From b0be0d28cc75abbf811481db5457955c4c6e74ad Mon Sep 17 00:00:00 2001
From: LIU Zhiwei <zhiwei_liu@c-sky.com>
Date: Fri, 26 Feb 2021 15:20:18 -0500
Subject: [PATCH] softfloat: Define operations for bfloat16

Backports 8282310d8535cc2a8431c516e907da79f92df6eb
---
 qemu/aarch64.h                     |  18 +++
 qemu/aarch64eb.h                   |  18 +++
 qemu/arm.h                         |  18 +++
 qemu/armeb.h                       |  18 +++
 qemu/fpu/softfloat.c               | 169 +++++++++++++++++++++++++++++
 qemu/header_gen.py                 |  18 +++
 qemu/include/fpu/softfloat-types.h |   5 +
 qemu/include/fpu/softfloat.h       |  38 +++++++
 qemu/m68k.h                        |  18 +++
 qemu/mips.h                        |  18 +++
 qemu/mips64.h                      |  18 +++
 qemu/mips64el.h                    |  18 +++
 qemu/mipsel.h                      |  18 +++
 qemu/powerpc.h                     |  18 +++
 qemu/riscv32.h                     |  18 +++
 qemu/riscv64.h                     |  18 +++
 qemu/sparc.h                       |  18 +++
 qemu/sparc64.h                     |  18 +++
 qemu/x86_64.h                      |  18 +++
 19 files changed, 500 insertions(+)

diff --git a/qemu/aarch64.h b/qemu/aarch64.h
index f70a0262..95a1232f 100644
--- a/qemu/aarch64.h
+++ b/qemu/aarch64.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_aarch64
 #define bad_mode_switch bad_mode_switch_aarch64
 #define bank_number bank_number_aarch64
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_aarch64
+#define bfloat16_add bfloat16_add_aarch64
+#define bfloat16_sub bfloat16_sub_aarch64
+#define bfloat16_mul bfloat16_mul_aarch64
+#define bfloat16_div bfloat16_div_aarch64
+#define bfloat16_muladd bfloat16_muladd_aarch64
+#define bfloat16_scalbn bfloat16_scalbn_aarch64
+#define bfloat16_min bfloat16_min_aarch64
+#define bfloat16_max bfloat16_max_aarch64
+#define bfloat16_minnum bfloat16_minnum_aarch64
+#define bfloat16_maxnum bfloat16_maxnum_aarch64
+#define bfloat16_minnummag bfloat16_minnummag_aarch64
+#define bfloat16_maxnummag bfloat16_maxnummag_aarch64
+#define bfloat16_sqrt bfloat16_sqrt_aarch64
+#define bfloat16_compare bfloat16_compare_aarch64
+#define bfloat16_compare_quiet bfloat16_compare_quiet_aarch64
+#define bfloat16_silence_nan bfloat16_silence_nan_aarch64
+#define bfloat16_default_nan bfloat16_default_nan_aarch64
 #define bitmap_zero_extend bitmap_zero_extend_aarch64
 #define bp_wp_matches bp_wp_matches_aarch64
 #define breakpoint_invalidate breakpoint_invalidate_aarch64
diff --git a/qemu/aarch64eb.h b/qemu/aarch64eb.h
index 5ba8fb7a..3101aa6b 100644
--- a/qemu/aarch64eb.h
+++ b/qemu/aarch64eb.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_aarch64eb
 #define bad_mode_switch bad_mode_switch_aarch64eb
 #define bank_number bank_number_aarch64eb
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_aarch64eb
+#define bfloat16_add bfloat16_add_aarch64eb
+#define bfloat16_sub bfloat16_sub_aarch64eb
+#define bfloat16_mul bfloat16_mul_aarch64eb
+#define bfloat16_div bfloat16_div_aarch64eb
+#define bfloat16_muladd bfloat16_muladd_aarch64eb
+#define bfloat16_scalbn bfloat16_scalbn_aarch64eb
+#define bfloat16_min bfloat16_min_aarch64eb
+#define bfloat16_max bfloat16_max_aarch64eb
+#define bfloat16_minnum bfloat16_minnum_aarch64eb
+#define bfloat16_maxnum bfloat16_maxnum_aarch64eb
+#define bfloat16_minnummag bfloat16_minnummag_aarch64eb
+#define bfloat16_maxnummag bfloat16_maxnummag_aarch64eb
+#define bfloat16_sqrt bfloat16_sqrt_aarch64eb
+#define bfloat16_compare bfloat16_compare_aarch64eb
+#define bfloat16_compare_quiet bfloat16_compare_quiet_aarch64eb
+#define bfloat16_silence_nan bfloat16_silence_nan_aarch64eb
+#define bfloat16_default_nan bfloat16_default_nan_aarch64eb
 #define bitmap_zero_extend bitmap_zero_extend_aarch64eb
 #define bp_wp_matches bp_wp_matches_aarch64eb
 #define breakpoint_invalidate breakpoint_invalidate_aarch64eb
diff --git a/qemu/arm.h b/qemu/arm.h
index e5863596..b9657fcf 100644
--- a/qemu/arm.h
+++ b/qemu/arm.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_arm
 #define bad_mode_switch bad_mode_switch_arm
 #define bank_number bank_number_arm
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_arm
+#define bfloat16_add bfloat16_add_arm
+#define bfloat16_sub bfloat16_sub_arm
+#define bfloat16_mul bfloat16_mul_arm
+#define bfloat16_div bfloat16_div_arm
+#define bfloat16_muladd bfloat16_muladd_arm
+#define bfloat16_scalbn bfloat16_scalbn_arm
+#define bfloat16_min bfloat16_min_arm
+#define bfloat16_max bfloat16_max_arm
+#define bfloat16_minnum bfloat16_minnum_arm
+#define bfloat16_maxnum bfloat16_maxnum_arm
+#define bfloat16_minnummag bfloat16_minnummag_arm
+#define bfloat16_maxnummag bfloat16_maxnummag_arm
+#define bfloat16_sqrt bfloat16_sqrt_arm
+#define bfloat16_compare bfloat16_compare_arm
+#define bfloat16_compare_quiet bfloat16_compare_quiet_arm
+#define bfloat16_silence_nan bfloat16_silence_nan_arm
+#define bfloat16_default_nan bfloat16_default_nan_arm
 #define bitmap_zero_extend bitmap_zero_extend_arm
 #define bp_wp_matches bp_wp_matches_arm
 #define breakpoint_invalidate breakpoint_invalidate_arm
diff --git a/qemu/armeb.h b/qemu/armeb.h
index de9b6584..a64464bd 100644
--- a/qemu/armeb.h
+++ b/qemu/armeb.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_armeb
 #define bad_mode_switch bad_mode_switch_armeb
 #define bank_number bank_number_armeb
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_armeb
+#define bfloat16_add bfloat16_add_armeb
+#define bfloat16_sub bfloat16_sub_armeb
+#define bfloat16_mul bfloat16_mul_armeb
+#define bfloat16_div bfloat16_div_armeb
+#define bfloat16_muladd bfloat16_muladd_armeb
+#define bfloat16_scalbn bfloat16_scalbn_armeb
+#define bfloat16_min bfloat16_min_armeb
+#define bfloat16_max bfloat16_max_armeb
+#define bfloat16_minnum bfloat16_minnum_armeb
+#define bfloat16_maxnum bfloat16_maxnum_armeb
+#define bfloat16_minnummag bfloat16_minnummag_armeb
+#define bfloat16_maxnummag bfloat16_maxnummag_armeb
+#define bfloat16_sqrt bfloat16_sqrt_armeb
+#define bfloat16_compare bfloat16_compare_armeb
+#define bfloat16_compare_quiet bfloat16_compare_quiet_armeb
+#define bfloat16_silence_nan bfloat16_silence_nan_armeb
+#define bfloat16_default_nan bfloat16_default_nan_armeb
 #define bitmap_zero_extend bitmap_zero_extend_armeb
 #define bp_wp_matches bp_wp_matches_armeb
 #define breakpoint_invalidate breakpoint_invalidate_armeb
diff --git a/qemu/fpu/softfloat.c b/qemu/fpu/softfloat.c
index c2d6f7ed..896d81a4 100644
--- a/qemu/fpu/softfloat.c
+++ b/qemu/fpu/softfloat.c
@@ -555,6 +555,10 @@ static const FloatFmt float16_params_ahp = {
     .arm_althp = true
 };
 
