On 5/21/2014 1:20 AM, Alexey Kardashevskiy wrote:
> This adds TIR/SIAR/SDAR/MMCRA/MMCR0/MMCR1.
>
> This redefines UMMCRA (was MCCRA) and defines hypv version of if.
>
> Signed-off-by: Alexey Kardashevskiy <a...@ozlabs.ru>
> ---
> target-ppc/cpu.h | 10 +++++++++-
> target-ppc/translate_init.c | 41 +++++++++++++++++++++++++++++++++++++----
> 2 files changed, 46 insertions(+), 5 deletions(-)
>
> diff --git a/target-ppc/cpu.h b/target-ppc/cpu.h
> index 13eea19..262cf0f 100644
> --- a/target-ppc/cpu.h
> +++ b/target-ppc/cpu.h
[...]
> @@ -1505,9 +1511,11 @@ static inline int cpu_mmu_index (CPUPPCState *env)
> #define SPR_UPERF9 (0x319)
> #define SPR_UPERFA (0x31A)
> #define SPR_UPERFB (0x31B)
> +#define SPR_POWER_MMCR0 (0x31B)
> #define SPR_UPERFC (0x31C)
> #define SPR_UPERFD (0x31D)
> #define SPR_UPERFE (0x31E)
> +#define SPR_POWER_MMCR1 (0x31E)
> #define SPR_UPERFF (0x31F)
Since MMCR0 and MMCR1 SPRs are assigned by the ISA, why "SPR_POWER_MMCR0"
instead of just
"SPR_MMCR0" (ditto for MMCR1)? Since you moved the outdated and unoffical 7xx
encodings
out of the way in patch 1, the extra "_POWER" seems extraneous.
> #define SPR_RCPU_MI_RA0 (0x320)
> #define SPR_MPC_MI_DBCAM (0x320)
> diff --git a/target-ppc/translate_init.c b/target-ppc/translate_init.c
> index e9c37fa..b92b447 100644
> --- a/target-ppc/translate_init.c
> +++ b/target-ppc/translate_init.c
[ ... ]
> static void gen_spr_book3s_purr(CPUPPCState *env)
> @@ -7711,16 +7716,20 @@ static void gen_spr_book3s_debug(CPUPPCState *env)
> SPR_NOACCESS, SPR_NOACCESS,
> &spr_read_cfar, &spr_write_cfar,
> 0x00000000);
> + spr_register_kvm(env, SPR_BOOK3S_SIAR, "SIAR",
> + &spr_read_generic, &spr_write_generic,
> + &spr_read_generic, &spr_write_generic,
> + KVM_REG_PPC_SIAR, 0x00000000);
> + spr_register_kvm(env, SPR_BOOK3S_SDAR, "SDAR",
> + &spr_read_generic, &spr_write_generic,
> + &spr_read_generic, &spr_write_generic,
> + KVM_REG_PPC_SDAR, 0x00000000);
> #endif
> }
SIAR and SDAR are read-only SPRs.
>
> static void gen_spr_book3s_pmu(CPUPPCState *env)
> {
> #if !defined(CONFIG_USER_ONLY)
> - spr_register_kvm(env, SPR_MMCRA, "SPR_MMCRA",
> - SPR_NOACCESS, SPR_NOACCESS,
> - &spr_read_generic, &spr_write_generic,
> - KVM_REG_PPC_MMCRA, 0x00000000);
> spr_register_kvm(env, SPR_PMC5, "SPR_PMC5",
> SPR_NOACCESS, SPR_NOACCESS,
> &spr_read_generic, &spr_write_generic,
> @@ -7729,6 +7738,30 @@ static void gen_spr_book3s_pmu(CPUPPCState *env)
> SPR_NOACCESS, SPR_NOACCESS,
> &spr_read_generic, &spr_write_generic,
> KVM_REG_PPC_PMC6, 0x00000000);
> + spr_register(env, SPR_MMCRA, "MMCRA",
> + &spr_read_generic, &spr_write_generic,
> + &spr_read_generic, &spr_write_generic,
> + 0x00000000);
> + spr_register_kvm(env, SPR_UMMCRA, "UMMCRA",
> + SPR_NOACCESS, SPR_NOACCESS,
> + &spr_read_generic, &spr_write_generic,
> + KVM_REG_PPC_MMCRA, 0x00000000);
> + spr_register(env, SPR_POWER_MMCR0, "MMCR0",
> + &spr_read_generic, &spr_write_generic,
> + &spr_read_generic, &spr_write_generic,
> + 0x00000000);
> + spr_register_kvm(env, SPR_POWER_UMMCR0, "UMMCR0",
> + SPR_NOACCESS, SPR_NOACCESS,
> + &spr_read_generic, &spr_write_generic,
> + KVM_REG_PPC_MMCR0, 0x00000000);
> + spr_register(env, SPR_POWER_MMCR1, "MMCR1",
> + &spr_read_generic, &spr_write_generic,
> + &spr_read_generic, &spr_write_generic,
> + 0x00000000);
> + spr_register_kvm(env, SPR_POWER_UMMCR1, "UMMCR1",
> + SPR_NOACCESS, SPR_NOACCESS,
> + &spr_read_generic, &spr_write_generic,
> + KVM_REG_PPC_MMCR1, 0x00000000);
> #endif
> }
>
>
This looks wrong to me .... the Uxxxx SPR numbers are accessible in user space
(they are Book I
SPRs) whereas the non-Uxxxx forms are not accessible from user space.
