diff mbox series

[v2,16/22] target/loongarch: Add floating point move instruction translation

Message ID 1626861198-6133-17-git-send-email-gaosong@loongson.cn
State New
Headers show
Series Add LoongArch linux-user emulation support | expand

Commit Message

Song Gao July 21, 2021, 9:53 a.m. UTC
This patch implement floationg point move instruction translation.

This includes:
- FMOV.{S/D}
- FSEL
- MOVGR2FR.{W/D}, MOVGR2FRH.W
- MOVFR2GR.{S/D}, MOVFRH2GR.S
- MOVGR2FCSR, MOVFCSR2GR
- MOVFR2CF, MOVCF2FR
- MOVGR2CF, MOVCF2GR

Signed-off-by: Song Gao <gaosong@loongson.cn>
---
 target/loongarch/fpu_helper.c |  80 +++++++++++++
 target/loongarch/helper.h     |   6 +
 target/loongarch/insns.decode |  41 +++++++
 target/loongarch/trans.inc.c  | 270 ++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 397 insertions(+)

Comments

Richard Henderson July 23, 2021, 6:29 a.m. UTC | #1
On 7/20/21 11:53 PM, Song Gao wrote:
> This patch implement floationg point move instruction translation.
> 
> This includes:
> - FMOV.{S/D}
> - FSEL
> - MOVGR2FR.{W/D}, MOVGR2FRH.W
> - MOVFR2GR.{S/D}, MOVFRH2GR.S
> - MOVGR2FCSR, MOVFCSR2GR
> - MOVFR2CF, MOVCF2FR
> - MOVGR2CF, MOVCF2GR
> 
> Signed-off-by: Song Gao <gaosong@loongson.cn>
> ---
>   target/loongarch/fpu_helper.c |  80 +++++++++++++
>   target/loongarch/helper.h     |   6 +
>   target/loongarch/insns.decode |  41 +++++++
>   target/loongarch/trans.inc.c  | 270 ++++++++++++++++++++++++++++++++++++++++++
>   4 files changed, 397 insertions(+)
> 
> diff --git a/target/loongarch/fpu_helper.c b/target/loongarch/fpu_helper.c
> index 162085a..7662715 100644
> --- a/target/loongarch/fpu_helper.c
> +++ b/target/loongarch/fpu_helper.c
> @@ -379,6 +379,11 @@ uint64_t helper_fp_logb_d(CPULoongArchState *env, uint64_t fp)
>       return fp1;
>   }
>   
> +void helper_movreg2cf(CPULoongArchState *env, uint32_t cd, target_ulong src)
> +{
> +    env->active_fpu.cf[cd & 0x7] = src & 0x1;
> +}

tcg_gen_andi_tl + tcg_gen_st8_tl.

> +target_ulong helper_fsel(CPULoongArchState *env, target_ulong fj,
> +                         target_ulong fk, uint32_t ca)
> +{
> +    if (env->active_fpu.cf[ca & 0x7]) {
> +        return fk;
> +    } else {
> +        return fj;
> +    }
> +}

tcg_gen_movcond_i64.

> +void helper_movgr2fcsr(CPULoongArchState *env, target_ulong arg1,
> +                       uint32_t fcsr)
> +{
> +    switch (fcsr) {
> +    case 0:
> +        env->active_fpu.fcsr0 = arg1;
> +        break;
> +    case 1:
> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M1) |
> +                                (env->active_fpu.fcsr0 & ~FCSR0_M1);
> +        break;
> +    case 2:
> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M2) |
> +                                (env->active_fpu.fcsr0 & ~FCSR0_M2);
> +        break;
> +    case 3:
> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M3) |
> +                                (env->active_fpu.fcsr0 & ~FCSR0_M3);
> +        break;

This is easily implemented inline, followed by a single helper call to re-load the 
rounding mode (if required by the mask).

> +    case 16:
> +        env->active_fpu.vcsr16 = arg1;
> +        break;

The documentation I have does not describe the vector stuff?

> +    default:
> +        printf("%s: warning, fcsr '%d' not supported\n", __func__, fcsr);
> +        assert(0);
> +        break;

No printfs, no assert.  This should have been caught by

> +target_ulong helper_movcf2reg(CPULoongArchState *env, uint32_t cj)
> +{
> +    return (target_ulong)env->active_fpu.cf[cj & 0x7];
> +}

tcg_gen_ld8u_tl.


r~
Song Gao July 27, 2021, 8:06 a.m. UTC | #2
Hi, Richard.

