Re: [PATCH v1 3/3] xen/arm: smmu: Renaming struct iommu_domain *domain to, struct iommu_domain *iommu_domain

[Julien Grall <julien.grall@linaro.org>]

On 27/03/2015 18:05, Jaggi, Manish wrote:
> From: Julien Grall <julien.grall@linaro.org>
> Sent: Friday, March 27, 2015 7:11 PM
> To: Jaggi, Manish; Xen Devel; Prasun.kapoor@cavium.com; Kumar, Vijaya; Ian Campbell; Stefano Stabellini
> Subject: Re: [PATCH v1 3/3] xen/arm: smmu: Renaming struct iommu_domain *domain to, struct iommu_domain *iommu_domain
>
> On 27/03/15 13:26, Jaggi, Manish wrote:
>> On 27/03/15 07:24, Manish Jaggi wrote:
>>> It is good for code readability as there are many structures ending with
>>> the name domain.
>>> Also a code like this one is now easy to understand with the rename
>>> old: dev_iommu_domain(dev) = domain;
>>> new: dev_iommu_domain(dev) = iommu_domain;
>> [manish] Did u see this line
>
> I don't care about the new vs old stuff. What I care is keeping the code
> as close as possible to the Linux code.
>
> [manish] then you are missing a point my friend. The line mentioned old is not at all intuitive and is confusing.

Please avoid familiarity... I'm not my friend and a such things will do 
the inverse effect you are trying to reach.

>   I am not proposing a design change. All I am asking to add a prefix to iommu_domain pointers as was already done for smmu_domain pointers.

I didn't talk about design change but code modification. If you change 
the name of variables, it will require more work to backport a patch 
from Linux because the patch won't apply cleanly.

I've spent quite a lots of time to change as little as possible the 
Linux code and justify any change. You can see the different mail I had 
with Ian & Stefano for this purpose.

IHMO, changing the code just for your own comprehension is not a valid 
justification.

Regards,

-- 
Julien Grall