[PATCH 0/7] target/arm: More trivial features, A76, N1

[PATCH 0/7] target/arm: More trivial features, A76, N1

[Richard Henderson]

Based-on: 20220409000742.293691-1-richard.henderson@linaro.org
("target/arm: Implement features Debugv8p4, RAS, IESB")

3 more completely trivial features, and a 4th that merely
needs to add some state words.

Add definitions for cortex-a76 (also requires gicv4) and
neoverse-n1 (also requires gicv4.1), but we now have all
of the in-cpu features implemented.


r~


Richard Henderson (7):
  target/arm: Enable FEAT_CSV2 for -cpu max
  target/arm: Update ISAR fields for ARMv8.8
  target/arm: Enable FEAT_CSV2_2 for -cpu max
  target/arm: Enable FEAT_CSV3 for -cpu max
  target/arm: Enable FEAT_DGH for -cpu max
  target/arm: Define cortex-a76
  target/arm: Define neoverse-n1

 target/arm/cpu.h           |  39 +++++++++++
 hw/arm/sbsa-ref.c          |   2 +
 hw/arm/virt.c              |   2 +
 target/arm/cpu64.c         | 131 +++++++++++++++++++++++++++++++++++++
 target/arm/cpu_tcg.c       |   2 +
 target/arm/helper.c        |  70 +++++++++++++++++++-
 target/arm/translate-a64.c |   1 +
 7 files changed, 246 insertions(+), 1 deletion(-)

-- 
2.25.1

[PATCH 1/7] target/arm: Enable FEAT_CSV2 for -cpu max

[Richard Henderson]

This extension concerns branch speculation, which TCG does
not implement.  Thus we can trivially enable this feature.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 target/arm/cpu64.c   | 1 +
 target/arm/cpu_tcg.c | 1 +
 2 files changed, 2 insertions(+)

diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index def0f1fdcb..c1006a067c 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -805,6 +805,7 @@ static void aarch64_max_initfn(Object *obj)
     t = FIELD_DP64(t, ID_AA64PFR0, SVE, 1);
     t = FIELD_DP64(t, ID_AA64PFR0, SEL2, 1);      /* FEAT_SEL2 */
     t = FIELD_DP64(t, ID_AA64PFR0, DIT, 1);       /* FEAT_DIT */
+    t = FIELD_DP64(t, ID_AA64PFR0, CSV2, 1);      /* FEAT_CSV2 */
     cpu->isar.id_aa64pfr0 = t;
 
     t = cpu->isar.id_aa64pfr1;
diff --git a/target/arm/cpu_tcg.c b/target/arm/cpu_tcg.c
index 5cce9116d0..2750cbebec 100644
--- a/target/arm/cpu_tcg.c
+++ b/target/arm/cpu_tcg.c
@@ -71,6 +71,7 @@ void arm32_max_features(ARMCPU *cpu)
     cpu->isar.id_mmfr4 = t;
 
     t = cpu->isar.id_pfr0;
+    t = FIELD_DP32(t, ID_PFR0, CSV2, 2);          /* FEAT_CVS2 */
     t = FIELD_DP32(t, ID_PFR0, DIT, 1);           /* FEAT_DIT */
     t = FIELD_DP32(t, ID_PFR0, RAS, 1);           /* FEAT_RAS */
     cpu->isar.id_pfr0 = t;
-- 
2.25.1

[PATCH 2/7] target/arm: Update ISAR fields for ARMv8.8

[Richard Henderson]

Update isar fields per ARM DDI0487 H.a.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 target/arm/cpu.h | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index eb8cb738b5..c6c6d89a69 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -1941,6 +1941,7 @@ FIELD(ID_MMFR4, CCIDX, 24, 4)
 FIELD(ID_MMFR4, EVT, 28, 4)
 
 FIELD(ID_MMFR5, ETS, 0, 4)
+FIELD(ID_MMFR5, NTLBPA, 4, 4)
 
 FIELD(ID_PFR0, STATE0, 0, 4)
 FIELD(ID_PFR0, STATE1, 4, 4)
@@ -1993,6 +1994,16 @@ FIELD(ID_AA64ISAR1, SPECRES, 40, 4)
 FIELD(ID_AA64ISAR1, BF16, 44, 4)
 FIELD(ID_AA64ISAR1, DGH, 48, 4)
 FIELD(ID_AA64ISAR1, I8MM, 52, 4)
+FIELD(ID_AA64ISAR1, XS, 56, 4)
+FIELD(ID_AA64ISAR1, LS64, 60, 4)
+
+FIELD(ID_AA64ISAR2, WFXT, 0, 4)
+FIELD(ID_AA64ISAR2, RPRES, 4, 4)
+FIELD(ID_AA64ISAR2, GPA3, 8, 4)
+FIELD(ID_AA64ISAR2, APA3, 12, 4)
+FIELD(ID_AA64ISAR2, MOPS, 16, 4)
+FIELD(ID_AA64ISAR2, BC, 20, 4)
+FIELD(ID_AA64ISAR2, PAC_FRAC, 24, 4)
 
 FIELD(ID_AA64PFR0, EL0, 0, 4)
 FIELD(ID_AA64PFR0, EL1, 4, 4)
@@ -2015,6 +2026,10 @@ FIELD(ID_AA64PFR1, SSBS, 4, 4)
 FIELD(ID_AA64PFR1, MTE, 8, 4)
 FIELD(ID_AA64PFR1, RAS_FRAC, 12, 4)
 FIELD(ID_AA64PFR1, MPAM_FRAC, 16, 4)
+FIELD(ID_AA64PFR1, SME, 24, 4)
+FIELD(ID_AA64PFR1, RNDR_TRAP, 28, 4)
+FIELD(ID_AA64PFR1, CSV2_FRAC, 32, 4)
+FIELD(ID_AA64PFR1, NMI, 36, 4)
 
 FIELD(ID_AA64MMFR0, PARANGE, 0, 4)
 FIELD(ID_AA64MMFR0, ASIDBITS, 4, 4)
@@ -2041,6 +2056,11 @@ FIELD(ID_AA64MMFR1, SPECSEI, 24, 4)
 FIELD(ID_AA64MMFR1, XNX, 28, 4)
 FIELD(ID_AA64MMFR1, TWED, 32, 4)
 FIELD(ID_AA64MMFR1, ETS, 36, 4)
+FIELD(ID_AA64MMFR1, HCX, 40, 4)
+FIELD(ID_AA64MMFR1, AFP, 44, 4)
+FIELD(ID_AA64MMFR1, NTLBPA, 48, 4)
+FIELD(ID_AA64MMFR1, TIDCP1, 52, 4)
+FIELD(ID_AA64MMFR1, CMOW, 56, 4)
 
