On 22/03/16 08:55, Christoffer Dall wrote:
On Mon, Mar 14, 2016 at 04:53:03PM +0000, Suzuki K Poulose wrote:
kvm_pmd_huge doesn't have any dependency on the page table
where the pmd lives (i.e, hyp vs. stage2). So, rename it to
huge_pmd() to make it explicit.
kvm_p.d_* wrappers will be used for helpers which differ
across hyp vs stage2.
I don't understand this commit message. Do you associate the kvm_
prefix specifically with one of hyp or stage2?
So the idea is kvm_ prefix will be used for handling either hyp or stage2
depending on what we are dealing with (i.e, kvm parameter to the helpers).
So here, we just want to know if a given pmd represents a huge page, either
via thp or via hugetlb and that doesn't have anything to do with hyp or stage2.
Hence the change. As
I remember reviewers in the past specifically asked to name anything
relating to pgtable macros in the kvm code with a kvm_ prefix to
distinguish them from logic used elsewhere in the kernel.
Correct. In this case it doesn't apply to kvm_pmd_huge().
I specifically do not like having huge_pmd() be significantly different
in logic from pmd_huge(), so defining pmd_thp_or_huge() for arm64 is a
much better option.
Yes, I have switched to that in the next version.
Thanks
Suzuki
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
