own DeviceState >> paremeter. Up to v2, all remaining set_irq() functions were changed this >> way. >> The patch resolving piix4's singleton variable got split into two patches: >> One which resolves the singleton variable and one which replaces magic >> constants. The split patches should be more comprehensible. >> Suggested by BALATON Zoltan, I took a chance to resolve gt64120_register(), >> a method akin to the legacy init functions we're trying to get rid of. >> >> v2: >> isa/piix4: Fix PCI IRQ levels to be preserved in VMState >> isa/piix4: Resolve redundant i8259[] attribute >> >> Bernhard Beschow (7): >> hw/mips/gt64xxx_pci: Fix PCI IRQ levels to be preserved during >> migration >> malta: Move PCI interrupt handling from gt64xxx_pci to piix4 >> hw/isa/piix4: Resolve redundant i8259[] attribute >> hw/isa/piix4: Pass PIIX4State as opaque parameter for piix4_set_irq() >> hw/isa/piix4: Resolve global instance variable >> hw/isa/piix4: Replace some magic IRQ constants >> hw/mips/gt64xxx_pci: Resolve gt64120_register() >> >> hw/isa/piix4.c | 54 +++++++++++++++++++++-- >> hw/mips/gt64xxx_pci.c | 80 +++-------------------------------- >> hw/mips/malta.c | 7 +-- >> include/hw/mips/mips.h | 3 -- >> include/hw/southbridge/piix.h | 2 - >> 5 files changed, 59 insertions(+), 87 deletions(-) >> > > Hi, > > all patches have at least one Reviewed-by tag. What's missing for merging? Time ;) Series queued to mips-next. Thanks for the nice - long due - cleanup! Phil.
© 2016 - 2026 Red Hat, Inc.