Skip to contentSkip to conversation

target/arm: honor HCR_E2H and HCR_TGE in ats_write64()

PATCH Applied Ake Koomsin 3 participants 3 messages Nov 1, 2022 – Nov 3, 2022

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 target/arm: honor HCR_E2H and HCR_TGE in ats_write64()

We need to check HCR_E2H and HCR_TGE to select the right MMU index for the correct translation regime.

To check for EL2&0 translation regime:
- For S1E0*, S1E1* and S12E* ops, check both HCR_E2H and HCR_TGE
- For S1E2* ops, check only HCR_E2H

---

v3:
- Avoid recomputing arm_hcr_el2_eff() as recommended by Richard H.
- Use ':?' for more compact code as recommended by Richard H.

v2:
- Rebase with the latest upstream
- It turns out that we need to check both HCR_E2H and HCR_TGE for S1E0*, S1E1* and S12E* address translation as well according to the Architecture Manual.
- https://lists.gnu.org/archive/html/qemu-devel/2022-10/msg06084.html

v1:
https://lists.gnu.org/archive/html/qemu-devel/2022-10/msg02627.html

 target/arm/helper.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)
target/arm/helper.c +9 -6
index b070a20f1a..c17962372d 100644
@@ -3501,19 +3501,22 @@ static void ats_write64(CPUARMState *env, const ARMCPRegInfo *ri,
     MMUAccessType access_type = ri->opc2 & 1 ? MMU_DATA_STORE : MMU_DATA_LOAD;
     ARMMMUIdx mmu_idx;
     int secure = arm_is_secure_below_el3(env);
+    uint64_t hcr_el2 = arm_hcr_el2_eff(env);
+    bool regime_e20 = (hcr_el2 & (HCR_E2H | HCR_TGE)) == (HCR_E2H | HCR_TGE);
 
     switch (ri->opc2 & 6) {
     case 0:
         switch (ri->opc1) {
         case 0: /* AT S1E1R, AT S1E1W, AT S1E1RP, AT S1E1WP */
             if (ri->crm == 9 && (env->pstate & PSTATE_PAN)) {
-                mmu_idx = ARMMMUIdx_Stage1_E1_PAN;
+                mmu_idx = regime_e20 ?
+                          ARMMMUIdx_E20_2_PAN : ARMMMUIdx_Stage1_E1_PAN;
             } else {
-                mmu_idx = ARMMMUIdx_Stage1_E1;
+                mmu_idx = regime_e20 ? ARMMMUIdx_E20_2 : ARMMMUIdx_Stage1_E1;
             }
             break;
         case 4: /* AT S1E2R, AT S1E2W */
-            mmu_idx = ARMMMUIdx_E2;
+            mmu_idx = hcr_el2 & HCR_E2H ? ARMMMUIdx_E20_2 : ARMMMUIdx_E2;
             break;
         case 6: /* AT S1E3R, AT S1E3W */
             mmu_idx = ARMMMUIdx_E3;
@@ -3524,13 +3527,13 @@ static void ats_write64(CPUARMState *env, const ARMCPRegInfo *ri,
         }
         break;
     case 2: /* AT S1E0R, AT S1E0W */
-        mmu_idx = ARMMMUIdx_Stage1_E0;
+        mmu_idx = regime_e20 ? ARMMMUIdx_E20_0 : ARMMMUIdx_Stage1_E0;
         break;
     case 4: /* AT S12E1R, AT S12E1W */
-        mmu_idx = ARMMMUIdx_E10_1;
+        mmu_idx = regime_e20 ? ARMMMUIdx_E20_2 : ARMMMUIdx_E10_1;
         break;
     case 6: /* AT S12E0R, AT S12E0W */
-        mmu_idx = ARMMMUIdx_E10_0;
+        mmu_idx = regime_e20 ? ARMMMUIdx_E20_0 : ARMMMUIdx_E10_0;
         break;
     default:
         g_assert_not_reached();
Signature · 1 line
2.25.1
Reviewed-by on v3 1/1
#1 lore raw
Quoted 21 lines · Ake Koomsin

We need to check HCR_E2H and HCR_TGE to select the right MMU index for the correct translation regime.

To check for EL2&0 translation regime:
- For S1E0*, S1E1* and S12E* ops, check both HCR_E2H and HCR_TGE
- For S1E2* ops, check only HCR_E2H

---

v3:
- Avoid recomputing arm_hcr_el2_eff() as recommended by Richard H.
- Use ':?' for more compact code as recommended by Richard H.

v2:
- Rebase with the latest upstream
- It turns out that we need to check both HCR_E2H and HCR_TGE for S1E0*, S1E1* and S12E* address translation as well according to the Architecture Manual.
-https://lists.gnu.org/archive/html/qemu-devel/2022-10/msg06084.html

Quoted 5 lines

v1:
https://lists.gnu.org/archive/html/qemu-devel/2022-10/msg02627.html

  target/arm/helper.c | 15 +++++++++------
  1 file changed, 9 insertions(+), 6 deletions(-)

r~

#2 lore raw
Quoted 23 lines · Richard Henderson
Quoted 21 lines · Ake Koomsin

We need to check HCR_E2H and HCR_TGE to select the right MMU index for the correct translation regime.

To check for EL2&0 translation regime:
- For S1E0*, S1E1* and S12E* ops, check both HCR_E2H and HCR_TGE
- For S1E2* ops, check only HCR_E2H

---

v3:
- Avoid recomputing arm_hcr_el2_eff() as recommended by Richard H.
- Use ':?' for more compact code as recommended by Richard H.

v2:
- Rebase with the latest upstream
- It turns out that we need to check both HCR_E2H and HCR_TGE for S1E0*, S1E1* and S12E* address translation as well according to the Architecture Manual.
-https://lists.gnu.org/archive/html/qemu-devel/2022-10/msg06084.html

Quoted 7 lines
Quoted 5 lines

v1:
https://lists.gnu.org/archive/html/qemu-devel/2022-10/msg02627.html

  target/arm/helper.c | 15 +++++++++------
  1 file changed, 9 insertions(+), 6 deletions(-)

Applied to target-arm.next, thanks.

-- PMM