On 07/23/2021 02:29 PM, Richard Henderson wrote:
> On 7/20/21 11:53 PM, Song Gao wrote:
>> This patch implement floationg point move instruction translation.
>>
>> This includes:
>> - FMOV.{S/D}
>> - FSEL
>> - MOVGR2FR.{W/D}, MOVGR2FRH.W
>> - MOVFR2GR.{S/D}, MOVFRH2GR.S
>> - MOVGR2FCSR, MOVFCSR2GR
>> - MOVFR2CF, MOVCF2FR
>> - MOVGR2CF, MOVCF2GR
>>
>> Signed-off-by: Song Gao <gaosong@loongson.cn>
>> ---
>>   target/loongarch/fpu_helper.c |  80 +++++++++++++
>>   target/loongarch/helper.h     |   6 +
>>   target/loongarch/insns.decode |  41 +++++++
>>   target/loongarch/trans.inc.c  | 270 ++++++++++++++++++++++++++++++++++++++++++
>>   4 files changed, 397 insertions(+)
>>
>> diff --git a/target/loongarch/fpu_helper.c b/target/loongarch/fpu_helper.c
>> index 162085a..7662715 100644
>> --- a/target/loongarch/fpu_helper.c
>> +++ b/target/loongarch/fpu_helper.c
>> @@ -379,6 +379,11 @@ uint64_t helper_fp_logb_d(CPULoongArchState *env, uint64_t fp)
>>       return fp1;
>>   }
>>   +void helper_movreg2cf(CPULoongArchState *env, uint32_t cd, target_ulong src)
>> +{
>> +    env->active_fpu.cf[cd & 0x7] = src & 0x1;
>> +}
> 
> tcg_gen_andi_tl + tcg_gen_st8_tl.
> 
OK.
>> +target_ulong helper_fsel(CPULoongArchState *env, target_ulong fj,
>> +                         target_ulong fk, uint32_t ca)
>> +{
>> +    if (env->active_fpu.cf[ca & 0x7]) {
>> +        return fk;
>> +    } else {
>> +        return fj;
>> +    }
>> +}
> 
> tcg_gen_movcond_i64.
> 
OK.
>> +void helper_movgr2fcsr(CPULoongArchState *env, target_ulong arg1,
>> +                       uint32_t fcsr)
>> +{
>> +    switch (fcsr) {
>> +    case 0:
>> +        env->active_fpu.fcsr0 = arg1;
>> +        break;
>> +    case 1:
>> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M1) |
>> +                                (env->active_fpu.fcsr0 & ~FCSR0_M1);
>> +        break;
>> +    case 2:
>> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M2) |
>> +                                (env->active_fpu.fcsr0 & ~FCSR0_M2);
>> +        break;
>> +    case 3:
>> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M3) |
>> +                                (env->active_fpu.fcsr0 & ~FCSR0_M3);
>> +        break;
> 
> This is easily implemented inline, followed by a single helper call to re-load the rounding mode (if required by the mask).
> 
OK.
>> +    case 16:
>> +        env->active_fpu.vcsr16 = arg1;
>> +        break;
> 
> The documentation I have does not describe the vector stuff?
> 

Yes, It is described in Volume II, but now  I need remove it .

>> +    default:
>> +        printf("%s: warning, fcsr '%d' not supported\n", __func__, fcsr);
>> +        assert(0);
>> +        break;
> 
> No printfs, no assert.  This should have been caught by
> 
>> +target_ulong helper_movcf2reg(CPULoongArchState *env, uint32_t cj)
>> +{
>> +    return (target_ulong)env->active_fpu.cf[cj & 0x7];
>> +}
> 
> tcg_gen_ld8u_tl.
>
OK.> 
> r~

Again. Thank you kindly help.

Thanks
Song Gao.
Song Gao Aug. 12, 2021, 9:20 a.m. UTC | #3
On 07/23/2021 02:29 PM, Richard Henderson wrote:
> 
>> +void helper_movgr2fcsr(CPULoongArchState *env, target_ulong arg1,
>> +                       uint32_t fcsr)
>> +{
>> +    switch (fcsr) {
>> +    case 0:
>> +        env->active_fpu.fcsr0 = arg1;
>> +        break;
>> +    case 1:
>> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M1) |
>> +                                (env->active_fpu.fcsr0 & ~FCSR0_M1);
>> +        break;
>> +    case 2:
>> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M2) |
>> +                                (env->active_fpu.fcsr0 & ~FCSR0_M2);
>> +        break;
>> +    case 3:
>> +        env->active_fpu.fcsr0 = (arg1 & FCSR0_M3) |
>> +                                (env->active_fpu.fcsr0 & ~FCSR0_M3);
>> +        break;
> 
> This is easily implemented inline, followed by a single helper call to re-load the rounding mode (if required by the mask).

Hi, Richard, 

Sorry to bother you, When I was revising this patch, I found that I didn't seem to understand your opinion. 
Could you explain it in detail?  thank you very much.

Thanks
Song Gao.
Richard Henderson Aug. 12, 2021, 7:31 p.m. UTC | #4
On 8/11/21 11:20 PM, Song Gao wrote:
>> This is easily implemented inline, followed by a single helper call to re-load the rounding mode (if required by the mask).
> 
> Hi, Richard,
>
> Sorry to bother you, When I was revising this patch, I found that I didn't seem to
> understand your opinion. Could you explain it in detail?  thank you very much.
---%<

static const uint32_t fcsr_mask[4] = {
     UINT32_MAX, FCSR0_M1, FCSR0_M2, FCSR0_M3
};

bool trans_movgr2fcsr(DisasContext *s, arg_movgr2fcsr *a)
{
     uint32_t mask = fcsr_mask[a->fcsr];

     if (mask == UINT32_MAX) {
         tcg_gen_extrl_i64_i32(fpu_fscr0, get_gpr(a->rj));
     } else {
         TCGv_i32 tmp = tcg_temp_new_i32();

         tcg_gen_extrl_i64_i32(tmp, get_gpr(a->rj));
         tcg_gen_andi_i32(tmp, tmp, mask);
         tcg_gen_andi_i32(fpu_fcsr0, cpu_fcsr0, ~mask);
         tcg_gen_or_i32(fpu_fcsr0, fpu_fcsr0, tmp);
         tcg_temp_free_i32(tmp);

         /*
          * Install the new rounding mode to fpu_status, if changed.
          * Note that FCSR3 is exactly the rounding mode field.
          */
         if (mask != FCSR0_M3) {
             return true;
         }
     }
     gen_helper_set_rounding_mode(cpu_env, fpu_fcsr0);
     return true;
}

void trans_movfcsr2gr(DisasContext *s, arg_movfcsr2gr *a)
{
     TCGv_i32 tmp = tcg_temp_new_i32();

     tcg_gen_andi_i32(tmp, fpu_fcsr0, fcsr_mask[a->fcsr]);
     tcg_gen_ext_i32_i64(dest_gpr(a->rd), tmp);
     tcg_temp_free_i32(tmp);
     return true;
}

---%<

DEF_HELPER_FLAGS_2(set_rounding_mode, TCG_CALL_NO_RWG, void, env, i32)

---%<

void HELPER(set_rounding_mode)(CPULoongArchState *env, uint32_t fcsr)
{
     set_float_rounding_mode(ieee_rm[(fcsr0 >> FCSR0_RM) & 0x3],
                             &env->fp_status);
}


r~
diff mbox series

Patch

diff --git a/target/loongarch/fpu_helper.c b/target/loongarch/fpu_helper.c
index 162085a..7662715 100644
--- a/target/loongarch/fpu_helper.c
+++ b/target/loongarch/fpu_helper.c
@@ -379,6 +379,11 @@  uint64_t helper_fp_logb_d(CPULoongArchState *env, uint64_t fp)
     return fp1;
 }
 
