On 05/03/2025 08:36, Jorge Ramirez wrote:
>> I think the patch context ("properties") is confusing here, but it looks
>> to me that these are in sync: interrupts is defined to have 8 items, and
>> interrupt-names is a list of msi0 through msi7.
>>
>> @Krishna, these two last patches (adding the global interrupt) doesn't
>> seem strongly connected to the switch patches. So, if Krzysztof agrees
>> with above assessment, please submit them separately (i.e. a new series,
>> 2 patches, v5).
>
> um, but without these two patches, the functionality is broken requiring
> users to manually rescan the pci bus (ie, via sysfs) to see what is
> behind the bridge.
>
> shouldnt the set include all the necessary patches?
Broken? Then the patchset should be NAKed because nothing can depend on
the DTS and the DTS should not depend on the drivers without some sort
of explanation, either.
Also nothing in the commit msg explains this dependency, so I don't get
the problem here.
Best regards,
Krzysztof
