diff mbox series

[v4,56/57] tcg/ppc: Support 128-bit load/store

Message ID 20230503070656.1746170-57-richard.henderson@linaro.org
State New
Headers show
Series tcg: Improve atomicity support | expand

Commit Message

Richard Henderson May 3, 2023, 7:06 a.m. UTC
Use LQ/STQ with ISA v2.07, and 16-byte atomicity is required.
Note that these instructions do not require 16-byte alignment.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 tcg/ppc/tcg-target-con-set.h |   2 +
 tcg/ppc/tcg-target-con-str.h |   1 +
 tcg/ppc/tcg-target.h         |   3 +-
 tcg/ppc/tcg-target.c.inc     | 173 +++++++++++++++++++++++++++++++----
 4 files changed, 158 insertions(+), 21 deletions(-)

Comments

Daniel Henrique Barboza May 8, 2023, 12:16 p.m. UTC | #1
On 5/3/23 04:06, Richard Henderson wrote:
> Use LQ/STQ with ISA v2.07, and 16-byte atomicity is required.
> Note that these instructions do not require 16-byte alignment.
> 
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---


Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>

>   tcg/ppc/tcg-target-con-set.h |   2 +
>   tcg/ppc/tcg-target-con-str.h |   1 +
>   tcg/ppc/tcg-target.h         |   3 +-
>   tcg/ppc/tcg-target.c.inc     | 173 +++++++++++++++++++++++++++++++----
>   4 files changed, 158 insertions(+), 21 deletions(-)
> 
> diff --git a/tcg/ppc/tcg-target-con-set.h b/tcg/ppc/tcg-target-con-set.h
> index f206b29205..bbd7b21247 100644
> --- a/tcg/ppc/tcg-target-con-set.h
> +++ b/tcg/ppc/tcg-target-con-set.h
> @@ -14,6 +14,7 @@ C_O0_I2(r, r)
>   C_O0_I2(r, ri)
>   C_O0_I2(v, r)
>   C_O0_I3(r, r, r)
> +C_O0_I3(o, m, r)
>   C_O0_I4(r, r, ri, ri)
>   C_O0_I4(r, r, r, r)
>   C_O1_I1(r, r)
> @@ -34,6 +35,7 @@ C_O1_I3(v, v, v, v)
>   C_O1_I4(r, r, ri, rZ, rZ)
>   C_O1_I4(r, r, r, ri, ri)
>   C_O2_I1(r, r, r)
> +C_O2_I1(o, m, r)
>   C_O2_I2(r, r, r, r)
>   C_O2_I4(r, r, rI, rZM, r, r)
>   C_O2_I4(r, r, r, r, rI, rZM)
> diff --git a/tcg/ppc/tcg-target-con-str.h b/tcg/ppc/tcg-target-con-str.h
> index 094613cbcb..20846901de 100644
> --- a/tcg/ppc/tcg-target-con-str.h
> +++ b/tcg/ppc/tcg-target-con-str.h
> @@ -9,6 +9,7 @@
>    * REGS(letter, register_mask)
>    */
>   REGS('r', ALL_GENERAL_REGS)
> +REGS('o', ALL_GENERAL_REGS & 0xAAAAAAAAu)  /* odd registers */
>   REGS('v', ALL_VECTOR_REGS)
>   
>   /*
> diff --git a/tcg/ppc/tcg-target.h b/tcg/ppc/tcg-target.h
> index 0914380bd7..204b70f86a 100644
> --- a/tcg/ppc/tcg-target.h
> +++ b/tcg/ppc/tcg-target.h
> @@ -149,7 +149,8 @@ extern bool have_vsx;
>   #define TCG_TARGET_HAS_mulsh_i64        1
>   #endif
>   
> -#define TCG_TARGET_HAS_qemu_ldst_i128   0
> +#define TCG_TARGET_HAS_qemu_ldst_i128   \
> +    (TCG_TARGET_REG_BITS == 64 && have_isa_2_07)
>   
>   /*
>    * While technically Altivec could support V64, it has no 64-bit store
> diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc
> index 60375804cd..682743a466 100644
> --- a/tcg/ppc/tcg-target.c.inc
> +++ b/tcg/ppc/tcg-target.c.inc
> @@ -295,25 +295,27 @@ static bool tcg_target_const_match(int64_t val, TCGType type, int ct)
>   
>   #define B      OPCD( 18)
>   #define BC     OPCD( 16)
> +
>   #define LBZ    OPCD( 34)
>   #define LHZ    OPCD( 40)
>   #define LHA    OPCD( 42)
>   #define LWZ    OPCD( 32)
>   #define LWZUX  XO31( 55)
> -#define STB    OPCD( 38)
> -#define STH    OPCD( 44)
> -#define STW    OPCD( 36)
> -
> -#define STD    XO62(  0)
> -#define STDU   XO62(  1)
> -#define STDX   XO31(149)
> -
>   #define LD     XO58(  0)
>   #define LDX    XO31( 21)
>   #define LDU    XO58(  1)
>   #define LDUX   XO31( 53)
>   #define LWA    XO58(  2)
>   #define LWAX   XO31(341)
> +#define LQ     OPCD( 56)
> +
> +#define STB    OPCD( 38)
> +#define STH    OPCD( 44)
> +#define STW    OPCD( 36)
> +#define STD    XO62(  0)
> +#define STDU   XO62(  1)
> +#define STDX   XO31(149)
> +#define STQ    XO62(  2)
>   
>   #define ADDIC  OPCD( 12)
>   #define ADDI   OPCD( 14)
> @@ -2015,11 +2017,25 @@ static bool tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *lb)
>   typedef struct {
>       TCGReg base;
>       TCGReg index;
> +    MemOp align;
> +    MemOp atom;
>   } HostAddress;
>   
>   bool tcg_target_has_memory_bswap(MemOp memop)
>   {
> -    return true;
> +    MemOp atom_a, atom_u;
> +
> +    if ((memop & MO_SIZE) <= MO_64) {
> +        return true;
> +    }
> +
> +    /*
> +     * Reject 16-byte memop with 16-byte atomicity,
> +     * but do allow a pair of 64-bit operations.
> +     */
> +    (void)atom_and_align_for_opc(tcg_ctx, &atom_a, &atom_u, memop,
> +                                 MO_ATOM_IFALIGN, true);
> +    return atom_a <= MO_64;
>   }
>   
>   /*
> @@ -2034,7 +2050,7 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
>   {
>       TCGLabelQemuLdst *ldst = NULL;
>       MemOp opc = get_memop(oi);
> -    MemOp a_bits, atom_a, atom_u;
> +    MemOp a_bits, atom_u, s_bits;
>   
>       /*
>        * Book II, Section 1.4, Single-Copy Atomicity, specifies:
> @@ -2046,10 +2062,19 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
>        * As of 3.0, "the non-atomic access is performed as described in
>        * the corresponding list", which matches MO_ATOM_SUBALIGN.
>        */
> -    a_bits = atom_and_align_for_opc(s, &atom_a, &atom_u, opc,
> +    s_bits = opc & MO_SIZE;
> +    a_bits = atom_and_align_for_opc(s, &h->atom, &atom_u, opc,
>                                       have_isa_3_00 ? MO_ATOM_SUBALIGN
>                                                     : MO_ATOM_IFALIGN,
> -                                    false);
> +                                    s_bits == MO_128);
> +
> +    if (TCG_TARGET_REG_BITS == 32) {
> +        /* We don't support unaligned accesses on 32-bits. */
> +        if (a_bits < s_bits) {
> +            a_bits = s_bits;
> +        }
> +    }
> +    h->align = a_bits;
>   
>   #ifdef CONFIG_SOFTMMU
>       int mem_index = get_mmuidx(oi);
> @@ -2058,7 +2083,6 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
>       int fast_off = TLB_MASK_TABLE_OFS(mem_index);
>       int mask_off = fast_off + offsetof(CPUTLBDescFast, mask);
>       int table_off = fast_off + offsetof(CPUTLBDescFast, table);
> -    unsigned s_bits = opc & MO_SIZE;
>   
>       ldst = new_ldst_label(s);
>       ldst->is_ld = is_ld;
> @@ -2108,13 +2132,6 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
>   
>       /* Clear the non-page, non-alignment bits from the address in R0. */
>       if (TCG_TARGET_REG_BITS == 32) {
> -        /* We don't support unaligned accesses on 32-bits.
> -         * Preserve the bottom bits and thus trigger a comparison
> -         * failure on unaligned accesses.
> -         */
> -        if (a_bits < s_bits) {
> -            a_bits = s_bits;
> -        }
>           tcg_out_rlw(s, RLWINM, TCG_REG_R0, addrlo, 0,
>                       (32 - a_bits) & 31, 31 - TARGET_PAGE_BITS);
>       } else {
> @@ -2299,6 +2316,108 @@ static void tcg_out_qemu_st(TCGContext *s, TCGReg datalo, TCGReg datahi,
>       }
>   }
>   
> +static TCGLabelQemuLdst *
> +prepare_host_addr_index_only(TCGContext *s, HostAddress *h, TCGReg addr_reg,
> +                             MemOpIdx oi, bool is_ld)
> +{
> +    TCGLabelQemuLdst *ldst;
> +
> +    ldst = prepare_host_addr(s, h, addr_reg, -1, oi, true);
> +
> +    /* Compose the final address, as LQ/STQ have no indexing. */
> +    if (h->base != 0) {
> +        tcg_out32(s, ADD | TAB(TCG_REG_TMP1, h->base, h->index));
> +        h->index = TCG_REG_TMP1;
> +        h->base = 0;
> +    }
> +
> +    return ldst;
> +}
> +
> +static void tcg_out_qemu_ld128(TCGContext *s, TCGReg datalo, TCGReg datahi,
> +                               TCGReg addr_reg, MemOpIdx oi)
> +{
> +    TCGLabelQemuLdst *ldst;
> +    HostAddress h;
> +    bool need_bswap;
> +
> +    ldst = prepare_host_addr_index_only(s, &h, addr_reg, oi, true);
> +    need_bswap = get_memop(oi) & MO_BSWAP;
> +
> +    if (h.atom == MO_128) {
> +        tcg_debug_assert(!need_bswap);
> +        tcg_debug_assert(datalo & 1);
> +        tcg_debug_assert(datahi == datalo - 1);
> +        tcg_out32(s, LQ | TAI(datahi, h.index, 0));
> +    } else {
> +        TCGReg d1, d2;
> +
> +        if (HOST_BIG_ENDIAN ^ need_bswap) {
> +            d1 = datahi, d2 = datalo;
> +        } else {
> +            d1 = datalo, d2 = datahi;
> +        }
> +
> +        if (need_bswap) {
> +            tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_R0, 8);
> +            tcg_out32(s, LDBRX | TAB(d1, 0, h.index));
> +            tcg_out32(s, LDBRX | TAB(d2, h.index, TCG_REG_R0));
> +        } else {
> +            tcg_out32(s, LD | TAI(d1, h.index, 0));
> +            tcg_out32(s, LD | TAI(d2, h.index, 8));
> +        }
> +    }
> +
> +    if (ldst) {
> +        ldst->type = TCG_TYPE_I128;
> +        ldst->datalo_reg = datalo;
> +        ldst->datahi_reg = datahi;
> +        ldst->raddr = tcg_splitwx_to_rx(s->code_ptr);
> +    }
> +}
> +
> +static void tcg_out_qemu_st128(TCGContext *s, TCGReg datalo, TCGReg datahi,
> +                               TCGReg addr_reg, MemOpIdx oi)
> +{
> +    TCGLabelQemuLdst *ldst;
> +    HostAddress h;
> +    bool need_bswap;
> +
> +    ldst = prepare_host_addr_index_only(s, &h, addr_reg, oi, false);
> +    need_bswap = get_memop(oi) & MO_BSWAP;
> +
> +    if (h.atom == MO_128) {
> +        tcg_debug_assert(!need_bswap);
> +        tcg_debug_assert(datalo & 1);
> +        tcg_debug_assert(datahi == datalo - 1);
> +        tcg_out32(s, STQ | TAI(datahi, h.index, 0));
> +    } else {
> +        TCGReg d1, d2;
> +
> +        if (HOST_BIG_ENDIAN ^ need_bswap) {
> +            d1 = datahi, d2 = datalo;
> +        } else {
> +            d1 = datalo, d2 = datahi;
> +        }
> +
> +        if (need_bswap) {
> +            tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_R0, 8);
> +            tcg_out32(s, STDBRX | TAB(d1, 0, h.index));
> +            tcg_out32(s, STDBRX | TAB(d2, h.index, TCG_REG_R0));
> +        } else {
> +            tcg_out32(s, STD | TAI(d1, h.index, 0));
> +            tcg_out32(s, STD | TAI(d2, h.index, 8));
> +        }
> +    }
> +
> +    if (ldst) {
> +        ldst->type = TCG_TYPE_I128;
> +        ldst->datalo_reg = datalo;
> +        ldst->datahi_reg = datahi;
> +        ldst->raddr = tcg_splitwx_to_rx(s->code_ptr);
> +    }
> +}
> +
>   static void tcg_out_nop_fill(tcg_insn_unit *p, int count)
>   {
>       int i;
> @@ -2849,6 +2968,11 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>                               args[4], TCG_TYPE_I64);
>           }
>           break;
> +    case INDEX_op_qemu_ld_i128:
> +        tcg_debug_assert(TCG_TARGET_REG_BITS == 64);
> +        tcg_out_qemu_ld128(s, args[0], args[1], args[2], args[3]);
> +        break;
> +
>       case INDEX_op_qemu_st_i32:
>           if (TCG_TARGET_REG_BITS >= TARGET_LONG_BITS) {
>               tcg_out_qemu_st(s, args[0], -1, args[1], -1,
> @@ -2870,6 +2994,10 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>                               args[4], TCG_TYPE_I64);
>           }
>           break;
> +    case INDEX_op_qemu_st_i128:
> +        tcg_debug_assert(TCG_TARGET_REG_BITS == 64);
> +        tcg_out_qemu_st128(s, args[0], args[1], args[2], args[3]);
> +        break;
>   
>       case INDEX_op_setcond_i32:
>           tcg_out_setcond(s, TCG_TYPE_I32, args[3], args[0], args[1], args[2],
> @@ -3705,6 +3833,11 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
>                   : TARGET_LONG_BITS == 32 ? C_O0_I3(r, r, r)
>                   : C_O0_I4(r, r, r, r));
>   
> +    case INDEX_op_qemu_ld_i128:
> +        return C_O2_I1(o, m, r);
> +    case INDEX_op_qemu_st_i128:
> +        return C_O0_I3(o, m, r);
> +
>       case INDEX_op_add_vec:
>       case INDEX_op_sub_vec:
>       case INDEX_op_mul_vec:
diff mbox series