 FIELD(ID_AA64MMFR2, CNP, 0, 4)
 FIELD(ID_AA64MMFR2, UAO, 4, 4)
@@ -2067,7 +2087,9 @@ FIELD(ID_AA64DFR0, CTX_CMPS, 28, 4)
 FIELD(ID_AA64DFR0, PMSVER, 32, 4)
 FIELD(ID_AA64DFR0, DOUBLELOCK, 36, 4)
 FIELD(ID_AA64DFR0, TRACEFILT, 40, 4)
+FIELD(ID_AA64DFR0, TRACEBUFFER, 44, 4)
 FIELD(ID_AA64DFR0, MTPMU, 48, 4)
+FIELD(ID_AA64DFR0, BRBE, 52, 4)
 
 FIELD(ID_AA64ZFR0, SVEVER, 0, 4)
 FIELD(ID_AA64ZFR0, AES, 4, 4)
@@ -2089,6 +2111,7 @@ FIELD(ID_DFR0, PERFMON, 24, 4)
 FIELD(ID_DFR0, TRACEFILT, 28, 4)
 
 FIELD(ID_DFR1, MTPMU, 0, 4)
+FIELD(ID_DFR1, HPMN0, 4, 4)
 
 FIELD(DBGDIDR, SE_IMP, 12, 1)
 FIELD(DBGDIDR, NSUHD_IMP, 14, 1)
-- 
2.25.1

[PATCH 3/7] target/arm: Enable FEAT_CSV2_2 for -cpu max

[Richard Henderson]

There is no branch prediction in TCG, therefore there is no
need to actually include the context number into the predictor.
Therefore all we need to do is add the state for SCXTNUM_ELx.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 target/arm/cpu.h    | 16 +++++++++++
 target/arm/cpu64.c  |  2 +-
 target/arm/helper.c | 70 ++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 86 insertions(+), 2 deletions(-)

diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index c6c6d89a69..0b89620662 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -688,6 +688,8 @@ typedef struct CPUArchState {
         ARMPACKey apdb;
         ARMPACKey apga;
     } keys;
+
+    uint64_t scxtnum_el[4];
 #endif
 
 #if defined(CONFIG_USER_ONLY)
@@ -1211,6 +1213,7 @@ void pmu_init(ARMCPU *cpu);
 #define SCTLR_WXN     (1U << 19)
 #define SCTLR_ST      (1U << 20) /* up to ??, RAZ in v6 */
 #define SCTLR_UWXN    (1U << 20) /* v7 onward, AArch32 only */
+#define SCTLR_TSCXT   (1U << 20) /* FEAT_CSV2_1p2, AArch64 only */
 #define SCTLR_FI      (1U << 21) /* up to v7, v8 RES0 */
 #define SCTLR_IESB    (1U << 21) /* v8.2-IESB, AArch64 only */
 #define SCTLR_U       (1U << 22) /* up to v6, RAO in v7 */
@@ -4368,6 +4371,19 @@ static inline bool isar_feature_aa64_dit(const ARMISARegisters *id)
     return FIELD_EX64(id->id_aa64pfr0, ID_AA64PFR0, DIT) != 0;
 }
 
+static inline bool isar_feature_aa64_scxtnum(const ARMISARegisters *id)
+{
+    int key = FIELD_EX64(id->id_aa64pfr0, ID_AA64PFR0, CSV2);
+    if (key >= 2) {
+        return true;      /* FEAT_CSV2_2 */
+    }
+    if (key == 1) {
+        key = FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, CSV2_FRAC);
+        return key >= 2;  /* FEAT_CSV2_1p2 */
+    }
+    return false;
+}
+
 static inline bool isar_feature_aa64_ssbs(const ARMISARegisters *id)
 {
     return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, SSBS) != 0;
diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index c1006a067c..9ff08bd995 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -805,7 +805,7 @@ static void aarch64_max_initfn(Object *obj)
     t = FIELD_DP64(t, ID_AA64PFR0, SVE, 1);
     t = FIELD_DP64(t, ID_AA64PFR0, SEL2, 1);      /* FEAT_SEL2 */
     t = FIELD_DP64(t, ID_AA64PFR0, DIT, 1);       /* FEAT_DIT */
-    t = FIELD_DP64(t, ID_AA64PFR0, CSV2, 1);      /* FEAT_CSV2 */
+    t = FIELD_DP64(t, ID_AA64PFR0, CSV2, 2);      /* FEAT_CSV2_2 */
     cpu->isar.id_aa64pfr0 = t;
 
     t = cpu->isar.id_aa64pfr1;
diff --git a/target/arm/helper.c b/target/arm/helper.c
index bd1c8e01cb..66af3397ee 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -1780,6 +1780,9 @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value)
         if (cpu_isar_feature(aa64_mte, cpu)) {
             valid_mask |= SCR_ATA;
         }
+        if (cpu_isar_feature(aa64_scxtnum, cpu)) {
+            valid_mask |= SCR_ENSCXT;
+        }
     } else {
         valid_mask &= ~(SCR_RW | SCR_ST);
         if (cpu_isar_feature(aa32_ras, cpu)) {
@@ -5312,6 +5315,9 @@ static void do_hcr_write(CPUARMState *env, uint64_t value, uint64_t valid_mask)
         if (cpu_isar_feature(aa64_mte, cpu)) {
             valid_mask |= HCR_ATA | HCR_DCT | HCR_TID5;
         }
+        if (cpu_isar_feature(aa64_scxtnum, cpu)) {
+            valid_mask |= HCR_ENSCXT;
+        }
     }
 
     /* Clear RES0 bits.  */
@@ -5965,6 +5971,10 @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu)
         { K(3, 0,  5, 6, 0), K(3, 4,  5, 6, 0), K(3, 5, 5, 6, 0),
           "TFSR_EL1", "TFSR_EL2", "TFSR_EL12", isar_feature_aa64_mte },
 
+        { K(3, 0, 0xd, 0, 7), K(3, 4, 0xd, 0, 7), K(3, 5, 0xd, 0, 7),
+          "SCXTNUM_EL1", "SCXTNUM_EL2", "SCXTNUM_EL12",
+          isar_feature_aa64_scxtnum },
+
         /* TODO: ARMv8.2-SPE -- PMSCR_EL2 */
         /* TODO: ARMv8.4-Trace -- TRFCR_EL2 */
     };
@@ -7434,7 +7444,61 @@ static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = {
     REGINFO_SENTINEL
 };
 
