Re: [PATCH 1/3] KVM: selftests: Delete the unused sparsebit copy and count helpers

[Sean Christopherson]

On Fri, Sep 18, 2026, Gokul K wrote:
> sparsebit_copy() and sparsebit_num_set() have no callers.  Nor does the
> FUZZ harness at the bottom of lib/sparsebit.c, which exercises most of
> the rest of the API and still builds and links with both removed.
> 
> sparsebit_copy() was the only non-recursive caller of
> node_copy_subtree(), which leaves that unreachable too.  It is static,
> so keeping it would trip -Wunused-function; delete it as well.
> 
> The comment on the num_set member gives sparsebit_num_set() as the
> reason the count is maintained.  The member stays - it is what lets
> sparsebit_all_set() answer in O(1), and the dump path prints it - so
> point the comment at what still uses it.
> 
> No functional change intended.
> 
> Signed-off-by: Gokul K <gokul02k@xxxxxxxxx>
> ---
>  .../testing/selftests/kvm/include/sparsebit.h |  2 -
>  tools/testing/selftests/kvm/lib/sparsebit.c   | 63 +------------------
>  2 files changed, 1 insertion(+), 64 deletions(-)
> 
> diff --git a/tools/testing/selftests/kvm/include/sparsebit.h b/tools/testing/selftests/kvm/include/sparsebit.h
> index e027e5790946..8a2ab11dae6f 100644
> --- a/tools/testing/selftests/kvm/include/sparsebit.h
> +++ b/tools/testing/selftests/kvm/include/sparsebit.h
> @@ -30,7 +30,6 @@ typedef u64 sparsebit_num_t;
>  
>  struct sparsebit *sparsebit_alloc(void);
>  void sparsebit_free(struct sparsebit **sbitp);
> -void sparsebit_copy(struct sparsebit *dstp, const struct sparsebit *src);

I'm somewhat tempted to keep sparsebit_copy() even though there are no users.
In the unlikely scenario someone wanted/needed to make a copy, it seems like a
potentially painful waste of time to have to recreate this or hunt down its prior
existence.

Anyone else have an opinion?