Reconciling qemu-kvm and qemu's PIT

DISCUSSION Avi Kivity 3 participants 10 messages Feb 28, 2012 – Feb 29, 2012 KVM

Share & tools

Avi Kivity

author

2012-02-28 14:33 UTC #0 lore raw

VMStateDescription vmstate_pit = {
.name = "i8254",
.version_id = 3,
.minimum_version_id = 2,
.minimum_version_id_old = 1,
.load_state_old = pit_load_old,
.fields = (VMStateField []) {
<<<<<<< HEAD
VMSTATE_UINT32(flags, PITState),
||||||| merged common ancestors
=======

        VMSTATE_UINT32_V(channels[0].irq_disabled, PITState, 3),

Quoted 1 line

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
ce967e2f33861b0e17753f97fa4527b5943c94b6

VMSTATE_STRUCT_ARRAY(channels, PITState, 3, 2,
vmstate_pit_channel, PITChannelState),

        VMSTATE_TIMER(channels[0].irq_timer, PITState),
        VMSTATE_END_OF_LIST()
    }
};

I'm guessing that flags and irq_disabled are equivalent, but do they have the same sense (that is, do the "1" values have the same meaning)? If not, we have a migration problem.

Is it save to just adopt the new version and drop the old one?

Signature · 1 line

error compiling committee.c: too many arguments to function

Jan Kiszka

2012-02-28 14:40 UTC #1 lore raw

Quoted 16 lines · Avi Kivity

VMStateDescription vmstate_pit = {
.name = "i8254",
.version_id = 3,
.minimum_version_id = 2,
.minimum_version_id_old = 1,
.load_state_old = pit_load_old,
.fields = (VMStateField []) {
<<<<<<< HEAD
VMSTATE_UINT32(flags, PITState),
||||||| merged common ancestors
=======
        VMSTATE_UINT32_V(channels[0].irq_disabled, PITState, 3),
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
ce967e2f33861b0e17753f97fa4527b5943c94b6
VMSTATE_STRUCT_ARRAY(channels, PITState, 3, 2,
vmstate_pit_channel, PITChannelState),

Quoted 7 lines

        VMSTATE_TIMER(channels[0].irq_timer, PITState),
        VMSTATE_END_OF_LIST()
    }
};
I'm guessing that flags and irq_disabled are equivalent, but do they have the same sense (that is, do the "1" values have the same meaning)?

Yes. qemu-kvm sets flags to 0 or PIT_FLAGS_HPET_LEGACY, which is 1. And the latter means "irq_disabled".

Quoted 3 lines

If not, we have a migration problem.
Is it save to just adopt the new version and drop the old one?

The new upstream code was designed to match qemu-kvm's migration format, so you can switch. Of course, the result needs a careful check.

Jan

Signature · 2 lines

Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux

Avi Kivity

author

2012-02-28 14:42 UTC #2 lore raw

Quoted 28 lines · Jan Kiszka

Quoted 16 lines · Avi Kivity
VMStateDescription vmstate_pit = {
.name = "i8254",
.version_id = 3,
.minimum_version_id = 2,
.minimum_version_id_old = 1,
.load_state_old = pit_load_old,
.fields = (VMStateField []) {
<<<<<<< HEAD
VMSTATE_UINT32(flags, PITState),
||||||| merged common ancestors
=======
        VMSTATE_UINT32_V(channels[0].irq_disabled, PITState, 3),
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
ce967e2f33861b0e17753f97fa4527b5943c94b6
VMSTATE_STRUCT_ARRAY(channels, PITState, 3, 2,
vmstate_pit_channel, PITChannelState),
Quoted 7 lines
        VMSTATE_TIMER(channels[0].irq_timer, PITState),
        VMSTATE_END_OF_LIST()
    }
};
I'm guessing that flags and irq_disabled are equivalent, but do they have the same sense (that is, do the "1" values have the same meaning)?
Yes. qemu-kvm sets flags to 0 or PIT_FLAGS_HPET_LEGACY, which is 1. And the latter means "irq_disabled".

Quoted 6 lines

Quoted 3 lines
If not, we have a migration problem.
Is it save to just adopt the new version and drop the old one?
The new upstream code was designed to match qemu-kvm's migration format, so you can switch.

Not entirely unexpected, but I wanted to make sure.

Quoted 2 lines

Of course, the result needs a careful check.

Thanks!