-#endif
+static CPAccessResult access_scxtnum(CPUARMState *env, const ARMCPRegInfo *ri,
+                                     bool isread)
+{
+    uint64_t hcr;
+
+    switch (arm_current_el(env)) {
+    case 0:
+        hcr = arm_hcr_el2_eff(env);
+        if ((hcr & (HCR_TGE | HCR_E2H)) != (HCR_TGE | HCR_E2H)) {
+            if (env->cp15.sctlr_el[1] & SCTLR_TSCXT) {
+                if (hcr & HCR_TGE) {
+                    return CP_ACCESS_TRAP_EL2;
+                }
+                return CP_ACCESS_TRAP;
+            }
+            if (arm_is_el2_enabled(env) && !(hcr & HCR_ENSCXT)) {
+                return CP_ACCESS_TRAP_EL2;
+            }
+        } else {
+            QEMU_FALLTHROUGH;
+    case 1:
+            if (env->cp15.sctlr_el[2] & SCTLR_TSCXT) {
+                return CP_ACCESS_TRAP_EL2;
+            }
+        }
+        QEMU_FALLTHROUGH;
+    case 2:
+        if (arm_feature(env, ARM_FEATURE_EL3)
+            && !(env->cp15.scr_el3 & SCR_ENSCXT)) {
+            return CP_ACCESS_TRAP_EL3;
+        }
+    }
+    return CP_ACCESS_OK;
+}
+
+static const ARMCPRegInfo scxtnum_reginfo[] = {
+    { .name = "SCXTNUM_EL0", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 3, .crn = 0xd, .crm = 0, .opc2 = 7,
+      .access = PL0_RW, .accessfn = access_scxtnum,
+      .fieldoffset = offsetof(CPUARMState, scxtnum_el[0]) },
+    { .name = "SCXTNUM_EL1", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 0, .crn = 0xd, .crm = 0, .opc2 = 7,
+      .access = PL1_RW, .accessfn = access_scxtnum,
+      .fieldoffset = offsetof(CPUARMState, scxtnum_el[1]) },
+    { .name = "SCXTNUM_EL2", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 4, .crn = 0xd, .crm = 0, .opc2 = 7,
+      .access = PL2_RW, .accessfn = access_scxtnum,
+      .fieldoffset = offsetof(CPUARMState, scxtnum_el[2]) },
+    { .name = "SCXTNUM_EL3", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 6, .crn = 0xd, .crm = 0, .opc2 = 7,
+      .access = PL3_RW,
+      .fieldoffset = offsetof(CPUARMState, scxtnum_el[3]) },
+    REGINFO_SENTINEL
+};
+#endif /* TARGET_AARCH64 */
 
 static CPAccessResult access_predinv(CPUARMState *env, const ARMCPRegInfo *ri,
                                      bool isread)
@@ -8634,6 +8698,10 @@ void register_cp_regs_for_features(ARMCPU *cpu)
         define_arm_cp_regs(cpu, mte_tco_ro_reginfo);
         define_arm_cp_regs(cpu, mte_el0_cacheop_reginfo);
     }
+
+    if (cpu_isar_feature(aa64_scxtnum, cpu)) {
+        define_arm_cp_regs(cpu, scxtnum_reginfo);
+    }
 #endif
 
     if (cpu_isar_feature(any_predinv, cpu)) {
-- 
2.25.1

[PATCH 4/7] target/arm: Enable FEAT_CSV3 for -cpu max

[Richard Henderson]

This extension concerns cache speculation, which TCG does
not implement.  Thus we can trivially enable this feature.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 target/arm/cpu64.c   | 1 +
 target/arm/cpu_tcg.c | 1 +
 2 files changed, 2 insertions(+)

diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index 9ff08bd995..a0429538cc 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -806,6 +806,7 @@ static void aarch64_max_initfn(Object *obj)
     t = FIELD_DP64(t, ID_AA64PFR0, SEL2, 1);      /* FEAT_SEL2 */
     t = FIELD_DP64(t, ID_AA64PFR0, DIT, 1);       /* FEAT_DIT */
     t = FIELD_DP64(t, ID_AA64PFR0, CSV2, 2);      /* FEAT_CSV2_2 */
+    t = FIELD_DP64(t, ID_AA64PFR0, CSV3, 1);      /* FEAT_CSV3 */
     cpu->isar.id_aa64pfr0 = t;
 
     t = cpu->isar.id_aa64pfr1;
diff --git a/target/arm/cpu_tcg.c b/target/arm/cpu_tcg.c
index 2750cbebec..31ea5a90e0 100644
--- a/target/arm/cpu_tcg.c
+++ b/target/arm/cpu_tcg.c
@@ -77,6 +77,7 @@ void arm32_max_features(ARMCPU *cpu)
     cpu->isar.id_pfr0 = t;
 
     t = cpu->isar.id_pfr2;
+    t = FIELD_DP32(t, ID_PFR2, CSV3, 1);          /* FEAT_CSV3 */
     t = FIELD_DP32(t, ID_PFR2, SSBS, 1);          /* FEAT_SSBS */
     cpu->isar.id_pfr2 = t;
 
-- 
2.25.1

[PATCH 5/7] target/arm: Enable FEAT_DGH for -cpu max

[Richard Henderson]

This extension concerns not merging memory access, which TCG does
not implement.  Thus we can trivially enable this feature.
Add a comment to handle_hint for the DGH instruction, but no code.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 target/arm/cpu64.c         | 1 +
 target/arm/translate-a64.c | 1 +
 2 files changed, 2 insertions(+)

diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index a0429538cc..199ca437a0 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -795,6 +795,7 @@ static void aarch64_max_initfn(Object *obj)
     t = FIELD_DP64(t, ID_AA64ISAR1, SB, 1);       /* FEAT_SB */
     t = FIELD_DP64(t, ID_AA64ISAR1, SPECRES, 1);  /* FEAT_SPECRES */
     t = FIELD_DP64(t, ID_AA64ISAR1, BF16, 1);     /* FEAT_BF16 */
+    t = FIELD_DP64(t, ID_AA64ISAR1, DGH, 1);      /* FEAT_DGH */
     t = FIELD_DP64(t, ID_AA64ISAR1, I8MM, 1);     /* FEAT_I8MM */
     cpu->isar.id_aa64isar1 = t;
 
diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c
index cc54dff83c..c3c1a19dea 100644
--- a/target/arm/translate-a64.c
+++ b/target/arm/translate-a64.c
@@ -1442,6 +1442,7 @@ static void handle_hint(DisasContext *s, uint32_t insn,
         break;
     case 0b00100: /* SEV */
     case 0b00101: /* SEVL */
+    case 0b00110: /* DGH */
         /* we treat all as NOP at least for now */
         break;
     case 0b00111: /* XPACLRI */
-- 
2.25.1

[PATCH 6/7] target/arm: Define cortex-a76

[Richard Henderson]

