Re: [Qemu-devel] [PATCH 3/7] s390x: fix error propagation in kvm-flic's realize
On 07/04/2017 06:59 PM, Cornelia Huck wrote:
> On Tue, 4 Jul 2017 17:08:52 +0200
> Halil Pasic wrote:
>
> cd.type = KVM_DEV_TYPE_FLIC;
> ret = kvm_vm_ioctl(kvm_state, KVM_CREATE_DEVICE, &cd);
> if (ret < 0) {
> -trace_flic_create_device(errno);
> -return;
> +error_setg_errno(&errp_local, errno, "Creating the KVM device
> failed");
> +trace_flic_no_device_api(errno);
Err... this should still be trace_flic_create_device(), no?
>>>
>>> I'm afraid you are right! Probably a copy paste error :/
>>>
>>
>> Do you think the traces are still appropriate once we have
>> proper error propagation?
>
> They add less value, but I'd just keep them.
>
>>
>> I did not feel comfortable removing them but thinking again,
>> than might be the thing to do.
>>
>> @Christian:
>> I think we should really fix this the one way or the other.
>> Can you tell me what is the proper procedure?
>
> I'd vote for just fixing the trace event. Smaller change, and we can
> revisit this later.
>
+1
Re: [Qemu-devel] [PATCH 3/7] s390x: fix error propagation in kvm-flic's realize
On Tue, 4 Jul 2017 17:08:52 +0200
Halil Pasic wrote:
> >>> cd.type = KVM_DEV_TYPE_FLIC;
> >>> ret = kvm_vm_ioctl(kvm_state, KVM_CREATE_DEVICE, &cd);
> >>> if (ret < 0) {
> >>> -trace_flic_create_device(errno);
> >>> -return;
> >>> +error_setg_errno(&errp_local, errno, "Creating the KVM device
> >>> failed");
> >>> +trace_flic_no_device_api(errno);
> >>
> >> Err... this should still be trace_flic_create_device(), no?
> >
> > I'm afraid you are right! Probably a copy paste error :/
> >
>
> Do you think the traces are still appropriate once we have
> proper error propagation?
They add less value, but I'd just keep them.
>
> I did not feel comfortable removing them but thinking again,
> than might be the thing to do.
>
> @Christian:
> I think we should really fix this the one way or the other.
> Can you tell me what is the proper procedure?
I'd vote for just fixing the trace event. Smaller change, and we can
revisit this later.
Re: [Qemu-devel] [PATCH 3/7] s390x: fix error propagation in kvm-flic's realize
On 07/04/2017 04:46 PM, Halil Pasic wrote:
>
>
> On 07/04/2017 04:31 PM, Cornelia Huck wrote:
>> On Tue, 4 Jul 2017 16:07:55 +0200
>> Christian Borntraeger wrote:
>>
>>> From: Halil Pasic
>>>
>>> From the moment it was introduced by commit a2875e6f98 ("s390x/kvm:
>>> implement floating-interrupt controller device", 2013-07-16) the kvm-flic
>>> is not making realize fail properly in case it's impossible to create the
>>> KVM device which basically serves as a backend and is absolutely
>>> essential for having an operational kvm-flic.
>>>
>>> Let's fix this by making sure we do proper error propagation in realize.
>>>
>>> Signed-off-by: Halil Pasic
>>> Fixes: a2875e6f98 "s390x/kvm: implement floating-interrupt controller
>>> device"
>>> Reviewed-by: Dong Jia Shi
>>> Reviewed-by: Yi Min Zhao
>>> Signed-off-by: Christian Borntraeger
>>> ---
>>> hw/intc/s390_flic_kvm.c | 14 +++---
>>> 1 file changed, 11 insertions(+), 3 deletions(-)
>>>
>>
>> (...)
