diff options
Diffstat (limited to 'target/mips/msa_helper.c')
-rw-r--r-- | target/mips/msa_helper.c | 1116 |
1 files changed, 940 insertions, 176 deletions
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c index c3b271934a..6865addaf6 100644 --- a/target/mips/msa_helper.c +++ b/target/mips/msa_helper.c @@ -2224,7 +2224,339 @@ void helper_msa_div_u_d(CPUMIPSState *env, * +---------------+----------------------------------------------------------+ */ -/* TODO: insert Int Dot Product group helpers here */ +#define SIGNED_EXTRACT(e, o, a, df) \ + do { \ + e = SIGNED_EVEN(a, df); \ + o = SIGNED_ODD(a, df); \ + } while (0) + +#define UNSIGNED_EXTRACT(e, o, a, df) \ + do { \ + e = UNSIGNED_EVEN(a, df); \ + o = UNSIGNED_ODD(a, df); \ + } while (0) + + +static inline int64_t msa_dotp_s_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + int64_t even_arg1; + int64_t even_arg2; + int64_t odd_arg1; + int64_t odd_arg2; + SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); + SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); + return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); +} + +void helper_msa_dotp_s_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_dotp_s_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_dotp_s_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_dotp_s_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_dotp_s_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_dotp_s_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_dotp_s_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_dotp_s_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_dotp_s_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_dotp_s_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_dotp_s_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_dotp_s_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_dotp_s_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_dotp_s_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_dotp_s_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_dotp_s_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_dotp_s_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_dotp_u_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + int64_t even_arg1; + int64_t even_arg2; + int64_t odd_arg1; + int64_t odd_arg2; + UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); + UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); + return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); +} + +void helper_msa_dotp_u_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_dotp_u_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_dotp_u_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_dotp_u_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_dotp_u_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_dotp_u_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_dotp_u_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_dotp_u_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_dotp_u_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_dotp_u_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_dotp_u_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_dotp_u_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_dotp_u_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_dotp_u_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_dotp_u_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_dotp_u_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_dotp_u_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_dpadd_s_df(uint32_t