Re: [PATCH] kvm-all.c: Move init of irqchip_inject_ioctl out of kvm_irqchip_create()

[Jan Kiszka <jan.kiszka@siemens.com>]

On 2012-08-14 15:14, Avi Kivity wrote:
> On 08/14/2012 02:05 PM, Jan Kiszka wrote:
>> On 2012-08-14 13:01, Avi Kivity wrote:
>>> On 08/14/2012 10:33 AM, Jan Kiszka wrote:
>>>>
>>>> KVM_IRQ_LINE is old-style, deprecated, KVM_IRQ_LINE_STATUS (i.e
>>>> injection with feedback to allow lost-tick compensation) is the current
>>>> standard that other archs should pick up.
>>>
>>> KVM_IRQ_LINE_STATUS may not make sense on all architectures.
>>>
>>> I don't think we're really deprecating KVM_IRQ_LINE or discouraging its
>>> use.  It's not like the kernel-allocated memory slot ioctls.
>>
>> I do not think it makes sense to provide both interfaces long term
>> (provided we ever do a cut). Also, it's almost trivial to provide the
>> add-on feature of KVM_IRQ_LINE_STATUS, and it keeps the door open for
>> IRQ decoalescing. If there is no way for an arch to detect coalescing,
>> it can still return >0 unconditionally.
> 
> That's lying.  I don't see how anything bad can come out of it, but we
> can always be surprised.  If we can't support something, let's not claim
> we do.

Fortunately, we are not in this position on ARM. It has in in-kernel
irqchip and can tell if some line is still being processed.

Jan

-- 
Siemens AG, Corporate Technology, CT RTC ITP SDP-DE
Corporate Competence Center Embedded Linux