Hi Will and Marc,
On 9/8/20 1:23 AM, Will Deacon wrote:
Hi all,
Here's version four of the KVM page-table rework I previously posted here:
v1: https://lore.kernel.org/r/20200730153406.25136-1-will@xxxxxxxxxx
v2: https://lore.kernel.org/r/20200818132818.16065-1-will@xxxxxxxxxx
v3: https://lore.kernel.org/r/20200825093953.26493-1-will@xxxxxxxxxx
Changes since v3 include:
* Removed some more unused code from the old implementation
* Renumbered KVM_PGTABLE_PROT_* to match file permissions in hex dumps
* Simplified kvm_granule_shift() and removed kvm_start_level()
* Only clear valid bit when invalidating a PTE, leaving phys addr intact
* Fail gracefully when attempting to allocate without a memory cache
* Rewrote stage2_apply_range() as a function instead of a macro
* Fixed fault address rounding with hugetlbfs
* Fixed permission fault handling when dirty logging is enabled
* Added comments
* Added reviewer tags
Many thanks to Alex and Gavin for their comments.
Will
[...]
Eventually, I'm lucky finding a machine from our pool where 16KB page
size is supported. I tried this series (v4) on this machine, everything
looks good as we expected:
Host: 5.9.0-rc3 + patchset (v4)
CONFIG_ARM64_VA_BITS=47
CONFIG_ARM64_PAGE_SHIFT=14
Hugepagesize: 32768 kB (from /proc/meminfo)
Config: -HugeTLBfs/-THP
+HugeTLBfs/-THP
-HugeTLBfs/THP
Guest: CONFIG_ARM64_PAGE_SHIFT=12/14/16
Let me know in case there are more tests needed from my side.
Thanks,
Gavin
_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