df, int64_t dest, int64_t arg1, + int64_t arg2) +{ + int64_t even_arg1; + int64_t even_arg2; + int64_t odd_arg1; + int64_t odd_arg2; + SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); + SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); + return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); +} + +void helper_msa_dpadd_s_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_dpadd_s_df(DF_HALF, pwd->h[0], pws->h[0], pwt->h[0]); + pwd->h[1] = msa_dpadd_s_df(DF_HALF, pwd->h[1], pws->h[1], pwt->h[1]); + pwd->h[2] = msa_dpadd_s_df(DF_HALF, pwd->h[2], pws->h[2], pwt->h[2]); + pwd->h[3] = msa_dpadd_s_df(DF_HALF, pwd->h[3], pws->h[3], pwt->h[3]); + pwd->h[4] = msa_dpadd_s_df(DF_HALF, pwd->h[4], pws->h[4], pwt->h[4]); + pwd->h[5] = msa_dpadd_s_df(DF_HALF, pwd->h[5], pws->h[5], pwt->h[5]); + pwd->h[6] = msa_dpadd_s_df(DF_HALF, pwd->h[6], pws->h[6], pwt->h[6]); + pwd->h[7] = msa_dpadd_s_df(DF_HALF, pwd->h[7], pws->h[7], pwt->h[7]); +} + +void helper_msa_dpadd_s_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_dpadd_s_df(DF_WORD, pwd->w[0], pws->w[0], pwt->w[0]); + pwd->w[1] = msa_dpadd_s_df(DF_WORD, pwd->w[1], pws->w[1], pwt->w[1]); + pwd->w[2] = msa_dpadd_s_df(DF_WORD, pwd->w[2], pws->w[2], pwt->w[2]); + pwd->w[3] = msa_dpadd_s_df(DF_WORD, pwd->w[3], pws->w[3], pwt->w[3]); +} + +void helper_msa_dpadd_s_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_dpadd_s_df(DF_DOUBLE, pwd->d[0], pws->d[0], pwt->d[0]); + pwd->d[1] = msa_dpadd_s_df(DF_DOUBLE, pwd->d[1], pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_dpadd_u_df(uint32_t df, int64_t dest, int64_t arg1, + int64_t arg2) +{ + int64_t even_arg1; + int64_t even_arg2; + int64_t odd_arg1; + int64_t odd_arg2; + UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); + UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); + return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); +} + +void helper_msa_dpadd_u_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_dpadd_u_df(DF_HALF, pwd->h[0], pws->h[0], pwt->h[0]); + pwd->h[1] = msa_dpadd_u_df(DF_HALF, pwd->h[1], pws->h[1], pwt->h[1]); + pwd->h[2] = msa_dpadd_u_df(DF_HALF, pwd->h[2], pws->h[2], pwt->h[2]); + pwd->h[3] = msa_dpadd_u_df(DF_HALF, pwd->h[3], pws->h[3], pwt->h[3]); + pwd->h[4] = msa_dpadd_u_df(DF_HALF, pwd->h[4], pws->h[4], pwt->h[4]); + pwd->h[5] = msa_dpadd_u_df(DF_HALF, pwd->h[5], pws->h[5], pwt->h[5]); + pwd->h[6] = msa_dpadd_u_df(DF_HALF, pwd->h[6], pws->h[6], pwt->h[6]); + pwd->h[7] = msa_dpadd_u_df(DF_HALF, pwd->h[7], pws->h[7], pwt->h[7]); +} + +void helper_msa_dpadd_u_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_dpadd_u_df(DF_WORD, pwd->w[0], pws->w[0], pwt->w[0]); + pwd->w[1] = msa_dpadd_u_df(DF_WORD, pwd->w[1], pws->w[1], pwt->w[1]); + pwd->w[2] = msa_dpadd_u_df(DF_WORD, pwd->w[2], pws->w[2], pwt->w[2]); + pwd->w[3] = msa_dpadd_u_df(DF_WORD, pwd->w[3], pws->w[3], pwt->w[3]); +} + +void helper_msa_dpadd_u_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_dpadd_u_df(DF_DOUBLE, pwd->d[0], pws->d[0], pwt->d[0]); + pwd->d[1] = msa_dpadd_u_df(DF_DOUBLE, pwd->d[1], pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_dpsub_s_df(uint32_t df, int64_t dest, int64_t arg1, + int64_t arg2) +{ + int64_t even_arg1; + int64_t even_arg2; + int64_t odd_arg1; + int64_t