Patchwork [v7,07/14] target-mips-ase-dsp: Add GPR-based shift instructions

login
register
mail settings
Submitter Jia Liu
Date Aug. 28, 2012, 6:36 a.m.
Message ID <1346135785-12119-8-git-send-email-proljc@gmail.com>
Download mbox | patch
Permalink /patch/180351/
State New
Headers show

Comments

Jia Liu - Aug. 28, 2012, 6:36 a.m.
Add MIPS ASE DSP GPR-Based Shift instructions.

Signed-off-by: Jia Liu <proljc@gmail.com>
---
 target-mips/dsp_helper.c |  991 ++++++++++++++++++++++++++++++++++++++++++++++
 target-mips/helper.h     |   62 +++
 target-mips/translate.c  |  359 +++++++++++++++++
 3 files changed, 1412 insertions(+)
Aurelien Jarno - Sept. 6, 2012, 9:11 a.m.
On Tue, Aug 28, 2012 at 02:36:18PM +0800, Jia Liu wrote:
> Add MIPS ASE DSP GPR-Based Shift instructions.
> 
> Signed-off-by: Jia Liu <proljc@gmail.com>
> ---
>  target-mips/dsp_helper.c |  991 ++++++++++++++++++++++++++++++++++++++++++++++
>  target-mips/helper.h     |   62 +++
>  target-mips/translate.c  |  359 +++++++++++++++++
>  3 files changed, 1412 insertions(+)

Acked-by: Aurelien Jarno <aurelien@aurel32.net>

> diff --git a/target-mips/dsp_helper.c b/target-mips/dsp_helper.c
> index 486b22f..8e7bccb 100644
> --- a/target-mips/dsp_helper.c
> +++ b/target-mips/dsp_helper.c
> @@ -3196,6 +3196,997 @@ target_ulong helper_preceu_qh_obra(target_ulong rt)
>  }
>  #endif
>  
> +/** DSP GPR-Based Shift Sub-class insns **/
> +target_ulong helper_shll_qb(CPUMIPSState *env, uint32_t sa, target_ulong rt)
> +{
> +    uint8_t  rt3, rt2, rt1, rt0;
> +    uint8_t  tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rt3 = (rt & MIPSDSP_Q3) >> 24;
> +    rt2 = (rt & MIPSDSP_Q2) >> 16;
> +    rt1 = (rt & MIPSDSP_Q1) >>  8;
> +    rt0 =  rt & MIPSDSP_Q0;
> +
> +    tempD = mipsdsp_lshift8(env, rt3, sa);
> +    tempC = mipsdsp_lshift8(env, rt2, sa);
> +    tempB = mipsdsp_lshift8(env, rt1, sa);
> +    tempA = mipsdsp_lshift8(env, rt0, sa);
> +    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
> +         ((uint32_t)tempB <<  8) | ((uint32_t)tempA);
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shllv_qb(CPUMIPSState *env,
> +                             target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs2_0;
> +    uint8_t  rt3, rt2, rt1, rt0;
> +    uint8_t  tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rs2_0 = rs & 0x07;
> +    rt3   = (rt & MIPSDSP_Q3) >> 24;
> +    rt2   = (rt & MIPSDSP_Q2) >> 16;
> +    rt1   = (rt & MIPSDSP_Q1) >>  8;
> +    rt0   =  rt & MIPSDSP_Q0;
> +
> +    tempD = mipsdsp_lshift8(env, rt3, rs2_0);
> +    tempC = mipsdsp_lshift8(env, rt2, rs2_0);
> +    tempB = mipsdsp_lshift8(env, rt1, rs2_0);
> +    tempA = mipsdsp_lshift8(env, rt0, rs2_0);
> +
> +    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
> +         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +#if defined(TARGET_MIPS64)
> +target_ulong helper_shll_ob(CPUMIPSState *env, target_ulong rt, uint32_t sa)
> +{
> +    int i;
> +    uint8_t rt_t[8];
> +    uint8_t temp_t[8];
> +    uint64_t temp;
> +
> +    sa = sa & 0x07;
> +    temp = 0;
> +
> +    for (i = 0; i < 8; i++) {
> +        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +        temp_t[i] = mipsdsp_lshift8(env, rt_t[i], sa);
> +
> +        temp |= (uint64_t)temp_t[i] << (8 * i);
> +    }
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shllv_ob(CPUMIPSState *env,
> +                             target_ulong rt, target_ulong sa)
> +{
> +    int i;
> +    uint8_t rt_t[8];
> +    uint8_t temp_t[8];
> +    uint64_t temp;
> +
> +    sa = sa & 0x07;
> +    temp = 0;
> +
> +    for (i = 0; i < 8; i++) {
> +        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +        temp_t[i] = mipsdsp_lshift8(env, rt_t[i], sa);
> +
> +        temp |= (uint64_t)temp_t[i] << (8 * i);
> +    }
> +
> +    return temp;
> +}
> +#endif
> +
> +target_ulong helper_shll_ph(CPUMIPSState *env, uint32_t sa, target_ulong rt)
> +{
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rth = (rt & MIPSDSP_HI) >> 16;
> +    rtl =  rt & MIPSDSP_LO;
> +    tempB = mipsdsp_lshift16(env, rth, sa);
> +    tempA = mipsdsp_lshift16(env, rtl, sa);
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shllv_ph(CPUMIPSState *env,
> +                             target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs3_0;
> +    uint16_t rth, rtl, tempB, tempA;
> +    uint32_t rd;
> +
> +    rth   = (rt & MIPSDSP_HI) >> 16;
> +    rtl   =  rt & MIPSDSP_LO;
> +    rs3_0 = rs & 0x0F;
> +
> +    tempB = mipsdsp_lshift16(env, rth, rs3_0);
> +    tempA = mipsdsp_lshift16(env, rtl, rs3_0);
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shll_s_ph(CPUMIPSState *env, uint32_t sa, target_ulong rt)
> +{
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rth = (rt & MIPSDSP_HI) >> 16;
> +    rtl =  rt & MIPSDSP_LO;
> +    tempB = mipsdsp_sat16_lshift(env, rth, sa);
> +    tempA = mipsdsp_sat16_lshift(env, rtl, sa);
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shllv_s_ph(CPUMIPSState *env,
> +                               target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs3_0;
> +    uint16_t rth, rtl, tempB, tempA;
> +    uint32_t rd;
> +
> +    rth   = (rt & MIPSDSP_HI) >> 16;
> +    rtl   =  rt & MIPSDSP_LO;
> +    rs3_0 = rs & 0x0F;
> +
> +    tempB = mipsdsp_sat16_lshift(env, rth, rs3_0);
> +    tempA = mipsdsp_sat16_lshift(env, rtl, rs3_0);
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +#if defined(TARGET_MIPS64)
> +target_ulong helper_shll_qh(CPUMIPSState *env, target_ulong rt, uint32_t sa)
> +{
> +    uint16_t rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x0F;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_lshift16(env, rt3, sa);
> +    tempC = mipsdsp_lshift16(env, rt2, sa);
> +    tempB = mipsdsp_lshift16(env, rt1, sa);
> +    tempA = mipsdsp_lshift16(env, rt0, sa);
> +
> +    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +           ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shllv_qh(CPUMIPSState *env,
> +                             target_ulong rt, target_ulong sa)
> +{
> +    uint16_t rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x0F;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_lshift16(env, rt3, sa);
> +    tempC = mipsdsp_lshift16(env, rt2, sa);
> +    tempB = mipsdsp_lshift16(env, rt1, sa);
> +    tempA = mipsdsp_lshift16(env, rt0, sa);
> +
> +    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +           ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shll_s_qh(CPUMIPSState *env, target_ulong rt, uint32_t sa)
> +{
> +    uint16_t rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x0F;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_sat16_lshift(env, rt3, sa);
> +    tempC = mipsdsp_sat16_lshift(env, rt2, sa);
> +    tempB = mipsdsp_sat16_lshift(env, rt1, sa);
> +    tempA = mipsdsp_sat16_lshift(env, rt0, sa);
> +
> +    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +           ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shllv_s_qh(CPUMIPSState *env,
> +                               target_ulong rt, target_ulong sa)
> +{
> +    uint16_t rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x0F;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_sat16_lshift(env, rt3, sa);
> +    tempC = mipsdsp_sat16_lshift(env, rt2, sa);
> +    tempB = mipsdsp_sat16_lshift(env, rt1, sa);
> +    tempA = mipsdsp_sat16_lshift(env, rt0, sa);
> +
> +    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +           ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +#endif
> +
> +target_ulong helper_shll_s_w(CPUMIPSState *env, uint32_t sa, target_ulong rt)
> +{
> +    uint32_t temp, rd;
> +
> +    temp = mipsdsp_sat32_lshift(env, rt, sa);
> +    rd   = temp;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shllv_s_w(CPUMIPSState *env,
> +                              target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs4_0;
> +    uint32_t rd;
> +
> +    rs4_0 = rs & 0x1F;
> +    rd = mipsdsp_sat32_lshift(env, rt, rs4_0);
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +#if defined(TARGET_MIPS64)
> +target_ulong helper_shll_pw(CPUMIPSState *env, target_ulong rt, uint32_t sa)
> +{
> +    uint32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t temp;
> +
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_lshift32(env, rt1, sa);
> +    tempA = mipsdsp_lshift32(env, rt0, sa);
> +    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shllv_pw(CPUMIPSState *env,
> +                             target_ulong rt, target_ulong sa)
> +{
> +    uint32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x1F;
> +
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_lshift32(env, rt1, sa);
> +    tempA = mipsdsp_lshift32(env, rt0, sa);
> +    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shll_s_pw(CPUMIPSState *env, target_ulong rt, uint32_t sa)
> +{
> +    uint32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t temp;
> +
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_sat32_lshift(env, rt1, sa);
> +    tempA = mipsdsp_sat32_lshift(env, rt0, sa);
> +    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shllv_s_pw(CPUMIPSState *env,
> +                               target_ulong rt, target_ulong sa)
> +{
> +    uint32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x1F;
> +
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_sat32_lshift(env, rt1, sa);
> +    tempA = mipsdsp_sat32_lshift(env, rt0, sa);
> +    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +#endif
> +
> +target_ulong helper_shrl_qb(uint32_t sa, target_ulong rt)
> +{
> +    uint8_t  rt3, rt2, rt1, rt0;
> +    uint8_t  tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rt3 = (rt & MIPSDSP_Q3) >> 24;
> +    rt2 = (rt & MIPSDSP_Q2) >> 16;
> +    rt1 = (rt & MIPSDSP_Q1) >>  8;
> +    rt0 =  rt & MIPSDSP_Q0;
> +
> +    tempD = mipsdsp_rshift8(rt3, sa);
> +    tempC = mipsdsp_rshift8(rt2, sa);
> +    tempB = mipsdsp_rshift8(rt1, sa);
> +    tempA = mipsdsp_rshift8(rt0, sa);
> +
> +    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
> +         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrlv_qb(target_ulong rs, target_ulong rt)
> +{
> +    uint8_t rs2_0;
> +    uint8_t rt3, rt2, rt1, rt0;
> +    uint8_t tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rs2_0 = rs & 0x07;
> +    rt3   = (rt & MIPSDSP_Q3) >> 24;
> +    rt2   = (rt & MIPSDSP_Q2) >> 16;
> +    rt1   = (rt & MIPSDSP_Q1) >>  8;
> +    rt0   =  rt & MIPSDSP_Q0;
> +
> +    tempD = mipsdsp_rshift8(rt3, rs2_0);
> +    tempC = mipsdsp_rshift8(rt2, rs2_0);
> +    tempB = mipsdsp_rshift8(rt1, rs2_0);
> +    tempA = mipsdsp_rshift8(rt0, rs2_0);
> +    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
> +         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrl_ph(uint32_t sa, target_ulong rt)
> +{
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rth = (rt & MIPSDSP_HI) >> 16;
> +    rtl =  rt & MIPSDSP_LO;
> +    tempB = rth >> sa;
> +    tempA = rtl >> sa;
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrlv_ph(target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs3_0;
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rs3_0 = rs & 0x0F;
> +    rth   = (rt & MIPSDSP_HI) >> 16;
> +    rtl   =  rt & MIPSDSP_LO;
> +
> +    tempB = rth >> rs3_0;
> +    tempA = rtl >> rs3_0;
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +#if defined(TARGET_MIPS64)
> +target_ulong helper_shrl_ob(target_ulong rt, uint32_t sa)
> +{
> +    int i;
> +    uint8_t rt_t[8];
> +    uint8_t temp[8];
> +    uint64_t result;
> +
> +    sa = sa & 0x07;
> +    result = 0;
> +
> +    for (i = 0; i < 8; i++) {
> +        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +        temp[i] = mipsdsp_rshift8(rt_t[i], sa);
> +        result |= (uint64_t)temp[i] << (8 * i);
> +    }
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrlv_ob(target_ulong rt, target_ulong sa)
> +{
> +    int i;
> +    uint8_t rt_t[8];
> +    uint8_t temp[8];
> +    uint64_t result;
> +
> +    sa = sa & 0x07;
> +    result = 0;
> +
> +    for (i = 0; i < 8; i++) {
> +        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +        temp[i] = mipsdsp_rshift8(rt_t[i], sa);
> +        result |= (uint64_t)temp[i] << (8 * i);
> +    }
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrl_qh(target_ulong rt, uint32_t sa)
> +{
> +    uint16_t rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x0F;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = rt3 >> sa;
> +    tempC = rt2 >> sa;
> +    tempB = rt1 >> sa;
> +    tempA = rt0 >> sa;
> +
> +    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +           ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +
> +target_ulong helper_shrlv_qh(target_ulong rt, target_ulong sa)
> +{
> +    uint16_t rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint64_t temp;
> +
> +    sa = sa & 0x0F;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = rt3 >> sa;
> +    tempC = rt2 >> sa;
> +    tempB = rt1 >> sa;
> +    tempA = rt0 >> sa;
> +
> +    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +           ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return temp;
> +}
> +#endif
> +
> +target_ulong helper_shra_qb(uint32_t sa, target_ulong rt)
> +{
> +    int8_t  rt3, rt2, rt1, rt0;
> +    uint8_t tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rt3 = (rt & MIPSDSP_Q3) >> 24;
> +    rt2 = (rt & MIPSDSP_Q2) >> 16;
> +    rt1 = (rt & MIPSDSP_Q1) >>  8;
> +    rt0 =  rt & MIPSDSP_Q0;
> +
> +    tempD = rt3 >> sa;
> +    tempC = rt2 >> sa;
> +    tempB = rt1 >> sa;
> +    tempA = rt0 >> sa;
> +
> +    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
> +          ((uint32_t)tempB << 8) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shra_r_qb(uint32_t sa, target_ulong rt)
> +{
> +    int8_t  rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rt3 = (rt & MIPSDSP_Q3) >> 24;
> +    rt2 = (rt & MIPSDSP_Q2) >> 16;
> +    rt1 = (rt & MIPSDSP_Q1) >>  8;
> +    rt0 =  rt & MIPSDSP_Q0;
> +
> +    if (sa == 0) {
> +        tempD = rt3 & 0x00FF;
> +        tempC = rt2 & 0x00FF;
> +        tempB = rt1 & 0x00FF;
> +        tempA = rt0 & 0x00FF;
> +    } else {
> +        tempD = ((int16_t)rt3 >> (sa - 1)) + 1;
> +        tempC = ((int16_t)rt2 >> (sa - 1)) + 1;
> +        tempB = ((int16_t)rt1 >> (sa - 1)) + 1;
> +        tempA = ((int16_t)rt0 >> (sa - 1)) + 1;
> +    }
> +
> +    rd = ((uint32_t)((tempD >> 1) & 0x00FF) << 24) |
> +         ((uint32_t)((tempC >> 1) & 0x00FF) << 16) |
> +         ((uint32_t)((tempB >> 1) & 0x00FF) <<  8) |
> +         (uint32_t)((tempA >>  1) & 0x00FF) ;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrav_qb(target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs2_0;
> +    int8_t   rt3, rt2, rt1, rt0;
> +    uint8_t  tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rs2_0 = rs & 0x07;
> +    rt3   = (rt & MIPSDSP_Q3) >> 24;
> +    rt2   = (rt & MIPSDSP_Q2) >> 16;
> +    rt1   = (rt & MIPSDSP_Q1) >>  8;
> +    rt0   =  rt & MIPSDSP_Q0;
> +
> +    if (rs2_0 == 0) {
> +        tempD = rt3;
> +        tempC = rt2;
> +        tempB = rt1;
> +        tempA = rt0;
> +    } else {
> +        tempD = rt3 >> rs2_0;
> +        tempC = rt2 >> rs2_0;
> +        tempB = rt1 >> rs2_0;
> +        tempA = rt0 >> rs2_0;
> +    }
> +
> +    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
> +         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrav_r_qb(target_ulong rs, target_ulong rt)
> +{
> +    uint8_t rs2_0;
> +    int8_t  rt3, rt2, rt1, rt0;
> +    uint16_t tempD, tempC, tempB, tempA;
> +    uint32_t rd;
> +
> +    rs2_0 = rs & 0x07;
> +    rt3 = (rt & MIPSDSP_Q3) >> 24;
> +    rt2 = (rt & MIPSDSP_Q2) >> 16;
> +    rt1 = (rt & MIPSDSP_Q1) >>  8;
> +    rt0 =  rt & MIPSDSP_Q0;
> +
> +    if (rs2_0 == 0) {
> +        tempD = (int16_t)rt3 << 1;
> +        tempC = (int16_t)rt2 << 1;
> +        tempB = (int16_t)rt1 << 1;
> +        tempA = (int16_t)rt0 << 1;
> +    } else {
> +        tempD = ((int16_t)rt3 >> (rs2_0 - 1)) + 1;
> +        tempC = ((int16_t)rt2 >> (rs2_0 - 1)) + 1;
> +        tempB = ((int16_t)rt1 >> (rs2_0 - 1)) + 1;
> +        tempA = ((int16_t)rt0 >> (rs2_0 - 1)) + 1;
> +    }
> +
> +    rd = ((uint32_t)((tempD >> 1) & 0x00FF) << 24) |
> +         ((uint32_t)((tempC >> 1) & 0x00FF) << 16) |
> +         ((uint32_t)((tempB >> 1) & 0x00FF) <<  8) |
> +         (uint32_t)((tempA >>  1) & 0x00FF) ;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +#if defined(TARGET_MIPS64)
> +target_ulong helper_shra_ob(target_ulong rt, uint32_t sa)
> +{
> +    int i;
> +    int8_t rt_t[8];
> +    uint8_t temp[8];
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0x07;
> +
> +    for (i = 0; i < 8; i++) {
> +        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +        temp[i] = rt_t[i] >> sa;
> +        result |= (uint64_t)temp[i] << (8 * i);
> +    }
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrav_ob(target_ulong rt, target_ulong sa)
> +{
> +    int i;
> +    int8_t rt_t[8];
> +    uint8_t temp[8];
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0x07;
> +
> +    for (i = 0; i < 8; i++) {
> +        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +        temp[i] = rt_t[i] >> sa;
> +        result |= (uint64_t)temp[i] << (8 * i);
> +      }
> +
> +    return result;
> +}
> +
> +target_ulong helper_shra_r_ob(target_ulong rt, uint32_t sa)
> +{
> +    int i;
> +    int8_t rt_t[8];
> +    int16_t rt_t_S[8];
> +    uint8_t temp[8];
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0x07;
> +
> +    if (sa == 0) {
> +        result = rt;
> +    } else {
> +        for (i = 0; i < 8; i++) {
> +            rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +            rt_t_S[i] = ((int16_t)rt_t[i] >> (sa - 1)) + 1;
> +            temp[i] = (rt_t_S[i] >> 1) & MIPSDSP_Q0;
> +            result |= (uint64_t)temp[i] << (8 * i);
> +        }
> +    }
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrav_r_ob(target_ulong rt, target_ulong sa)
> +{
> +    int i;
> +    int8_t rt_t[8];
> +    int16_t rt_t_S[8];
> +    uint8_t temp[8];
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0x07;
> +
> +    if (sa == 0) {
> +        result = rt;
> +    } else {
> +        for (i = 0; i < 8; i++) {
> +            rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
> +            rt_t_S[i] = ((int16_t)rt_t[i] >> (sa - 1)) + 1;
> +            temp[i] = (rt_t_S[i] >> 1) & MIPSDSP_Q0;
> +            result |= (uint64_t)temp[i] << (8 * i);
> +        }
> +    }
> +
> +    return result;
> +}
> +#endif
> +
> +target_ulong helper_shra_ph(uint32_t sa, target_ulong rt)
> +{
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rth = (rt & MIPSDSP_HI) >> 16;
> +    rtl =  rt & MIPSDSP_LO;
> +    tempB = mipsdsp_rashift16(rth, sa);
> +    tempA = mipsdsp_rashift16(rtl, sa);
> +    rd = ((uint32_t)tempB << 16) | (uint32_t) tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrav_ph(target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs3_0;
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rs3_0 = rs & 0x0F;
> +    rth   = (rt & MIPSDSP_HI) >> 16;
> +    rtl   =  rt & MIPSDSP_LO;
> +    tempB = mipsdsp_rashift16(rth, rs3_0);
> +    tempA = mipsdsp_rashift16(rtl, rs3_0);
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shra_r_ph(uint32_t sa, target_ulong rt)
> +{
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rth = (rt & MIPSDSP_HI) >> 16;
> +    rtl =  rt & MIPSDSP_LO;
> +    tempB = mipsdsp_rnd16_rashift(rth, sa);
> +    tempA = mipsdsp_rnd16_rashift(rtl, sa);
> +    rd = ((uint32_t)tempB << 16) | (uint32_t) tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrav_r_ph(target_ulong rs, target_ulong rt)
> +{
> +    uint8_t  rs3_0;
> +    uint16_t rth, rtl;
> +    uint16_t tempB, tempA;
> +    uint32_t rd;
> +
> +    rs3_0 = rs & 0x0F;
> +    rth   = (rt & MIPSDSP_HI) >> 16;
> +    rtl   =  rt & MIPSDSP_LO;
> +    tempB = mipsdsp_rnd16_rashift(rth, rs3_0);
> +    tempA = mipsdsp_rnd16_rashift(rtl, rs3_0);
> +
> +    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shra_r_w(uint32_t sa, target_ulong rt)
> +{
> +    uint32_t rd;
> +
> +    rd = mipsdsp_rnd32_rashift(rt, sa);
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +target_ulong helper_shrav_r_w(target_ulong rs, target_ulong rt)
> +{
> +    uint8_t rs4_0;
> +    uint32_t rd;
> +
> +    rs4_0 = rs & 0x1F;
> +    rd = mipsdsp_rnd32_rashift(rt, rs4_0);
> +
> +    return (target_long)(int32_t)rd;
> +}
> +
> +#if defined(TARGET_MIPS64)
> +target_ulong helper_shra_qh(target_ulong rt, uint32_t sa)
> +{
> +    int32_t rt3, rt2, rt1, rt0;
> +    uint32_t tempD, tempC, tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0xF;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_rashift16(rt3, sa);
