Re: [Xen-devel] [PATCH 5/7] xen/iommu: smmu-v3: Add Xen specific code to enable the ported driver

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 01.03.18 at 15:33, <julien.grall@xxxxxxx> wrote:
> What Sameer has been doing for SMMUv3 is following the word we did on 
> the ARM SMMUv2 driver. The header is a suggestion for consolidating the 
> macros over the files here.

In which case - why isn't the patch introducing this compat header
removing the equivalent stuff from that other file. Instead it goes
and introduces the new smmu-v3.c right here. If it was more visible
what code is merely moved, it might be easier to agree to the
introduction.

Otoh this is still all very ARM IOMMU specific - perhaps the header
would then better be local to xen/drivers/passthrough/arm/ (and
you won't see me complain about all this random stuff anymore)?

> Overall I think it makes sense to keep "Linux wording" as much as 
> possible because those files are using Linux philosophy. But I would be 
> open to suggestion here. So would you mind giving a list of things you 
> consider "basic aspects to 'Xen wording'"?

- as already said, anything gfp-ish
- enum irqreturn, and in particular IRQ_WAKE_THREAD, the
  semantics of which surely aren't matched in Xen

Otoh I'm having less of a problem with e.g. dev_printk() and friends
or ktime_t et al, even if I would prefer them to not appear.


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel