Re: [Qemu-devel] [PATCH 2/2] ppc: Fix 64K pages support in full emulation

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Thu, 2016-06-30 at 18:01 +0200, Cédric Le Goater wrote:
> +static uint32_t ppc_hash64_pte_size_decode(PowerPCCPU *cpu, uint64_t
> pte0,
> +                                           uint64_t pte1, uint32_t
> slb_pshift)
>  {
> -    switch (slb_pshift) {
> -    case 12:
> -        return 12;
> -    case 16:
> -        if ((pte1 & 0xf000) == 0x1000) {
> -            return 16;
> -        }
> -        return 0;
> -    case 24:
> -        if ((pte1 & 0xff000) == 0) {
> -            return 24;
> -        }
> -        return 0;
> -    }
> -    return 0;
> +    unsigned spshift;
> +
> +    return ppc_hash64_hpte_page_shift_noslb(cpu, pte0, pte1,
> &spshift);
>  }

Why not call ppc_hash64_hpte_page_shift_noslb() directly from the call
site ? That or rename it to ppc_hash64_pte_size_decode :-)

Otherwise yes, your patch looks correct as in what
doesppc_hash64_hpte_page_shift_noslb() is definitely more correct than
what ppc_hash64_pte_size_decode() is doing.

Cheers,
Ben.