Enable the a76 for virt and sbsa board use.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 hw/arm/sbsa-ref.c  |  1 +
 hw/arm/virt.c      |  1 +
 target/arm/cpu64.c | 64 ++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 66 insertions(+)

diff --git a/hw/arm/sbsa-ref.c b/hw/arm/sbsa-ref.c
index 2387401963..2ddde88f5e 100644
--- a/hw/arm/sbsa-ref.c
+++ b/hw/arm/sbsa-ref.c
@@ -145,6 +145,7 @@ static const int sbsa_ref_irqmap[] = {
 static const char * const valid_cpus[] = {
     ARM_CPU_TYPE_NAME("cortex-a57"),
     ARM_CPU_TYPE_NAME("cortex-a72"),
+    ARM_CPU_TYPE_NAME("cortex-a76"),
     ARM_CPU_TYPE_NAME("max"),
 };
 
diff --git a/hw/arm/virt.c b/hw/arm/virt.c
index d2e5ecd234..ce15c36a7f 100644
--- a/hw/arm/virt.c
+++ b/hw/arm/virt.c
@@ -202,6 +202,7 @@ static const char *valid_cpus[] = {
     ARM_CPU_TYPE_NAME("cortex-a53"),
     ARM_CPU_TYPE_NAME("cortex-a57"),
     ARM_CPU_TYPE_NAME("cortex-a72"),
+    ARM_CPU_TYPE_NAME("cortex-a76"),
     ARM_CPU_TYPE_NAME("a64fx"),
     ARM_CPU_TYPE_NAME("host"),
     ARM_CPU_TYPE_NAME("max"),
diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index 199ca437a0..a00c02be2c 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -251,6 +251,69 @@ static void aarch64_a72_initfn(Object *obj)
     define_arm_cp_regs(cpu, cortex_a72_a57_a53_cp_reginfo);
 }
 
+static void aarch64_a76_initfn(Object *obj)
+{
+    ARMCPU *cpu = ARM_CPU(obj);
+
+    cpu->dtb_compatible = "arm,cortex-a76";
+    set_feature(&cpu->env, ARM_FEATURE_V8);
+    set_feature(&cpu->env, ARM_FEATURE_NEON);
+    set_feature(&cpu->env, ARM_FEATURE_GENERIC_TIMER);
+    set_feature(&cpu->env, ARM_FEATURE_AARCH64);
+    set_feature(&cpu->env, ARM_FEATURE_CBAR_RO);
+    set_feature(&cpu->env, ARM_FEATURE_EL2);
+    set_feature(&cpu->env, ARM_FEATURE_EL3);
+    set_feature(&cpu->env, ARM_FEATURE_PMU);
+
+    /* Ordered by B2.4 AArch64 registers by functional group */
+    cpu->clidr = 0x82000023;
+    cpu->ctr = 0x8444C004;
+    cpu->dcz_blocksize = 4;
+    cpu->isar.id_aa64dfr0  = 0x0000000010305408ull;
+    cpu->isar.id_aa64isar0 = 0x0000100010211120ull;
+    cpu->isar.id_aa64isar1 = 0x0000000000100001ull;
+    cpu->isar.id_aa64mmfr0 = 0x0000000000101122ull;
+    cpu->isar.id_aa64mmfr1 = 0x0000000010212122ull;
+    cpu->isar.id_aa64mmfr2 = 0x0000000000001011ull;
+    cpu->isar.id_aa64pfr0  = 0x1100000010111112ull;
+    cpu->isar.id_aa64pfr1  = 0x0000000000000010ull;
+    cpu->isar.id_dfr0  = 0x04010088;
+    cpu->isar.id_isar0 = 0x02101110;
+    cpu->isar.id_isar1 = 0x13112111;
+    cpu->isar.id_isar2 = 0x21232042;
+    cpu->isar.id_isar3 = 0x01112131;
+    cpu->isar.id_isar4 = 0x00010142;
+    cpu->isar.id_isar5 = 0x01011121;
+    cpu->isar.id_isar6 = 0x00000010;
+    cpu->isar.id_mmfr0 = 0x10201105;
+    cpu->isar.id_mmfr1 = 0x40000000;
+    cpu->isar.id_mmfr2 = 0x01260000;
+    cpu->isar.id_mmfr3 = 0x02122211;
+    cpu->isar.id_mmfr4 = 0x00021110;
+    cpu->isar.id_pfr0  = 0x10010131;
+    cpu->isar.id_pfr1  = 0x10010000;
+    cpu->isar.id_pfr2  = 0x00000011;
+    cpu->midr = 0x413fd0b1;
+
+    /* From B2.18 CCSIDR_EL1 */
+    cpu->ccsidr[0] = 0x701fe01a; /* 64KB L1 dcache */
+    cpu->ccsidr[1] = 0x201fe01a; /* 64KB L1 icache */
+    cpu->ccsidr[2] = 0x707fe03a; /* 512KB L2 cache */
+
+    /* From B2.93 SCTLR_EL3 */
+    cpu->reset_sctlr = 0x30c50838;
+
+    /* From B4.23 ICH_VTR_EL2 */
+    cpu->gic_num_lrs = 4;
+    cpu->gic_vpribits = 5;
+    cpu->gic_vprebits = 5;
+
+    /* From B5.1 AdvSIMD AArch64 register summary */
+    cpu->isar.mvfr0 = 0x10110222;
+    cpu->isar.mvfr1 = 0x13211111;
+    cpu->isar.mvfr2 = 0x00000043;
+}
+
 void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp)
 {
     /*
@@ -935,6 +998,7 @@ static const ARMCPUInfo aarch64_cpus[] = {
     { .name = "cortex-a57",         .initfn = aarch64_a57_initfn },
     { .name = "cortex-a53",         .initfn = aarch64_a53_initfn },
     { .name = "cortex-a72",         .initfn = aarch64_a72_initfn },
+    { .name = "cortex-a76",         .initfn = aarch64_a76_initfn },
     { .name = "a64fx",              .initfn = aarch64_a64fx_initfn },
     { .name = "max",                .initfn = aarch64_max_initfn },
 #if defined(CONFIG_KVM) || defined(CONFIG_HVF)
-- 
2.25.1

[PATCH 7/7] target/arm: Define neoverse-n1

[Richard Henderson]

