Re: [musl] [PATCH] Add clock_gettime VDSO support for Riscv64

[Rich Felker <dalias@libc.org>]

On Wed, Sep 18, 2024 at 02:08:02AM +0800, Kevin wrote:
> Hello,
> 
> Currently, musl on Riscv64 does not implement VDSO for the clock_gettime
> function call. The following patch should fix it. (Rudimentary testing on a
> Riscv64 virtual machine confirms this works.) A similar patch could
> probably be made for Riscv32. This patch was made based on information
> from: https://man7.org/linux/man-pages/man7/vdso.7.html
> 
> Please CC me on replies.
> 
> ---
>  arch/riscv64/syscall_arch.h | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/arch/riscv64/syscall_arch.h b/arch/riscv64/syscall_arch.h
> index 7fd042cd..81993fc8 100644
> --- a/arch/riscv64/syscall_arch.h
> +++ b/arch/riscv64/syscall_arch.h
> @@ -71,8 +71,7 @@ static inline long __syscall6(long n, long a, long b,
> long c, long d, long e, lo
>  }
> 
>  #define VDSO_USEFUL
> -/* We don't have a clock_gettime function.
>  #define VDSO_CGT_SYM "__vdso_clock_gettime"
> -#define VDSO_CGT_VER "LINUX_2.6" */
> +#define VDSO_CGT_VER "LINUX_4.15"
> 
>  #define IPC_64 0
> -- 
> 2.34.1

If I understand correctly, the problem isn't that we didn't have it
but that the symbol version string was wrong so it wasn't used?

I have another pending patch for ppc[64] and s390x to add vdso, so it
would be nice to get rv fixed too. Thanks.

Rich