[Qemu-devel] [PATCH] Check if the i8254 timer is active before deactivating it

[Qemu-devel] [PATCH] Check if the i8254 timer is active before deactivating it

[Alexander Graf]

The HPET emulation can disable the i8254 when the HPET is
in legacy mode, thus emulating the i8254's behavior.

But if it does, the i8254 doesn't have to be running, so
let's check to see if the timer works and not disable it
if it's not.

This fixes a segmentation fault when running Mac OS X as
guest os.

Signed-off-by: Alexander Graf <agraf@suse.de>
---
 hw/i8254.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/hw/i8254.c b/hw/i8254.c
index a4a1efe..44e4531 100644
--- a/hw/i8254.c
+++ b/hw/i8254.c
@@ -467,7 +467,8 @@ static void pit_reset(void *opaque)
 void hpet_pit_disable(void) {
     PITChannelState *s;
     s = &pit_state.channels[0];
-    qemu_del_timer(s->irq_timer);
+    if (s->irq_timer)
+        qemu_del_timer(s->irq_timer);
 }
 
 /* When HPET is reset or leaving legacy mode, it must reenable i8254
-- 
1.6.0.2

Re: [Qemu-devel] [PATCH] Check if the i8254 timer is active before deactivating it

[Anthony Liguori]

Alexander Graf wrote:
> The HPET emulation can disable the i8254 when the HPET is
> in legacy mode, thus emulating the i8254's behavior.
>
> But if it does, the i8254 doesn't have to be running, so
> let's check to see if the timer works and not disable it
> if it's not.
>
> This fixes a segmentation fault when running Mac OS X as
> guest os.
>   

So the HPET works with Mac OS X with this patch?  That's good to know 
since there was a fair bit of change from your original patch.

> Signed-off-by: Alexander Graf <agraf@suse.de>
>   

Applied.  Thanks.

Regards,

Anthony Liguori

> ---
>  hw/i8254.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/hw/i8254.c b/hw/i8254.c
> index a4a1efe..44e4531 100644
> --- a/hw/i8254.c
> +++ b/hw/i8254.c
> @@ -467,7 +467,8 @@ static void pit_reset(void *opaque)
>  void hpet_pit_disable(void) {
>      PITChannelState *s;
>      s = &pit_state.channels[0];
> -    qemu_del_timer(s->irq_timer);
> +    if (s->irq_timer)
> +        qemu_del_timer(s->irq_timer);
>  }
>  
>  /* When HPET is reset or leaving legacy mode, it must reenable i8254
>   

Re: [Qemu-devel] [PATCH] Check if the i8254 timer is active before deactivating it

[Alexander Graf]

On 26.01.2009, at 21:33, Anthony Liguori <aliguori@us.ibm.com> wrote:

> Alexander Graf wrote:
>> The HPET emulation can disable the i8254 when the HPET is
>> in legacy mode, thus emulating the i8254's behavior.
>>
>> But if it does, the i8254 doesn't have to be running, so
>> let's check to see if the timer works and not disable it
>> if it's not.
>>
>> This fixes a segmentation fault when running Mac OS X as
>> guest os.
>>
>
> So the HPET works with Mac OS X with this patch?  That's good to  
> know since there was a fair bit of change from your original patch.

Yes, works perfectly fine as long as you add the hpet descriptor in  
acpi.

So far the only user in osx that uses the hpet I've encountered is the  
power management though - and so far that only checks if 1 ibterrupt  
arrives on bootup.

But beth's changes do look sane, so I'd guess if anything the new  
implementation is more likely to work ;). My old was was actually just  
a hack to get osx running at all, so I'm really grateful someons  
cleaned it up.

Alex

>
>
>> Signed-off-by: Alexander Graf <agraf@suse.de>
>>
>
> Applied.  Thanks.
>
> Regards,
>
> Anthony Liguori
>
>> ---
>> hw/i8254.c |    3 ++-
>> 1 files changed, 2 insertions(+), 1 deletions(-)
>>
>> diff --git a/hw/i8254.c b/hw/i8254.c
>> index a4a1efe..44e4531 100644
>> --- a/hw/i8254.c
>> +++ b/hw/i8254.c
>> @@ -467,7 +467,8 @@ static void pit_reset(void *opaque)
>> void hpet_pit_disable(void) {
>>     PITChannelState *s;
>>     s = &pit_state.channels[0];
>> -    qemu_del_timer(s->irq_timer);
>> +    if (s->irq_timer)
>> +        qemu_del_timer(s->irq_timer);
>> }
>>  /* When HPET is reset or leaving legacy mode, it must reenable i8254
>>
>
>
>

Re: [Qemu-devel] [PATCH] Check if the i8254 timer is active before deactivating it

[Alexander Graf]

On 27.01.2009, at 08:18, Alexander Graf <agraf@suse.de> wrote:

>
>
>
>
> On 26.01.2009, at 21:33, Anthony Liguori <aliguori@us.ibm.com> wrote:
>
>> Alexander Graf wrote:
>>> The HPET emulation can disable the i8254 when the HPET is
>>> in legacy mode, thus emulating the i8254's behavior.
>>>
>>> But if it does, the i8254 doesn't have to be running, so
>>> let's check to see if the timer works and not disable it
>>> if it's not.
>>>
>>> This fixes a segmentation fault when running Mac OS X as
>>> guest os.
>>>
>>
>> So the HPET works with Mac OS X with this patch?  That's good to  
>> know since there was a fair bit of change from your original patch.
>
> Yes, works perfectly fine as long as you add the hpet descriptor in  
> acpi.

Oh and the LPC of course, which tells osx where to look for the hpet,  
before it reads the acpi tables.

Alex

>
>
> So far the only user in osx that uses the hpet I've encountered is  
> the power management though - and so far that only checks if 1  
> ibterrupt arrives on bootup.
>
> But beth's changes do look sane, so I'd guess if anything the new  
> implementation is more likely to work ;). My old was was actually  
> just a hack to get osx running at all, so I'm really grateful  
> someons cleaned it up.
>
> Alex
>
>>
>>
>>> Signed-off-by: Alexander Graf <agraf@suse.de>
>>>
>>
>> Applied.  Thanks.
>>
>> Regards,
>>
>> Anthony Liguori
>>
>>> ---
>>> hw/i8254.c |    3 ++-
>>> 1 files changed, 2 insertions(+), 1 deletions(-)
>>>
>>> diff --git a/hw/i8254.c b/hw/i8254.c
>>> index a4a1efe..44e4531 100644
>>> --- a/hw/i8254.c
>>> +++ b/hw/i8254.c
>>> @@ -467,7 +467,8 @@ static void pit_reset(void *opaque)
>>> void hpet_pit_disable(void) {
>>>    PITChannelState *s;
>>>    s = &pit_state.channels[0];
>>> -    qemu_del_timer(s->irq_timer);
>>> +    if (s->irq_timer)
>>> +        qemu_del_timer(s->irq_timer);
>>> }
>>> /* When HPET is reset or leaving legacy mode, it must reenable i8254
>>>
>>
>>
>>
>
>