[Qemu-devel] [RFC 0/3] split core mmu_idx from ARMMMUIdx values

[Qemu-devel] [RFC 0/3] split core mmu_idx from ARMMMUIdx values

[Peter Maydell]

For M profile, we're eventually going to want some
new MMU index values:
     non secure User
     non secure Privileged
     non secure Privileged, execution priority < 0
     secure User
     secure Privileged
     secure Privileged, execution priority < 0

(currently we have ns User and ns Priv). Although we
can reasonably cleanly think of 4 of those as being
like A profile NS EL0/NS EL1/S EL0/S EL1, the "privileged
and execution priority is < 0" cases are entirely M profile
specific. This gives us two choices of implementation:

(1) Add new ARMMMUIdx values, bringing NB_MMU_MODES up to 9.
This is large enough that it would mean the size of a TLB on
MIPS, PPC and 32-bit ARM hosts would be halved.

(2) Allow A profile and M profile cores to allocate the mmu_idx
values to different semantics, since any particular CPU will
be definitely one kind or the other and doesn't need all 9.

This patchset is an RFC for choice (2), to see whether people
think it looks sufficiently maintainable. The idea is that
rather than just having
  ARMMMUIdx_S12NSE0 = 0,
  ARMMMUIdx_MUser = 0,
and relying on all the target/arm code to remember to disambiguate
(which seemed very error prone to me), we have different values:
    ARMMMUIdx_S12NSE0 = 0 | ARM_MMU_IDX_A,
    ARMMMUIdx_MUser = 0 | ARM_MMU_IDX_M,
so instead of ARMMMUIdx enum values being identical to core QEMU
MMU index values, they are now the core index values with some
high bits set. Any particular CPU always uses the same high
bits. New functions arm_to_core_mmu_idx() and core_to_arm_mmu_idx()
convert between the two.

This means that if a function that does
  switch (mmu_idx) {
  case ARMMMUIdx_S12NSE0:
  ...
  }
is inadvertently called with an M profile mmu index and it wasn't
expecting it, it will at least assert rather than just doing
something odd, and it doesn't have to explicitly look at the
feature bits to make a decision.

In general core index values are stored in 'int' types, and
ARM values are stored in ARMMMUIdx types. QEMU core code,
the TB flags value and the translate.c code which passes
mmu_idx values to gen_qemu_ld/st_* all use core index values,
and the rest uses ARMMMUIdx values.
    

Overall questions:
 * does this look like it's workable enough to justify taking
   it rather than the simple-but-worse-perf-on-some-archs
   option 1 ?
 * this does risk breaking A profile in corner cases, because
   C unfortunately doesn't typecheck int vs enum so if I forgot
   a call to an index conversion function somewhere things will
   break. are we happy the risk isn't too huge?


thanks
-- PMM

Peter Maydell (3):
  arm: Use the mmu_idx we're passed in arm_cpu_do_unaligned_access()
  arm: Add support for M profile CPUs having different MMU index
    semantics
  arm: Use different ARMMMUIdx values for M profile

 target/arm/cpu.h           |  77 ++++++++++++++++++----
 target/arm/translate.h     |   2 +-
 target/arm/helper.c        | 156 ++++++++++++++++++++++++---------------------
 target/arm/op_helper.c     |   3 +-
 target/arm/translate-a64.c |  18 ++++--
 target/arm/translate.c     |  13 ++--
 6 files changed, 171 insertions(+), 98 deletions(-)

-- 
2.7.4

[Qemu-devel] [RFC 1/3] arm: Use the mmu_idx we're passed in arm_cpu_do_unaligned_access()

[Peter Maydell]

When identifying the DFSR format for an alignment fault, use
the mmu index that we are passed, rather than calling cpu_mmu_index()
to get the mmu index for the current CPU state. This doesn't actually
make any difference since the only cases where the current MMU index
differs from the index used for the load are the "unprivileged
load/store" instructions, and in that case the mmu index may
differ but the translation regime is the same (apart from the
"use from Hyp mode" case which is UNPREDICTABLE).
However it's the more logical thing to do.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target/arm/op_helper.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c
index 156b825..de24815 100644
--- a/target/arm/op_helper.c
+++ b/target/arm/op_helper.c
@@ -208,7 +208,7 @@ void arm_cpu_do_unaligned_access(CPUState *cs, vaddr vaddr,
     /* the DFSR for an alignment fault depends on whether we're using
      * the LPAE long descriptor format, or the short descriptor format
      */
-    if (arm_s1_regime_using_lpae_format(env, cpu_mmu_index(env, false))) {
+    if (arm_s1_regime_using_lpae_format(env, mmu_idx)) {
         env->exception.fsr = (1 << 9) | 0x21;
     } else {
         env->exception.fsr = 0x1;
-- 
2.7.4

[Qemu-devel] [RFC 2/3] arm: Add support for M profile CPUs having different MMU index semantics

[Peter Maydell]

The M profile CPU's MPU has an awkward corner case which we
would like to implement with a different MMU index.

We can avoid having to bump the number of MMU modes ARM
uses, because some of our existing MMU indexes are only
used by non-M-profile CPUs, so we can borrow one.
To avoid that getting too confusing, clean up the code
to try to keep the two meanings of the index separate.

Instead of ARMMMUIdx enum values being identical to core QEMU
MMU index values, they are now the core index values with some
high bits set. Any particular CPU always uses the same high
bits (so eventually A profile cores and M profile cores will
use different bits). New functions arm_to_core_mmu_idx()
and core_to_arm_mmu_idx() convert between the two.