odd_arg2; + SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); + SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); + return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2)); +} + +void helper_msa_dpsub_s_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_dpsub_s_df(DF_HALF, pwd->h[0], pws->h[0], pwt->h[0]); + pwd->h[1] = msa_dpsub_s_df(DF_HALF, pwd->h[1], pws->h[1], pwt->h[1]); + pwd->h[2] = msa_dpsub_s_df(DF_HALF, pwd->h[2], pws->h[2], pwt->h[2]); + pwd->h[3] = msa_dpsub_s_df(DF_HALF, pwd->h[3], pws->h[3], pwt->h[3]); + pwd->h[4] = msa_dpsub_s_df(DF_HALF, pwd->h[4], pws->h[4], pwt->h[4]); + pwd->h[5] = msa_dpsub_s_df(DF_HALF, pwd->h[5], pws->h[5], pwt->h[5]); + pwd->h[6] = msa_dpsub_s_df(DF_HALF, pwd->h[6], pws->h[6], pwt->h[6]); + pwd->h[7] = msa_dpsub_s_df(DF_HALF, pwd->h[7], pws->h[7], pwt->h[7]); +} + +void helper_msa_dpsub_s_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_dpsub_s_df(DF_WORD, pwd->w[0], pws->w[0], pwt->w[0]); + pwd->w[1] = msa_dpsub_s_df(DF_WORD, pwd->w[1], pws->w[1], pwt->w[1]); + pwd->w[2] = msa_dpsub_s_df(DF_WORD, pwd->w[2], pws->w[2], pwt->w[2]); + pwd->w[3] = msa_dpsub_s_df(DF_WORD, pwd->w[3], pws->w[3], pwt->w[3]); +} + +void helper_msa_dpsub_s_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_dpsub_s_df(DF_DOUBLE, pwd->d[0], pws->d[0], pwt->d[0]); + pwd->d[1] = msa_dpsub_s_df(DF_DOUBLE, pwd->d[1], pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_dpsub_u_df(uint32_t df, int64_t dest, int64_t arg1, + int64_t arg2) +{ + int64_t even_arg1; + int64_t even_arg2; + int64_t odd_arg1; + int64_t odd_arg2; + UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); + UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); + return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2)); +} + +void helper_msa_dpsub_u_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_dpsub_u_df(DF_HALF, pwd->h[0], pws->h[0], pwt->h[0]); + pwd->h[1] = msa_dpsub_u_df(DF_HALF, pwd->h[1], pws->h[1], pwt->h[1]); + pwd->h[2] = msa_dpsub_u_df(DF_HALF, pwd->h[2], pws->h[2], pwt->h[2]); + pwd->h[3] = msa_dpsub_u_df(DF_HALF, pwd->h[3], pws->h[3], pwt->h[3]); + pwd->h[4] = msa_dpsub_u_df(DF_HALF, pwd->h[4], pws->h[4], pwt->h[4]); + pwd->h[5] = msa_dpsub_u_df(DF_HALF, pwd->h[5], pws->h[5], pwt->h[5]); + pwd->h[6] = msa_dpsub_u_df(DF_HALF, pwd->h[6], pws->h[6], pwt->h[6]); + pwd->h[7] = msa_dpsub_u_df(DF_HALF, pwd->h[7], pws->h[7], pwt->h[7]); +} + +void helper_msa_dpsub_u_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_dpsub_u_df(DF_WORD, pwd->w[0], pws->w[0], pwt->w[0]); + pwd->w[1] = msa_dpsub_u_df(DF_WORD, pwd->w[1], pws->w[1], pwt->w[1]); + pwd->w[2] = msa_dpsub_u_df(DF_WORD, pwd->w[2], pws->w[2], pwt->w[2]); + pwd->w[3] = msa_dpsub_u_df(DF_WORD, pwd->w[3], pws->w[3], pwt->w[3]); +} + +void helper_msa_dpsub_u_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_dpsub_u_df(DF_DOUBLE, pwd->d[0], pws->d[0], pwt->d[0]); + pwd->d[1] = msa_dpsub_u_df(DF_DOUBLE, pwd->d[1], pws->d[1], pwt->d[1]); +} /* @@ -2883,7 +3215,221 @@ void helper_msa_mod_u_d(CPUMIPSState *env, * +---------------+----------------------------------------------------------+ */ -/* TODO: insert Int Multiply group helpers here */ +static inline int64_t msa_maddv_df(uint32_t df, int64_t dest, int64_t arg1, + int64_t arg2) +{ + return dest + arg1 * arg2; +} + +void