>>
>>> cd.type = KVM_DEV_TYPE_FLIC;
>>> ret = kvm_vm_ioctl(kvm_state, KVM_CREATE_DEVICE, &cd);
>>> if (ret < 0) {
>>> -trace_flic_create_device(errno);
>>> -return;
>>> +error_setg_errno(&errp_local, errno, "Creating the KVM device
>>> failed");
>>> +trace_flic_no_device_api(errno);
>>
>> Err... this should still be trace_flic_create_device(), no?
>
> I'm afraid you are right! Probably a copy paste error :/
>
Do you think the traces are still appropriate once we have
proper error propagation?
I did not feel comfortable removing them but thinking again,
than might be the thing to do.
@Christian:
I think we should really fix this the one way or the other.
Can you tell me what is the proper procedure?
>>
>>> +goto fail;
>>> }
>>> flic_state->fd = cd.fd;
>>
>
>
Re: [Qemu-devel] [PATCH 3/7] s390x: fix error propagation in kvm-flic's realize
On 07/04/2017 04:31 PM, Cornelia Huck wrote:
> On Tue, 4 Jul 2017 16:07:55 +0200
> Christian Borntraeger wrote:
>
>> From: Halil Pasic
>>
>> From the moment it was introduced by commit a2875e6f98 ("s390x/kvm:
>> implement floating-interrupt controller device", 2013-07-16) the kvm-flic
>> is not making realize fail properly in case it's impossible to create the
>> KVM device which basically serves as a backend and is absolutely
>> essential for having an operational kvm-flic.
>>
>> Let's fix this by making sure we do proper error propagation in realize.
>>
>> Signed-off-by: Halil Pasic
>> Fixes: a2875e6f98 "s390x/kvm: implement floating-interrupt controller device"
>> Reviewed-by: Dong Jia Shi
>> Reviewed-by: Yi Min Zhao
>> Signed-off-by: Christian Borntraeger
>> ---
>> hw/intc/s390_flic_kvm.c | 14 +++---
>> 1 file changed, 11 insertions(+), 3 deletions(-)
>>
>
> (...)
>
>> cd.type = KVM_DEV_TYPE_FLIC;
>> ret = kvm_vm_ioctl(kvm_state, KVM_CREATE_DEVICE, &cd);
>> if (ret < 0) {
>> -trace_flic_create_device(errno);
>> -return;
>> +error_setg_errno(&errp_local, errno, "Creating the KVM device
>> failed");
>> +trace_flic_no_device_api(errno);
>
> Err... this should still be trace_flic_create_device(), no?
I'm afraid you are right! Probably a copy paste error :/
>
>> +goto fail;
>> }
>> flic_state->fd = cd.fd;
>
Re: [Qemu-devel] [PATCH 3/7] s390x: fix error propagation in kvm-flic's realize
On Tue, 4 Jul 2017 16:07:55 +0200
Christian Borntraeger wrote:
> From: Halil Pasic
>
> From the moment it was introduced by commit a2875e6f98 ("s390x/kvm:
> implement floating-interrupt controller device", 2013-07-16) the kvm-flic
> is not making realize fail properly in case it's impossible to create the
> KVM device which basically serves as a backend and is absolutely
> essential for having an operational kvm-flic.
>
> Let's fix this by making sure we do proper error propagation in realize.
>
> Signed-off-by: Halil Pasic
> Fixes: a2875e6f98 "s390x/kvm: implement floating-interrupt controller device"
> Reviewed-by: Dong Jia Shi
> Reviewed-by: Yi Min Zhao
> Signed-off-by: Christian Borntraeger
> ---
> hw/intc/s390_flic_kvm.c | 14 +++---
> 1 file changed, 11 insertions(+), 3 deletions(-)
>
(...)
> cd.type = KVM_DEV_TYPE_FLIC;
> ret = kvm_vm_ioctl(kvm_state, KVM_CREATE_DEVICE, &cd);
> if (ret < 0) {
> -trace_flic_create_device(errno);
> -return;
> +error_setg_errno(&errp_local, errno, "Creating the KVM device
> failed");
> +trace_flic_no_device_api(errno);
Err... this should still be trace_flic_create_device(), no?
> +goto fail;
> }
> flic_state->fd = cd.fd;