In general core index values are stored in 'int' types, and
ARM values are stored in ARMMMUIdx types.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target/arm/cpu.h           |  60 ++++++++++++++----
 target/arm/translate.h     |   2 +-
 target/arm/helper.c        | 151 ++++++++++++++++++++++++---------------------
 target/arm/op_helper.c     |   3 +-
 target/arm/translate-a64.c |  18 ++++--
 target/arm/translate.c     |  10 +--
 6 files changed, 147 insertions(+), 97 deletions(-)

diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index 1055bfe..c162d43 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -2037,6 +2037,16 @@ static inline bool arm_excp_unmasked(CPUState *cs, unsigned int excp_idx,
  * for the accesses done as part of a stage 1 page table walk, rather than
  * having to walk the stage 2 page table over and over.)
  *
+ * The ARMMMUIdx and the mmu index value used by the core QEMU TLB code
+ * are not quite the same -- different CPU types (most notably M profile
+ * vs A/R profile) would like to use MMU indexes with different semantics,
+ * but since we don't ever need to use all of those in a single CPU we
+ * can avoid setting NB_MMU_MODES to more than 8. The lower bits of
+ * ARMMMUIdx are the core TLB mmu index, and the higher bits are always
+ * the same for any particular CPU.
+ * Variables of type ARMMUIdx are always full values, and the core
+ * index values are in variables of type 'int'.
+ *
  * Our enumeration includes at the end some entries which are not "true"
  * mmu_idx values in that they don't have corresponding TLBs and are only
  * valid for doing slow path page table walks.
@@ -2045,23 +2055,49 @@ static inline bool arm_excp_unmasked(CPUState *cs, unsigned int excp_idx,
  * of the AT/ATS operations.
  * The values used are carefully arranged to make mmu_idx => EL lookup easy.
  */
+#define ARM_MMU_IDX_A 0x10 /* A profile (and M profile, for the moment) */
+#define ARM_MMU_IDX_NOTLB 0x20 /* does not have a TLB */
+
 typedef enum ARMMMUIdx {
-    ARMMMUIdx_S12NSE0 = 0,
-    ARMMMUIdx_S12NSE1 = 1,
-    ARMMMUIdx_S1E2 = 2,
-    ARMMMUIdx_S1E3 = 3,
-    ARMMMUIdx_S1SE0 = 4,
-    ARMMMUIdx_S1SE1 = 5,
-    ARMMMUIdx_S2NS = 6,
+    ARMMMUIdx_S12NSE0 = 0 | ARM_MMU_IDX_A,
+    ARMMMUIdx_S12NSE1 = 1 | ARM_MMU_IDX_A,
+    ARMMMUIdx_S1E2 = 2 | ARM_MMU_IDX_A,
+    ARMMMUIdx_S1E3 = 3 | ARM_MMU_IDX_A,
+    ARMMMUIdx_S1SE0 = 4 | ARM_MMU_IDX_A,
+    ARMMMUIdx_S1SE1 = 5 | ARM_MMU_IDX_A,
+    ARMMMUIdx_S2NS = 6 | ARM_MMU_IDX_A,
     /* Indexes below here don't have TLBs and are used only for AT system
      * instructions or for the first stage of an S12 page table walk.
      */
-    ARMMMUIdx_S1NSE0 = 7,
-    ARMMMUIdx_S1NSE1 = 8,
+    ARMMMUIdx_S1NSE0 = 0 | ARM_MMU_IDX_NOTLB,
+    ARMMMUIdx_S1NSE1 = 1 | ARM_MMU_IDX_NOTLB,
 } ARMMMUIdx;
 
+/* Bit macros for the core-mmu-index values for each index,
+ * for use when calling tlb_flush_by_mmuidx() and friends.
+ */
+typedef enum ARMMMUIdxBit {
+    ARMMMUIdxBit_S12NSE0 = 1 << 0,
+    ARMMMUIdxBit_S12NSE1 = 1 << 1,
+    ARMMMUIdxBit_S1E2 = 1 << 2,
+    ARMMMUIdxBit_S1E3 = 1 << 3,
+    ARMMMUIdxBit_S1SE0 = 1 << 4,
+    ARMMMUIdxBit_S1SE1 = 1 << 5,
+    ARMMMUIdxBit_S2NS = 1 << 6,
+} ARMMMUIdxBit;
+
 #define MMU_USER_IDX 0
 
+static inline int arm_to_core_mmu_idx(ARMMMUIdx mmu_idx)
+{
+    return mmu_idx & 7;
+}
+
+static inline ARMMMUIdx core_to_arm_mmu_idx(CPUARMState *env, int mmu_idx)
+{
+    return mmu_idx | ARM_MMU_IDX_A;
+}
+
 /* Return the exception level we're running at if this is our mmu_idx */
 static inline int arm_mmu_idx_to_el(ARMMMUIdx mmu_idx)
 {
@@ -2075,7 +2111,7 @@ static inline int cpu_mmu_index(CPUARMState *env, bool ifetch)
     int el = arm_current_el(env);
 
     if (el < 2 && arm_is_secure_below_el3(env)) {
-        return ARMMMUIdx_S1SE0 + el;
+        return arm_to_core_mmu_idx(ARMMMUIdx_S1SE0 + el);
     }
     return el;
 }
@@ -2471,7 +2507,7 @@ static inline uint32_t arm_regime_tbi1(CPUARMState *env, ARMMMUIdx mmu_idx)
 static inline void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc,
                                         target_ulong *cs_base, uint32_t *flags)
 {
-    ARMMMUIdx mmu_idx = cpu_mmu_index(env, false);
+    ARMMMUIdx mmu_idx = core_to_arm_mmu_idx(env, cpu_mmu_index(env, false));
     if (is_a64(env)) {
         *pc = env->pc;
         *flags = ARM_TBFLAG_AARCH64_STATE_MASK;
@@ -2496,7 +2532,7 @@ static inline void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc,
                    << ARM_TBFLAG_XSCALE_CPAR_SHIFT);
     }
 
-    *flags |= (mmu_idx << ARM_TBFLAG_MMUIDX_SHIFT);
+    *flags |= (arm_to_core_mmu_idx(mmu_idx) << ARM_TBFLAG_MMUIDX_SHIFT);
 