+static const FloatFmt bfloat16_params = {
+    FLOAT_PARAMS(8, 7)
+};
+
 static const FloatFmt float32_params = {
     FLOAT_PARAMS(8, 23)
 };
@@ -581,6 +585,11 @@ static inline FloatParts float16_unpack_raw(float16 f)
     return unpack_raw(float16_params, f);
 }
 
+static inline FloatParts bfloat16_unpack_raw(bfloat16 f)
+{
+    return unpack_raw(bfloat16_params, f);
+}
+
 static inline FloatParts float32_unpack_raw(float32 f)
 {
     return unpack_raw(float32_params, f);
@@ -604,6 +613,11 @@ static inline float16 float16_pack_raw(FloatParts p)
     return make_float16(pack_raw(float16_params, p));
 }
 
+static inline bfloat16 bfloat16_pack_raw(FloatParts p)
+{
+    return pack_raw(bfloat16_params, p);
+}
+
 static inline float32 float32_pack_raw(FloatParts p)
 {
     return make_float32(pack_raw(float32_params, p));
@@ -821,6 +835,11 @@ static FloatParts float16_unpack_canonical(float16 f, float_status *s)
     return float16a_unpack_canonical(f, s, &float16_params);
 }
 
+static FloatParts bfloat16_unpack_canonical(bfloat16 f, float_status *s)
+{
+    return sf_canonicalize(bfloat16_unpack_raw(f), &bfloat16_params, s);
+}
+
 static float16 float16a_round_pack_canonical(FloatParts p, float_status *s,
                                              const FloatFmt *params)
 {
@@ -832,6 +851,11 @@ static float16 float16_round_pack_canonical(FloatParts p, float_status *s)
     return float16a_round_pack_canonical(p, s, &float16_params);
 }
 
+static bfloat16 bfloat16_round_pack_canonical(FloatParts p, float_status *s)
+{
+    return bfloat16_pack_raw(round_canonical(p, s, &bfloat16_params));
+}
+
 static FloatParts float32_unpack_canonical(float32 f, float_status *s)
 {
     return sf_canonicalize(float32_unpack_raw(f), &float32_params, s);
@@ -1159,6 +1183,28 @@ float64_sub(float64 a, float64 b, float_status *s)
     return float64_addsub(a, b, s, hard_f64_sub, soft_f64_sub);
 }
 
+/*
+ * Returns the result of adding or subtracting the bfloat16
+ * values `a' and `b'.
+ */
+bfloat16 QEMU_FLATTEN bfloat16_add(bfloat16 a, bfloat16 b, float_status *status)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, status);
+    FloatParts pb = bfloat16_unpack_canonical(b, status);
+    FloatParts pr = addsub_floats(pa, pb, false, status);
+
+    return bfloat16_round_pack_canonical(pr, status);
+}
+
+bfloat16 QEMU_FLATTEN bfloat16_sub(bfloat16 a, bfloat16 b, float_status *status)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, status);
+    FloatParts pb = bfloat16_unpack_canonical(b, status);
+    FloatParts pr = addsub_floats(pa, pb, true, status);
+
+    return bfloat16_round_pack_canonical(pr, status);
+}
+
 /*
  * Returns the result of multiplying the floating-point values `a' and
  * `b'. The operation is performed according to the IEC/IEEE Standard
@@ -1261,6 +1307,20 @@ float64_mul(float64 a, float64 b, float_status *s)
                         f64_is_zon2, f64_addsubmul_post);
 }
 
+/*
+ * Returns the result of multiplying the bfloat16
+ * values `a' and `b'.
+ */
+
+bfloat16 QEMU_FLATTEN bfloat16_mul(bfloat16 a, bfloat16 b, float_status *status)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, status);
+    FloatParts pb = bfloat16_unpack_canonical(b, status);
+    FloatParts pr = mul_floats(pa, pb, status);
+
+    return bfloat16_round_pack_canonical(pr, status);
+}
+
 /*
  * Returns the result of multiplying the floating-point values `a' and
  * `b' then adding 'c', with no intermediate rounding step after the
@@ -1621,6 +1681,23 @@ float64_muladd(float64 xa, float64 xb, float64 xc, int flags, float_status *s)
     return soft_f64_muladd(ua.s, ub.s, uc.s, flags, s);
 }
 
+/*
+ * Returns the result of multiplying the bfloat16 values `a'
+ * and `b' then adding 'c', with no intermediate rounding step after the
+ * multiplication.
+ */
+
+bfloat16 QEMU_FLATTEN bfloat16_muladd(bfloat16 a, bfloat16 b, bfloat16 c,
+                                      int flags, float_status *status)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, status);
+    FloatParts pb = bfloat16_unpack_canonical(b, status);
+    FloatParts pc = bfloat16_unpack_canonical(c, status);
+    FloatParts pr = muladd_floats(pa, pb, pc, flags, status);
+
+    return bfloat16_round_pack_canonical(pr, status);
+}
+
 /*
  * Returns the result of dividing the floating-point value `a' by the
  * corresponding value `b'. The operation is performed according to
@@ -1787,6 +1864,20 @@ float64_div(float64 a, float64 b, float_status *s)
                         f64_div_pre, f64_div_post);
 }
 
+/*
+ * Returns the result of dividing the bfloat16
+ * value `a' by the corresponding value `b'.
+ */
+
+bfloat16 bfloat16_div(bfloat16 a, bfloat16 b, float_status *status)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, status);
+    FloatParts pb = bfloat16_unpack_canonical(b, status);
+    FloatParts pr = div_floats(pa, pb, status);
+
+    return bfloat16_round_pack_canonical(pr, status);
+}
+
 /*
  * Float to Float conversions
  *
@@ -2843,6 +2934,25 @@ MINMAX(64, maxnummag, false, true, true)
 
 #undef MINMAX
 
+#define BF16_MINMAX(name, ismin, isiee, ismag)                          \
+bfloat16 bfloat16_ ## name(bfloat16 a, bfloat16 b, float_status *s)     \
+{                                                                       \
+    FloatParts pa = bfloat16_unpack_canonical(a, s);                    \
+    FloatParts pb = bfloat16_unpack_canonical(b, s);                    \
+    FloatParts pr = minmax_floats(pa, pb, ismin, isiee, ismag, s);      \
+                                                                        \
+    return bfloat16_round_pack_canonical(pr, s);                        \
+}
+
+BF16_MINMAX(min, true, false, false)
+BF16_MINMAX(minnum, true, true, false)
+BF16_MINMAX(minnummag, true, true, true)
+BF16_MINMAX(max, false, false, false)
+BF16_MINMAX(maxnum, false, true, false)
+BF16_MINMAX(maxnummag, false, true, true)
+
+#undef BF16_MINMAX
+
 /* Floating point compare */
 static FloatRelation compare_floats(FloatParts a, FloatParts b, bool is_quiet,
                                     float_status *s)
@@ -3004,6 +3114,24 @@ FloatRelation float64_compare_quiet(float64 a, float64 b, float_status *s)
     return f64_compare(a, b, true, s);
 }
 
+static FloatRelation QEMU_FLATTEN
+soft_bf16_compare(bfloat16 a, bfloat16 b, bool is_quiet, float_status *s)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, s);
+    FloatParts pb = bfloat16_unpack_canonical(b, s);
+    return compare_floats(pa, pb, is_quiet, s);
+}
+
+FloatRelation bfloat16_compare(bfloat16 a, bfloat16 b, float_status *s)
+{
+    return soft_bf16_compare(a, b, false, s);
+}
+
+FloatRelation bfloat16_compare_quiet(bfloat16 a, bfloat16 b, float_status *s)
+{
+    return soft_bf16_compare(a, b, true, s);
+}
+
 /* Multiply A by 2 raised to the power N.  */
 static FloatParts scalbn_decomposed(FloatParts a, int n, float_status *s)
 {
@@ -3043,6 +3171,13 @@ float64 float64_scalbn(float64 a, int n, float_status *status)
     return float64_round_pack_canonical(pr, status);
 }
 
+bfloat16 bfloat16_scalbn(bfloat16 a, int n, float_status *status)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, status);
+    FloatParts pr = scalbn_decomposed(pa, n, status);
+    return bfloat16_round_pack_canonical(pr, status);
+}
+
 /*
  * Square Root
  *
@@ -3193,6 +3328,13 @@ float64 QEMU_FLATTEN float64_sqrt(float64 xa, float_status *s)
     return soft_f64_sqrt(ua.s, s);
 }
 
+bfloat16 QEMU_FLATTEN bfloat16_sqrt(bfloat16 a, float_status *status)
+{
+    FloatParts pa = bfloat16_unpack_canonical(a, status);
+    FloatParts pr = sqrt_float(pa, status, &bfloat16_params);
+    return bfloat16_round_pack_canonical(pr, status);
+}
+
 /*----------------------------------------------------------------------------
 | The pattern for a default generated NaN.
 *----------------------------------------------------------------------------*/
@@ -3235,6 +3377,13 @@ float128 float128_default_nan(float_status *status)
     return r;
 }
 
+bfloat16 bfloat16_default_nan(float_status *status)
+{
+    FloatParts p = parts_default_nan(status);
+    p.frac >>= bfloat16_params.frac_shift;
+    return bfloat16_pack_raw(p);
+}
+
 /*----------------------------------------------------------------------------
 | Returns a quiet NaN from a signalling NaN for the floating point value `a'.
 *----------------------------------------------------------------------------*/
@@ -3266,6 +3415,15 @@ float64 float64_silence_nan(float64 a, float_status *status)
     return float64_pack_raw(p);
 }
 
+bfloat16 bfloat16_silence_nan(bfloat16 a, float_status *status)
+{
+    FloatParts p = bfloat16_unpack_raw(a);
+    p.frac <<= bfloat16_params.frac_shift;
+    p = parts_silence_nan(p, status);
+    p.frac >>= bfloat16_params.frac_shift;
+    return bfloat16_pack_raw(p);
+}
+
 /*----------------------------------------------------------------------------
 | If `a' is denormal and we are in flush-to-zero mode then set the
 | input-denormal exception and return zero. Otherwise just return the value.
@@ -3314,6 +3472,17 @@ float64 float64_squash_input_denormal(float64 a, float_status *status)
     return a;
 }
 
+bfloat16 bfloat16_squash_input_denormal(bfloat16 a, float_status *status)
+{
+    if (status->flush_inputs_to_zero) {
+        FloatParts p = bfloat16_unpack_raw(a);
+        if (parts_squash_denormal(p, status)) {
+            return bfloat16_set_sign(bfloat16_zero, p.sign);
+        }
+    }
+    return a;
+}
+
 /*----------------------------------------------------------------------------
 | Takes a 64-bit fixed-point value `absZ' with binary point between bits 6
 | and 7, and returns the properly rounded 32-bit integer corresponding to the
diff --git a/qemu/header_gen.py b/qemu/header_gen.py
index fb49bfcb..83bd433d 100644
--- a/qemu/header_gen.py
+++ b/qemu/header_gen.py
@@ -203,6 +203,24 @@ symbols = (
     'ats_write',
     'bad_mode_switch',
     'bank_number',
+    'bfloat16_squash_input_denormal',
+    'bfloat16_add',
+    'bfloat16_sub',
+    'bfloat16_mul',
+    'bfloat16_div',
+    'bfloat16_muladd',
+    'bfloat16_scalbn',
+    'bfloat16_min',
+    'bfloat16_max',
+    'bfloat16_minnum',
+    'bfloat16_maxnum',
+    'bfloat16_minnummag',
+    'bfloat16_maxnummag',
+    'bfloat16_sqrt',
+    'bfloat16_compare',
+    'bfloat16_compare_quiet',
+    'bfloat16_silence_nan',
+    'bfloat16_default_nan',
     'bitmap_zero_extend',
     'bp_wp_matches',
     'breakpoint_invalidate',
diff --git a/qemu/include/fpu/softfloat-types.h b/qemu/include/fpu/softfloat-types.h
index c7ddcab8..8a3f20fa 100644
--- a/qemu/include/fpu/softfloat-types.h
+++ b/qemu/include/fpu/softfloat-types.h
@@ -112,6 +112,11 @@ typedef struct {
 #define make_float128(high_, low_) ((float128) { .high = high_, .low = low_ })
 #define make_float128_init(high_, low_) { .high = high_, .low = low_ }
 
+/*
+ * Software neural-network floating-point types.
+ */
+typedef uint16_t bfloat16;
+
 /*
  * Software IEC/IEEE floating-point underflow tininess-detection mode.
  */
diff --git a/qemu/include/fpu/softfloat.h b/qemu/include/fpu/softfloat.h
index 0b27ad93..06abba6b 100644
--- a/qemu/include/fpu/softfloat.h
+++ b/qemu/include/fpu/softfloat.h
@@ -112,6 +112,7 @@ void float_raise(uint8_t flags, float_status *status);
 float16 float16_squash_input_denormal(float16 a, float_status *status);
 float32 float32_squash_input_denormal(float32 a, float_status *status);
 float64 float64_squash_input_denormal(float64 a, float_status *status);
+bfloat16 bfloat16_squash_input_denormal(bfloat16 a, float_status *status);
 
 /*----------------------------------------------------------------------------
 | Options to indicate which negations to perform in float*_muladd()
@@ -350,6 +351,43 @@ static inline bool float16_unordered_quiet(float16 a, float16 b,
 #define float16_three make_float16(0x4200)
 #define float16_infinity make_float16(0x7c00)
 
+/*----------------------------------------------------------------------------
+| Software bfloat16 operations.
+*----------------------------------------------------------------------------*/
+
+bfloat16 bfloat16_add(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_sub(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_mul(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_div(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_muladd(bfloat16, bfloat16, bfloat16, int,
+                         float_status *status);
+float16 bfloat16_scalbn(bfloat16, int, float_status *status);
+bfloat16 bfloat16_min(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_max(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_minnum(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_maxnum(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_minnummag(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_maxnummag(bfloat16, bfloat16, float_status *status);
+bfloat16 bfloat16_sqrt(bfloat16, float_status *status);
+FloatRelation bfloat16_compare(bfloat16, bfloat16, float_status *status);
+FloatRelation bfloat16_compare_quiet(bfloat16, bfloat16, float_status *status);
+
+bfloat16 bfloat16_silence_nan(bfloat16, float_status *status);
+bfloat16 bfloat16_default_nan(float_status *status);
+
+static inline bfloat16 bfloat16_set_sign(bfloat16 a, int sign)
+{
+    return (a & 0x7fff) | (sign << 15);
+}
+
+#define bfloat16_zero 0
+#define bfloat16_half 0x3f00
+#define bfloat16_one 0x3f80
+#define bfloat16_one_point_five 0x3fc0
+#define bfloat16_two 0x4000
+#define bfloat16_three 0x4040
+#define bfloat16_infinity 0x7f80
+
 /*----------------------------------------------------------------------------
 | The pattern for a default generated half-precision NaN.
 *----------------------------------------------------------------------------*/
diff --git a/qemu/m68k.h b/qemu/m68k.h
index 27f65aa5..8b5a0f17 100644
--- a/qemu/m68k.h
+++ b/qemu/m68k.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_m68k
 #define bad_mode_switch bad_mode_switch_m68k
 #define bank_number bank_number_m68k
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_m68k
+#define bfloat16_add bfloat16_add_m68k
+#define bfloat16_sub bfloat16_sub_m68k
+#define bfloat16_mul bfloat16_mul_m68k
+#define bfloat16_div bfloat16_div_m68k
+#define bfloat16_muladd bfloat16_muladd_m68k
+#define bfloat16_scalbn bfloat16_scalbn_m68k
+#define bfloat16_min bfloat16_min_m68k
+#define bfloat16_max bfloat16_max_m68k
+#define bfloat16_minnum bfloat16_minnum_m68k
+#define bfloat16_maxnum bfloat16_maxnum_m68k
+#define bfloat16_minnummag bfloat16_minnummag_m68k
+#define bfloat16_maxnummag bfloat16_maxnummag_m68k
+#define bfloat16_sqrt bfloat16_sqrt_m68k
+#define bfloat16_compare bfloat16_compare_m68k
+#define bfloat16_compare_quiet bfloat16_compare_quiet_m68k
+#define bfloat16_silence_nan bfloat16_silence_nan_m68k
+#define bfloat16_default_nan bfloat16_default_nan_m68k
 #define bitmap_zero_extend bitmap_zero_extend_m68k
 #define bp_wp_matches bp_wp_matches_m68k
 #define breakpoint_invalidate breakpoint_invalidate_m68k
diff --git a/qemu/mips.h b/qemu/mips.h
index ea7ddd7e..fa503cbf 100644
--- a/qemu/mips.h
+++ b/qemu/mips.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_mips