Enable the n1 for virt and sbsa board use.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 hw/arm/sbsa-ref.c  |  1 +
 hw/arm/virt.c      |  1 +
 target/arm/cpu64.c | 64 ++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 66 insertions(+)

diff --git a/hw/arm/sbsa-ref.c b/hw/arm/sbsa-ref.c
index 2ddde88f5e..dac8860f2d 100644
--- a/hw/arm/sbsa-ref.c
+++ b/hw/arm/sbsa-ref.c
@@ -146,6 +146,7 @@ static const char * const valid_cpus[] = {
     ARM_CPU_TYPE_NAME("cortex-a57"),
     ARM_CPU_TYPE_NAME("cortex-a72"),
     ARM_CPU_TYPE_NAME("cortex-a76"),
+    ARM_CPU_TYPE_NAME("neoverse-n1"),
     ARM_CPU_TYPE_NAME("max"),
 };
 
diff --git a/hw/arm/virt.c b/hw/arm/virt.c
index ce15c36a7f..82dd934de6 100644
--- a/hw/arm/virt.c
+++ b/hw/arm/virt.c
@@ -204,6 +204,7 @@ static const char *valid_cpus[] = {
     ARM_CPU_TYPE_NAME("cortex-a72"),
     ARM_CPU_TYPE_NAME("cortex-a76"),
     ARM_CPU_TYPE_NAME("a64fx"),
+    ARM_CPU_TYPE_NAME("neoverse-n1"),
     ARM_CPU_TYPE_NAME("host"),
     ARM_CPU_TYPE_NAME("max"),
 };
diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index a00c02be2c..6a67b3d840 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -314,6 +314,69 @@ static void aarch64_a76_initfn(Object *obj)
     cpu->isar.mvfr2 = 0x00000043;
 }
 
+static void aarch64_neoverse_n1_initfn(Object *obj)
+{
+    ARMCPU *cpu = ARM_CPU(obj);
+
+    cpu->dtb_compatible = "arm,neoverse-n1";
+    set_feature(&cpu->env, ARM_FEATURE_V8);
+    set_feature(&cpu->env, ARM_FEATURE_NEON);
+    set_feature(&cpu->env, ARM_FEATURE_GENERIC_TIMER);
+    set_feature(&cpu->env, ARM_FEATURE_AARCH64);
+    set_feature(&cpu->env, ARM_FEATURE_CBAR_RO);
+    set_feature(&cpu->env, ARM_FEATURE_EL2);
+    set_feature(&cpu->env, ARM_FEATURE_EL3);
+    set_feature(&cpu->env, ARM_FEATURE_PMU);
+
+    /* Ordered by B2.4 AArch64 registers by functional group */
+    cpu->clidr = 0x82000023;
+    cpu->ctr = 0x8444c004;
+    cpu->dcz_blocksize = 4;
+    cpu->isar.id_aa64dfr0  = 0x0000000110305408ull;
+    cpu->isar.id_aa64isar0 = 0x0000100010211120ull;
+    cpu->isar.id_aa64isar1 = 0x0000000000100001ull;
+    cpu->isar.id_aa64mmfr0 = 0x0000000000101125ull;
+    cpu->isar.id_aa64mmfr1 = 0x0000000010212122ull;
+    cpu->isar.id_aa64mmfr2 = 0x0000000000001011ull;
+    cpu->isar.id_aa64pfr0  = 0x1100000010111112ull;
+    cpu->isar.id_aa64pfr1  = 0x0000000000000020ull;
+    cpu->isar.id_dfr0  = 0x04010088;
+    cpu->isar.id_isar0 = 0x02101110;
+    cpu->isar.id_isar1 = 0x13112111;
+    cpu->isar.id_isar2 = 0x21232042;
+    cpu->isar.id_isar3 = 0x01112131;
+    cpu->isar.id_isar4 = 0x00010142;
+    cpu->isar.id_isar5 = 0x01011121;
+    cpu->isar.id_isar6 = 0x00000010;
+    cpu->isar.id_mmfr0 = 0x10201105;
+    cpu->isar.id_mmfr1 = 0x40000000;
+    cpu->isar.id_mmfr2 = 0x01260000;
+    cpu->isar.id_mmfr3 = 0x02122211;
+    cpu->isar.id_mmfr4 = 0x00021110;
+    cpu->isar.id_pfr0  = 0x10010131;
+    cpu->isar.id_pfr1  = 0x10010000;
+    cpu->isar.id_pfr2  = 0x00000011;
+    cpu->midr = 0x414fd0c1;
+
+    /* From B2.23 CCSIDR_EL1 */
+    cpu->ccsidr[0] = 0x701fe01a; /* 64KB L1 dcache */
+    cpu->ccsidr[1] = 0x201fe01a; /* 64KB L1 icache */
+    cpu->ccsidr[2] = 0x70ffe03a; /* 1MB L2 cache */
+
+    /* From B2.98 SCTLR_EL3 */
+    cpu->reset_sctlr = 0x30c50838;
+
+    /* From B4.23 ICH_VTR_EL2 */
+    cpu->gic_num_lrs = 4;
+    cpu->gic_vpribits = 5;
+    cpu->gic_vprebits = 5;
+
+    /* From B5.1 AdvSIMD AArch64 register summary */
+    cpu->isar.mvfr0 = 0x10110222;
+    cpu->isar.mvfr1 = 0x12111111;
+    cpu->isar.mvfr2 = 0x00000043;
+}
+
 void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp)
 {
     /*
@@ -1000,6 +1063,7 @@ static const ARMCPUInfo aarch64_cpus[] = {
     { .name = "cortex-a72",         .initfn = aarch64_a72_initfn },
     { .name = "cortex-a76",         .initfn = aarch64_a76_initfn },
     { .name = "a64fx",              .initfn = aarch64_a64fx_initfn },
+    { .name = "neoverse-n1",        .initfn = aarch64_neoverse_n1_initfn },
     { .name = "max",                .initfn = aarch64_max_initfn },
 #if defined(CONFIG_KVM) || defined(CONFIG_HVF)
     { .name = "host",               .initfn = aarch64_host_initfn },
-- 
2.25.1

Re: [PATCH 1/7] target/arm: Enable FEAT_CSV2 for -cpu max

[Peter Maydell]

On Sun, 10 Apr 2022 at 06:58, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> This extension concerns branch speculation, which TCG does
> not implement.  Thus we can trivially enable this feature.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
>  target/arm/cpu64.c   | 1 +
>  target/arm/cpu_tcg.c | 1 +


Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH 2/7] target/arm: Update ISAR fields for ARMv8.8

[Peter Maydell]

On Sun, 10 Apr 2022 at 06:58, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> Update isar fields per ARM DDI0487 H.a.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
>  target/arm/cpu.h | 23 +++++++++++++++++++++++
>  1 file changed, 23 insertions(+)

> @@ -2067,7 +2087,9 @@ FIELD(ID_AA64DFR0, CTX_CMPS, 28, 4)
>  FIELD(ID_AA64DFR0, PMSVER, 32, 4)
>  FIELD(ID_AA64DFR0, DOUBLELOCK, 36, 4)
>  FIELD(ID_AA64DFR0, TRACEFILT, 40, 4)
> +FIELD(ID_AA64DFR0, TRACEBUFFER, 44, 4)
>  FIELD(ID_AA64DFR0, MTPMU, 48, 4)
> +FIELD(ID_AA64DFR0, BRBE, 52, 4)

Also
   FIELD(ID_AA64DFR0, HPMN0, 60, 4)

Otherwise
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH 4/7] target/arm: Enable FEAT_CSV3 for -cpu max

[Peter Maydell]

On Sun, 10 Apr 2022 at 07:07, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> This extension concerns cache speculation, which TCG does
> not implement.  Thus we can trivially enable this feature.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
>  target/arm/cpu64.c   | 1 +
>  target/arm/cpu_tcg.c | 1 +
>  2 files changed, 2 insertions(+)
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH 5/7] target/arm: Enable FEAT_DGH for -cpu max

[Peter Maydell]

On Sun, 10 Apr 2022 at 07:07, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> This extension concerns not merging memory access, which TCG does
> not implement.  Thus we can trivially enable this feature.
> Add a comment to handle_hint for the DGH instruction, but no code.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH 3/7] target/arm: Enable FEAT_CSV2_2 for -cpu max

[Peter Maydell]

On Sun, 10 Apr 2022 at 07:03, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> There is no branch prediction in TCG, therefore there is no
> need to actually include the context number into the predictor.
> Therefore all we need to do is add the state for SCXTNUM_ELx.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
>  target/arm/cpu.h    | 16 +++++++++++
>  target/arm/cpu64.c  |  2 +-
>  target/arm/helper.c | 70 ++++++++++++++++++++++++++++++++++++++++++++-
>  3 files changed, 86 insertions(+), 2 deletions(-)
>
> diff --git a/target/arm/cpu.h b/target/arm/cpu.h
> index c6c6d89a69..0b89620662 100644
> --- a/target/arm/cpu.h
> +++ b/target/arm/cpu.h
> @@ -688,6 +688,8 @@ typedef struct CPUArchState {
>          ARMPACKey apdb;
>          ARMPACKey apga;
>      } keys;
> +
> +    uint64_t scxtnum_el[4];
>  #endif
>
>  #if defined(CONFIG_USER_ONLY)
> @@ -1211,6 +1213,7 @@ void pmu_init(ARMCPU *cpu);
>  #define SCTLR_WXN     (1U << 19)
>  #define SCTLR_ST      (1U << 20) /* up to ??, RAZ in v6 */
>  #define SCTLR_UWXN    (1U << 20) /* v7 onward, AArch32 only */
> +#define SCTLR_TSCXT   (1U << 20) /* FEAT_CSV2_1p2, AArch64 only */
>  #define SCTLR_FI      (1U << 21) /* up to v7, v8 RES0 */
>  #define SCTLR_IESB    (1U << 21) /* v8.2-IESB, AArch64 only */
>  #define SCTLR_U       (1U << 22) /* up to v6, RAO in v7 */
> @@ -4368,6 +4371,19 @@ static inline bool isar_feature_aa64_dit(const ARMISARegisters *id)
>      return FIELD_EX64(id->id_aa64pfr0, ID_AA64PFR0, DIT) != 0;
>  }
>
> +static inline bool isar_feature_aa64_scxtnum(const ARMISARegisters *id)
> +{
> +    int key = FIELD_EX64(id->id_aa64pfr0, ID_AA64PFR0, CSV2);
> +    if (key >= 2) {
> +        return true;      /* FEAT_CSV2_2 */
> +    }
> +    if (key == 1) {
> +        key = FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, CSV2_FRAC);
> +        return key >= 2;  /* FEAT_CSV2_1p2 */
> +    }
> +    return false;
> +}
> +
>  static inline bool isar_feature_aa64_ssbs(const ARMISARegisters *id)
>  {
>      return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, SSBS) != 0;
> diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
> index c1006a067c..9ff08bd995 100644
> --- a/target/arm/cpu64.c
> +++ b/target/arm/cpu64.c
> @@ -805,7 +805,7 @@ static void aarch64_max_initfn(Object *obj)
>      t = FIELD_DP64(t, ID_AA64PFR0, SVE, 1);
>      t = FIELD_DP64(t, ID_AA64PFR0, SEL2, 1);      /* FEAT_SEL2 */
>      t = FIELD_DP64(t, ID_AA64PFR0, DIT, 1);       /* FEAT_DIT */
> -    t = FIELD_DP64(t, ID_AA64PFR0, CSV2, 1);      /* FEAT_CSV2 */
> +    t = FIELD_DP64(t, ID_AA64PFR0, CSV2, 2);      /* FEAT_CSV2_2 */

I think we should probably explicitly zero ID_AA64PFR1.CSV2_frac.
Doesn't matter now, but might if in future we move 'max' from
"a57 with extras" to "some CPU with CSV2_1p1 plus extras".

>      cpu->isar.id_aa64pfr0 = t;
>
>      t = cpu->isar.id_aa64pfr1;
> diff --git a/target/arm/helper.c b/target/arm/helper.c
> index bd1c8e01cb..66af3397ee 100644
> --- a/target/arm/helper.c
> +++ b/target/arm/helper.c
> @@ -1780,6 +1780,9 @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value)
>          if (cpu_isar_feature(aa64_mte, cpu)) {
>              valid_mask |= SCR_ATA;
>          }
> +        if (cpu_isar_feature(aa64_scxtnum, cpu)) {
> +            valid_mask |= SCR_ENSCXT;
> +        }
>      } else {
>          valid_mask &= ~(SCR_RW | SCR_ST);
>          if (cpu_isar_feature(aa32_ras, cpu)) {
> @@ -5312,6 +5315,9 @@ static void do_hcr_write(CPUARMState *env, uint64_t value, uint64_t valid_mask)
>          if (cpu_isar_feature(aa64_mte, cpu)) {
>              valid_mask |= HCR_ATA | HCR_DCT | HCR_TID5;
>          }
> +        if (cpu_isar_feature(aa64_scxtnum, cpu)) {
> +            valid_mask |= HCR_ENSCXT;
> +        }
>      }
>
>      /* Clear RES0 bits.  */
> @@ -5965,6 +5971,10 @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu)
>          { K(3, 0,  5, 6, 0), K(3, 4,  5, 6, 0), K(3, 5, 5, 6, 0),
>            "TFSR_EL1", "TFSR_EL2", "TFSR_EL12", isar_feature_aa64_mte },
>
> +        { K(3, 0, 0xd, 0, 7), K(3, 4, 0xd, 0, 7), K(3, 5, 0xd, 0, 7),

Use decimal to follow the style of the rest of this table, please.