Patch

diff --git a/tcg/ppc/tcg-target-con-set.h b/tcg/ppc/tcg-target-con-set.h
index f206b29205..bbd7b21247 100644
--- a/tcg/ppc/tcg-target-con-set.h
+++ b/tcg/ppc/tcg-target-con-set.h
@@ -14,6 +14,7 @@  C_O0_I2(r, r)
 C_O0_I2(r, ri)
 C_O0_I2(v, r)
 C_O0_I3(r, r, r)
+C_O0_I3(o, m, r)
 C_O0_I4(r, r, ri, ri)
 C_O0_I4(r, r, r, r)
 C_O1_I1(r, r)
@@ -34,6 +35,7 @@  C_O1_I3(v, v, v, v)
 C_O1_I4(r, r, ri, rZ, rZ)
 C_O1_I4(r, r, r, ri, ri)
 C_O2_I1(r, r, r)
+C_O2_I1(o, m, r)
 C_O2_I2(r, r, r, r)
 C_O2_I4(r, r, rI, rZM, r, r)
 C_O2_I4(r, r, r, r, rI, rZM)
diff --git a/tcg/ppc/tcg-target-con-str.h b/tcg/ppc/tcg-target-con-str.h
index 094613cbcb..20846901de 100644
--- a/tcg/ppc/tcg-target-con-str.h
+++ b/tcg/ppc/tcg-target-con-str.h
@@ -9,6 +9,7 @@ 
  * REGS(letter, register_mask)
  */
 REGS('r', ALL_GENERAL_REGS)
+REGS('o', ALL_GENERAL_REGS & 0xAAAAAAAAu)  /* odd registers */
 REGS('v', ALL_VECTOR_REGS)
 
