Re: [PATCH v3 07/17] x86/apic: Support LAPIC timer for Secure AVIC

[Neeraj Upadhyay <Neeraj.Upadhyay@amd.com>]

On 4/3/2025 5:43 PM, Thomas Gleixner wrote:
> On Tue, Apr 01 2025 at 17:06, Neeraj Upadhyay wrote:
>> In addition, add a static call for apic's update_vector() callback,
>> to configure ALLOWED_IRR for the hypervisor to inject timer interrupt
>> using LOCAL_TIMER_VECTOR.
> 
> How is this static call related to the timer vector? It just works with
> the conditional callback. apic_update_vector() is not used in a
> hotpath.
>

No it is not related. I missed the point that static call is only for
callbacks which are used in hotpaths. I thought that they are for
callbacks which are called from multiple call sites. So, when I added
second caller here (first was added in 05/17, in vector.c), I converted
local apic_update_vector() in vector.c to static call.

 
> Even if there is a valid reason for the static call, why is this not
> part of the patch, which adds the update_vector() callback?
> 

Again, this was based on incorrect understanding mentioned above.

> It's well documented that you should not do random unrelated things in
> patches.
> 
> You really try hard to make review a pain.
> 

Apologies for that.


- Neeraj

> Thanks,
> 
>         tglx
> 
>