     /* The SS_ACTIVE and PSTATE_SS bits correspond to the state machine
      * states defined in the ARM ARM for software singlestep:
diff --git a/target/arm/translate.h b/target/arm/translate.h
index 629dab9..6b2cc34 100644
--- a/target/arm/translate.h
+++ b/target/arm/translate.h
@@ -88,7 +88,7 @@ static inline int arm_dc_feature(DisasContext *dc, int feature)
 
 static inline int get_mem_index(DisasContext *s)
 {
-    return s->mmu_idx;
+    return arm_to_core_mmu_idx(s->mmu_idx);
 }
 
 /* Function used to determine the target exception EL when otherwise not known
diff --git a/target/arm/helper.c b/target/arm/helper.c
index 8a3e448..520adcc 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -571,9 +571,9 @@ static void tlbiall_nsnh_write(CPUARMState *env, const ARMCPRegInfo *ri,
     CPUState *cs = ENV_GET_CPU(env);
 
     tlb_flush_by_mmuidx(cs,
-                        (1 << ARMMMUIdx_S12NSE1) |
-                        (1 << ARMMMUIdx_S12NSE0) |
-                        (1 << ARMMMUIdx_S2NS));
+                        ARMMMUIdxBit_S12NSE1 |
+                        ARMMMUIdxBit_S12NSE0 |
+                        ARMMMUIdxBit_S2NS);
 }
 
 static void tlbiall_nsnh_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -582,9 +582,9 @@ static void tlbiall_nsnh_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
     CPUState *cs = ENV_GET_CPU(env);
 
     tlb_flush_by_mmuidx_all_cpus_synced(cs,
-                                        (1 << ARMMMUIdx_S12NSE1) |
-                                        (1 << ARMMMUIdx_S12NSE0) |
-                                        (1 << ARMMMUIdx_S2NS));
+                                        ARMMMUIdxBit_S12NSE1 |
+                                        ARMMMUIdxBit_S12NSE0 |
+                                        ARMMMUIdxBit_S2NS);
 }
 
 static void tlbiipas2_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -605,7 +605,7 @@ static void tlbiipas2_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     pageaddr = sextract64(value << 12, 0, 40);
 
-    tlb_flush_page_by_mmuidx(cs, pageaddr, (1 << ARMMMUIdx_S2NS));
+    tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_S2NS);
 }
 
 static void tlbiipas2_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -621,7 +621,7 @@ static void tlbiipas2_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
     pageaddr = sextract64(value << 12, 0, 40);
 
     tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr,
-                                             (1 << ARMMMUIdx_S2NS));
+                                             ARMMMUIdxBit_S2NS);
 }
 
 static void tlbiall_hyp_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -629,7 +629,7 @@ static void tlbiall_hyp_write(CPUARMState *env, const ARMCPRegInfo *ri,
 {
     CPUState *cs = ENV_GET_CPU(env);
 
-    tlb_flush_by_mmuidx(cs, (1 << ARMMMUIdx_S1E2));
+    tlb_flush_by_mmuidx(cs, ARMMMUIdxBit_S1E2);
 }
 
 static void tlbiall_hyp_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -637,7 +637,7 @@ static void tlbiall_hyp_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
 {
     CPUState *cs = ENV_GET_CPU(env);
 
-    tlb_flush_by_mmuidx_all_cpus_synced(cs, (1 << ARMMMUIdx_S1E2));
+    tlb_flush_by_mmuidx_all_cpus_synced(cs, ARMMMUIdxBit_S1E2);
 }
 
 static void tlbimva_hyp_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -646,7 +646,7 @@ static void tlbimva_hyp_write(CPUARMState *env, const ARMCPRegInfo *ri,
     CPUState *cs = ENV_GET_CPU(env);
     uint64_t pageaddr = value & ~MAKE_64BIT_MASK(0, 12);
 
-    tlb_flush_page_by_mmuidx(cs, pageaddr, (1 << ARMMMUIdx_S1E2));
+    tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_S1E2);
 }
 
 static void tlbimva_hyp_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -656,7 +656,7 @@ static void tlbimva_hyp_is_write(CPUARMState *env, const ARMCPRegInfo *ri,
     uint64_t pageaddr = value & ~MAKE_64BIT_MASK(0, 12);
 
     tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr,
-                                             (1 << ARMMMUIdx_S1E2));
+                                             ARMMMUIdxBit_S1E2);
 }
 
 static const ARMCPRegInfo cp_reginfo[] = {
@@ -2596,9 +2596,9 @@ static void vttbr_write(CPUARMState *env, const ARMCPRegInfo *ri,
     /* Accesses to VTTBR may change the VMID so we must flush the TLB.  */
     if (raw_read(env, ri) != value) {
         tlb_flush_by_mmuidx(cs,
-                            (1 << ARMMMUIdx_S12NSE1) |
-                            (1 << ARMMMUIdx_S12NSE0) |
-                            (1 << ARMMMUIdx_S2NS));
+                            ARMMMUIdxBit_S12NSE1 |
+                            ARMMMUIdxBit_S12NSE0 |
+                            ARMMMUIdxBit_S2NS);
         raw_write(env, ri, value);
     }
 }
@@ -2957,12 +2957,12 @@ static void tlbi_aa64_vmalle1_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     if (arm_is_secure_below_el3(env)) {
         tlb_flush_by_mmuidx(cs,
-                            (1 << ARMMMUIdx_S1SE1) |
-                            (1 << ARMMMUIdx_S1SE0));
+                            ARMMMUIdxBit_S1SE1 |
+                            ARMMMUIdxBit_S1SE0);
     } else {
         tlb_flush_by_mmuidx(cs,
-                            (1 << ARMMMUIdx_S12NSE1) |
-                            (1 << ARMMMUIdx_S12NSE0));
+                            ARMMMUIdxBit_S12NSE1 |
+                            ARMMMUIdxBit_S12NSE0);
     }
 }
 