 /*
diff --git a/tcg/ppc/tcg-target.h b/tcg/ppc/tcg-target.h
index 0914380bd7..204b70f86a 100644
--- a/tcg/ppc/tcg-target.h
+++ b/tcg/ppc/tcg-target.h
@@ -149,7 +149,8 @@  extern bool have_vsx;
 #define TCG_TARGET_HAS_mulsh_i64        1
 #endif
 
-#define TCG_TARGET_HAS_qemu_ldst_i128   0
+#define TCG_TARGET_HAS_qemu_ldst_i128   \
+    (TCG_TARGET_REG_BITS == 64 && have_isa_2_07)
 
 /*
  * While technically Altivec could support V64, it has no 64-bit store
diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc
index 60375804cd..682743a466 100644
--- a/tcg/ppc/tcg-target.c.inc
+++ b/tcg/ppc/tcg-target.c.inc
@@ -295,25 +295,27 @@  static bool tcg_target_const_match(int64_t val, TCGType type, int ct)
 
 #define B      OPCD( 18)
 #define BC     OPCD( 16)
+
 #define LBZ    OPCD( 34)
 #define LHZ    OPCD( 40)
 #define LHA    OPCD( 42)
 #define LWZ    OPCD( 32)
 #define LWZUX  XO31( 55)
-#define STB    OPCD( 38)
-#define STH    OPCD( 44)
-#define STW    OPCD( 36)
-
-#define STD    XO62(  0)
-#define STDU   XO62(  1)
-#define STDX   XO31(149)
-
 #define LD     XO58(  0)
 #define LDX    XO31( 21)
 #define LDU    XO58(  1)
 #define LDUX   XO31( 53)
 #define LWA    XO58(  2)
 #define LWAX   XO31(341)
+#define LQ     OPCD( 56)
+
+#define STB    OPCD( 38)
+#define STH    OPCD( 44)
+#define STW    OPCD( 36)
+#define STD    XO62(  0)
+#define STDU   XO62(  1)
+#define STDX   XO31(149)
+#define STQ    XO62(  2)
 
 #define ADDIC  OPCD( 12)
 #define ADDI   OPCD( 14)
@@ -2015,11 +2017,25 @@  static bool tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *lb)
 typedef struct {
     TCGReg base;
     TCGReg index;
+    MemOp align;
+    MemOp atom;
 } HostAddress;
 
 bool tcg_target_has_memory_bswap(MemOp memop)
 {
-    return true;
+    MemOp atom_a, atom_u;
+
+    if ((memop & MO_SIZE) <= MO_64) {
+        return true;
+    }
+
+    /*
+     * Reject 16-byte memop with 16-byte atomicity,
+     * but do allow a pair of 64-bit operations.
+     */
+    (void)atom_and_align_for_opc(tcg_ctx, &atom_a, &atom_u, memop,
+                                 MO_ATOM_IFALIGN, true);
+    return atom_a <= MO_64;
 }
 
 /*
@@ -2034,7 +2050,7 @@  static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
 {
     TCGLabelQemuLdst *ldst = NULL;
     MemOp opc = get_memop(oi);
-    MemOp a_bits, atom_a, atom_u;
+    MemOp a_bits, atom_u, s_bits;
 
