The 1st patch (new in v2) has the effect of the 2nd one no longer
affecting Arm.

1: restrict concept of pIRQ to x86
2: cmdline: document and enforce "extra_guest_irqs" upper bounds

Jan

On 03.05.2023 17:31, Jan Beulich wrote:
> The 1st patch (new in v2) has the effect of the 2nd one no longer
> affecting Arm.
> 
> 1: restrict concept of pIRQ to x86
> 2: cmdline: document and enforce "extra_guest_irqs" upper bounds

REST- and Arm-maintainers,

may I please ask for some feedback here? Roger did supply some, resulting
in me considering to revert back to the earlier version (i.e. patch 1
dropped again and patch 2 minimally adjusted to address feedback there),
but which route to go wants input from other maintainers. What I think is
unacceptable is that we continue to not document and enforce the upper
bound, leading to people needlessly running into issues on big enough
systems.

Just to clarify, "the earlier version" was mistakenly not tagged v2, so
what is meant is the 2nd submission ('cmdline: document and enforce
"extra_guest_irqs" upper bounds'), not the 1st one ('cmdline: document
"extra_guest_irqs" upper bound').

Jan

Hi Jan,

On 10/07/2023 08:33, Jan Beulich wrote:
> On 03.05.2023 17:31, Jan Beulich wrote:
>> The 1st patch (new in v2) has the effect of the 2nd one no longer
>> affecting Arm.
>>
>> 1: restrict concept of pIRQ to x86
>> 2: cmdline: document and enforce "extra_guest_irqs" upper bounds
> 
> REST- and Arm-maintainers,
> 
> may I please ask for some feedback here? Roger did supply some, resulting
> in me considering to revert back to the earlier version (i.e. patch 1
> dropped again and patch 2 minimally adjusted to address feedback there),
> but which route to go wants input from other maintainers. What I think is
> unacceptable is that we continue to not document and enforce the upper
> bound, leading to people needlessly running into issues on big enough
> systems.

I looked at patch #1, I think this the right approach for Arm. Do you 
need further input?

Cheers,

-- 
Julien Grall

On 11.07.2023 09:27, Julien Grall wrote:
> On 10/07/2023 08:33, Jan Beulich wrote:
>> On 03.05.2023 17:31, Jan Beulich wrote:
>>> The 1st patch (new in v2) has the effect of the 2nd one no longer
>>> affecting Arm.
>>>
>>> 1: restrict concept of pIRQ to x86
>>> 2: cmdline: document and enforce "extra_guest_irqs" upper bounds
>>
>> REST- and Arm-maintainers,
>>
>> may I please ask for some feedback here? Roger did supply some, resulting
>> in me considering to revert back to the earlier version (i.e. patch 1
>> dropped again and patch 2 minimally adjusted to address feedback there),
>> but which route to go wants input from other maintainers. What I think is
>> unacceptable is that we continue to not document and enforce the upper
>> bound, leading to people needlessly running into issues on big enough
>> systems.
> 
> I looked at patch #1, I think this the right approach for Arm. Do you 
> need further input?

I guess I'll reply to your comments on the patch itself. Feedback
there was what I was pinging for. So thanks for getting back.

Jan