@@ -2974,12 +2974,12 @@ static void tlbi_aa64_vmalle1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     if (sec) {
         tlb_flush_by_mmuidx_all_cpus_synced(cs,
-                                            (1 << ARMMMUIdx_S1SE1) |
-                                            (1 << ARMMMUIdx_S1SE0));
+                                            ARMMMUIdxBit_S1SE1 |
+                                            ARMMMUIdxBit_S1SE0);
     } else {
         tlb_flush_by_mmuidx_all_cpus_synced(cs,
-                                            (1 << ARMMMUIdx_S12NSE1) |
-                                            (1 << ARMMMUIdx_S12NSE0));
+                                            ARMMMUIdxBit_S12NSE1 |
+                                            ARMMMUIdxBit_S12NSE0);
     }
 }
 
@@ -2995,18 +2995,18 @@ static void tlbi_aa64_alle1_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     if (arm_is_secure_below_el3(env)) {
         tlb_flush_by_mmuidx(cs,
-                            (1 << ARMMMUIdx_S1SE1) |
-                            (1 << ARMMMUIdx_S1SE0));
+                            ARMMMUIdxBit_S1SE1 |
+                            ARMMMUIdxBit_S1SE0);
     } else {
         if (arm_feature(env, ARM_FEATURE_EL2)) {
             tlb_flush_by_mmuidx(cs,
-                                (1 << ARMMMUIdx_S12NSE1) |
-                                (1 << ARMMMUIdx_S12NSE0) |
-                                (1 << ARMMMUIdx_S2NS));
+                                ARMMMUIdxBit_S12NSE1 |
+                                ARMMMUIdxBit_S12NSE0 |
+                                ARMMMUIdxBit_S2NS);
         } else {
             tlb_flush_by_mmuidx(cs,
-                                (1 << ARMMMUIdx_S12NSE1) |
-                                (1 << ARMMMUIdx_S12NSE0));
+                                ARMMMUIdxBit_S12NSE1 |
+                                ARMMMUIdxBit_S12NSE0);
         }
     }
 }
@@ -3017,7 +3017,7 @@ static void tlbi_aa64_alle2_write(CPUARMState *env, const ARMCPRegInfo *ri,
     ARMCPU *cpu = arm_env_get_cpu(env);
     CPUState *cs = CPU(cpu);
 
-    tlb_flush_by_mmuidx(cs, (1 << ARMMMUIdx_S1E2));
+    tlb_flush_by_mmuidx(cs, ARMMMUIdxBit_S1E2);
 }
 
 static void tlbi_aa64_alle3_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3026,7 +3026,7 @@ static void tlbi_aa64_alle3_write(CPUARMState *env, const ARMCPRegInfo *ri,
     ARMCPU *cpu = arm_env_get_cpu(env);
     CPUState *cs = CPU(cpu);
 
-    tlb_flush_by_mmuidx(cs, (1 << ARMMMUIdx_S1E3));
+    tlb_flush_by_mmuidx(cs, ARMMMUIdxBit_S1E3);
 }
 
 static void tlbi_aa64_alle1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3042,17 +3042,17 @@ static void tlbi_aa64_alle1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     if (sec) {
         tlb_flush_by_mmuidx_all_cpus_synced(cs,
-                                            (1 << ARMMMUIdx_S1SE1) |
-                                            (1 << ARMMMUIdx_S1SE0));
+                                            ARMMMUIdxBit_S1SE1 |
+                                            ARMMMUIdxBit_S1SE0);
     } else if (has_el2) {
         tlb_flush_by_mmuidx_all_cpus_synced(cs,
-                                            (1 << ARMMMUIdx_S12NSE1) |
-                                            (1 << ARMMMUIdx_S12NSE0) |
-                                            (1 << ARMMMUIdx_S2NS));
+                                            ARMMMUIdxBit_S12NSE1 |
+                                            ARMMMUIdxBit_S12NSE0 |
+                                            ARMMMUIdxBit_S2NS);
     } else {
           tlb_flush_by_mmuidx_all_cpus_synced(cs,
-                                              (1 << ARMMMUIdx_S12NSE1) |
-                                              (1 << ARMMMUIdx_S12NSE0));
+                                              ARMMMUIdxBit_S12NSE1 |
+                                              ARMMMUIdxBit_S12NSE0);
     }
 }
 
@@ -3061,7 +3061,7 @@ static void tlbi_aa64_alle2is_write(CPUARMState *env, const ARMCPRegInfo *ri,
 {
     CPUState *cs = ENV_GET_CPU(env);
 
-    tlb_flush_by_mmuidx_all_cpus_synced(cs, (1 << ARMMMUIdx_S1E2));
+    tlb_flush_by_mmuidx_all_cpus_synced(cs, ARMMMUIdxBit_S1E2);
 }
 
 static void tlbi_aa64_alle3is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3069,7 +3069,7 @@ static void tlbi_aa64_alle3is_write(CPUARMState *env, const ARMCPRegInfo *ri,
 {
     CPUState *cs = ENV_GET_CPU(env);
 
-    tlb_flush_by_mmuidx_all_cpus_synced(cs, (1 << ARMMMUIdx_S1E3));
+    tlb_flush_by_mmuidx_all_cpus_synced(cs, ARMMMUIdxBit_S1E3);
 }
 
 static void tlbi_aa64_vae1_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3086,12 +3086,12 @@ static void tlbi_aa64_vae1_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     if (arm_is_secure_below_el3(env)) {
         tlb_flush_page_by_mmuidx(cs, pageaddr,
-                                 (1 << ARMMMUIdx_S1SE1) |
-                                 (1 << ARMMMUIdx_S1SE0));
+                                 ARMMMUIdxBit_S1SE1 |
+                                 ARMMMUIdxBit_S1SE0);
     } else {
         tlb_flush_page_by_mmuidx(cs, pageaddr,
-                                 (1 << ARMMMUIdx_S12NSE1) |
-                                 (1 << ARMMMUIdx_S12NSE0));
+                                 ARMMMUIdxBit_S12NSE1 |
+                                 ARMMMUIdxBit_S12NSE0);
     }
 }
 
@@ -3106,7 +3106,7 @@ static void tlbi_aa64_vae2_write(CPUARMState *env, const ARMCPRegInfo *ri,
     CPUState *cs = CPU(cpu);
     uint64_t pageaddr = sextract64(value << 12, 0, 56);
 