> +    tempC = mipsdsp_rashift16(rt2, sa);
> +    tempB = mipsdsp_rashift16(rt1, sa);
> +    tempA = mipsdsp_rashift16(rt0, sa);
> +
> +    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +             ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrav_qh(target_ulong rt, target_ulong sa)
> +{
> +    int32_t rt3, rt2, rt1, rt0;
> +    uint32_t tempD, tempC, tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0xF;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_rashift16(rt3, sa);
> +    tempC = mipsdsp_rashift16(rt2, sa);
> +    tempB = mipsdsp_rashift16(rt1, sa);
> +    tempA = mipsdsp_rashift16(rt0, sa);
> +
> +    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +             ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +
> +
> +target_ulong helper_shra_r_qh(target_ulong rt, uint32_t sa)
> +{
> +    int32_t rt3, rt2, rt1, rt0;
> +    uint32_t tempD, tempC, tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0xF;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_rnd16_rashift(rt3, sa);
> +    tempC = mipsdsp_rnd16_rashift(rt2, sa);
> +    tempB = mipsdsp_rnd16_rashift(rt1, sa);
> +    tempA = mipsdsp_rnd16_rashift(rt0, sa);
> +
> +    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +             ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrav_r_qh(target_ulong rt, target_ulong sa)
> +{
> +    int32_t rt3, rt2, rt1, rt0;
> +    uint32_t tempD, tempC, tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0xF;
> +
> +    rt3 = (rt >> 48) & MIPSDSP_LO;
> +    rt2 = (rt >> 32) & MIPSDSP_LO;
> +    rt1 = (rt >> 16) & MIPSDSP_LO;
> +    rt0 = rt & MIPSDSP_LO;
> +
> +    tempD = mipsdsp_rnd16_rashift(rt3, sa);
> +    tempC = mipsdsp_rnd16_rashift(rt2, sa);
> +    tempB = mipsdsp_rnd16_rashift(rt1, sa);
> +    tempA = mipsdsp_rnd16_rashift(rt0, sa);
> +
> +    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
> +             ((uint64_t)tempB << 16) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +
> +target_ulong helper_shra_pw(target_ulong rt, uint32_t sa)
> +{
> +    int32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +
> +    /* FIXME, it seems like document error. */
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_rashift32(rt1, sa);
> +    tempA = mipsdsp_rashift32(rt0, sa);
> +
> +    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrav_pw(target_ulong rt, target_ulong sa)
> +{
> +    int32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0x1F;
> +
> +    /* FIXME, it seems like document error. */
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_rashift32(rt1, sa);
> +    tempA = mipsdsp_rashift32(rt0, sa);
> +
> +    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +
> +target_ulong helper_shra_r_pw(target_ulong rt, uint32_t sa)
> +{
> +    int32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_rnd32_rashift(rt1, sa);
> +    tempA = mipsdsp_rnd32_rashift(rt0, sa);
> +
> +    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +
> +target_ulong helper_shrav_r_pw(target_ulong rt, target_ulong sa)
> +{
> +    int32_t rt1, rt0;
> +    uint32_t tempB, tempA;
> +    uint64_t result;
> +
> +    result = 0;
> +    sa = sa & 0x1F;
> +
> +    rt1 = (rt >> 32) & MIPSDSP_LLO;
> +    rt0 = rt & MIPSDSP_LLO;
> +
> +    tempB = mipsdsp_rnd32_rashift(rt1, sa);
> +    tempA = mipsdsp_rnd32_rashift(rt0, sa);
> +
> +    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
> +
> +    return result;
> +}
> +#endif
> +
>  #undef MIPSDSP_LHI
>  #undef MIPSDSP_LLO
>  #undef MIPSDSP_HI
> diff --git a/target-mips/helper.h b/target-mips/helper.h
> index 0607105..7046380 100644
> --- a/target-mips/helper.h
> +++ b/target-mips/helper.h
> @@ -429,4 +429,66 @@ DEF_HELPER_FLAGS_1(preceu_qh_obla, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl)
>  DEF_HELPER_FLAGS_1(preceu_qh_obra, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl)
>  #endif
>  
> +/* DSP GPR-Based Shift Sub-class insns */
> +DEF_HELPER_FLAGS_3(shll_qb, 0, tl, env, i32, tl)
> +DEF_HELPER_FLAGS_3(shllv_qb, 0, tl, env, tl, tl)
> +#if defined(TARGET_MIPS64)
> +DEF_HELPER_FLAGS_3(shll_ob, 0, tl, env, tl, i32)
> +DEF_HELPER_FLAGS_3(shllv_ob, 0, tl, env, tl, tl)
> +#endif
> +DEF_HELPER_FLAGS_3(shll_ph, 0, tl, env, i32, tl)
> +DEF_HELPER_FLAGS_3(shllv_ph, 0, tl, env, tl, tl)
> +DEF_HELPER_FLAGS_3(shll_s_ph, 0, tl, env, i32, tl)
> +DEF_HELPER_FLAGS_3(shllv_s_ph, 0, tl, env, tl, tl)
> +#if defined(TARGET_MIPS64)
> +DEF_HELPER_FLAGS_3(shll_qh, 0, tl, env, tl, i32)
> +DEF_HELPER_FLAGS_3(shllv_qh, 0, tl, env, tl, tl)
> +DEF_HELPER_FLAGS_3(shll_s_qh, 0, tl, env, tl, i32)
> +DEF_HELPER_FLAGS_3(shllv_s_qh, 0, tl, env, tl, tl)
> +#endif
> +DEF_HELPER_FLAGS_3(shll_s_w, 0, tl, env, i32, tl)
> +DEF_HELPER_FLAGS_3(shllv_s_w, 0, tl, env, tl, tl)
> +#if defined(TARGET_MIPS64)
> +DEF_HELPER_FLAGS_3(shll_pw, 0, tl, env, tl, i32)
> +DEF_HELPER_FLAGS_3(shllv_pw, 0, tl, env, tl, tl)
> +DEF_HELPER_FLAGS_3(shll_s_pw, 0, tl, env, tl, i32)
> +DEF_HELPER_FLAGS_3(shllv_s_pw, 0, tl, env, tl, tl)
> +#endif
> +DEF_HELPER_FLAGS_2(shrl_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
> +DEF_HELPER_FLAGS_2(shrlv_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shrl_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
> +DEF_HELPER_FLAGS_2(shrlv_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +#if defined(TARGET_MIPS64)
> +DEF_HELPER_FLAGS_2(shrl_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrlv_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shrl_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrlv_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +#endif
> +DEF_HELPER_FLAGS_2(shra_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
> +DEF_HELPER_FLAGS_2(shra_r_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
> +DEF_HELPER_FLAGS_2(shrav_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shrav_r_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +#if defined(TARGET_MIPS64)
> +DEF_HELPER_FLAGS_2(shra_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrav_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shra_r_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrav_r_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +#endif
> +DEF_HELPER_FLAGS_2(shra_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
> +DEF_HELPER_FLAGS_2(shrav_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shra_r_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
> +DEF_HELPER_FLAGS_2(shrav_r_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shra_r_w, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
> +DEF_HELPER_FLAGS_2(shrav_r_w, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +#if defined(TARGET_MIPS64)
> +DEF_HELPER_FLAGS_2(shra_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrav_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shra_r_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrav_r_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shra_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrav_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +DEF_HELPER_FLAGS_2(shra_r_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
> +DEF_HELPER_FLAGS_2(shrav_r_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
> +#endif
> +
>  #include "def-helper.h"
> diff --git a/target-mips/translate.c b/target-mips/translate.c
> index 724c162..d87dbee 100644
> --- a/target-mips/translate.c
> +++ b/target-mips/translate.c
> @@ -331,6 +331,18 @@ enum {
>  #if defined(TARGET_MIPS64)
>      OPC_CMPU_EQ_OB_DSP = 0x15 | OPC_SPECIAL3,
>  #endif
> +    /* MIPS DSP GPR-Based Shift Sub-class */
> +    OPC_SHLL_QB_DSP    = 0x13 | OPC_SPECIAL3,
> +#if defined(TARGET_MIPS64)
> +    OPC_SHLL_OB_DSP    = 0x17 | OPC_SPECIAL3,
> +#endif
> +    /* MIPS DSP Multiply Sub-class insns */
> +    /* OPC_MUL_PH_DSP is same as OPC_ADDUH_QB_DSP.  */
> +    /* OPC_MUL_PH_DSP     = 0x18 | OPC_SPECIAL3,  */
> +    OPC_DPA_W_PH_DSP   = 0x30 | OPC_SPECIAL3,
> +#if defined(TARGET_MIPS64)
> +    OPC_DPAQ_W_QH_DSP  = 0x34 | OPC_SPECIAL3,
> +#endif
>  };
>  
>  /* BSHFL opcodes */
> @@ -439,6 +451,32 @@ enum {
>      OPC_PRECRQ_RS_PH_W   = (0x15 << 6) | OPC_CMPU_EQ_QB_DSP,
>      OPC_PRECRQU_S_QB_PH  = (0x0F << 6) | OPC_CMPU_EQ_QB_DSP,
>  };
> +#define MASK_SHLL_QB(op) (MASK_SPECIAL3(op) | (op & (0x1F << 6)))
> +enum {
> +    /* MIPS DSP GPR-Based Shift Sub-class */
> +    OPC_SHLL_QB    = (0x00 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHLLV_QB   = (0x02 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHLL_PH    = (0x08 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHLLV_PH   = (0x0A << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHLL_S_PH  = (0x0C << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHLLV_S_PH = (0x0E << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHLL_S_W   = (0x14 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHLLV_S_W  = (0x16 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRL_QB    = (0x01 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRLV_QB   = (0x03 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRL_PH    = (0x19 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRLV_PH   = (0x1B << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRA_QB    = (0x04 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRA_R_QB  = (0x05 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRAV_QB   = (0x06 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRAV_R_QB = (0x07 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRA_PH    = (0x09 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRAV_PH   = (0x0B << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRA_R_PH  = (0x0D << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRAV_R_PH = (0x0F << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRA_R_W   = (0x15 << 6) | OPC_SHLL_QB_DSP,
> +    OPC_SHRAV_R_W  = (0x17 << 6) | OPC_SHLL_QB_DSP,
> +};
>  
>  #if defined(TARGET_MIPS64)
>  #define MASK_ABSQ_S_QH(op) (MASK_SPECIAL3(op) | (op & (0x1F << 6)))
> @@ -507,6 +545,39 @@ enum {
>  };
>  #endif
>  
> +#if defined(TARGET_MIPS64)
> +#define MASK_SHLL_OB(op) (MASK_SPECIAL3(op) | (op & (0x1F << 6)))
> +enum {
> +    /* MIPS DSP GPR-Based Shift Sub-class */
> +    OPC_SHLL_PW    = (0x10 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLL_S_PW  = (0x14 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLLV_OB   = (0x02 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLLV_PW   = (0x12 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLLV_S_PW = (0x16 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLLV_QH   = (0x0A << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLLV_S_QH = (0x0E << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRA_PW    = (0x11 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRA_R_PW  = (0x15 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRAV_OB   = (0x06 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRAV_R_OB = (0x07 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRAV_PW   = (0x13 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRAV_R_PW = (0x17 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRAV_QH   = (0x0B << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRAV_R_QH = (0x0F << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRLV_OB   = (0x03 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRLV_QH   = (0x1B << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLL_OB    = (0x00 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLL_QH    = (0x08 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHLL_S_QH  = (0x0C << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRA_OB    = (0x04 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRA_R_OB  = (0x05 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRA_QH    = (0x09 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRA_R_QH  = (0x0D << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRL_OB    = (0x01 << 6) | OPC_SHLL_OB_DSP,
> +    OPC_SHRL_QH    = (0x19 << 6) | OPC_SHLL_OB_DSP,
> +};
> +#endif
> +
>  /* Coprocessor 0 (rs field) */
>  #define MASK_CP0(op)       MASK_OP_MAJOR(op) | (op & (0x1F << 21))
>  
> @@ -12660,6 +12731,117 @@ static void decode_opc (CPUMIPSState *env, DisasContext *ctx, int *is_branch)
>                  break;
>              }
>              break;
> +        case OPC_SHLL_QB_DSP:
> +            {
> +                TCGv_i32 temp_rs = tcg_const_i32(rs);
> +                op2 = MASK_SHLL_QB(ctx->opcode);
> +                switch (op2) {
> +                case OPC_SHLL_QB:
> +                    check_dsp(ctx);
> +                    gen_helper_shll_qb(cpu_gpr[rd], cpu_env,
> +                                       temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHLLV_QB:
> +                    check_dsp(ctx);
> +                    gen_helper_shllv_qb(cpu_gpr[rd], cpu_env,
> +                                        cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHLL_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shll_ph(cpu_gpr[rd], cpu_env,
> +                                       temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHLLV_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shllv_ph(cpu_gpr[rd], cpu_env,
> +                                        cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHLL_S_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shll_s_ph(cpu_gpr[rd], cpu_env,
> +                                         temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHLLV_S_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shllv_s_ph(cpu_gpr[rd], cpu_env,
> +                                          cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHLL_S_W:
> +                    check_dsp(ctx);
> +                    gen_helper_shll_s_w(cpu_gpr[rd], cpu_env,
> +                                        temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHLLV_S_W:
> +                    check_dsp(ctx);
> +                    gen_helper_shllv_s_w(cpu_gpr[rd], cpu_env,
> +                                         cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRL_QB:
> +                    check_dsp(ctx);
> +                    gen_helper_shrl_qb(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRLV_QB:
> +                    check_dsp(ctx);
> +                    gen_helper_shrlv_qb(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRL_PH:
> +                    check_dspr2(ctx);
> +                    gen_helper_shrl_ph(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRLV_PH:
> +                    check_dspr2(ctx);
> +                    gen_helper_shrlv_ph(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRA_QB:
> +                    check_dspr2(ctx);
> +                    gen_helper_shra_qb(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRA_R_QB:
> +                    check_dspr2(ctx);
> +                    gen_helper_shra_r_qb(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRAV_QB:
> +                    check_dspr2(ctx);
> +                    gen_helper_shrav_qb(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRAV_R_QB:
> +                    check_dspr2(ctx);
> +                    gen_helper_shrav_r_qb(cpu_gpr[rd], cpu_gpr[rs],
> +                                          cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRA_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shra_ph(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRA_R_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shra_r_ph(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRAV_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shrav_ph(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRAV_R_PH:
> +                    check_dsp(ctx);
> +                    gen_helper_shrav_r_ph(cpu_gpr[rd], cpu_gpr[rs],
> +                                          cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRA_R_W:
> +                    check_dsp(ctx);
> +                    gen_helper_shra_r_w(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
> +                    break;
> +                case OPC_SHRAV_R_W:
> +                    check_dsp(ctx);
> +                    gen_helper_shrav_r_w(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
> +                    break;
> +                default:            /* Invalid */
> +                    MIPS_INVAL("MASK SHLL.QB");
> +                    generate_exception(ctx, EXCP_RI);
> +                    break;
> +                }
> +                tcg_temp_free_i32(temp_rs);
> +                break;
> +            }
>  #if defined(TARGET_MIPS64)
>          case OPC_DEXTM ... OPC_DEXT:
>          case OPC_DINSM ... OPC_DINS:
> @@ -12929,6 +13111,183 @@ static void decode_opc (CPUMIPSState *env, DisasContext *ctx, int *is_branch)
>              }
>              break;
>  #endif
> +#if defined(TARGET_MIPS64)
> +        case OPC_SHLL_OB_DSP:
> +            op2 = MASK_SHLL_OB(ctx->opcode);
> +            switch (op2) {
> +            case OPC_SHLL_PW:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shll_pw(cpu_gpr[rd], cpu_env,
> +                                       cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHLLV_PW:
> +                check_dsp(ctx);
> +                gen_helper_shllv_pw(cpu_gpr[rd], cpu_env,
> +                                    cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHLL_S_PW:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shll_s_pw(cpu_gpr[rd], cpu_env,
> +                                         cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHLLV_S_PW:
> +                check_dsp(ctx);
> +                gen_helper_shllv_s_pw(cpu_gpr[rd], cpu_env,
> +                                      cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHLL_OB:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shll_ob(cpu_gpr[rd], cpu_env,
> +                                       cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHLLV_OB:
> +                check_dsp(ctx);
> +                gen_helper_shllv_ob(cpu_gpr[rd], cpu_env,
> +                                    cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHLL_QH:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shll_qh(cpu_gpr[rd], cpu_env,
> +                                       cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHLLV_QH:
> +                check_dsp(ctx);
> +                gen_helper_shllv_qh(cpu_gpr[rd], cpu_env,
> +                                    cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHLL_S_QH:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shll_s_qh(cpu_gpr[rd], cpu_env,
> +                                         cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHLLV_S_QH:
> +                check_dsp(ctx);
> +                gen_helper_shllv_s_qh(cpu_gpr[rd], cpu_env,
> +                                      cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRA_OB:
> +                check_dspr2(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shra_ob(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRAV_OB:
> +                check_dspr2(ctx);
> +                gen_helper_shrav_ob(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRA_R_OB:
> +                check_dspr2(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shra_r_ob(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRAV_R_OB:
> +                check_dspr2(ctx);
> +                gen_helper_shrav_r_ob(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRA_PW:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shra_pw(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRAV_PW:
> +                check_dsp(ctx);
> +                gen_helper_shrav_pw(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRA_R_PW:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shra_r_pw(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRAV_R_PW:
> +                check_dsp(ctx);
> +                gen_helper_shrav_r_pw(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRA_QH:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shra_qh(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRAV_QH:
> +                check_dsp(ctx);
> +                gen_helper_shrav_qh(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRA_R_QH:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shra_r_qh(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRAV_R_QH:
> +                check_dsp(ctx);
> +                gen_helper_shrav_r_qh(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRL_OB:
> +                check_dsp(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shrl_ob(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRLV_OB:
> +                check_dsp(ctx);
> +                gen_helper_shrlv_ob(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            case OPC_SHRL_QH:
> +                check_dspr2(ctx);
> +                {
> +                    TCGv_i32 sa_v = tcg_const_i32(rs);
> +                    gen_helper_shrl_qh(cpu_gpr[rd], cpu_gpr[rt], sa_v);
> +                    tcg_temp_free_i32(sa_v);
> +                    break;
> +                }
> +            case OPC_SHRLV_QH:
> +                check_dspr2(ctx);
> +                gen_helper_shrlv_qh(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
> +                break;
> +            default:            /* Invalid */
> +                MIPS_INVAL("MASK SHLL.OB");
> +                generate_exception(ctx, EXCP_RI);
> +                break;
> +            }
> +          break;
> +#endif
>          default:            /* Invalid */
>              MIPS_INVAL("special3");
>              generate_exception(ctx, EXCP_RI);
> -- 
> 1.7.9.5
> 
>

Patch

diff --git a/target-mips/dsp_helper.c b/target-mips/dsp_helper.c
index 486b22f..8e7bccb 100644
--- a/target-mips/dsp_helper.c
+++ b/target-mips/dsp_helper.c
@@ -3196,6 +3196,997 @@  target_ulong helper_preceu_qh_obra(target_ulong rt)
 }
 #endif
 
+/** DSP GPR-Based Shift Sub-class insns **/
+target_ulong helper_shll_qb(CPUMIPSState *env, uint32_t sa, target_ulong rt)
+{
+    uint8_t  rt3, rt2, rt1, rt0;
+    uint8_t  tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rt3 = (rt & MIPSDSP_Q3) >> 24;
+    rt2 = (rt & MIPSDSP_Q2) >> 16;
+    rt1 = (rt & MIPSDSP_Q1) >>  8;
+    rt0 =  rt & MIPSDSP_Q0;
+
+    tempD = mipsdsp_lshift8(env, rt3, sa);
+    tempC = mipsdsp_lshift8(env, rt2, sa);
+    tempB = mipsdsp_lshift8(env, rt1, sa);
+    tempA = mipsdsp_lshift8(env, rt0, sa);
+    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
+         ((uint32_t)tempB <<  8) | ((uint32_t)tempA);
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shllv_qb(CPUMIPSState *env,
+                             target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs2_0;
+    uint8_t  rt3, rt2, rt1, rt0;
+    uint8_t  tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rs2_0 = rs & 0x07;
+    rt3   = (rt & MIPSDSP_Q3) >> 24;
+    rt2   = (rt & MIPSDSP_Q2) >> 16;
+    rt1   = (rt & MIPSDSP_Q1) >>  8;
+    rt0   =  rt & MIPSDSP_Q0;
+
+    tempD = mipsdsp_lshift8(env, rt3, rs2_0);
+    tempC = mipsdsp_lshift8(env, rt2, rs2_0);
+    tempB = mipsdsp_lshift8(env, rt1, rs2_0);
+    tempA = mipsdsp_lshift8(env, rt0, rs2_0);
+
+    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
+         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+#if defined(TARGET_MIPS64)
+target_ulong helper_shll_ob(CPUMIPSState *env, target_ulong rt, uint32_t sa)
+{
+    int i;
+    uint8_t rt_t[8];
+    uint8_t temp_t[8];
+    uint64_t temp;
+
+    sa = sa & 0x07;
+    temp = 0;
+
+    for (i = 0; i < 8; i++) {
+        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+        temp_t[i] = mipsdsp_lshift8(env, rt_t[i], sa);
+
+        temp |= (uint64_t)temp_t[i] << (8 * i);
+    }
+
+    return temp;
+}
+
+target_ulong helper_shllv_ob(CPUMIPSState *env,
+                             target_ulong rt, target_ulong sa)
+{
+    int i;
+    uint8_t rt_t[8];
+    uint8_t temp_t[8];
+    uint64_t temp;
+
+    sa = sa & 0x07;
+    temp = 0;
+
+    for (i = 0; i < 8; i++) {
+        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+        temp_t[i] = mipsdsp_lshift8(env, rt_t[i], sa);
+
+        temp |= (uint64_t)temp_t[i] << (8 * i);
+    }
+
+    return temp;
+}
+#endif
+
+target_ulong helper_shll_ph(CPUMIPSState *env, uint32_t sa, target_ulong rt)
+{
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rth = (rt & MIPSDSP_HI) >> 16;
+    rtl =  rt & MIPSDSP_LO;
+    tempB = mipsdsp_lshift16(env, rth, sa);
+    tempA = mipsdsp_lshift16(env, rtl, sa);
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shllv_ph(CPUMIPSState *env,
+                             target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs3_0;
+    uint16_t rth, rtl, tempB, tempA;
+    uint32_t rd;
+
+    rth   = (rt & MIPSDSP_HI) >> 16;
+    rtl   =  rt & MIPSDSP_LO;
+    rs3_0 = rs & 0x0F;
+
+    tempB = mipsdsp_lshift16(env, rth, rs3_0);
+    tempA = mipsdsp_lshift16(env, rtl, rs3_0);
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shll_s_ph(CPUMIPSState *env, uint32_t sa, target_ulong rt)
+{
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rth = (rt & MIPSDSP_HI) >> 16;
+    rtl =  rt & MIPSDSP_LO;
+    tempB = mipsdsp_sat16_lshift(env, rth, sa);
+    tempA = mipsdsp_sat16_lshift(env, rtl, sa);
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shllv_s_ph(CPUMIPSState *env,
+                               target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs3_0;
+    uint16_t rth, rtl, tempB, tempA;
+    uint32_t rd;
+
+    rth   = (rt & MIPSDSP_HI) >> 16;
+    rtl   =  rt & MIPSDSP_LO;
+    rs3_0 = rs & 0x0F;
+
+    tempB = mipsdsp_sat16_lshift(env, rth, rs3_0);
+    tempA = mipsdsp_sat16_lshift(env, rtl, rs3_0);
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+#if defined(TARGET_MIPS64)
+target_ulong helper_shll_qh(CPUMIPSState *env, target_ulong rt, uint32_t sa)
+{
+    uint16_t rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x0F;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_lshift16(env, rt3, sa);
+    tempC = mipsdsp_lshift16(env, rt2, sa);
+    tempB = mipsdsp_lshift16(env, rt1, sa);
+    tempA = mipsdsp_lshift16(env, rt0, sa);
+
+    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+           ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return temp;
+}
+
+target_ulong helper_shllv_qh(CPUMIPSState *env,
+                             target_ulong rt, target_ulong sa)
+{
+    uint16_t rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x0F;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_lshift16(env, rt3, sa);
+    tempC = mipsdsp_lshift16(env, rt2, sa);
+    tempB = mipsdsp_lshift16(env, rt1, sa);
+    tempA = mipsdsp_lshift16(env, rt0, sa);
+
+    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+           ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return temp;
+}
+
+target_ulong helper_shll_s_qh(CPUMIPSState *env, target_ulong rt, uint32_t sa)
+{
+    uint16_t rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x0F;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_sat16_lshift(env, rt3, sa);
+    tempC = mipsdsp_sat16_lshift(env, rt2, sa);
+    tempB = mipsdsp_sat16_lshift(env, rt1, sa);
+    tempA = mipsdsp_sat16_lshift(env, rt0, sa);
+
+    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+           ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return temp;
+}
+
+target_ulong helper_shllv_s_qh(CPUMIPSState *env,
+                               target_ulong rt, target_ulong sa)
+{
+    uint16_t rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x0F;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_sat16_lshift(env, rt3, sa);
+    tempC = mipsdsp_sat16_lshift(env, rt2, sa);
+    tempB = mipsdsp_sat16_lshift(env, rt1, sa);
+    tempA = mipsdsp_sat16_lshift(env, rt0, sa);
+
+    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+           ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return temp;
+}
+#endif
+
+target_ulong helper_shll_s_w(CPUMIPSState *env, uint32_t sa, target_ulong rt)
+{
+    uint32_t temp, rd;
+
+    temp = mipsdsp_sat32_lshift(env, rt, sa);
+    rd   = temp;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shllv_s_w(CPUMIPSState *env,
+                              target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs4_0;
+    uint32_t rd;
+
+    rs4_0 = rs & 0x1F;
+    rd = mipsdsp_sat32_lshift(env, rt, rs4_0);
+
+    return (target_long)(int32_t)rd;
+}
+
+#if defined(TARGET_MIPS64)
+target_ulong helper_shll_pw(CPUMIPSState *env, target_ulong rt, uint32_t sa)
+{
+    uint32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t temp;
+
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_lshift32(env, rt1, sa);
+    tempA = mipsdsp_lshift32(env, rt0, sa);
+    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return temp;
+}
+
+target_ulong helper_shllv_pw(CPUMIPSState *env,
+                             target_ulong rt, target_ulong sa)
+{
+    uint32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x1F;
+
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_lshift32(env, rt1, sa);
+    tempA = mipsdsp_lshift32(env, rt0, sa);
+    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return temp;
+}
+
+target_ulong helper_shll_s_pw(CPUMIPSState *env, target_ulong rt, uint32_t sa)
+{
+    uint32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t temp;
+
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_sat32_lshift(env, rt1, sa);
+    tempA = mipsdsp_sat32_lshift(env, rt0, sa);
+    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return temp;
+}
+
+target_ulong helper_shllv_s_pw(CPUMIPSState *env,
+                               target_ulong rt, target_ulong sa)
+{
+    uint32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x1F;
+
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_sat32_lshift(env, rt1, sa);
+    tempA = mipsdsp_sat32_lshift(env, rt0, sa);
+    temp = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return temp;
+}
+#endif
+
+target_ulong helper_shrl_qb(uint32_t sa, target_ulong rt)
+{
+    uint8_t  rt3, rt2, rt1, rt0;
+    uint8_t  tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rt3 = (rt & MIPSDSP_Q3) >> 24;
+    rt2 = (rt & MIPSDSP_Q2) >> 16;
+    rt1 = (rt & MIPSDSP_Q1) >>  8;
+    rt0 =  rt & MIPSDSP_Q0;
+
+    tempD = mipsdsp_rshift8(rt3, sa);
+    tempC = mipsdsp_rshift8(rt2, sa);
+    tempB = mipsdsp_rshift8(rt1, sa);
+    tempA = mipsdsp_rshift8(rt0, sa);
+
+    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
+         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrlv_qb(target_ulong rs, target_ulong rt)
+{
+    uint8_t rs2_0;
+    uint8_t rt3, rt2, rt1, rt0;
+    uint8_t tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rs2_0 = rs & 0x07;
+    rt3   = (rt & MIPSDSP_Q3) >> 24;
+    rt2   = (rt & MIPSDSP_Q2) >> 16;
+    rt1   = (rt & MIPSDSP_Q1) >>  8;
+    rt0   =  rt & MIPSDSP_Q0;
+
+    tempD = mipsdsp_rshift8(rt3, rs2_0);
+    tempC = mipsdsp_rshift8(rt2, rs2_0);
+    tempB = mipsdsp_rshift8(rt1, rs2_0);
+    tempA = mipsdsp_rshift8(rt0, rs2_0);
+    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
+         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrl_ph(uint32_t sa, target_ulong rt)
+{
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rth = (rt & MIPSDSP_HI) >> 16;
+    rtl =  rt & MIPSDSP_LO;
+    tempB = rth >> sa;
+    tempA = rtl >> sa;
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrlv_ph(target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs3_0;
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rs3_0 = rs & 0x0F;
+    rth   = (rt & MIPSDSP_HI) >> 16;
+    rtl   =  rt & MIPSDSP_LO;
+
+    tempB = rth >> rs3_0;
+    tempA = rtl >> rs3_0;
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+#if defined(TARGET_MIPS64)
+target_ulong helper_shrl_ob(target_ulong rt, uint32_t sa)
+{
+    int i;
+    uint8_t rt_t[8];
+    uint8_t temp[8];
+    uint64_t result;
+
+    sa = sa & 0x07;
+    result = 0;
+
+    for (i = 0; i < 8; i++) {
+        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+        temp[i] = mipsdsp_rshift8(rt_t[i], sa);
+        result |= (uint64_t)temp[i] << (8 * i);
+    }
+
+    return result;
+}
+
+target_ulong helper_shrlv_ob(target_ulong rt, target_ulong sa)
+{
+    int i;
+    uint8_t rt_t[8];
+    uint8_t temp[8];
+    uint64_t result;
+
+    sa = sa & 0x07;
+    result = 0;
+
+    for (i = 0; i < 8; i++) {
+        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+        temp[i] = mipsdsp_rshift8(rt_t[i], sa);
+        result |= (uint64_t)temp[i] << (8 * i);
+    }
+
+    return result;
+}
+
+target_ulong helper_shrl_qh(target_ulong rt, uint32_t sa)
+{
+    uint16_t rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x0F;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = rt3 >> sa;
+    tempC = rt2 >> sa;
+    tempB = rt1 >> sa;
+    tempA = rt0 >> sa;
+
+    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+           ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return temp;
+}
+
+target_ulong helper_shrlv_qh(target_ulong rt, target_ulong sa)
+{
+    uint16_t rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint64_t temp;
+
+    sa = sa & 0x0F;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = rt3 >> sa;
+    tempC = rt2 >> sa;
+    tempB = rt1 >> sa;
+    tempA = rt0 >> sa;
+
+    temp = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+           ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return temp;
+}
+#endif
+
+target_ulong helper_shra_qb(uint32_t sa, target_ulong rt)
+{
+    int8_t  rt3, rt2, rt1, rt0;
+    uint8_t tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rt3 = (rt & MIPSDSP_Q3) >> 24;
+    rt2 = (rt & MIPSDSP_Q2) >> 16;
+    rt1 = (rt & MIPSDSP_Q1) >>  8;
+    rt0 =  rt & MIPSDSP_Q0;
+
+    tempD = rt3 >> sa;
+    tempC = rt2 >> sa;
+    tempB = rt1 >> sa;
+    tempA = rt0 >> sa;
+
+    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
+          ((uint32_t)tempB << 8) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shra_r_qb(uint32_t sa, target_ulong rt)
+{
+    int8_t  rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rt3 = (rt & MIPSDSP_Q3) >> 24;
+    rt2 = (rt & MIPSDSP_Q2) >> 16;
+    rt1 = (rt & MIPSDSP_Q1) >>  8;
+    rt0 =  rt & MIPSDSP_Q0;
+
+    if (sa == 0) {
+        tempD = rt3 & 0x00FF;
+        tempC = rt2 & 0x00FF;
+        tempB = rt1 & 0x00FF;
+        tempA = rt0 & 0x00FF;
+    } else {
+        tempD = ((int16_t)rt3 >> (sa - 1)) + 1;
+        tempC = ((int16_t)rt2 >> (sa - 1)) + 1;
+        tempB = ((int16_t)rt1 >> (sa - 1)) + 1;
+        tempA = ((int16_t)rt0 >> (sa - 1)) + 1;
+    }
+
+    rd = ((uint32_t)((tempD >> 1) & 0x00FF) << 24) |
+         ((uint32_t)((tempC >> 1) & 0x00FF) << 16) |
+         ((uint32_t)((tempB >> 1) & 0x00FF) <<  8) |
+         (uint32_t)((tempA >>  1) & 0x00FF) ;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrav_qb(target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs2_0;
+    int8_t   rt3, rt2, rt1, rt0;
+    uint8_t  tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rs2_0 = rs & 0x07;
+    rt3   = (rt & MIPSDSP_Q3) >> 24;
+    rt2   = (rt & MIPSDSP_Q2) >> 16;
+    rt1   = (rt & MIPSDSP_Q1) >>  8;
+    rt0   =  rt & MIPSDSP_Q0;
+
+    if (rs2_0 == 0) {
+        tempD = rt3;
+        tempC = rt2;
+        tempB = rt1;
+        tempA = rt0;
+    } else {
+        tempD = rt3 >> rs2_0;
+        tempC = rt2 >> rs2_0;
+        tempB = rt1 >> rs2_0;
+        tempA = rt0 >> rs2_0;
+    }
+
+    rd = ((uint32_t)tempD << 24) | ((uint32_t)tempC << 16) |
+         ((uint32_t)tempB <<  8) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrav_r_qb(target_ulong rs, target_ulong rt)
+{
+    uint8_t rs2_0;
+    int8_t  rt3, rt2, rt1, rt0;
+    uint16_t tempD, tempC, tempB, tempA;
+    uint32_t rd;
+
+    rs2_0 = rs & 0x07;
+    rt3 = (rt & MIPSDSP_Q3) >> 24;
+    rt2 = (rt & MIPSDSP_Q2) >> 16;
+    rt1 = (rt & MIPSDSP_Q1) >>  8;
+    rt0 =  rt & MIPSDSP_Q0;
+
+    if (rs2_0 == 0) {
+        tempD = (int16_t)rt3 << 1;
+        tempC = (int16_t)rt2 << 1;
+        tempB = (int16_t)rt1 << 1;
+        tempA = (int16_t)rt0 << 1;
+    } else {
+        tempD = ((int16_t)rt3 >> (rs2_0 - 1)) + 1;
+        tempC = ((int16_t)rt2 >> (rs2_0 - 1)) + 1;
+        tempB = ((int16_t)rt1 >> (rs2_0 - 1)) + 1;
+        tempA = ((int16_t)rt0 >> (rs2_0 - 1)) + 1;
+    }
+
+    rd = ((uint32_t)((tempD >> 1) & 0x00FF) << 24) |
+         ((uint32_t)((tempC >> 1) & 0x00FF) << 16) |
+         ((uint32_t)((tempB >> 1) & 0x00FF) <<  8) |
+         (uint32_t)((tempA >>  1) & 0x00FF) ;
+
+    return (target_long)(int32_t)rd;
+}
+
+#if defined(TARGET_MIPS64)
+target_ulong helper_shra_ob(target_ulong rt, uint32_t sa)
+{
+    int i;
+    int8_t rt_t[8];
+    uint8_t temp[8];
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0x07;
+
+    for (i = 0; i < 8; i++) {
+        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+        temp[i] = rt_t[i] >> sa;
+        result |= (uint64_t)temp[i] << (8 * i);
+    }
+
+    return result;
+}
+
+target_ulong helper_shrav_ob(target_ulong rt, target_ulong sa)
+{
+    int i;
+    int8_t rt_t[8];
+    uint8_t temp[8];
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0x07;
+
+    for (i = 0; i < 8; i++) {
+        rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+        temp[i] = rt_t[i] >> sa;
+        result |= (uint64_t)temp[i] << (8 * i);
+      }
+
+    return result;
+}
+
+target_ulong helper_shra_r_ob(target_ulong rt, uint32_t sa)
+{
+    int i;
+    int8_t rt_t[8];
+    int16_t rt_t_S[8];
+    uint8_t temp[8];
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0x07;
+
+    if (sa == 0) {
+        result = rt;
+    } else {
+        for (i = 0; i < 8; i++) {
+            rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+            rt_t_S[i] = ((int16_t)rt_t[i] >> (sa - 1)) + 1;
+            temp[i] = (rt_t_S[i] >> 1) & MIPSDSP_Q0;
+            result |= (uint64_t)temp[i] << (8 * i);
+        }
+    }
+
+    return result;
+}
+
+target_ulong helper_shrav_r_ob(target_ulong rt, target_ulong sa)
+{
+    int i;
+    int8_t rt_t[8];
+    int16_t rt_t_S[8];
+    uint8_t temp[8];
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0x07;
+
+    if (sa == 0) {
+        result = rt;
+    } else {
+        for (i = 0; i < 8; i++) {
+            rt_t[i] = (rt >> (8 * i)) & MIPSDSP_Q0;
+            rt_t_S[i] = ((int16_t)rt_t[i] >> (sa - 1)) + 1;
+            temp[i] = (rt_t_S[i] >> 1) & MIPSDSP_Q0;
+            result |= (uint64_t)temp[i] << (8 * i);
+        }
+    }
+
+    return result;
+}
+#endif
+
+target_ulong helper_shra_ph(uint32_t sa, target_ulong rt)
+{
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rth = (rt & MIPSDSP_HI) >> 16;
+    rtl =  rt & MIPSDSP_LO;
+    tempB = mipsdsp_rashift16(rth, sa);
+    tempA = mipsdsp_rashift16(rtl, sa);
+    rd = ((uint32_t)tempB << 16) | (uint32_t) tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrav_ph(target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs3_0;
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rs3_0 = rs & 0x0F;
+    rth   = (rt & MIPSDSP_HI) >> 16;
+    rtl   =  rt & MIPSDSP_LO;
+    tempB = mipsdsp_rashift16(rth, rs3_0);
+    tempA = mipsdsp_rashift16(rtl, rs3_0);
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shra_r_ph(uint32_t sa, target_ulong rt)
+{
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rth = (rt & MIPSDSP_HI) >> 16;
+    rtl =  rt & MIPSDSP_LO;
+    tempB = mipsdsp_rnd16_rashift(rth, sa);
+    tempA = mipsdsp_rnd16_rashift(rtl, sa);
+    rd = ((uint32_t)tempB << 16) | (uint32_t) tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrav_r_ph(target_ulong rs, target_ulong rt)
+{
+    uint8_t  rs3_0;
+    uint16_t rth, rtl;
+    uint16_t tempB, tempA;
+    uint32_t rd;
+
+    rs3_0 = rs & 0x0F;
+    rth   = (rt & MIPSDSP_HI) >> 16;
+    rtl   =  rt & MIPSDSP_LO;
+    tempB = mipsdsp_rnd16_rashift(rth, rs3_0);
+    tempA = mipsdsp_rnd16_rashift(rtl, rs3_0);
+
+    rd = ((uint32_t)tempB << 16) | (uint32_t)tempA;
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shra_r_w(uint32_t sa, target_ulong rt)
+{
+    uint32_t rd;
+
+    rd = mipsdsp_rnd32_rashift(rt, sa);
+
+    return (target_long)(int32_t)rd;
+}
+
+target_ulong helper_shrav_r_w(target_ulong rs, target_ulong rt)
+{
+    uint8_t rs4_0;
+    uint32_t rd;
+
+    rs4_0 = rs & 0x1F;
+    rd = mipsdsp_rnd32_rashift(rt, rs4_0);
+
+    return (target_long)(int32_t)rd;
+}
+
+#if defined(TARGET_MIPS64)
+target_ulong helper_shra_qh(target_ulong rt, uint32_t sa)
+{
+    int32_t rt3, rt2, rt1, rt0;
+    uint32_t tempD, tempC, tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0xF;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_rashift16(rt3, sa);
+    tempC = mipsdsp_rashift16(rt2, sa);
+    tempB = mipsdsp_rashift16(rt1, sa);
+    tempA = mipsdsp_rashift16(rt0, sa);
+
+    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+             ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return result;
+}
+
+target_ulong helper_shrav_qh(target_ulong rt, target_ulong sa)
+{
+    int32_t rt3, rt2, rt1, rt0;
+    uint32_t tempD, tempC, tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0xF;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_rashift16(rt3, sa);
+    tempC = mipsdsp_rashift16(rt2, sa);
+    tempB = mipsdsp_rashift16(rt1, sa);
+    tempA = mipsdsp_rashift16(rt0, sa);
+