Signature · 1 line

error compiling committee.c: too many arguments to function

Avi Kivity

author

2012-02-28 19:50 UTC #3 lore raw

On 02/28/2012 04:42 PM, Avi Kivity wrote:

<snip>

I'm getting a crash now:

#0  0x00007ffff4ea0285 in raise    () from /lib64/libc.so.6
#1  0x00007ffff4ea1b9b in abort    () from /lib64/libc.so.6
#2  0x00007ffff4e98e9e in __assert_fail_base ()    from /lib64/libc.so.6
#3  0x00007ffff4e98f42 in __assert_fail    () from /lib64/libc.so.6
#4  0x000055555571c69b in qdev_connect_gpio_out    (dev=0x555556913a70,

n=0, pin=0x5555568ef770) at /home/tlv/akivity/qemu/hw/qdev.c:297
#5 0x000055555582beae in pit_init (bus=0x5555568d5610, base=64,
isa_irq=-1, alt_irq=0x5555568ef770) at /home/tlv/akivity/qemu/hw/i8254.h:85 #6 0x000055555582ebbc in pc_basic_device_init
(isa_bus=0x5555568d5610, gsi=0x5555568cadb0,
rtc_state=0x7fffffffdea0, floppy=0x7fffffffdea8, no_vmport=false)

    at /home/tlv/akivity/qemu/hw/pc.c:1182
#7  0x000055555582f454 in pc_init1 (system_memory=0x55555646f270,
system_io=0x55555646f360, ram_size=1073741824,
boot_device=0x7fffffffe300 "cad", kernel_filename=0x0,   
    kernel_cmdline=0x5555558baa12 "", initrd_filename=0x0,

cpu_model=0x0, pci_enabled=1, kvmclock_enabled=1) at /home/tlv/akivity/qemu/hw/pc_piix.c:256
#8 0x000055555582f8d7 in pc_init_pci (ram_size=1073741824, boot_device=0x7fffffffe300 "cad", kernel_filename=0x0, kernel_cmdline=0x5555558baa12 "", initrd_filename=0x0,

    cpu_model=0x0) at /home/tlv/akivity/qemu/hw/pc_piix.c:335

#9 0x00005555556f1349 in main (argc=6, argv=0x7fffffffe428,
envp=0x7fffffffe460) at /home/tlv/akivity/qemu/vl.c:3431

Looks like isa-pit has zero gpio pins, so it fails when crashing. I must have mismerged it, but where is the gpio pin count set?

Signature · 1 line

error compiling committee.c: too many arguments to function

Jan Kiszka

2012-02-28 21:48 UTC #4 lore raw

Quoted 27 lines · Avi Kivity

On 02/28/2012 04:42 PM, Avi Kivity wrote:
<snip>
I'm getting a crash now:
#0  0x00007ffff4ea0285 in raise    () from /lib64/libc.so.6
#1  0x00007ffff4ea1b9b in abort    () from /lib64/libc.so.6
#2  0x00007ffff4e98e9e in __assert_fail_base ()    from /lib64/libc.so.6
#3  0x00007ffff4e98f42 in __assert_fail    () from /lib64/libc.so.6
#4  0x000055555571c69b in qdev_connect_gpio_out    (dev=0x555556913a70,
n=0, pin=0x5555568ef770) at /home/tlv/akivity/qemu/hw/qdev.c:297
#5 0x000055555582beae in pit_init (bus=0x5555568d5610, base=64,
isa_irq=-1, alt_irq=0x5555568ef770) at /home/tlv/akivity/qemu/hw/i8254.h:85 #6 0x000055555582ebbc in pc_basic_device_init
(isa_bus=0x5555568d5610, gsi=0x5555568cadb0,
rtc_state=0x7fffffffdea0, floppy=0x7fffffffdea8, no_vmport=false)
    at /home/tlv/akivity/qemu/hw/pc.c:1182
#7  0x000055555582f454 in pc_init1 (system_memory=0x55555646f270,
system_io=0x55555646f360, ram_size=1073741824,
boot_device=0x7fffffffe300 "cad", kernel_filename=0x0,   
    kernel_cmdline=0x5555558baa12 "", initrd_filename=0x0,
cpu_model=0x0, pci_enabled=1, kvmclock_enabled=1) at /home/tlv/akivity/qemu/hw/pc_piix.c:256
#8 0x000055555582f8d7 in pc_init_pci (ram_size=1073741824, boot_device=0x7fffffffe300 "cad", kernel_filename=0x0, kernel_cmdline=0x5555558baa12 "", initrd_filename=0x0,

Quoted 6 lines

    cpu_model=0x0) at /home/tlv/akivity/qemu/hw/pc_piix.c:335
#9 0x00005555556f1349 in main (argc=6, argv=0x7fffffffe428,
envp=0x7fffffffe460) at /home/tlv/akivity/qemu/vl.c:3431
Looks like isa-pit has zero gpio pins, so it fails when crashing. I must have mismerged it, but where is the gpio pin count set?