-    tlb_flush_page_by_mmuidx(cs, pageaddr, (1 << ARMMMUIdx_S1E2));
+    tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_S1E2);
 }
 
 static void tlbi_aa64_vae3_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3120,7 +3120,7 @@ static void tlbi_aa64_vae3_write(CPUARMState *env, const ARMCPRegInfo *ri,
     CPUState *cs = CPU(cpu);
     uint64_t pageaddr = sextract64(value << 12, 0, 56);
 
-    tlb_flush_page_by_mmuidx(cs, pageaddr, (1 << ARMMMUIdx_S1E3));
+    tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_S1E3);
 }
 
 static void tlbi_aa64_vae1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3133,12 +3133,12 @@ static void tlbi_aa64_vae1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     if (sec) {
         tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr,
-                                                 (1 << ARMMMUIdx_S1SE1) |
-                                                 (1 << ARMMMUIdx_S1SE0));
+                                                 ARMMMUIdxBit_S1SE1 |
+                                                 ARMMMUIdxBit_S1SE0);
     } else {
         tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr,
-                                                 (1 << ARMMMUIdx_S12NSE1) |
-                                                 (1 << ARMMMUIdx_S12NSE0));
+                                                 ARMMMUIdxBit_S12NSE1 |
+                                                 ARMMMUIdxBit_S12NSE0);
     }
 }
 
@@ -3149,7 +3149,7 @@ static void tlbi_aa64_vae2is_write(CPUARMState *env, const ARMCPRegInfo *ri,
     uint64_t pageaddr = sextract64(value << 12, 0, 56);
 
     tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr,
-                                             (1 << ARMMMUIdx_S1E2));
+                                             ARMMMUIdxBit_S1E2);
 }
 
 static void tlbi_aa64_vae3is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3159,7 +3159,7 @@ static void tlbi_aa64_vae3is_write(CPUARMState *env, const ARMCPRegInfo *ri,
     uint64_t pageaddr = sextract64(value << 12, 0, 56);
 
     tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr,
-                                             (1 << ARMMMUIdx_S1E3));
+                                             ARMMMUIdxBit_S1E3);
 }
 
 static void tlbi_aa64_ipas2e1_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3181,7 +3181,7 @@ static void tlbi_aa64_ipas2e1_write(CPUARMState *env, const ARMCPRegInfo *ri,
 
     pageaddr = sextract64(value << 12, 0, 48);
 
-    tlb_flush_page_by_mmuidx(cs, pageaddr, (1 << ARMMMUIdx_S2NS));
+    tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_S2NS);
 }
 
 static void tlbi_aa64_ipas2e1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -3197,7 +3197,7 @@ static void tlbi_aa64_ipas2e1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
     pageaddr = sextract64(value << 12, 0, 48);
 
     tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr,
-                                             (1 << ARMMMUIdx_S2NS));
+                                             ARMMMUIdxBit_S2NS);
 }
 
 static CPAccessResult aa64_zva_access(CPUARMState *env, const ARMCPRegInfo *ri,
@@ -7049,6 +7049,17 @@ static inline TCR *regime_tcr(CPUARMState *env, ARMMMUIdx mmu_idx)
     return &env->cp15.tcr_el[regime_el(env, mmu_idx)];
 }
 
+/* Convert a possible stage1+2 MMU index into the appropriate
+ * stage 1 MMU index
+ */
+static inline ARMMMUIdx stage_1_mmu_idx(ARMMMUIdx mmu_idx)
+{
+    if (mmu_idx == ARMMMUIdx_S12NSE0 || mmu_idx == ARMMMUIdx_S12NSE1) {
+        mmu_idx += (ARMMMUIdx_S1NSE0 - ARMMMUIdx_S12NSE0);
+    }
+    return mmu_idx;
+}
+
 /* Returns TBI0 value for current regime el */
 uint32_t arm_regime_tbi0(CPUARMState *env, ARMMMUIdx mmu_idx)
 {
@@ -7056,11 +7067,9 @@ uint32_t arm_regime_tbi0(CPUARMState *env, ARMMMUIdx mmu_idx)
     uint32_t el;
 
     /* For EL0 and EL1, TBI is controlled by stage 1's TCR, so convert
-       * a stage 1+2 mmu index into the appropriate stage 1 mmu index.
-       */
-    if (mmu_idx == ARMMMUIdx_S12NSE0 || mmu_idx == ARMMMUIdx_S12NSE1) {
-        mmu_idx += ARMMMUIdx_S1NSE0;
-    }
+     * a stage 1+2 mmu index into the appropriate stage 1 mmu index.
+     */
+    mmu_idx = stage_1_mmu_idx(mmu_idx);
 
     tcr = regime_tcr(env, mmu_idx);
     el = regime_el(env, mmu_idx);
@@ -7079,11 +7088,9 @@ uint32_t arm_regime_tbi1(CPUARMState *env, ARMMMUIdx mmu_idx)
     uint32_t el;
 
     /* For EL0 and EL1, TBI is controlled by stage 1's TCR, so convert
-       * a stage 1+2 mmu index into the appropriate stage 1 mmu index.
-       */
-    if (mmu_idx == ARMMMUIdx_S12NSE0 || mmu_idx == ARMMMUIdx_S12NSE1) {
-        mmu_idx += ARMMMUIdx_S1NSE0;
-    }
+     * a stage 1+2 mmu index into the appropriate stage 1 mmu index.
+     */
+    mmu_idx = stage_1_mmu_idx(mmu_idx);
 
     tcr = regime_tcr(env, mmu_idx);
     el = regime_el(env, mmu_idx);
@@ -7129,9 +7136,7 @@ static inline bool regime_using_lpae_format(CPUARMState *env,
  * on whether the long or short descriptor format is in use. */
 bool arm_s1_regime_using_lpae_format(CPUARMState *env, ARMMMUIdx mmu_idx)
 {
-    if (mmu_idx == ARMMMUIdx_S12NSE0 || mmu_idx == ARMMMUIdx_S12NSE1) {
-        mmu_idx += ARMMMUIdx_S1NSE0;
-    }
+    mmu_idx = stage_1_mmu_idx(mmu_idx);
 
     return regime_using_lpae_format(env, mmu_idx);
 }
@@ -8385,7 +8390,7 @@ static bool get_phys_addr(CPUARMState *env, target_ulong address,
             int ret;
 
             ret = get_phys_addr(env, address, access_type,
-                                mmu_idx + ARMMMUIdx_S1NSE0, &ipa, attrs,
+                                stage_1_mmu_idx(mmu_idx), &ipa, attrs,
                                 prot, page_size, fsr, fi);
 
             /* If S1 fails or S2 is disabled, return early.  */
@@ -8406,7 +8411,7 @@ static bool get_phys_addr(CPUARMState *env, target_ulong address,
             /*
              * For non-EL2 CPUs a stage1+stage2 translation is just stage 1.
              */
-            mmu_idx += ARMMMUIdx_S1NSE0;
+            mmu_idx = stage_1_mmu_idx(mmu_idx);
         }
     }
 
@@ -8482,7 +8487,8 @@ bool arm_tlb_fill(CPUState *cs, vaddr address,
     int ret;
     MemTxAttrs attrs = {};
 
-    ret = get_phys_addr(env, address, access_type, mmu_idx, &phys_addr,
+    ret = get_phys_addr(env, address, access_type,
+                        core_to_arm_mmu_idx(env, mmu_idx), &phys_addr,
                         &attrs, &prot, &page_size, fsr, fi);
     if (!ret) {
         /* Map a single [sub]page.  */
@@ -8507,10 +8513,11 @@ hwaddr arm_cpu_get_phys_page_attrs_debug(CPUState *cs, vaddr addr,
     bool ret;
     uint32_t fsr;
     ARMMMUFaultInfo fi = {};
+    ARMMMUIdx mmu_idx = core_to_arm_mmu_idx(env, cpu_mmu_index(env, false));
 
     *attrs = (MemTxAttrs) {};
 
-    ret = get_phys_addr(env, addr, 0, cpu_mmu_index(env, false), &phys_addr,
+    ret = get_phys_addr(env, addr, 0, mmu_idx, &phys_addr,
                         attrs, &prot, &page_size, &fsr, &fi);
 
     if (ret) {
diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c
index de24815..2a85666 100644
--- a/target/arm/op_helper.c
+++ b/target/arm/op_helper.c
@@ -194,6 +194,7 @@ void arm_cpu_do_unaligned_access(CPUState *cs, vaddr vaddr,
     int target_el;
     bool same_el;
     uint32_t syn;
+    ARMMMUIdx arm_mmu_idx = core_to_arm_mmu_idx(env, mmu_idx);
 
     if (retaddr) {
         /* now we have a real cpu fault */
@@ -208,7 +209,7 @@ void arm_cpu_do_unaligned_access(CPUState *cs, vaddr vaddr,
     /* the DFSR for an alignment fault depends on whether we're using
      * the LPAE long descriptor format, or the short descriptor format
      */
-    if (arm_s1_regime_using_lpae_format(env, mmu_idx)) {
+    if (arm_s1_regime_using_lpae_format(env, arm_mmu_idx)) {
         env->exception.fsr = (1 << 9) | 0x21;
     } else {
         env->exception.fsr = 0x1;
diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c
index 24de30d..a82ab49 100644
--- a/target/arm/translate-a64.c
+++ b/target/arm/translate-a64.c
@@ -101,21 +101,27 @@ void a64_translate_init(void)
         offsetof(CPUARMState, exclusive_high), "exclusive_high");
 }
 
-static inline ARMMMUIdx get_a64_user_mem_index(DisasContext *s)
+static inline int get_a64_user_mem_index(DisasContext *s)
 {
-    /* Return the mmu_idx to use for A64 "unprivileged load/store" insns:
+    /* Return the core mmu_idx to use for A64 "unprivileged load/store" insns:
      *  if EL1, access as if EL0; otherwise access at current EL
      */
+    ARMMMUIdx useridx;
+
     switch (s->mmu_idx) {
     case ARMMMUIdx_S12NSE1:
-        return ARMMMUIdx_S12NSE0;
+        useridx = ARMMMUIdx_S12NSE0;
+        break;
     case ARMMMUIdx_S1SE1:
-        return ARMMMUIdx_S1SE0;
+        useridx = ARMMMUIdx_S1SE0;
+        break;
     case ARMMMUIdx_S2NS:
         g_assert_not_reached();
     default:
-        return s->mmu_idx;
+        useridx = s->mmu_idx;
+        break;
     }
+    return arm_to_core_mmu_idx(useridx);
 }
 
 void aarch64_cpu_dump_state(CPUState *cs, FILE *f,
@@ -11212,7 +11218,7 @@ void gen_intermediate_code_a64(ARMCPU *cpu, TranslationBlock *tb)
     dc->be_data = ARM_TBFLAG_BE_DATA(tb->flags) ? MO_BE : MO_LE;
     dc->condexec_mask = 0;
     dc->condexec_cond = 0;
-    dc->mmu_idx = ARM_TBFLAG_MMUIDX(tb->flags);
+    dc->mmu_idx = core_to_arm_mmu_idx(env, ARM_TBFLAG_MMUIDX(tb->flags));
     dc->tbi0 = ARM_TBFLAG_TBI0(tb->flags);
     dc->tbi1 = ARM_TBFLAG_TBI1(tb->flags);
     dc->current_el = arm_mmu_idx_to_el(dc->mmu_idx);
diff --git a/target/arm/translate.c b/target/arm/translate.c
index 0b5a0bc..8d509a2 100644
--- a/target/arm/translate.c
+++ b/target/arm/translate.c
@@ -145,9 +145,9 @@ static void disas_set_da_iss(DisasContext *s, TCGMemOp memop, ISSInfo issinfo)
     disas_set_insn_syndrome(s, syn);
 }
 
