On Sat, Jul 09, 2016 at 01:41:31PM +1000, Benjamin Herrenschmidt wrote:
> MacOS uses an architecturally illegal MSR combination that
> seems nonetheless supported by 32-bit processors, which is
> to have MSR[PR]=1 and one or more of MSR[DR/IR/EE]=0.
>
> This adds support for it. To work properly we need to also
> properly include support for PR=1,{I,D}R=0 to the MMU index
> used by the qemu TLB.
>
> Signed-off-by: Benjamin Herrenschmidt <[email protected]>
Applied to ppc-for-2.7, thanks.
> ---
>
> v2. Use the correct flags
>
> target-ppc/helper_regs.h | 46 ++++++++++++++++++++++------------------------
> 1 file changed, 22 insertions(+), 24 deletions(-)
>
> diff --git a/target-ppc/helper_regs.h b/target-ppc/helper_regs.h
> index 8fdfa5c..466ad67 100644
> --- a/target-ppc/helper_regs.h
> +++ b/target-ppc/helper_regs.h
> @@ -41,17 +41,19 @@ static inline void hreg_swap_gpr_tgpr(CPUPPCState *env)
>
> static inline void hreg_compute_mem_idx(CPUPPCState *env)
> {
> - /* This is our encoding for server processors
> + /* This is our encoding for server processors. The architecture
> + * specifies that there is no such thing as userspace with
> + * translation off, however it appears that MacOS does it and
> + * some 32-bit CPUs support it. Weird...
> *
> * 0 = Guest User space virtual mode
> * 1 = Guest Kernel space virtual mode
> - * 2 = Guest Kernel space real mode
> - * 3 = HV User space virtual mode
> - * 4 = HV Kernel space virtual mode
> - * 5 = HV Kernel space real mode
> - *
> - * The combination PR=1 IR&DR=0 is invalid, we will treat
> - * it as IR=DR=1
> + * 2 = Guest User space real mode
> + * 3 = Guest Kernel space real mode
> + * 4 = HV User space virtual mode
> + * 5 = HV Kernel space virtual mode
> + * 6 = HV User space real mode
> + * 7 = HV Kernel space real mode
> *
> * For BookE, we need 8 MMU modes as follow:
> *
> @@ -71,20 +73,11 @@ static inline void hreg_compute_mem_idx(CPUPPCState *env)
> env->immu_idx += msr_gs ? 4 : 0;
> env->dmmu_idx += msr_gs ? 4 : 0;
> } else {
> - /* First calucalte a base value independent of HV */
> - if (msr_pr != 0) {
> - /* User space, ignore IR and DR */
> - env->immu_idx = env->dmmu_idx = 0;
> - } else {
> - /* Kernel, setup a base I/D value */
> - env->immu_idx = msr_ir ? 1 : 2;
> - env->dmmu_idx = msr_dr ? 1 : 2;
> - }
> - /* Then offset it for HV */
> - if (msr_hv) {
> - env->immu_idx += 3;
> - env->dmmu_idx += 3;
> - }
> + env->immu_idx = env->dmmu_idx = msr_pr ? 0 : 1;
> + env->immu_idx += msr_ir ? 0 : 2;
> + env->dmmu_idx += msr_dr ? 0 : 2;
> + env->immu_idx += msr_hv ? 4 : 0;
> + env->dmmu_idx += msr_hv ? 4 : 0;
> }
> }
>
> @@ -136,8 +129,13 @@ static inline int hreg_store_msr(CPUPPCState *env,
> target_ulong value,
> /* Change the exception prefix on PowerPC 601 */
> env->excp_prefix = ((value >> MSR_EP) & 1) * 0xFFF00000;
> }
> - /* If PR=1 then EE, IR and DR must be 1 */
> - if ((value >> MSR_PR) & 1) {
> + /* If PR=1 then EE, IR and DR must be 1
> + *
> + * Note: We only enforce this on 64-bit processors. It appears that
> + * 32-bit implementations supports PR=1 and EE/DR/IR=0 and MacOS
> + * exploits it.
> + */
> + if ((env->insns_flags & PPC_64B) && ((value >> MSR_PR) & 1)) {
> value |= (1 << MSR_EE) | (1 << MSR_DR) | (1 << MSR_IR);
> }
> #endif
>
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
signature.asc
Description: PGP signature