In pit_initfn. Does -no-kvm-irqchip work fine? It's a bit tricky to discuss this without seeing your code.

Jan

Avi Kivity

author

2012-02-29 09:29 UTC #5 lore raw

Quoted 7 lines · Jan Kiszka

Quoted 3 lines
Looks like isa-pit has zero gpio pins, so it fails when crashing. I must have mismerged it, but where is the gpio pin count set?
In pit_initfn. Does -no-kvm-irqchip work fine? It's a bit tricky to discuss this without seeing your code.

pushed it into qemu-kvm.git usptream-merge

Signature · 1 line

error compiling committee.c: too many arguments to function

Jan Kiszka

2012-02-29 10:14 UTC #6 lore raw

Quoted 11 lines · Avi Kivity

Quoted 7 lines · Jan Kiszka
Quoted 3 lines
Looks like isa-pit has zero gpio pins, so it fails when crashing. I must have mismerged it, but where is the gpio pin count set?
In pit_initfn. Does -no-kvm-irqchip work fine? It's a bit tricky to discuss this without seeing your code.
pushed it into qemu-kvm.git usptream-merge

As I assumed: You try to initialize the kvm-pit like the userspace pit, that cannot work. Don't run the qdev_connect_gpio_out in pit_init, just like my kvm-pit patches does in their kvm_pit_init.

Jan

Signature · 2 lines

Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux

Avi Kivity

author

2012-02-29 13:45 UTC #7 lore raw

Quoted 10 lines · Jan Kiszka

Quoted 9 lines · Avi Kivity
Quoted 7 lines · Jan Kiszka
Quoted 3 lines
Looks like isa-pit has zero gpio pins, so it fails when crashing. I must have mismerged it, but where is the gpio pin count set?
In pit_initfn. Does -no-kvm-irqchip work fine? It's a bit tricky to discuss this without seeing your code.

Quoted 7 lines

Quoted 2 lines
pushed it into qemu-kvm.git usptream-merge
As I assumed: You try to initialize the kvm-pit like the userspace pit, that cannot work. Don't run the qdev_connect_gpio_out in pit_init, just like my kvm-pit patches does in their kvm_pit_init.

The merge is now in 'next', hopefully not too many regressions.

Signature · 1 line

error compiling committee.c: too many arguments to function

Jan Kiszka

2012-02-29 14:24 UTC #8 lore raw

Quoted 14 lines · Avi Kivity

Quoted 13 lines · Jan Kiszka
Quoted 11 lines · Avi Kivity
Quoted 7 lines · Jan Kiszka
Quoted 3 lines
Looks like isa-pit has zero gpio pins, so it fails when crashing. I must have mismerged it, but where is the gpio pin count set?
In pit_initfn. Does -no-kvm-irqchip work fine? It's a bit tricky to discuss this without seeing your code.
pushed it into qemu-kvm.git usptream-merge

Quoted 6 lines

Quoted 4 lines
As I assumed: You try to initialize the kvm-pit like the userspace pit, that cannot work. Don't run the qdev_connect_gpio_out in pit_init, just like my kvm-pit patches does in their kvm_pit_init.
The merge is now in 'next', hopefully not too many regressions.

Hope we can quickly finish the switch to the new kvm-pit.

When do you plan to push or rebase uq/master? Then I could check/refresh my kvm-pit series on top of it.

Jan

Signature · 2 lines

Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux

Avi Kivity

author

2012-02-29 14:42 UTC #9 lore raw

Quoted 20 lines · Jan Kiszka

Quoted 19 lines · Avi Kivity
Quoted 17 lines · Jan Kiszka
Quoted 11 lines · Avi Kivity
Quoted 7 lines · Jan Kiszka
Quoted 3 lines
Looks like isa-pit has zero gpio pins, so it fails when crashing. I must have mismerged it, but where is the gpio pin count set?
In pit_initfn. Does -no-kvm-irqchip work fine? It's a bit tricky to discuss this without seeing your code.
pushed it into qemu-kvm.git usptream-merge
As I assumed: You try to initialize the kvm-pit like the userspace pit, that cannot work. Don't run the qdev_connect_gpio_out in pit_init, just like my kvm-pit patches does in their kvm_pit_init.

Quoted 6 lines

Quoted 1 line
The merge is now in 'next', hopefully not too many regressions.
Hope we can quickly finish the switch to the new kvm-pit.
When do you plan to push or rebase uq/master? Then I could check/refresh my kvm-pit series on top of it.

Real soon... I hoped to do it after the upstream merge, but it doesn't pass autotest due to a screendump regression.

Signature · 1 line

error compiling committee.c: too many arguments to function