     /*
      * Book II, Section 1.4, Single-Copy Atomicity, specifies:
@@ -2046,10 +2062,19 @@  static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
      * As of 3.0, "the non-atomic access is performed as described in
      * the corresponding list", which matches MO_ATOM_SUBALIGN.
      */
-    a_bits = atom_and_align_for_opc(s, &atom_a, &atom_u, opc,
+    s_bits = opc & MO_SIZE;
+    a_bits = atom_and_align_for_opc(s, &h->atom, &atom_u, opc,
                                     have_isa_3_00 ? MO_ATOM_SUBALIGN
                                                   : MO_ATOM_IFALIGN,
-                                    false);
+                                    s_bits == MO_128);
+
+    if (TCG_TARGET_REG_BITS == 32) {
+        /* We don't support unaligned accesses on 32-bits. */
+        if (a_bits < s_bits) {
+            a_bits = s_bits;
+        }
+    }
+    h->align = a_bits;
 
 #ifdef CONFIG_SOFTMMU
     int mem_index = get_mmuidx(oi);
@@ -2058,7 +2083,6 @@  static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
     int fast_off = TLB_MASK_TABLE_OFS(mem_index);
     int mask_off = fast_off + offsetof(CPUTLBDescFast, mask);
     int table_off = fast_off + offsetof(CPUTLBDescFast, table);
-    unsigned s_bits = opc & MO_SIZE;
 
     ldst = new_ldst_label(s);
     ldst->is_ld = is_ld;
@@ -2108,13 +2132,6 @@  static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
 
     /* Clear the non-page, non-alignment bits from the address in R0. */
     if (TCG_TARGET_REG_BITS == 32) {
-        /* We don't support unaligned accesses on 32-bits.
-         * Preserve the bottom bits and thus trigger a comparison
-         * failure on unaligned accesses.
-         */
-        if (a_bits < s_bits) {
-            a_bits = s_bits;
-        }
         tcg_out_rlw(s, RLWINM, TCG_REG_R0, addrlo, 0,
                     (32 - a_bits) & 31, 31 - TARGET_PAGE_BITS);
     } else {
@@ -2299,6 +2316,108 @@  static void tcg_out_qemu_st(TCGContext *s, TCGReg datalo, TCGReg datahi,
     }
 }
 