+void helper_movreg2cf(CPULoongArchState *env, uint32_t cd, target_ulong src)
+{
+    env->active_fpu.cf[cd & 0x7] = src & 0x1;
+}
+
 void helper_movreg2cf_i32(CPULoongArchState *env, uint32_t cd, uint32_t src)
 {
     env->active_fpu.cf[cd & 0x7] = src & 0x1;
@@ -1353,3 +1358,78 @@  uint64_t helper_fp_rint_d(CPULoongArchState *env, uint64_t src)
     update_fcsr0(env, GETPC());
     return dest;
 }
+
+target_ulong helper_fsel(CPULoongArchState *env, target_ulong fj,
+                         target_ulong fk, uint32_t ca)
+{
+    if (env->active_fpu.cf[ca & 0x7]) {
+        return fk;
+    } else {
+        return fj;
+    }
+}
+
+void helper_movgr2fcsr(CPULoongArchState *env, target_ulong arg1,
+                       uint32_t fcsr)
+{
+    switch (fcsr) {
+    case 0:
+        env->active_fpu.fcsr0 = arg1;
+        break;
+    case 1:
+        env->active_fpu.fcsr0 = (arg1 & FCSR0_M1) |
+                                (env->active_fpu.fcsr0 & ~FCSR0_M1);
+        break;
+    case 2:
+        env->active_fpu.fcsr0 = (arg1 & FCSR0_M2) |
+                                (env->active_fpu.fcsr0 & ~FCSR0_M2);
+        break;
+    case 3:
+        env->active_fpu.fcsr0 = (arg1 & FCSR0_M3) |
+                                (env->active_fpu.fcsr0 & ~FCSR0_M3);
+        break;
+    case 16:
+        env->active_fpu.vcsr16 = arg1;
+        break;
+    default:
+        printf("%s: warning, fcsr '%d' not supported\n", __func__, fcsr);
+        assert(0);
+        break;
+    }
+    restore_fp_status(env);
+    set_float_exception_flags(0, &env->active_fpu.fp_status);
+}
+
+target_ulong helper_movfcsr2gr(CPULoongArchState *env, uint32_t reg)
+{
+    target_ulong r = 0;
+
+    switch (reg) {
+    case 0:
+        r = (uint32_t)env->active_fpu.fcsr0;
+        break;
+    case 1:
+        r = (env->active_fpu.fcsr0 & FCSR0_M1);
+        break;
+    case 2:
+        r = (env->active_fpu.fcsr0 & FCSR0_M2);
+        break;
+    case 3:
+        r = (env->active_fpu.fcsr0 & FCSR0_M3);
+        break;
+    case 16:
+        r = (uint32_t)env->active_fpu.vcsr16;
+        break;
+    default:
+        printf("%s: warning, fcsr '%d' not supported\n", __func__, reg);
+        assert(0);
+        break;
+    }
+
+    return r;
+}
+
+target_ulong helper_movcf2reg(CPULoongArchState *env, uint32_t cj)
+{
+    return (target_ulong)env->active_fpu.cf[cj & 0x7];
+}
diff --git a/target/loongarch/helper.h b/target/loongarch/helper.h
index 9ec2b53..eedf174 100644
--- a/target/loongarch/helper.h
+++ b/target/loongarch/helper.h
@@ -150,3 +150,9 @@  DEF_HELPER_2(fp_tint_l_s, i64, env, i32)
 DEF_HELPER_2(fp_tint_l_d, i64, env, i64)
 DEF_HELPER_2(fp_tint_w_s, i32, env, i32)
 DEF_HELPER_2(fp_tint_w_d, i32, env, i64)
+
+DEF_HELPER_4(fsel, i64, env, i64, i64, i32)
+DEF_HELPER_3(movreg2cf, void, env, i32, tl)
+DEF_HELPER_2(movcf2reg, tl, env, i32)
+DEF_HELPER_2(movfcsr2gr, tl, env, i32)
+DEF_HELPER_3(movgr2fcsr, void, env, tl, i32)
diff --git a/target/loongarch/insns.decode b/target/loongarch/insns.decode
index c6fd762..febf89a 100644
--- a/target/loongarch/insns.decode
+++ b/target/loongarch/insns.decode
@@ -34,6 +34,10 @@ 
 %fa      15:5
 %cd      0:3
 %fcond   15:5
+%cj      5:3
+%ca      15:3
+%fcsrd   0:5
+%fcsrs   5:5
 
 #
 # Argument sets
@@ -59,6 +63,15 @@ 
 &fmt_fdfjfkfa       fd fj fk fa
 &fmt_fdfj           fd fj
 &fmt_cdfjfkfcond    cd fj fk fcond
+&fmt_fdfjfkca       fd fj fk ca
+&fmt_fdrj           fd rj
+&fmt_rdfj           rd fj
+&fmt_fcsrdrj        fcsrd rj
+&fmt_rdfcsrs        rd fcsrs
+&fmt_cdfj           cd fj
+&fmt_fdcj           fd cj
+&fmt_cdrj           cd rj
+&fmt_rdcj           rd cj
 
 #
 # Formats
@@ -84,6 +97,15 @@ 
 @fmt_fdfjfkfa        .... ........ ..... ..... ..... .....    &fmt_fdfjfkfa       %fd %fj %fk %fa
 @fmt_fdfj            .... ........ ..... ..... ..... .....    &fmt_fdfj           %fd %fj
 @fmt_cdfjfkfcond     .... ........ ..... ..... ..... .. ...   &fmt_cdfjfkfcond    %cd %fj %fk %fcond
+@fmt_fdfjfkca        .... ........ .. ... ..... ..... .....   &fmt_fdfjfkca       %fd %fj %fk %ca
+@fmt_fdrj            .... ........ ..... ..... ..... .....    &fmt_fdrj           %fd %rj
+@fmt_rdfj            .... ........ ..... ..... ..... .....    &fmt_rdfj           %rd %fj
+@fmt_fcsrdrj         .... ........ ..... ..... ..... .....    &fmt_fcsrdrj        %fcsrd %rj
+@fmt_rdfcsrs         .... ........ ..... ..... ..... .....    &fmt_rdfcsrs        %rd %fcsrs
+@fmt_cdfj            .... ........ ..... ..... ..... .. ...   &fmt_cdfj           %cd %fj
+@fmt_fdcj            .... ........ ..... ..... .. ... .....   &fmt_fdcj           %fd %cj
+@fmt_cdrj            .... ........ ..... ..... ..... .. ...   &fmt_cdrj           %cd %rj
+@fmt_rdcj            .... ........ ..... ..... .. ... .....   &fmt_rdcj           %rd %cj
 
 #
 # Fixed point arithmetic operation instruction
@@ -383,3 +405,22 @@  ffint_d_w        0000 00010001 11010 01000 ..... .....    @fmt_fdfj
 ffint_d_l        0000 00010001 11010 01010 ..... .....    @fmt_fdfj
 frint_s          0000 00010001 11100 10001 ..... .....    @fmt_fdfj
 frint_d          0000 00010001 11100 10010 ..... .....    @fmt_fdfj
+
+#
+# Floating point move instruction
+#
+fmov_s           0000 00010001 01001 00101 ..... .....    @fmt_fdfj
+fmov_d           0000 00010001 01001 00110 ..... .....    @fmt_fdfj
+fsel             0000 11010000 00 ... ..... ..... .....   @fmt_fdfjfkca
+movgr2fr_w       0000 00010001 01001 01001 ..... .....    @fmt_fdrj
+movgr2fr_d       0000 00010001 01001 01010 ..... .....    @fmt_fdrj
+movgr2frh_w      0000 00010001 01001 01011 ..... .....    @fmt_fdrj
+movfr2gr_s       0000 00010001 01001 01101 ..... .....    @fmt_rdfj
+movfr2gr_d       0000 00010001 01001 01110 ..... .....    @fmt_rdfj
+movfrh2gr_s      0000 00010001 01001 01111 ..... .....    @fmt_rdfj
+movgr2fcsr       0000 00010001 01001 10000 ..... .....    @fmt_fcsrdrj
+movfcsr2gr       0000 00010001 01001 10010 ..... .....    @fmt_rdfcsrs
+movfr2cf         0000 00010001 01001 10100 ..... 00 ...   @fmt_cdfj
+movcf2fr         0000 00010001 01001 10101 00 ... .....   @fmt_fdcj
+movgr2cf         0000 00010001 01001 10110 ..... 00 ...   @fmt_cdrj
+movcf2gr         0000 00010001 01001 10111 00 ... .....   @fmt_rdcj
diff --git a/target/loongarch/trans.inc.c b/target/loongarch/trans.inc.c
index aa9920e..56677f8 100644
--- a/target/loongarch/trans.inc.c
+++ b/target/loongarch/trans.inc.c
@@ -4758,3 +4758,273 @@  static bool trans_frint_d(DisasContext *ctx, arg_frint_d *a)
 
     return true;
 }
+
+/* Floating point move instruction translation */
+static bool trans_fmov_s(DisasContext *ctx, arg_fmov_s *a)
+{
+    TCGv_i32 fp0;
+
+    fp0 = tcg_temp_new_i32();
+
+    check_fpu_enabled(ctx);
+    gen_load_fpr32(fp0, a->fj);
+    gen_store_fpr32(fp0, a->fd);
+
+    tcg_temp_free_i32(fp0);
+
+    return true;
+}
+
+static bool trans_fmov_d(DisasContext *ctx, arg_fmov_d *a)
+{
+    TCGv_i64 fp0;
+
+    fp0 = tcg_temp_new_i64();
+
+    check_fpu_enabled(ctx);
+    gen_load_fpr64(fp0, a->fj);
+    gen_store_fpr64(fp0, a->fd);
+
+    tcg_temp_free_i64(fp0);
+
+    return true;
+}
+
+static bool trans_fsel(DisasContext *ctx, arg_fsel *a)
+{
+    TCGv_i64 fj, fk, fd;
+    TCGv_i32 ca;
+
+    fj = tcg_temp_new_i64();
+    fk = tcg_temp_new_i64();
+    fd = tcg_temp_new_i64();
+    ca = tcg_const_i32(a->ca);
+
+    check_fpu_enabled(ctx);
+    gen_load_fpr64(fj, a->fj);
+    gen_load_fpr64(fk, a->fk);
+    gen_helper_fsel(fd, cpu_env, fj, fk, ca);
+    gen_store_fpr64(fd, a->fd);
+
+    tcg_temp_free_i64(fj);
+    tcg_temp_free_i64(fk);
+    tcg_temp_free_i64(fd);
+    tcg_temp_free_i32(ca);
+
+    return true;
+}
+
+static bool trans_movgr2fr_w(DisasContext *ctx, arg_movgr2fr_w *a)
+{
+    TCGv t0;
+    TCGv_i32 fp0;
+
+    t0 = get_gpr(a->rj);
+    fp0 = tcg_temp_new_i32();
+
+    check_fpu_enabled(ctx);
+    tcg_gen_trunc_tl_i32(fp0, t0);
+    gen_store_fpr32(fp0, a->fd);
+
+    tcg_temp_free_i32(fp0);
+
+    return true;
+}
+
+static bool trans_movgr2fr_d(DisasContext *ctx, arg_movgr2fr_d *a)
+{
+    TCGv t0;
+
+    t0 = get_gpr(a->rj);
+
+    check_fpu_enabled(ctx);
+    gen_store_fpr64(t0, a->fd);
+
+    return true;
+}
+
+static bool trans_movgr2frh_w(DisasContext *ctx, arg_movgr2frh_w *a)
+{
+    TCGv t0;
+    TCGv_i32 fp0;
+
+    t0 = get_gpr(a->rj);
+    fp0 = tcg_temp_new_i32();
+
+    check_fpu_enabled(ctx);
+    tcg_gen_trunc_tl_i32(fp0, t0);
+    gen_store_fpr32h(fp0, a->fd);
+
+    tcg_temp_free_i32(fp0);
+
+    return true;
+}
+
+static bool trans_movfr2gr_s(DisasContext *ctx, arg_movfr2gr_s *a)
+{
+    TCGv t0;
+    TCGv_i32 fp0;
+    TCGv Rd = cpu_gpr[a->rd];
+
+    if (a->rd == 0) {
+        /* Nop */
+        return true;
+    }
+
+    t0 = tcg_temp_new();
+    fp0 = tcg_temp_new_i32();
+
+    check_fpu_enabled(ctx);
+    gen_load_fpr32(fp0, a->fj);
+    tcg_gen_ext_i32_tl(t0, fp0);
+    tcg_gen_mov_tl(Rd, t0);
+
+    tcg_temp_free(t0);
+    tcg_temp_free_i32(fp0);
+
+    return true;
+}
+
+static bool trans_movfr2gr_d(DisasContext *ctx, arg_movfr2gr_d *a)
+{
+    TCGv t0;
+    TCGv Rd = cpu_gpr[a->rd];
+
+    if (a->rd == 0) {
+        /* Nop */
+        return true;
+    }
+
+    t0 = tcg_temp_new();
+
+    check_fpu_enabled(ctx);
+    gen_load_fpr64(t0, a->fj);
+    tcg_gen_mov_tl(Rd, t0);
+
+    tcg_temp_free(t0);
+
+    return true;
+}
+
+static bool trans_movfrh2gr_s(DisasContext *ctx, arg_movfrh2gr_s *a)
+{
+    TCGv t0;
+    TCGv_i32 fp0;
+    TCGv Rd = cpu_gpr[a->rd];
+
+    if (a->rd == 0) {
+        /* Nop */
+        return true;
+    }
+
+    t0 = tcg_temp_new();
+    fp0 = tcg_temp_new_i32();
+
+    check_fpu_enabled(ctx);
+    gen_load_fpr32h(fp0, a->fj);
+    tcg_gen_ext_i32_tl(t0, fp0);
+    tcg_gen_mov_tl(Rd, t0);
+
+    tcg_temp_free_i32(fp0);
+    tcg_temp_free(t0);
+
+    return true;
+}
+
+static bool trans_movgr2fcsr(DisasContext *ctx, arg_movgr2fcsr *a)
+{
+    TCGv t0 = tcg_temp_new();
+    TCGv_i32 t1 = tcg_const_i32(a->fcsrd);
+
+    check_fpu_enabled(ctx);
+    gen_load_gpr(t0, a->rj);
+    save_cpu_state(ctx, 0);
+    gen_helper_movgr2fcsr(cpu_env, t0, t1);
+    /* Stop translation as we may have changed hflags */
+    ctx->base.is_jmp = DISAS_STOP;
+
+    tcg_temp_free(t0);
+    tcg_temp_free_i32(t1);
+
+    return true;
+}
+
+static bool trans_movfcsr2gr(DisasContext *ctx, arg_movfcsr2gr *a)
+{
+    TCGv t0;
+    TCGv_i32 t1;
+    TCGv Rd = cpu_gpr[a->rd];
+
+    if (a->rd == 0) {
+        /* Nop */
+        return true;
+    }
+
+    t0 = tcg_temp_new();
+    t1 = tcg_const_i32(a->fcsrs);
+
+    gen_helper_movfcsr2gr(t0, cpu_env, t1);
+    tcg_gen_mov_tl(Rd, t0);
+
+    tcg_temp_free(t0);
+    tcg_temp_free_i32(t1);
+
+    return true;
+}
+
+static bool trans_movfr2cf(DisasContext *ctx, arg_movfr2cf *a)
+{
+    TCGv_i64 fp0 = tcg_temp_new_i64();
+    TCGv_i32 cd  = tcg_const_i32(a->cd);
+
+    check_fpu_enabled(ctx);
+    gen_load_fpr64(fp0, a->fj);
+    gen_helper_movreg2cf(cpu_env, cd, fp0);
+
+    tcg_temp_free_i64(fp0);
+    tcg_temp_free_i32(cd);
+
+    return true;
+}
+
+static bool trans_movcf2fr(DisasContext *ctx, arg_movcf2fr *a)
+{
+    TCGv t0 = tcg_temp_new();
+    TCGv_i32 cj = tcg_const_i32(a->cj);
+
+    check_fpu_enabled(ctx);
+    gen_helper_movcf2reg(t0, cpu_env, cj);
+    gen_store_fpr64(t0, a->fd);
+
+    tcg_temp_free(t0);
+    tcg_temp_free_i32(cj);
+    return true;
+}
+
+static bool trans_movgr2cf(DisasContext *ctx, arg_movgr2cf *a)
+{
+    TCGv t0 = tcg_temp_new();
+    TCGv_i32 cd = tcg_const_i32(a->cd);
+
+    check_fpu_enabled(ctx);
+    gen_load_gpr(t0, a->rj);
+    gen_helper_movreg2cf(cpu_env, cd, t0);
+
+    tcg_temp_free(t0);
+    tcg_temp_free_i32(cd);
+
+    return true;
+}
+
+static bool trans_movcf2gr(DisasContext *ctx, arg_movcf2gr *a)
+{
+    TCGv Rd = cpu_gpr[a->rd];
+    TCGv_i32 cj = tcg_const_i32(a->cj);
+
+    check_fpu_enabled(ctx);
+    gen_helper_movcf2reg(Rd, cpu_env, cj);
+
+    tcg_temp_free_i32(cj);
+
+    return true;
+}