> +          "SCXTNUM_EL1", "SCXTNUM_EL2", "SCXTNUM_EL12",
> +          isar_feature_aa64_scxtnum },
> +
>          /* TODO: ARMv8.2-SPE -- PMSCR_EL2 */
>          /* TODO: ARMv8.4-Trace -- TRFCR_EL2 */
>      };
> @@ -7434,7 +7444,61 @@ static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = {
>      REGINFO_SENTINEL
>  };
>
> -#endif
> +static CPAccessResult access_scxtnum(CPUARMState *env, const ARMCPRegInfo *ri,
> +                                     bool isread)
> +{
> +    uint64_t hcr;
> +
> +    switch (arm_current_el(env)) {
> +    case 0:
> +        hcr = arm_hcr_el2_eff(env);
> +        if ((hcr & (HCR_TGE | HCR_E2H)) != (HCR_TGE | HCR_E2H)) {
> +            if (env->cp15.sctlr_el[1] & SCTLR_TSCXT) {
> +                if (hcr & HCR_TGE) {
> +                    return CP_ACCESS_TRAP_EL2;
> +                }
> +                return CP_ACCESS_TRAP;
> +            }
> +            if (arm_is_el2_enabled(env) && !(hcr & HCR_ENSCXT)) {
> +                return CP_ACCESS_TRAP_EL2;
> +            }
> +        } else {
> +            QEMU_FALLTHROUGH;
> +    case 1:

I'm sure this is valid C, but firm 'no' to interleaving
case labels and if/else statements.

> +            if (env->cp15.sctlr_el[2] & SCTLR_TSCXT) {
> +                return CP_ACCESS_TRAP_EL2;
> +            }
> +        }
> +        QEMU_FALLTHROUGH;
> +    case 2:
> +        if (arm_feature(env, ARM_FEATURE_EL3)
> +            && !(env->cp15.scr_el3 & SCR_ENSCXT)) {
> +            return CP_ACCESS_TRAP_EL3;
> +        }
> +    }
> +    return CP_ACCESS_OK;
> +}
> +
> +static const ARMCPRegInfo scxtnum_reginfo[] = {
> +    { .name = "SCXTNUM_EL0", .state = ARM_CP_STATE_AA64,
> +      .opc0 = 3, .opc1 = 3, .crn = 0xd, .crm = 0, .opc2 = 7,
> +      .access = PL0_RW, .accessfn = access_scxtnum,
> +      .fieldoffset = offsetof(CPUARMState, scxtnum_el[0]) },
> +    { .name = "SCXTNUM_EL1", .state = ARM_CP_STATE_AA64,
> +      .opc0 = 3, .opc1 = 0, .crn = 0xd, .crm = 0, .opc2 = 7,
> +      .access = PL1_RW, .accessfn = access_scxtnum,
> +      .fieldoffset = offsetof(CPUARMState, scxtnum_el[1]) },
> +    { .name = "SCXTNUM_EL2", .state = ARM_CP_STATE_AA64,
> +      .opc0 = 3, .opc1 = 4, .crn = 0xd, .crm = 0, .opc2 = 7,
> +      .access = PL2_RW, .accessfn = access_scxtnum,
> +      .fieldoffset = offsetof(CPUARMState, scxtnum_el[2]) },
> +    { .name = "SCXTNUM_EL3", .state = ARM_CP_STATE_AA64,
> +      .opc0 = 3, .opc1 = 6, .crn = 0xd, .crm = 0, .opc2 = 7,
> +      .access = PL3_RW,
> +      .fieldoffset = offsetof(CPUARMState, scxtnum_el[3]) },
> +    REGINFO_SENTINEL

Decimal not hex here too, please.

thanks
-- PMM

Re: [PATCH 6/7] target/arm: Define cortex-a76

[Peter Maydell]

On Sun, 10 Apr 2022 at 07:13, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> Enable the a76 for virt and sbsa board use.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
>  hw/arm/sbsa-ref.c  |  1 +
>  hw/arm/virt.c      |  1 +
>  target/arm/cpu64.c | 64 ++++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 66 insertions(+)
>
> diff --git a/hw/arm/sbsa-ref.c b/hw/arm/sbsa-ref.c
> index 2387401963..2ddde88f5e 100644
> --- a/hw/arm/sbsa-ref.c
> +++ b/hw/arm/sbsa-ref.c
> @@ -145,6 +145,7 @@ static const int sbsa_ref_irqmap[] = {
>  static const char * const valid_cpus[] = {
>      ARM_CPU_TYPE_NAME("cortex-a57"),
>      ARM_CPU_TYPE_NAME("cortex-a72"),
> +    ARM_CPU_TYPE_NAME("cortex-a76"),
>      ARM_CPU_TYPE_NAME("max"),
>  };
>
> diff --git a/hw/arm/virt.c b/hw/arm/virt.c
> index d2e5ecd234..ce15c36a7f 100644
> --- a/hw/arm/virt.c
> +++ b/hw/arm/virt.c
> @@ -202,6 +202,7 @@ static const char *valid_cpus[] = {
>      ARM_CPU_TYPE_NAME("cortex-a53"),
>      ARM_CPU_TYPE_NAME("cortex-a57"),
>      ARM_CPU_TYPE_NAME("cortex-a72"),
> +    ARM_CPU_TYPE_NAME("cortex-a76"),
>      ARM_CPU_TYPE_NAME("a64fx"),
>      ARM_CPU_TYPE_NAME("host"),
>      ARM_CPU_TYPE_NAME("max"),

Need to update the list in docs/system/arm/virt.rst as well.

> diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
> index 199ca437a0..a00c02be2c 100644
> --- a/target/arm/cpu64.c
> +++ b/target/arm/cpu64.c
> @@ -251,6 +251,69 @@ static void aarch64_a72_initfn(Object *obj)
>      define_arm_cp_regs(cpu, cortex_a72_a57_a53_cp_reginfo);
>  }
>
> +static void aarch64_a76_initfn(Object *obj)
> +{
> +    ARMCPU *cpu = ARM_CPU(obj);
> +
> +    cpu->dtb_compatible = "arm,cortex-a76";
> +    set_feature(&cpu->env, ARM_FEATURE_V8);
> +    set_feature(&cpu->env, ARM_FEATURE_NEON);
> +    set_feature(&cpu->env, ARM_FEATURE_GENERIC_TIMER);
> +    set_feature(&cpu->env, ARM_FEATURE_AARCH64);
> +    set_feature(&cpu->env, ARM_FEATURE_CBAR_RO);
> +    set_feature(&cpu->env, ARM_FEATURE_EL2);
> +    set_feature(&cpu->env, ARM_FEATURE_EL3);
> +    set_feature(&cpu->env, ARM_FEATURE_PMU);
> +
> +    /* Ordered by B2.4 AArch64 registers by functional group */
> +    cpu->clidr = 0x82000023;
> +    cpu->ctr = 0x8444C004;
> +    cpu->dcz_blocksize = 4;
> +    cpu->isar.id_aa64dfr0  = 0x0000000010305408ull;
> +    cpu->isar.id_aa64isar0 = 0x0000100010211120ull;
> +    cpu->isar.id_aa64isar1 = 0x0000000000100001ull;
> +    cpu->isar.id_aa64mmfr0 = 0x0000000000101122ull;
> +    cpu->isar.id_aa64mmfr1 = 0x0000000010212122ull;
> +    cpu->isar.id_aa64mmfr2 = 0x0000000000001011ull;
> +    cpu->isar.id_aa64pfr0  = 0x1100000010111112ull;

This has the GIC field clear. On the one hand this is true
also of all our other CPU implementations. On the other hand
if we wire up a GICv3 in the board code then we will be
presenting the GIC CPU interface registers, which is what this
ID register field is supposed to be telling software.
I guess we should handle this by having gicv3_init_cpuif()
arrange for the ID field to be set. Which is mildly painful,
because that will mean the value for the cpreg isn't constant.

(In real hardware the cpuif really is part of the CPU proper,
so it (ie the sysregs for the GICv3) exists whether the
GICv3 proper is present somewhere outside the CPU in the
SoC. But our implementation doesn't put the dividing line
quite where the h/w does.)

> +    cpu->isar.id_aa64pfr1  = 0x0000000000000010ull;
> +    cpu->isar.id_dfr0  = 0x04010088;
> +    cpu->isar.id_isar0 = 0x02101110;
> +    cpu->isar.id_isar1 = 0x13112111;
> +    cpu->isar.id_isar2 = 0x21232042;
> +    cpu->isar.id_isar3 = 0x01112131;
> +    cpu->isar.id_isar4 = 0x00010142;
> +    cpu->isar.id_isar5 = 0x01011121;
> +    cpu->isar.id_isar6 = 0x00000010;
> +    cpu->isar.id_mmfr0 = 0x10201105;
> +    cpu->isar.id_mmfr1 = 0x40000000;
> +    cpu->isar.id_mmfr2 = 0x01260000;
> +    cpu->isar.id_mmfr3 = 0x02122211;
> +    cpu->isar.id_mmfr4 = 0x00021110;
> +    cpu->isar.id_pfr0  = 0x10010131;
> +    cpu->isar.id_pfr1  = 0x10010000;

Here you have set the GIC field. We should at least be consistent :-)

> +    cpu->isar.id_pfr2  = 0x00000011;
> +    cpu->midr = 0x413fd0b1;

This is r3p1; the r4p1 TRM is available so we might as
well claim to be that.

> +
> +    /* From B2.18 CCSIDR_EL1 */
> +    cpu->ccsidr[0] = 0x701fe01a; /* 64KB L1 dcache */
> +    cpu->ccsidr[1] = 0x201fe01a; /* 64KB L1 icache */
> +    cpu->ccsidr[2] = 0x707fe03a; /* 512KB L2 cache */
> +
> +    /* From B2.93 SCTLR_EL3 */
> +    cpu->reset_sctlr = 0x30c50838;
> +
> +    /* From B4.23 ICH_VTR_EL2 */
> +    cpu->gic_num_lrs = 4;
> +    cpu->gic_vpribits = 5;
> +    cpu->gic_vprebits = 5;

(side note, we should really support having the non-virtual
pribits/prebits for the GIC be configurable too. I have a
half-written set of patches for that.)

> +
> +    /* From B5.1 AdvSIMD AArch64 register summary */
> +    cpu->isar.mvfr0 = 0x10110222;
> +    cpu->isar.mvfr1 = 0x13211111;
> +    cpu->isar.mvfr2 = 0x00000043;

Missing (just cross-checking what other CPUs are setting):
  cpu->revidr
  cpu->reset_fpsid
  cpu->id_afr0
  cpu->isar.dbgdidr

> +}
> +
>  void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp)
>  {
>      /*
> @@ -935,6 +998,7 @@ static const ARMCPUInfo aarch64_cpus[] = {
>      { .name = "cortex-a57",         .initfn = aarch64_a57_initfn },
>      { .name = "cortex-a53",         .initfn = aarch64_a53_initfn },
>      { .name = "cortex-a72",         .initfn = aarch64_a72_initfn },
> +    { .name = "cortex-a76",         .initfn = aarch64_a76_initfn },
>      { .name = "a64fx",              .initfn = aarch64_a64fx_initfn },
>      { .name = "max",                .initfn = aarch64_max_initfn },
>  #if defined(CONFIG_KVM) || defined(CONFIG_HVF)
> --

thanks
-- PMM

Re: [PATCH 7/7] target/arm: Define neoverse-n1

[Peter Maydell]

On Sun, 10 Apr 2022 at 07:15, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> Enable the n1 for virt and sbsa board use.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---

Comments on previous patch apply here too, by and large.

thanks
-- PMM

Re: [PATCH 6/7] target/arm: Define cortex-a76

[Richard Henderson]

On 4/11/22 11:09, Peter Maydell wrote:
>> +    cpu->isar.id_aa64pfr0  = 0x1100000010111112ull;
> 
> This has the GIC field clear. On the one hand this is true
> also of all our other CPU implementations. On the other hand
> if we wire up a GICv3 in the board code then we will be
> presenting the GIC CPU interface registers, which is what this
> ID register field is supposed to be telling software.
> I guess we should handle this by having gicv3_init_cpuif()
> arrange for the ID field to be set. Which is mildly painful,
> because that will mean the value for the cpreg isn't constant.

We already handle this via id_aa64pfr0_read().

>> +    cpu->isar.id_pfr1  = 0x10010000;
> 
> Here you have set the GIC field. We should at least be consistent :-)

Oops.  :-)  We have id_pfr1_read() for that as well.

>> +    cpu->midr = 0x413fd0b1;
> 
> This is r3p1; the r4p1 TRM is available so we might as
> well claim to be that.

Ok.

> Missing (just cross-checking what other CPUs are setting):
>    cpu->revidr
>    cpu->reset_fpsid
>    cpu->id_afr0

revidr and id_afr0 are 0 -- explicit zero required?  We certainly assume we begin with 
zero for other registers for other cpus...

FPSID is deprecated, and not mentioned in the a76 or n1 specs.  Presumably this is because 
it is not accessible at all: FPSID is not accessible from EL0 for VFPv3, and the A76 does 
not support aa32 at other than EL0.

>    cpu->isar.dbgdidr

"If EL1 cannot use AArch32 then the implementation of this register is OPTIONAL and 
deprecated."


r~

Re: [PATCH 6/7] target/arm: Define cortex-a76

[Richard Henderson]

On 4/11/22 11:30, Richard Henderson wrote:
>>    cpu->isar.dbgdidr
> 
> "If EL1 cannot use AArch32 then the implementation of this register is OPTIONAL and 
> deprecated."

Which we already handle in define_debug_regs by not defining DBGDIDR if this value is 0.


r~