 #define bad_mode_switch bad_mode_switch_mips
 #define bank_number bank_number_mips
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_mips
+#define bfloat16_add bfloat16_add_mips
+#define bfloat16_sub bfloat16_sub_mips
+#define bfloat16_mul bfloat16_mul_mips
+#define bfloat16_div bfloat16_div_mips
+#define bfloat16_muladd bfloat16_muladd_mips
+#define bfloat16_scalbn bfloat16_scalbn_mips
+#define bfloat16_min bfloat16_min_mips
+#define bfloat16_max bfloat16_max_mips
+#define bfloat16_minnum bfloat16_minnum_mips
+#define bfloat16_maxnum bfloat16_maxnum_mips
+#define bfloat16_minnummag bfloat16_minnummag_mips
+#define bfloat16_maxnummag bfloat16_maxnummag_mips
+#define bfloat16_sqrt bfloat16_sqrt_mips
+#define bfloat16_compare bfloat16_compare_mips
+#define bfloat16_compare_quiet bfloat16_compare_quiet_mips
+#define bfloat16_silence_nan bfloat16_silence_nan_mips
+#define bfloat16_default_nan bfloat16_default_nan_mips
 #define bitmap_zero_extend bitmap_zero_extend_mips
 #define bp_wp_matches bp_wp_matches_mips
 #define breakpoint_invalidate breakpoint_invalidate_mips
diff --git a/qemu/mips64.h b/qemu/mips64.h
index f714ce35..30290e80 100644
--- a/qemu/mips64.h
+++ b/qemu/mips64.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_mips64
 #define bad_mode_switch bad_mode_switch_mips64
 #define bank_number bank_number_mips64
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_mips64
+#define bfloat16_add bfloat16_add_mips64
+#define bfloat16_sub bfloat16_sub_mips64
+#define bfloat16_mul bfloat16_mul_mips64
+#define bfloat16_div bfloat16_div_mips64
+#define bfloat16_muladd bfloat16_muladd_mips64
+#define bfloat16_scalbn bfloat16_scalbn_mips64
+#define bfloat16_min bfloat16_min_mips64
+#define bfloat16_max bfloat16_max_mips64
+#define bfloat16_minnum bfloat16_minnum_mips64
+#define bfloat16_maxnum bfloat16_maxnum_mips64
+#define bfloat16_minnummag bfloat16_minnummag_mips64
+#define bfloat16_maxnummag bfloat16_maxnummag_mips64
+#define bfloat16_sqrt bfloat16_sqrt_mips64
+#define bfloat16_compare bfloat16_compare_mips64
+#define bfloat16_compare_quiet bfloat16_compare_quiet_mips64
+#define bfloat16_silence_nan bfloat16_silence_nan_mips64
+#define bfloat16_default_nan bfloat16_default_nan_mips64
 #define bitmap_zero_extend bitmap_zero_extend_mips64
 #define bp_wp_matches bp_wp_matches_mips64
 #define breakpoint_invalidate breakpoint_invalidate_mips64
diff --git a/qemu/mips64el.h b/qemu/mips64el.h
index 6fb367cb..98190251 100644
--- a/qemu/mips64el.h
+++ b/qemu/mips64el.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_mips64el
 #define bad_mode_switch bad_mode_switch_mips64el
 #define bank_number bank_number_mips64el
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_mips64el
+#define bfloat16_add bfloat16_add_mips64el
+#define bfloat16_sub bfloat16_sub_mips64el
+#define bfloat16_mul bfloat16_mul_mips64el
+#define bfloat16_div bfloat16_div_mips64el
+#define bfloat16_muladd bfloat16_muladd_mips64el
+#define bfloat16_scalbn bfloat16_scalbn_mips64el
+#define bfloat16_min bfloat16_min_mips64el
+#define bfloat16_max bfloat16_max_mips64el
+#define bfloat16_minnum bfloat16_minnum_mips64el
+#define bfloat16_maxnum bfloat16_maxnum_mips64el
+#define bfloat16_minnummag bfloat16_minnummag_mips64el
+#define bfloat16_maxnummag bfloat16_maxnummag_mips64el
+#define bfloat16_sqrt bfloat16_sqrt_mips64el
+#define bfloat16_compare bfloat16_compare_mips64el
+#define bfloat16_compare_quiet bfloat16_compare_quiet_mips64el
+#define bfloat16_silence_nan bfloat16_silence_nan_mips64el
+#define bfloat16_default_nan bfloat16_default_nan_mips64el
 #define bitmap_zero_extend bitmap_zero_extend_mips64el
 #define bp_wp_matches bp_wp_matches_mips64el
 #define breakpoint_invalidate breakpoint_invalidate_mips64el
diff --git a/qemu/mipsel.h b/qemu/mipsel.h
index f0e20914..75c7a3bc 100644
--- a/qemu/mipsel.h
+++ b/qemu/mipsel.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_mipsel
 #define bad_mode_switch bad_mode_switch_mipsel
 #define bank_number bank_number_mipsel
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_mipsel
+#define bfloat16_add bfloat16_add_mipsel
+#define bfloat16_sub bfloat16_sub_mipsel
+#define bfloat16_mul bfloat16_mul_mipsel
+#define bfloat16_div bfloat16_div_mipsel
+#define bfloat16_muladd bfloat16_muladd_mipsel
+#define bfloat16_scalbn bfloat16_scalbn_mipsel
+#define bfloat16_min bfloat16_min_mipsel
+#define bfloat16_max bfloat16_max_mipsel
+#define bfloat16_minnum bfloat16_minnum_mipsel
+#define bfloat16_maxnum bfloat16_maxnum_mipsel
+#define bfloat16_minnummag bfloat16_minnummag_mipsel
+#define bfloat16_maxnummag bfloat16_maxnummag_mipsel
+#define bfloat16_sqrt bfloat16_sqrt_mipsel
+#define bfloat16_compare bfloat16_compare_mipsel
+#define bfloat16_compare_quiet bfloat16_compare_quiet_mipsel
+#define bfloat16_silence_nan bfloat16_silence_nan_mipsel
+#define bfloat16_default_nan bfloat16_default_nan_mipsel
 #define bitmap_zero_extend bitmap_zero_extend_mipsel
 #define bp_wp_matches bp_wp_matches_mipsel
 #define breakpoint_invalidate breakpoint_invalidate_mipsel
diff --git a/qemu/powerpc.h b/qemu/powerpc.h
index b7890a3c..61de8961 100644
--- a/qemu/powerpc.h
+++ b/qemu/powerpc.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_powerpc
 #define bad_mode_switch bad_mode_switch_powerpc
 #define bank_number bank_number_powerpc
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_powerpc
+#define bfloat16_add bfloat16_add_powerpc
+#define bfloat16_sub bfloat16_sub_powerpc
+#define bfloat16_mul bfloat16_mul_powerpc
+#define bfloat16_div bfloat16_div_powerpc
+#define bfloat16_muladd bfloat16_muladd_powerpc
+#define bfloat16_scalbn bfloat16_scalbn_powerpc
+#define bfloat16_min bfloat16_min_powerpc
+#define bfloat16_max bfloat16_max_powerpc
+#define bfloat16_minnum bfloat16_minnum_powerpc
+#define bfloat16_maxnum bfloat16_maxnum_powerpc
+#define bfloat16_minnummag bfloat16_minnummag_powerpc
+#define bfloat16_maxnummag bfloat16_maxnummag_powerpc
+#define bfloat16_sqrt bfloat16_sqrt_powerpc
+#define bfloat16_compare bfloat16_compare_powerpc
+#define bfloat16_compare_quiet bfloat16_compare_quiet_powerpc
+#define bfloat16_silence_nan bfloat16_silence_nan_powerpc
+#define bfloat16_default_nan bfloat16_default_nan_powerpc
 #define bitmap_zero_extend bitmap_zero_extend_powerpc
 #define bp_wp_matches bp_wp_matches_powerpc
 #define breakpoint_invalidate breakpoint_invalidate_powerpc
diff --git a/qemu/riscv32.h b/qemu/riscv32.h
index 435090bc..a0dc0c57 100644
--- a/qemu/riscv32.h
+++ b/qemu/riscv32.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_riscv32
 #define bad_mode_switch bad_mode_switch_riscv32
 #define bank_number bank_number_riscv32
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_riscv32
+#define bfloat16_add bfloat16_add_riscv32
+#define bfloat16_sub bfloat16_sub_riscv32
+#define bfloat16_mul bfloat16_mul_riscv32
+#define bfloat16_div bfloat16_div_riscv32
+#define bfloat16_muladd bfloat16_muladd_riscv32
+#define bfloat16_scalbn bfloat16_scalbn_riscv32
+#define bfloat16_min bfloat16_min_riscv32
+#define bfloat16_max bfloat16_max_riscv32
+#define bfloat16_minnum bfloat16_minnum_riscv32
+#define bfloat16_maxnum bfloat16_maxnum_riscv32
+#define bfloat16_minnummag bfloat16_minnummag_riscv32
+#define bfloat16_maxnummag bfloat16_maxnummag_riscv32
+#define bfloat16_sqrt bfloat16_sqrt_riscv32
+#define bfloat16_compare bfloat16_compare_riscv32
+#define bfloat16_compare_quiet bfloat16_compare_quiet_riscv32
+#define bfloat16_silence_nan bfloat16_silence_nan_riscv32
+#define bfloat16_default_nan bfloat16_default_nan_riscv32
 #define bitmap_zero_extend bitmap_zero_extend_riscv32
 #define bp_wp_matches bp_wp_matches_riscv32
 #define breakpoint_invalidate breakpoint_invalidate_riscv32
diff --git a/qemu/riscv64.h b/qemu/riscv64.h
index ba5af45f..e8bb608d 100644
--- a/qemu/riscv64.h
+++ b/qemu/riscv64.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_riscv64
 #define bad_mode_switch bad_mode_switch_riscv64
 #define bank_number bank_number_riscv64
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_riscv64
+#define bfloat16_add bfloat16_add_riscv64
+#define bfloat16_sub bfloat16_sub_riscv64
+#define bfloat16_mul bfloat16_mul_riscv64
+#define bfloat16_div bfloat16_div_riscv64
+#define bfloat16_muladd bfloat16_muladd_riscv64
+#define bfloat16_scalbn bfloat16_scalbn_riscv64
+#define bfloat16_min bfloat16_min_riscv64
+#define bfloat16_max bfloat16_max_riscv64
+#define bfloat16_minnum bfloat16_minnum_riscv64
+#define bfloat16_maxnum bfloat16_maxnum_riscv64
+#define bfloat16_minnummag bfloat16_minnummag_riscv64
+#define bfloat16_maxnummag bfloat16_maxnummag_riscv64
+#define bfloat16_sqrt bfloat16_sqrt_riscv64
+#define bfloat16_compare bfloat16_compare_riscv64
+#define bfloat16_compare_quiet bfloat16_compare_quiet_riscv64
+#define bfloat16_silence_nan bfloat16_silence_nan_riscv64
+#define bfloat16_default_nan bfloat16_default_nan_riscv64
 #define bitmap_zero_extend bitmap_zero_extend_riscv64
 #define bp_wp_matches bp_wp_matches_riscv64
 #define breakpoint_invalidate breakpoint_invalidate_riscv64
diff --git a/qemu/sparc.h b/qemu/sparc.h
index 0211882a..cb39403e 100644
--- a/qemu/sparc.h
+++ b/qemu/sparc.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_sparc
 #define bad_mode_switch bad_mode_switch_sparc
 #define bank_number bank_number_sparc
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_sparc
+#define bfloat16_add bfloat16_add_sparc
+#define bfloat16_sub bfloat16_sub_sparc
+#define bfloat16_mul bfloat16_mul_sparc
+#define bfloat16_div bfloat16_div_sparc
+#define bfloat16_muladd bfloat16_muladd_sparc
+#define bfloat16_scalbn bfloat16_scalbn_sparc
+#define bfloat16_min bfloat16_min_sparc
+#define bfloat16_max bfloat16_max_sparc
+#define bfloat16_minnum bfloat16_minnum_sparc
+#define bfloat16_maxnum bfloat16_maxnum_sparc
+#define bfloat16_minnummag bfloat16_minnummag_sparc
+#define bfloat16_maxnummag bfloat16_maxnummag_sparc
+#define bfloat16_sqrt bfloat16_sqrt_sparc
+#define bfloat16_compare bfloat16_compare_sparc
+#define bfloat16_compare_quiet bfloat16_compare_quiet_sparc
+#define bfloat16_silence_nan bfloat16_silence_nan_sparc
+#define bfloat16_default_nan bfloat16_default_nan_sparc
 #define bitmap_zero_extend bitmap_zero_extend_sparc
 #define bp_wp_matches bp_wp_matches_sparc
 #define breakpoint_invalidate breakpoint_invalidate_sparc
diff --git a/qemu/sparc64.h b/qemu/sparc64.h
index 242d85f9..982e3ec5 100644
--- a/qemu/sparc64.h
+++ b/qemu/sparc64.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_sparc64
 #define bad_mode_switch bad_mode_switch_sparc64
 #define bank_number bank_number_sparc64
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_sparc64
+#define bfloat16_add bfloat16_add_sparc64
+#define bfloat16_sub bfloat16_sub_sparc64
+#define bfloat16_mul bfloat16_mul_sparc64
+#define bfloat16_div bfloat16_div_sparc64
+#define bfloat16_muladd bfloat16_muladd_sparc64
+#define bfloat16_scalbn bfloat16_scalbn_sparc64
+#define bfloat16_min bfloat16_min_sparc64
+#define bfloat16_max bfloat16_max_sparc64
+#define bfloat16_minnum bfloat16_minnum_sparc64
+#define bfloat16_maxnum bfloat16_maxnum_sparc64
+#define bfloat16_minnummag bfloat16_minnummag_sparc64
+#define bfloat16_maxnummag bfloat16_maxnummag_sparc64
+#define bfloat16_sqrt bfloat16_sqrt_sparc64
+#define bfloat16_compare bfloat16_compare_sparc64
+#define bfloat16_compare_quiet bfloat16_compare_quiet_sparc64
+#define bfloat16_silence_nan bfloat16_silence_nan_sparc64
+#define bfloat16_default_nan bfloat16_default_nan_sparc64
 #define bitmap_zero_extend bitmap_zero_extend_sparc64
 #define bp_wp_matches bp_wp_matches_sparc64
 #define breakpoint_invalidate breakpoint_invalidate_sparc64
diff --git a/qemu/x86_64.h b/qemu/x86_64.h
index 114de5a3..c4247435 100644
--- a/qemu/x86_64.h
+++ b/qemu/x86_64.h
@@ -197,6 +197,24 @@
 #define ats_write ats_write_x86_64
 #define bad_mode_switch bad_mode_switch_x86_64
 #define bank_number bank_number_x86_64
+#define bfloat16_squash_input_denormal bfloat16_squash_input_denormal_x86_64
+#define bfloat16_add bfloat16_add_x86_64
+#define bfloat16_sub bfloat16_sub_x86_64
+#define bfloat16_mul bfloat16_mul_x86_64
+#define bfloat16_div bfloat16_div_x86_64
+#define bfloat16_muladd bfloat16_muladd_x86_64
+#define bfloat16_scalbn bfloat16_scalbn_x86_64
+#define bfloat16_min bfloat16_min_x86_64
+#define bfloat16_max bfloat16_max_x86_64
+#define bfloat16_minnum bfloat16_minnum_x86_64
+#define bfloat16_maxnum bfloat16_maxnum_x86_64
+#define bfloat16_minnummag bfloat16_minnummag_x86_64
+#define bfloat16_maxnummag bfloat16_maxnummag_x86_64
+#define bfloat16_sqrt bfloat16_sqrt_x86_64
+#define bfloat16_compare bfloat16_compare_x86_64
+#define bfloat16_compare_quiet bfloat16_compare_quiet_x86_64
+#define bfloat16_silence_nan bfloat16_silence_nan_x86_64
+#define bfloat16_default_nan bfloat16_default_nan_x86_64
 #define bitmap_zero_extend bitmap_zero_extend_x86_64
 #define bp_wp_matches bp_wp_matches_x86_64
 #define breakpoint_invalidate breakpoint_invalidate_x86_64