Re: [v2 PATCH] RISC-V: Optimize tlb flush path.

[Andreas Schwab]

On Aug 20 2019, Atish Patra <Atish.Patra@xxxxxxx> wrote:

> +
> +       cpuid = get_cpu();
> +	if (!cmask) {
> +               riscv_cpuid_to_hartid_mask(cpu_online_mask, &hmask);
> +               goto issue_sfence;
> +       }
> +
> +       
> +       if (cpumask_test_cpu(cpuid, cmask) && cpumask_weight(cmask) ==
> 1) {
> +               /* Save trap cost by issuing a local tlb flush here */
> +               if ((start == 0 && size == -1) || (size > PAGE_SIZE))
> +                       local_flush_tlb_all();
> +               else if (size == PAGE_SIZE)
> +                       local_flush_tlb_page(start);
> +               goto done;
> +       }
> +
>         riscv_cpuid_to_hartid_mask(cmask, &hmask);
> +
> +issue_sfence:
>         sbi_remote_sfence_vma(hmask.bits, start, size);
> +done:
> +       put_cpu();
>  }
>
> This is much simpler than what I had done in v2. I will address the if
> condition around size as well.

I still think that this function should be moved out of the header.

Andreas.

-- 
Andreas Schwab, schwab@xxxxxxxxxxxxxx
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."