-static inline ARMMMUIdx get_a32_user_mem_index(DisasContext *s)
+static inline int get_a32_user_mem_index(DisasContext *s)
 {
-    /* Return the mmu_idx to use for A32/T32 "unprivileged load/store"
+    /* Return the core mmu_idx to use for A32/T32 "unprivileged load/store"
      * insns:
      *  if PL2, UNPREDICTABLE (we choose to implement as if PL0)
      *  otherwise, access as if at PL0.
@@ -156,11 +156,11 @@ static inline ARMMMUIdx get_a32_user_mem_index(DisasContext *s)
     case ARMMMUIdx_S1E2:        /* this one is UNPREDICTABLE */
     case ARMMMUIdx_S12NSE0:
     case ARMMMUIdx_S12NSE1:
-        return ARMMMUIdx_S12NSE0;
+        return arm_to_core_mmu_idx(ARMMMUIdx_S12NSE0);
     case ARMMMUIdx_S1E3:
     case ARMMMUIdx_S1SE0:
     case ARMMMUIdx_S1SE1:
-        return ARMMMUIdx_S1SE0;
+        return arm_to_core_mmu_idx(ARMMMUIdx_S1SE0);
     case ARMMMUIdx_S2NS:
     default:
         g_assert_not_reached();
@@ -11816,7 +11816,7 @@ void gen_intermediate_code(CPUARMState *env, TranslationBlock *tb)
     dc->be_data = ARM_TBFLAG_BE_DATA(tb->flags) ? MO_BE : MO_LE;
     dc->condexec_mask = (ARM_TBFLAG_CONDEXEC(tb->flags) & 0xf) << 1;
     dc->condexec_cond = ARM_TBFLAG_CONDEXEC(tb->flags) >> 4;
-    dc->mmu_idx = ARM_TBFLAG_MMUIDX(tb->flags);
+    dc->mmu_idx = core_to_arm_mmu_idx(env, ARM_TBFLAG_MMUIDX(tb->flags));
     dc->current_el = arm_mmu_idx_to_el(dc->mmu_idx);
 #if !defined(CONFIG_USER_ONLY)
     dc->user = (dc->current_el == 0);
-- 
2.7.4

[Qemu-devel] [RFC 3/3] arm: Use different ARMMMUIdx values for M profile

[Peter Maydell]

Make M profile use completely separate ARMMMUIdx values from
those that A profile CPUs use. This is a prelude to adding
support for the MPU and for v8M, which together will require
6 MMU indexes which don't map cleanly onto the A profile
uses:
 non secure User
 non secure Privileged
 non secure Privileged, execution priority < 0
 secure User
 secure Privileged
 secure Privileged, execution priority < 0

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target/arm/cpu.h       | 21 ++++++++++++++++++---
 target/arm/helper.c    |  5 +++++
 target/arm/translate.c |  3 +++
 3 files changed, 26 insertions(+), 3 deletions(-)

diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index c162d43..226ddd9 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -2055,8 +2055,9 @@ static inline bool arm_excp_unmasked(CPUState *cs, unsigned int excp_idx,
  * of the AT/ATS operations.
  * The values used are carefully arranged to make mmu_idx => EL lookup easy.
  */
-#define ARM_MMU_IDX_A 0x10 /* A profile (and M profile, for the moment) */
+#define ARM_MMU_IDX_A 0x10 /* A profile */
 #define ARM_MMU_IDX_NOTLB 0x20 /* does not have a TLB */
+#define ARM_MMU_IDX_M 0x40 /* M profile */
 
 typedef enum ARMMMUIdx {
     ARMMMUIdx_S12NSE0 = 0 | ARM_MMU_IDX_A,
@@ -2066,6 +2067,8 @@ typedef enum ARMMMUIdx {
     ARMMMUIdx_S1SE0 = 4 | ARM_MMU_IDX_A,
     ARMMMUIdx_S1SE1 = 5 | ARM_MMU_IDX_A,
     ARMMMUIdx_S2NS = 6 | ARM_MMU_IDX_A,
+    ARMMMUIdx_MUser = 0 | ARM_MMU_IDX_M,
+    ARMMMUIdx_MPriv = 1 | ARM_MMU_IDX_M,
     /* Indexes below here don't have TLBs and are used only for AT system
      * instructions or for the first stage of an S12 page table walk.
      */
@@ -2084,6 +2087,8 @@ typedef enum ARMMMUIdxBit {
     ARMMMUIdxBit_S1SE0 = 1 << 4,
     ARMMMUIdxBit_S1SE1 = 1 << 5,
     ARMMMUIdxBit_S2NS = 1 << 6,
+    ARMMMUIdxBit_MUser = 1 << 0,
+    ARMMMUIdxBit_MPriv = 1 << 1,
 } ARMMMUIdxBit;
 
 #define MMU_USER_IDX 0
@@ -2095,13 +2100,17 @@ static inline int arm_to_core_mmu_idx(ARMMMUIdx mmu_idx)
 
 static inline ARMMMUIdx core_to_arm_mmu_idx(CPUARMState *env, int mmu_idx)
 {
-    return mmu_idx | ARM_MMU_IDX_A;
+    if (arm_feature(env, ARM_FEATURE_M)) {
+        return mmu_idx | ARM_MMU_IDX_M;
+    } else {
+        return mmu_idx | ARM_MMU_IDX_A;
+    }
 }
 
 /* Return the exception level we're running at if this is our mmu_idx */
 static inline int arm_mmu_idx_to_el(ARMMMUIdx mmu_idx)
 {
-    assert(mmu_idx < ARMMMUIdx_S2NS);
+    assert((mmu_idx & ~7) != ARM_MMU_IDX_NOTLB);
     return mmu_idx & 3;
 }
 
@@ -2110,6 +2119,12 @@ static inline int cpu_mmu_index(CPUARMState *env, bool ifetch)
 {
     int el = arm_current_el(env);
 
