Re: [PATCH v1 0/4] KVM selftests enhancements

[Sean Christopherson <seanjc@google.com>]

On Mon, Oct 21, 2024, Manali Shukla wrote:
> This series was prompted by feedback given in [1].
> Patch 1   : Adds safe_hlt() and cli() helpers.
> Patch 2, 3: Adds an interface to read vcpu stat in selftest.  Adds
>             a macro to generate compiler error to detect typos at
>             compile time while parsing vcpu and vm stats.
> Patch 4   : Fix few of the selftests based on newly defined macro.
> 
> This series was split from the Idle HLT intercept support series [2]
> because the series has a few changes in the vm_get_stat() interface
> as suggested in [1] and a few changes in two of the self-tests
> (nx_huge_pages_test.c and dirty_log_page_splitting_test.c) which use
> vm_get_stat() functionality to retrieve specified VM stats. These
> changes are unrelated to the Idle HLT intercept support series [2].
> 
> [1] https://lore.kernel.org/kvm/ZruDweYzQRRcJeTO@google.com/T/#m7cd7a110f0fcff9a03176935ed016f49088df6b0
> 
> [2] https://lore.kernel.org/kvm/ZruDweYzQRRcJeTO@google.com/T/#m6c67ca8ccb226e5f357d943da2a2965a4224f69a
> 
> 
> Manali Shukla (4):
>   KVM: selftests: Add safe_halt() and cli() helpers to common code
>   KVM: selftests: Add an interface to read the data of named vcpu stat
>   KVM: selftests: convert vm_get_stat to macro
>   KVM: selftests: Replace previously used vm_get_stat() to macro

Thanks for giving this a shot.  I appreciate the effort, especially on the stats
code.  But unfortunately, very little of the code in this series moves things in
the right direction.  There's too much copy+paste, and too much unnecessary
complexity.

I'll post a patch for the STI/CLI helpers and a series for the stats changes,
and will review the Idle HLT series.  Please plan on posting v5 of that series
on top of kvm-x86/next plus the to-be-posted patches (I highly doubt I will get
them queued before you are ready to post v5).