+    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+             ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return result;
+}
+
+
+target_ulong helper_shra_r_qh(target_ulong rt, uint32_t sa)
+{
+    int32_t rt3, rt2, rt1, rt0;
+    uint32_t tempD, tempC, tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0xF;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_rnd16_rashift(rt3, sa);
+    tempC = mipsdsp_rnd16_rashift(rt2, sa);
+    tempB = mipsdsp_rnd16_rashift(rt1, sa);
+    tempA = mipsdsp_rnd16_rashift(rt0, sa);
+
+    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+             ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return result;
+}
+
+target_ulong helper_shrav_r_qh(target_ulong rt, target_ulong sa)
+{
+    int32_t rt3, rt2, rt1, rt0;
+    uint32_t tempD, tempC, tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0xF;
+
+    rt3 = (rt >> 48) & MIPSDSP_LO;
+    rt2 = (rt >> 32) & MIPSDSP_LO;
+    rt1 = (rt >> 16) & MIPSDSP_LO;
+    rt0 = rt & MIPSDSP_LO;
+
+    tempD = mipsdsp_rnd16_rashift(rt3, sa);
+    tempC = mipsdsp_rnd16_rashift(rt2, sa);
+    tempB = mipsdsp_rnd16_rashift(rt1, sa);
+    tempA = mipsdsp_rnd16_rashift(rt0, sa);
+
+    result = ((uint64_t)tempD << 48) | ((uint64_t)tempC << 32) |
+             ((uint64_t)tempB << 16) | (uint64_t)tempA;
+
+    return result;
+}
+
+target_ulong helper_shra_pw(target_ulong rt, uint32_t sa)
+{
+    int32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+
+    /* FIXME, it seems like document error. */
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_rashift32(rt1, sa);
+    tempA = mipsdsp_rashift32(rt0, sa);
+
+    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return result;
+}
+
+target_ulong helper_shrav_pw(target_ulong rt, target_ulong sa)
+{
+    int32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0x1F;
+
+    /* FIXME, it seems like document error. */
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_rashift32(rt1, sa);
+    tempA = mipsdsp_rashift32(rt0, sa);
+
+    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return result;
+}
+
+target_ulong helper_shra_r_pw(target_ulong rt, uint32_t sa)
+{
+    int32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_rnd32_rashift(rt1, sa);
+    tempA = mipsdsp_rnd32_rashift(rt0, sa);
+
+    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return result;
+}
+
+target_ulong helper_shrav_r_pw(target_ulong rt, target_ulong sa)
+{
+    int32_t rt1, rt0;
+    uint32_t tempB, tempA;
+    uint64_t result;
+
+    result = 0;
+    sa = sa & 0x1F;
+
+    rt1 = (rt >> 32) & MIPSDSP_LLO;
+    rt0 = rt & MIPSDSP_LLO;
+
+    tempB = mipsdsp_rnd32_rashift(rt1, sa);
+    tempA = mipsdsp_rnd32_rashift(rt0, sa);
+
+    result = ((uint64_t)tempB << 32) | (uint64_t)tempA;
+
+    return result;
+}
+#endif
+
 #undef MIPSDSP_LHI
 #undef MIPSDSP_LLO
 #undef MIPSDSP_HI
diff --git a/target-mips/helper.h b/target-mips/helper.h
index 0607105..7046380 100644
--- a/target-mips/helper.h
+++ b/target-mips/helper.h
@@ -429,4 +429,66 @@  DEF_HELPER_FLAGS_1(preceu_qh_obla, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl)
 DEF_HELPER_FLAGS_1(preceu_qh_obra, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl)
 #endif
 
+/* DSP GPR-Based Shift Sub-class insns */
+DEF_HELPER_FLAGS_3(shll_qb, 0, tl, env, i32, tl)
+DEF_HELPER_FLAGS_3(shllv_qb, 0, tl, env, tl, tl)
+#if defined(TARGET_MIPS64)
+DEF_HELPER_FLAGS_3(shll_ob, 0, tl, env, tl, i32)
+DEF_HELPER_FLAGS_3(shllv_ob, 0, tl, env, tl, tl)
+#endif
+DEF_HELPER_FLAGS_3(shll_ph, 0, tl, env, i32, tl)
+DEF_HELPER_FLAGS_3(shllv_ph, 0, tl, env, tl, tl)
+DEF_HELPER_FLAGS_3(shll_s_ph, 0, tl, env, i32, tl)
+DEF_HELPER_FLAGS_3(shllv_s_ph, 0, tl, env, tl, tl)
+#if defined(TARGET_MIPS64)
+DEF_HELPER_FLAGS_3(shll_qh, 0, tl, env, tl, i32)
+DEF_HELPER_FLAGS_3(shllv_qh, 0, tl, env, tl, tl)
+DEF_HELPER_FLAGS_3(shll_s_qh, 0, tl, env, tl, i32)
+DEF_HELPER_FLAGS_3(shllv_s_qh, 0, tl, env, tl, tl)
+#endif
+DEF_HELPER_FLAGS_3(shll_s_w, 0, tl, env, i32, tl)
+DEF_HELPER_FLAGS_3(shllv_s_w, 0, tl, env, tl, tl)
+#if defined(TARGET_MIPS64)
+DEF_HELPER_FLAGS_3(shll_pw, 0, tl, env, tl, i32)
+DEF_HELPER_FLAGS_3(shllv_pw, 0, tl, env, tl, tl)
+DEF_HELPER_FLAGS_3(shll_s_pw, 0, tl, env, tl, i32)
+DEF_HELPER_FLAGS_3(shllv_s_pw, 0, tl, env, tl, tl)
+#endif
+DEF_HELPER_FLAGS_2(shrl_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
+DEF_HELPER_FLAGS_2(shrlv_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shrl_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
+DEF_HELPER_FLAGS_2(shrlv_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+#if defined(TARGET_MIPS64)
+DEF_HELPER_FLAGS_2(shrl_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrlv_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shrl_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrlv_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+#endif
+DEF_HELPER_FLAGS_2(shra_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
+DEF_HELPER_FLAGS_2(shra_r_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
+DEF_HELPER_FLAGS_2(shrav_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shrav_r_qb, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+#if defined(TARGET_MIPS64)
+DEF_HELPER_FLAGS_2(shra_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrav_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shra_r_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrav_r_ob, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+#endif
+DEF_HELPER_FLAGS_2(shra_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
+DEF_HELPER_FLAGS_2(shrav_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shra_r_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
+DEF_HELPER_FLAGS_2(shrav_r_ph, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shra_r_w, TCG_CALL_CONST | TCG_CALL_PURE, tl, i32, tl)
+DEF_HELPER_FLAGS_2(shrav_r_w, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+#if defined(TARGET_MIPS64)
+DEF_HELPER_FLAGS_2(shra_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrav_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shra_r_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrav_r_qh, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shra_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrav_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+DEF_HELPER_FLAGS_2(shra_r_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, i32)
+DEF_HELPER_FLAGS_2(shrav_r_pw, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
+#endif
+
 #include "def-helper.h"
diff --git a/target-mips/translate.c b/target-mips/translate.c
index 724c162..d87dbee 100644
--- a/target-mips/translate.c
+++ b/target-mips/translate.c
@@ -331,6 +331,18 @@  enum {
 #if defined(TARGET_MIPS64)
     OPC_CMPU_EQ_OB_DSP = 0x15 | OPC_SPECIAL3,
 #endif
+    /* MIPS DSP GPR-Based Shift Sub-class */
+    OPC_SHLL_QB_DSP    = 0x13 | OPC_SPECIAL3,
+#if defined(TARGET_MIPS64)
+    OPC_SHLL_OB_DSP    = 0x17 | OPC_SPECIAL3,
+#endif
+    /* MIPS DSP Multiply Sub-class insns */
+    /* OPC_MUL_PH_DSP is same as OPC_ADDUH_QB_DSP.  */
+    /* OPC_MUL_PH_DSP     = 0x18 | OPC_SPECIAL3,  */
+    OPC_DPA_W_PH_DSP   = 0x30 | OPC_SPECIAL3,
+#if defined(TARGET_MIPS64)
+    OPC_DPAQ_W_QH_DSP  = 0x34 | OPC_SPECIAL3,
+#endif
 };
 
 /* BSHFL opcodes */
@@ -439,6 +451,32 @@  enum {
     OPC_PRECRQ_RS_PH_W   = (0x15 << 6) | OPC_CMPU_EQ_QB_DSP,
     OPC_PRECRQU_S_QB_PH  = (0x0F << 6) | OPC_CMPU_EQ_QB_DSP,
 };
+#define MASK_SHLL_QB(op) (MASK_SPECIAL3(op) | (op & (0x1F << 6)))
+enum {
+    /* MIPS DSP GPR-Based Shift Sub-class */
+    OPC_SHLL_QB    = (0x00 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHLLV_QB   = (0x02 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHLL_PH    = (0x08 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHLLV_PH   = (0x0A << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHLL_S_PH  = (0x0C << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHLLV_S_PH = (0x0E << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHLL_S_W   = (0x14 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHLLV_S_W  = (0x16 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRL_QB    = (0x01 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRLV_QB   = (0x03 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRL_PH    = (0x19 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRLV_PH   = (0x1B << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRA_QB    = (0x04 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRA_R_QB  = (0x05 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRAV_QB   = (0x06 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRAV_R_QB = (0x07 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRA_PH    = (0x09 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRAV_PH   = (0x0B << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRA_R_PH  = (0x0D << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRAV_R_PH = (0x0F << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRA_R_W   = (0x15 << 6) | OPC_SHLL_QB_DSP,
+    OPC_SHRAV_R_W  = (0x17 << 6) | OPC_SHLL_QB_DSP,
+};
 
 #if defined(TARGET_MIPS64)
 #define MASK_ABSQ_S_QH(op) (MASK_SPECIAL3(op) | (op & (0x1F << 6)))
@@ -507,6 +545,39 @@  enum {
 };
 #endif
 
+#if defined(TARGET_MIPS64)
+#define MASK_SHLL_OB(op) (MASK_SPECIAL3(op) | (op & (0x1F << 6)))
+enum {
+    /* MIPS DSP GPR-Based Shift Sub-class */
+    OPC_SHLL_PW    = (0x10 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLL_S_PW  = (0x14 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLLV_OB   = (0x02 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLLV_PW   = (0x12 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLLV_S_PW = (0x16 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLLV_QH   = (0x0A << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLLV_S_QH = (0x0E << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRA_PW    = (0x11 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRA_R_PW  = (0x15 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRAV_OB   = (0x06 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRAV_R_OB = (0x07 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRAV_PW   = (0x13 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRAV_R_PW = (0x17 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRAV_QH   = (0x0B << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRAV_R_QH = (0x0F << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRLV_OB   = (0x03 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRLV_QH   = (0x1B << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLL_OB    = (0x00 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLL_QH    = (0x08 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHLL_S_QH  = (0x0C << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRA_OB    = (0x04 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRA_R_OB  = (0x05 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRA_QH    = (0x09 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRA_R_QH  = (0x0D << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRL_OB    = (0x01 << 6) | OPC_SHLL_OB_DSP,
+    OPC_SHRL_QH    = (0x19 << 6) | OPC_SHLL_OB_DSP,
+};
+#endif
+
 /* Coprocessor 0 (rs field) */
 #define MASK_CP0(op)       MASK_OP_MAJOR(op) | (op & (0x1F << 21))
 
@@ -12660,6 +12731,117 @@  static void decode_opc (CPUMIPSState *env, DisasContext *ctx, int *is_branch)
                 break;
             }
             break;
+        case OPC_SHLL_QB_DSP:
+            {
+                TCGv_i32 temp_rs = tcg_const_i32(rs);
+                op2 = MASK_SHLL_QB(ctx->opcode);
+                switch (op2) {
+                case OPC_SHLL_QB:
+                    check_dsp(ctx);
+                    gen_helper_shll_qb(cpu_gpr[rd], cpu_env,
+                                       temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHLLV_QB:
+                    check_dsp(ctx);
+                    gen_helper_shllv_qb(cpu_gpr[rd], cpu_env,
+                                        cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHLL_PH:
+                    check_dsp(ctx);
+                    gen_helper_shll_ph(cpu_gpr[rd], cpu_env,
+                                       temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHLLV_PH:
+                    check_dsp(ctx);
+                    gen_helper_shllv_ph(cpu_gpr[rd], cpu_env,
+                                        cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHLL_S_PH:
+                    check_dsp(ctx);
+                    gen_helper_shll_s_ph(cpu_gpr[rd], cpu_env,
+                                         temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHLLV_S_PH:
+                    check_dsp(ctx);
+                    gen_helper_shllv_s_ph(cpu_gpr[rd], cpu_env,
+                                          cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHLL_S_W:
+                    check_dsp(ctx);
+                    gen_helper_shll_s_w(cpu_gpr[rd], cpu_env,
+                                        temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHLLV_S_W:
+                    check_dsp(ctx);
+                    gen_helper_shllv_s_w(cpu_gpr[rd], cpu_env,
+                                         cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHRL_QB:
+                    check_dsp(ctx);
+                    gen_helper_shrl_qb(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHRLV_QB:
+                    check_dsp(ctx);
+                    gen_helper_shrlv_qb(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHRL_PH:
+                    check_dspr2(ctx);
+                    gen_helper_shrl_ph(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHRLV_PH:
+                    check_dspr2(ctx);
+                    gen_helper_shrlv_ph(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHRA_QB:
+                    check_dspr2(ctx);
+                    gen_helper_shra_qb(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHRA_R_QB:
+                    check_dspr2(ctx);
+                    gen_helper_shra_r_qb(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHRAV_QB:
+                    check_dspr2(ctx);
+                    gen_helper_shrav_qb(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHRAV_R_QB:
+                    check_dspr2(ctx);
+                    gen_helper_shrav_r_qb(cpu_gpr[rd], cpu_gpr[rs],
+                                          cpu_gpr[rt]);
+                    break;
+                case OPC_SHRA_PH:
+                    check_dsp(ctx);
+                    gen_helper_shra_ph(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHRA_R_PH:
+                    check_dsp(ctx);
+                    gen_helper_shra_r_ph(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHRAV_PH:
+                    check_dsp(ctx);
+                    gen_helper_shrav_ph(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                case OPC_SHRAV_R_PH:
+                    check_dsp(ctx);
+                    gen_helper_shrav_r_ph(cpu_gpr[rd], cpu_gpr[rs],
+                                          cpu_gpr[rt]);
+                    break;
+                case OPC_SHRA_R_W:
+                    check_dsp(ctx);
+                    gen_helper_shra_r_w(cpu_gpr[rd], temp_rs, cpu_gpr[rt]);
+                    break;
+                case OPC_SHRAV_R_W:
+                    check_dsp(ctx);
+                    gen_helper_shrav_r_w(cpu_gpr[rd], cpu_gpr[rs], cpu_gpr[rt]);
+                    break;
+                default:            /* Invalid */
+                    MIPS_INVAL("MASK SHLL.QB");
+                    generate_exception(ctx, EXCP_RI);
+                    break;
+                }
+                tcg_temp_free_i32(temp_rs);
+                break;
+            }
 #if defined(TARGET_MIPS64)
         case OPC_DEXTM ... OPC_DEXT:
         case OPC_DINSM ... OPC_DINS:
@@ -12929,6 +13111,183 @@  static void decode_opc (CPUMIPSState *env, DisasContext *ctx, int *is_branch)
             }
             break;
 #endif
+#if defined(TARGET_MIPS64)
+        case OPC_SHLL_OB_DSP:
+            op2 = MASK_SHLL_OB(ctx->opcode);
+            switch (op2) {
+            case OPC_SHLL_PW:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shll_pw(cpu_gpr[rd], cpu_env,
+                                       cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHLLV_PW:
+                check_dsp(ctx);
+                gen_helper_shllv_pw(cpu_gpr[rd], cpu_env,
+                                    cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHLL_S_PW:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shll_s_pw(cpu_gpr[rd], cpu_env,
+                                         cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHLLV_S_PW:
+                check_dsp(ctx);
+                gen_helper_shllv_s_pw(cpu_gpr[rd], cpu_env,
+                                      cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHLL_OB:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shll_ob(cpu_gpr[rd], cpu_env,
+                                       cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHLLV_OB:
+                check_dsp(ctx);
+                gen_helper_shllv_ob(cpu_gpr[rd], cpu_env,
+                                    cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHLL_QH:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shll_qh(cpu_gpr[rd], cpu_env,
+                                       cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHLLV_QH:
+                check_dsp(ctx);
+                gen_helper_shllv_qh(cpu_gpr[rd], cpu_env,
+                                    cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHLL_S_QH:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shll_s_qh(cpu_gpr[rd], cpu_env,
+                                         cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHLLV_S_QH:
+                check_dsp(ctx);
+                gen_helper_shllv_s_qh(cpu_gpr[rd], cpu_env,
+                                      cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRA_OB:
+                check_dspr2(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shra_ob(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRAV_OB:
+                check_dspr2(ctx);
+                gen_helper_shrav_ob(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRA_R_OB:
+                check_dspr2(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shra_r_ob(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRAV_R_OB:
+                check_dspr2(ctx);
+                gen_helper_shrav_r_ob(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRA_PW:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shra_pw(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRAV_PW:
+                check_dsp(ctx);
+                gen_helper_shrav_pw(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRA_R_PW:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shra_r_pw(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRAV_R_PW:
+                check_dsp(ctx);
+                gen_helper_shrav_r_pw(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRA_QH:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shra_qh(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRAV_QH:
+                check_dsp(ctx);
+                gen_helper_shrav_qh(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRA_R_QH:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shra_r_qh(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRAV_R_QH:
+                check_dsp(ctx);
+                gen_helper_shrav_r_qh(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRL_OB:
+                check_dsp(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shrl_ob(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRLV_OB:
+                check_dsp(ctx);
+                gen_helper_shrlv_ob(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            case OPC_SHRL_QH:
+                check_dspr2(ctx);
+                {
+                    TCGv_i32 sa_v = tcg_const_i32(rs);
+                    gen_helper_shrl_qh(cpu_gpr[rd], cpu_gpr[rt], sa_v);
+                    tcg_temp_free_i32(sa_v);
+                    break;
+                }
+            case OPC_SHRLV_QH:
+                check_dspr2(ctx);
+                gen_helper_shrlv_qh(cpu_gpr[rd], cpu_gpr[rt], cpu_gpr[rs]);
+                break;
+            default:            /* Invalid */
+                MIPS_INVAL("MASK SHLL.OB");
+                generate_exception(ctx, EXCP_RI);
+                break;
+            }
+          break;
+#endif
         default:            /* Invalid */
             MIPS_INVAL("special3");
             generate_exception(ctx, EXCP_RI);