+static TCGLabelQemuLdst *
+prepare_host_addr_index_only(TCGContext *s, HostAddress *h, TCGReg addr_reg,
+                             MemOpIdx oi, bool is_ld)
+{
+    TCGLabelQemuLdst *ldst;
+
+    ldst = prepare_host_addr(s, h, addr_reg, -1, oi, true);
+
+    /* Compose the final address, as LQ/STQ have no indexing. */
+    if (h->base != 0) {
+        tcg_out32(s, ADD | TAB(TCG_REG_TMP1, h->base, h->index));
+        h->index = TCG_REG_TMP1;
+        h->base = 0;
+    }
+
+    return ldst;
+}
+
+static void tcg_out_qemu_ld128(TCGContext *s, TCGReg datalo, TCGReg datahi,
+                               TCGReg addr_reg, MemOpIdx oi)
+{
+    TCGLabelQemuLdst *ldst;
+    HostAddress h;
+    bool need_bswap;
+
+    ldst = prepare_host_addr_index_only(s, &h, addr_reg, oi, true);
+    need_bswap = get_memop(oi) & MO_BSWAP;
+
+    if (h.atom == MO_128) {
+        tcg_debug_assert(!need_bswap);
+        tcg_debug_assert(datalo & 1);
+        tcg_debug_assert(datahi == datalo - 1);
+        tcg_out32(s, LQ | TAI(datahi, h.index, 0));
+    } else {
+        TCGReg d1, d2;
+
+        if (HOST_BIG_ENDIAN ^ need_bswap) {
+            d1 = datahi, d2 = datalo;
+        } else {
+            d1 = datalo, d2 = datahi;
+        }
+
+        if (need_bswap) {
+            tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_R0, 8);
+            tcg_out32(s, LDBRX | TAB(d1, 0, h.index));
+            tcg_out32(s, LDBRX | TAB(d2, h.index, TCG_REG_R0));
+        } else {
+            tcg_out32(s, LD | TAI(d1, h.index, 0));
+            tcg_out32(s, LD | TAI(d2, h.index, 8));
+        }
+    }
+
+    if (ldst) {
+        ldst->type = TCG_TYPE_I128;
+        ldst->datalo_reg = datalo;
+        ldst->datahi_reg = datahi;
+        ldst->raddr = tcg_splitwx_to_rx(s->code_ptr);
+    }
+}
+
+static void tcg_out_qemu_st128(TCGContext *s, TCGReg datalo, TCGReg datahi,
+                               TCGReg addr_reg, MemOpIdx oi)
+{
+    TCGLabelQemuLdst *ldst;
+    HostAddress h;
+    bool need_bswap;
+
+    ldst = prepare_host_addr_index_only(s, &h, addr_reg, oi, false);
+    need_bswap = get_memop(oi) & MO_BSWAP;
+
+    if (h.atom == MO_128) {
+        tcg_debug_assert(!need_bswap);
+        tcg_debug_assert(datalo & 1);
+        tcg_debug_assert(datahi == datalo - 1);
+        tcg_out32(s, STQ | TAI(datahi, h.index, 0));
+    } else {
+        TCGReg d1, d2;
+
+        if (HOST_BIG_ENDIAN ^ need_bswap) {
+            d1 = datahi, d2 = datalo;
+        } else {
+            d1 = datalo, d2 = datahi;
+        }
+
+        if (need_bswap) {
+            tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_R0, 8);
+            tcg_out32(s, STDBRX | TAB(d1, 0, h.index));
+            tcg_out32(s, STDBRX | TAB(d2, h.index, TCG_REG_R0));
+        } else {
+            tcg_out32(s, STD | TAI(d1, h.index, 0));
+            tcg_out32(s, STD | TAI(d2, h.index, 8));
+        }
+    }
+
+    if (ldst) {
+        ldst->type = TCG_TYPE_I128;
+        ldst->datalo_reg = datalo;
+        ldst->datahi_reg = datahi;
+        ldst->raddr = tcg_splitwx_to_rx(s->code_ptr);
+    }
+}
+
 static void tcg_out_nop_fill(tcg_insn_unit *p, int count)
 {
     int i;
@@ -2849,6 +2968,11 @@  static void tcg_out_op(TCGContext *s, TCGOpcode opc,
                             args[4], TCG_TYPE_I64);
         }
         break;
+    case INDEX_op_qemu_ld_i128:
+        tcg_debug_assert(TCG_TARGET_REG_BITS == 64);
+        tcg_out_qemu_ld128(s, args[0], args[1], args[2], args[3]);
+        break;
+
     case INDEX_op_qemu_st_i32:
         if (TCG_TARGET_REG_BITS >= TARGET_LONG_BITS) {
             tcg_out_qemu_st(s, args[0], -1, args[1], -1,
@@ -2870,6 +2994,10 @@  static void tcg_out_op(TCGContext *s, TCGOpcode opc,
                             args[4], TCG_TYPE_I64);
         }
         break;
+    case INDEX_op_qemu_st_i128:
+        tcg_debug_assert(TCG_TARGET_REG_BITS == 64);
+        tcg_out_qemu_st128(s, args[0], args[1], args[2], args[3]);
+        break;
 
     case INDEX_op_setcond_i32:
         tcg_out_setcond(s, TCG_TYPE_I32, args[3], args[0], args[1], args[2],
@@ -3705,6 +3833,11 @@  static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode op)
                 : TARGET_LONG_BITS == 32 ? C_O0_I3(r, r, r)
                 : C_O0_I4(r, r, r, r));
 
+    case INDEX_op_qemu_ld_i128:
+        return C_O2_I1(o, m, r);
+    case INDEX_op_qemu_st_i128:
+        return C_O0_I3(o, m, r);
+
     case INDEX_op_add_vec:
     case INDEX_op_sub_vec:
     case INDEX_op_mul_vec: