[PATCH v6 0/7] KVM: arm64: nv: Implement nested stage-2 reverse map

[Shuai Xue]

On 9/15/26 11:42 PM, Wei-Lin Chang wrote:
> Hi,
> 
> This is v6 of optimizing the shadow s2 mmu unmapping during MMU
> notifiers.
> 
> This version fixes a few issues, and adds Marc's space optimization
> for kvm_guest_s2_mapping [1]. Please see the changelog for the details.
> 
> Tested by booting L3, and running in-kernel targetted tests described
> in [2]. v5 got some tested-by's from Itaru and Wang (thanks), I didn't
> carry them over since some bugs fixed in v6 are non-obvious. Some
> retest would be much appreciated!
> 
> Series based on v7.3-rc3 + Marc's nested mmu lifecycle fixes [3].
>

Hi Wei-Lin,

Thanks for the updated series.

I tested all seven patches on top of v7.3-rc3 with Marc's two
nested MMU lifetime fixes, using the L1 boot workload that previously
triggered soft lockups during host NUMA migration.

The setup was:
   - ARM64 host with nested virtualization enabled
   - Automatic NUMA balancing enabled, with numad running
   - QEMU 10.2.3, virtualization=on
   - Ubuntu 26.04 L1 guest, 8 vCPUs and 32 GiB RAM

Each trial used a fresh overlay backed by the same read-only golden
image. I ran 15 boot trials, observing each for at least 600 seconds.

   Result                         Count
   -----------------------------  -----
   Trials reaching guest login    15/15
   Trials with guest soft lockups   0/15
   Guest RCU stalls                   0
   Host hung-task reports             0
   Host soft lockups                  0

Tracing confirmed that host page migration and the reverse-map
recording path were exercised. I did not reproduce the original
soft lockup in this batch.

For the series:

Tested-by: Shuai Xue <xueshuai at linux.alibaba.com>

Thanks,
Shuai