Re: [PATCH] KVM: arm64: Avoid naming collision in tracing

[Fuad Tabba]

Hi Mostafa,

On Sun, 12 Jul 2026 at 16:38, Mostafa Saleh <smostafa@xxxxxxxxxx> wrote:
>
> When the hypervisor tracing (CONFIG_NVHE_EL2_TRACING) is disabled, it
> defines a static inline stub for trace_clock().
>
> However, trace_clock() is already declared as an extern function in
> linux/trace_clock.h which is pulled in EL2 compilation.

The rename itself looks correct and complete: all three sites (the
declarations and stubs in nvhe/clock.h, the definitions in clock.c,
and the two callers in trace.c) are updated, and no other reference to
the old names remains under arch/arm64/kvm.

Two small things...

>
> If the file <nvhe/clock.h> is included when CONFIG_NVHE_EL2_TRACING
> is disabled (by including it manually in setup.c) it will cause:
> In file included from arch/arm64/kvm/hyp/nvhe/setup.c:22:
>
> ./arch/arm64/kvm/hyp/include/nvhe/clock.h:14:19: error: static declaration of ‘trace_clock’ follows non-static declaration
>
>    14 | static inline u64 trace_clock(void) { return 0; }
>
>       |                   ^~~~~~~~~~~
>
> on GCC and a linker error on LLVM (it seems to change the linkage to
> global)
>
> Although that is not a problem at the moment, as no other files
> include <nvhe/clock.h>. That does not seem to be the intent of
> this code and that will cause issues with more users as the SMMUv3
> driver.
>
> Signed-off-by: Mostafa Saleh <smostafa@xxxxxxxxxx>
>
> --

git am only treats a three-dash "---" line as the scissors; the "--"
here is two dashes, so mailinfo keeps everything above the diffstat.
Worth a "---" on the repost so the note stays out of the log.

> I did not add Fixes tag as this is currently dormant and not breaking
> anything.
> ---
>  arch/arm64/kvm/hyp/include/nvhe/clock.h | 8 ++++----
>  arch/arm64/kvm/hyp/nvhe/clock.c         | 4 ++--
>  arch/arm64/kvm/hyp/nvhe/trace.c         | 4 ++--
>  3 files changed, 8 insertions(+), 8 deletions(-)
>
> diff --git a/arch/arm64/kvm/hyp/include/nvhe/clock.h b/arch/arm64/kvm/hyp/include/nvhe/clock.h
> index 9f429f5c0664..c2ccd0e8bf22 100644
> --- a/arch/arm64/kvm/hyp/include/nvhe/clock.h
> +++ b/arch/arm64/kvm/hyp/include/nvhe/clock.h
> @@ -6,11 +6,11 @@
>  #include <asm/kvm_hyp.h>
>
>  #ifdef CONFIG_NVHE_EL2_TRACING
> -void trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc);
> -u64 trace_clock(void);
> +void hyp_trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc);
> +u64 hyp_trace_clock(void);

hyp_trace_clock overlaps the host side: arch/arm64/kvm/hyp_trace.c
already has a struct hyp_trace_clock and static helpers
hyp_trace_clock_enable() / hyp_trace_clock_show() for the debugfs view
of the same clock. No actual collision, so this is only a readability
point, but a reader grepping hyp_trace_clock now gets two unrelated
things. Maybe you'd want to consider a different name, but naming is
hard :)

Reviewed-by: Fuad Tabba <fuad.tabba@xxxxxxxxx>
Tested-by: Fuad Tabba < fuad.tabba@xxxxxxxxx>

Test: builds fine with the different config enables.

Cheers,
/fuad


>  #else
>  static inline void
> -trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc) { }
> -static inline u64 trace_clock(void) { return 0; }
> +hyp_trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc) { }
> +static inline u64 hyp_trace_clock(void) { return 0; }
>  #endif
>  #endif
> diff --git a/arch/arm64/kvm/hyp/nvhe/clock.c b/arch/arm64/kvm/hyp/nvhe/clock.c
> index a7fc61976fd0..8adefb0b696c 100644
> --- a/arch/arm64/kvm/hyp/nvhe/clock.c
> +++ b/arch/arm64/kvm/hyp/nvhe/clock.c
> @@ -30,7 +30,7 @@ static u64 __clock_mult_uint128(u64 cyc, u32 mult, u32 shift)
>  }
>
>  /* Does not guarantee no reader on the modified bank. */
> -void trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc)
> +void hyp_trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc)
>  {
>         struct clock_data *clock = &trace_clock_data;
>         u64 bank = clock->cur ^ 1;
> @@ -48,7 +48,7 @@ void trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc)
>  }
>
>  /* Use untrusted host data */
> -u64 trace_clock(void)
> +u64 hyp_trace_clock(void)
>  {
>         struct clock_data *clock = &trace_clock_data;
>         u64 bank = smp_load_acquire(&clock->cur);
> diff --git a/arch/arm64/kvm/hyp/nvhe/trace.c b/arch/arm64/kvm/hyp/nvhe/trace.c
> index e7e150ab265f..f2a32463c6c2 100644
> --- a/arch/arm64/kvm/hyp/nvhe/trace.c
> +++ b/arch/arm64/kvm/hyp/nvhe/trace.c
> @@ -35,7 +35,7 @@ static bool hyp_trace_buffer_loaded(struct hyp_trace_buffer *trace_buffer)
>  void *tracing_reserve_entry(unsigned long length)
>  {
>         return simple_ring_buffer_reserve(this_cpu_ptr(trace_buffer.simple_rbs), length,
> -                                         trace_clock());
> +                                         hyp_trace_clock());
>  }
>
>  void tracing_commit_entry(void)
> @@ -290,7 +290,7 @@ void __tracing_update_clock(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc)
>         }
>
>         /* ...we can now override the old one and swap. */
> -       trace_clock_update(mult, shift, epoch_ns, epoch_cyc);
> +       hyp_trace_clock_update(mult, shift, epoch_ns, epoch_cyc);
>  }
>
>  int __tracing_reset(unsigned int cpu)
> --
> 2.55.0.795.g602f6c329a-goog
>