+    if (arm_feature(env, ARM_FEATURE_M)) {
+        ARMMMUIdx mmu_idx = el == 0 ? ARMMMUIdx_MUser : ARMMMUIdx_MPriv;
+
+        return arm_to_core_mmu_idx(mmu_idx);
+    }
+
     if (el < 2 && arm_is_secure_below_el3(env)) {
         return arm_to_core_mmu_idx(ARMMMUIdx_S1SE0 + el);
     }
diff --git a/target/arm/helper.c b/target/arm/helper.c
index 520adcc..791332c 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -6992,6 +6992,8 @@ static inline uint32_t regime_el(CPUARMState *env, ARMMMUIdx mmu_idx)
     case ARMMMUIdx_S1SE1:
     case ARMMMUIdx_S1NSE0:
     case ARMMMUIdx_S1NSE1:
+    case ARMMMUIdx_MPriv:
+    case ARMMMUIdx_MUser:
         return 1;
     default:
         g_assert_not_reached();
@@ -7008,6 +7010,8 @@ static inline bool regime_is_secure(CPUARMState *env, ARMMMUIdx mmu_idx)
     case ARMMMUIdx_S1NSE1:
     case ARMMMUIdx_S1E2:
     case ARMMMUIdx_S2NS:
+    case ARMMMUIdx_MPriv:
+    case ARMMMUIdx_MUser:
         return false;
     case ARMMMUIdx_S1E3:
     case ARMMMUIdx_S1SE0:
@@ -7146,6 +7150,7 @@ static inline bool regime_is_user(CPUARMState *env, ARMMMUIdx mmu_idx)
     switch (mmu_idx) {
     case ARMMMUIdx_S1SE0:
     case ARMMMUIdx_S1NSE0:
+    case ARMMMUIdx_MUser:
         return true;
     default:
         return false;
diff --git a/target/arm/translate.c b/target/arm/translate.c
index 8d509a2..ac905dd 100644
--- a/target/arm/translate.c
+++ b/target/arm/translate.c
@@ -161,6 +161,9 @@ static inline int get_a32_user_mem_index(DisasContext *s)
     case ARMMMUIdx_S1SE0:
     case ARMMMUIdx_S1SE1:
         return arm_to_core_mmu_idx(ARMMMUIdx_S1SE0);
+    case ARMMMUIdx_MUser:
+    case ARMMMUIdx_MPriv:
+        return arm_to_core_mmu_idx(ARMMMUIdx_MUser);
     case ARMMMUIdx_S2NS:
     default:
         g_assert_not_reached();
-- 
2.7.4

Re: [Qemu-devel] [Qemu-arm] [RFC 0/3] split core mmu_idx from ARMMMUIdx values

[Peter Maydell]

On 24 April 2017 at 17:33, Peter Maydell <peter.maydell@linaro.org> wrote:
> For M profile, we're eventually going to want some
> new MMU index values:
>      non secure User
>      non secure Privileged
>      non secure Privileged, execution priority < 0
>      secure User
>      secure Privileged
>      secure Privileged, execution priority < 0

Hmm, Alex and Edgar were definitely on the cc when I sent it but
seem to have fallen off it. (If you got a copy anyway then you
should probably check whether you have accidentally told mailman to
"filter out duplicate messages": you never want this setting
enabled, because it means "silently drop me from the cc list".)

thanks
-- PMM

Re: [Qemu-devel] [RFC 1/3] arm: Use the mmu_idx we're passed in arm_cpu_do_unaligned_access()

[Richard Henderson]

On 04/24/2017 06:33 PM, Peter Maydell wrote:
> When identifying the DFSR format for an alignment fault, use
> the mmu index that we are passed, rather than calling cpu_mmu_index()
> to get the mmu index for the current CPU state. This doesn't actually
> make any difference since the only cases where the current MMU index
> differs from the index used for the load are the "unprivileged
> load/store" instructions, and in that case the mmu index may
> differ but the translation regime is the same (apart from the
> "use from Hyp mode" case which is UNPREDICTABLE).
> However it's the more logical thing to do.
> 
> Signed-off-by: Peter Maydell<peter.maydell@linaro.org>
> ---
>   target/arm/op_helper.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)

Reviewed-by: Richard Henderson <rth@twiddle.net>


r~

Re: [Qemu-devel] [RFC 0/3] split core mmu_idx from ARMMMUIdx values

[Richard Henderson]

On 04/24/2017 06:33 PM, Peter Maydell wrote:
> Overall questions:
>   * does this look like it's workable enough to justify taking
>     it rather than the simple-but-worse-perf-on-some-archs
>     option 1 ?

I think it's workable.

>   * this does risk breaking A profile in corner cases, because
>     C unfortunately doesn't typecheck int vs enum so if I forgot
>     a call to an index conversion function somewhere things will
>     break. are we happy the risk isn't too huge?

If you like, we could add a tcg_debug assert somewhere that verifies that the 
given index is < NB_MMU_MODES...


r~

Re: [Qemu-devel] [RFC 0/3] split core mmu_idx from ARMMMUIdx values

[Peter Maydell]

On 26 April 2017 at 11:07, Richard Henderson <rth@twiddle.net> wrote:
> On 04/24/2017 06:33 PM, Peter Maydell wrote:
>>
>> Overall questions:
>>   * does this look like it's workable enough to justify taking
>>     it rather than the simple-but-worse-perf-on-some-archs
>>     option 1 ?
>
>
> I think it's workable.

Yes, I think I agree overall.

>>   * this does risk breaking A profile in corner cases, because
>>     C unfortunately doesn't typecheck int vs enum so if I forgot
>>     a call to an index conversion function somewhere things will
>>     break. are we happy the risk isn't too huge?
>
>
> If you like, we could add a tcg_debug assert somewhere that verifies that
> the given index is < NB_MMU_MODES...

There is one already, because in my testing I ran into it.
The problem is that since it's a runtime assert, it's not
certain that I caught everything.

thanks
-- PMM