helper_msa_maddv_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_maddv_df(DF_BYTE, pwt->b[0], pws->b[0], pwt->b[0]); + pwd->b[1] = msa_maddv_df(DF_BYTE, pwt->b[1], pws->b[1], pwt->b[1]); + pwd->b[2] = msa_maddv_df(DF_BYTE, pwt->b[2], pws->b[2], pwt->b[2]); + pwd->b[3] = msa_maddv_df(DF_BYTE, pwt->b[3], pws->b[3], pwt->b[3]); + pwd->b[4] = msa_maddv_df(DF_BYTE, pwt->b[4], pws->b[4], pwt->b[4]); + pwd->b[5] = msa_maddv_df(DF_BYTE, pwt->b[5], pws->b[5], pwt->b[5]); + pwd->b[6] = msa_maddv_df(DF_BYTE, pwt->b[6], pws->b[6], pwt->b[6]); + pwd->b[7] = msa_maddv_df(DF_BYTE, pwt->b[7], pws->b[7], pwt->b[7]); + pwd->b[8] = msa_maddv_df(DF_BYTE, pwt->b[8], pws->b[8], pwt->b[8]); + pwd->b[9] = msa_maddv_df(DF_BYTE, pwt->b[9], pws->b[9], pwt->b[9]); + pwd->b[10] = msa_maddv_df(DF_BYTE, pwt->b[10], pws->b[10], pwt->b[10]); + pwd->b[11] = msa_maddv_df(DF_BYTE, pwt->b[11], pws->b[11], pwt->b[11]); + pwd->b[12] = msa_maddv_df(DF_BYTE, pwt->b[12], pws->b[12], pwt->b[12]); + pwd->b[13] = msa_maddv_df(DF_BYTE, pwt->b[13], pws->b[13], pwt->b[13]); + pwd->b[14] = msa_maddv_df(DF_BYTE, pwt->b[14], pws->b[14], pwt->b[14]); + pwd->b[15] = msa_maddv_df(DF_BYTE, pwt->b[15], pws->b[15], pwt->b[15]); +} + +void helper_msa_maddv_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_maddv_df(DF_HALF, pwd->h[0], pws->h[0], pwt->h[0]); + pwd->h[1] = msa_maddv_df(DF_HALF, pwd->h[1], pws->h[1], pwt->h[1]); + pwd->h[2] = msa_maddv_df(DF_HALF, pwd->h[2], pws->h[2], pwt->h[2]); + pwd->h[3] = msa_maddv_df(DF_HALF, pwd->h[3], pws->h[3], pwt->h[3]); + pwd->h[4] = msa_maddv_df(DF_HALF, pwd->h[4], pws->h[4], pwt->h[4]); + pwd->h[5] = msa_maddv_df(DF_HALF, pwd->h[5], pws->h[5], pwt->h[5]); + pwd->h[6] = msa_maddv_df(DF_HALF, pwd->h[6], pws->h[6], pwt->h[6]); + pwd->h[7] = msa_maddv_df(DF_HALF, pwd->h[7], pws->h[7], pwt->h[7]); +} + +void helper_msa_maddv_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_maddv_df(DF_WORD, pwd->w[0], pws->w[0], pwt->w[0]); + pwd->w[1] = msa_maddv_df(DF_WORD, pwd->w[1], pws->w[1], pwt->w[1]); + pwd->w[2] = msa_maddv_df(DF_WORD, pwd->w[2], pws->w[2], pwt->w[2]); + pwd->w[3] = msa_maddv_df(DF_WORD, pwd->w[3], pws->w[3], pwt->w[3]); +} + +void helper_msa_maddv_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_maddv_df(DF_DOUBLE, pwd->d[0], pws->d[0], pwt->d[0]); + pwd->d[1] = msa_maddv_df(DF_DOUBLE, pwd->d[1], pws->d[1], pwt->d[1]); +} + +static inline int64_t msa_msubv_df(uint32_t df, int64_t dest, int64_t arg1, + int64_t arg2) +{ + return dest - arg1 * arg2; +} + +void helper_msa_msubv_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_msubv_df(DF_BYTE, pwt->b[0], pws->b[0], pwt->b[0]); + pwd->b[1] = msa_msubv_df(DF_BYTE, pwt->b[1], pws->b[1], pwt->b[1]); + pwd->b[2] = msa_msubv_df(DF_BYTE, pwt->b[2], pws->b[2], pwt->b[2]); + pwd->b[3] = msa_msubv_df(DF_BYTE, pwt->b[3], pws->b[3], pwt->b[3]); + pwd->b[4] = msa_msubv_df(DF_BYTE, pwt->b[4], pws->b[4], pwt->b[4]); + pwd->b[5] = msa_msubv_df(DF_BYTE, pwt->b[5], pws->b[5], pwt->b[5]); + pwd->b[6] = msa_msubv_df(DF_BYTE, pwt->b[6], pws->b[6], pwt->b[6]); + pwd->b[7] = msa_msubv_df(DF_BYTE, pwt->b[7], pws->b[7], pwt->b[7]); + pwd->b[8] = msa_msubv_df(DF_BYTE, pwt->b[8], pws->b[8], pwt->b[8]); + pwd->b[9] = msa_msubv_df(DF_BYTE, pwt->b[9], pws->b[9], pwt->b[9]); + pwd->b[10] = msa_msubv_df(DF_BYTE, pwt->b[10], pws->b[10], pwt->b[10]); + pwd->b[11] = msa_msubv_df(DF_BYTE, pwt->b[11], pws->b[11], pwt->b[11]); + pwd->b[12] = msa_msubv_df(DF_BYTE, pwt->b[12], pws->b[12], pwt->b[12]); + pwd->b[13] = msa_msubv_df(DF_BYTE, pwt->b[13], pws->b[13], pwt->b[13]); + pwd->b[14] = msa_msubv_df(DF_BYTE, pwt->b[14], pws->b[14], pwt->b[14]); + pwd->b[15] = msa_msubv_df(DF_BYTE, pwt->b[15], pws->b[15], pwt->b[15]); +} + +void helper_msa_msubv_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_msubv_df(DF_HALF, pwd->h[0], pws->h[0], pwt->h[0]); + pwd->h[1] = msa_msubv_df(DF_HALF, pwd->h[1], pws->h[1], pwt->h[1]); + pwd->h[2] = msa_msubv_df(DF_HALF, pwd->h[2], pws->h[2], pwt->h[2]); + pwd->h[3] = msa_msubv_df(DF_HALF, pwd->h[3], pws->h[3], pwt->h[3]); + pwd->h[4] = msa_msubv_df(DF_HALF, pwd->h[4], pws->h[4], pwt->h[4]); + pwd->h[5] = msa_msubv_df(DF_HALF, pwd->h[5], pws->h[5], pwt->h[5]); + pwd->h[6] = msa_msubv_df(DF_HALF, pwd->h[6], pws->h[6], pwt->h[6]); + pwd->h[7] = msa_msubv_df(DF_HALF, pwd->h[7], pws->h[7], pwt->h[7]); +} + +void helper_msa_msubv_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_msubv_df(DF_WORD, pwd->w[0], pws->w[0], pwt->w[0]); + pwd->w[1] = msa_msubv_df(DF_WORD, pwd->w[1], pws->w[1], pwt->w[1]); + pwd->w[2] = msa_msubv_df(DF_WORD, pwd->w[2], pws->w[2], pwt->w[2]); + pwd->w[3] = msa_msubv_df(DF_WORD, pwd->w[3], pws->w[3], pwt->w[3]); +} + +void helper_msa_msubv_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_msubv_df(DF_DOUBLE, pwd->d[0], pws->d[0], pwt->d[0]); + pwd->d[1] = msa_msubv_df(DF_DOUBLE, pwd->d[1], pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + return arg1 * arg2; +} + +void helper_msa_mulv_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_mulv_df(DF_BYTE, pws->b[0], pwt->b[0]); + pwd->b[1] = msa_mulv_df(DF_BYTE, pws->b[1], pwt->b[1]); + pwd->b[2] = msa_mulv_df(DF_BYTE, pws->b[2], pwt->b[2]); + pwd->b[3] = msa_mulv_df(DF_BYTE, pws->b[3], pwt->b[3]); + pwd->b[4] = msa_mulv_df(DF_BYTE, pws->b[4], pwt->b[4]); + pwd->b[5] = msa_mulv_df(DF_BYTE, pws->b[5], pwt->b[5]); + pwd->b[6] = msa_mulv_df(DF_BYTE, pws->b[6], pwt->b[6]); + pwd->b[7] = msa_mulv_df(DF_BYTE, pws->b[7], pwt->b[7]); + pwd->b[8] = msa_mulv_df(DF_BYTE, pws->b[8], pwt->b[8]); + pwd->b[9] = msa_mulv_df(DF_BYTE, pws->b[9], pwt->b[9]); + pwd->b[10] = msa_mulv_df(DF_BYTE, pws->b[10], pwt->b[10]); + pwd->b[11] = msa_mulv_df(DF_BYTE, pws->b[11], pwt->b[11]); + pwd->b[12] = msa_mulv_df(DF_BYTE, pws->b[12], pwt->b[12]); + pwd->b[13] = msa_mulv_df(DF_BYTE, pws->b[13], pwt->b[13]); + pwd->b[14] = msa_mulv_df(DF_BYTE, pws->b[14], pwt->b[14]); + pwd->b[15] = msa_mulv_df(DF_BYTE, pws->b[15], pwt->b[15]); +} + +void helper_msa_mulv_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_mulv_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_mulv_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_mulv_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_mulv_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_mulv_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_mulv_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_mulv_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_mulv_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_mulv_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_mulv_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_mulv_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_mulv_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_mulv_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_mulv_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_mulv_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_mulv_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} /* @@ -3079,9 +3625,6 @@ void helper_msa_asub_u_d(CPUMIPSState *env, } -/* TODO: insert the rest of Int Subtract group helpers here */ - - static inline int64_t msa_hsub_s_df(uint32_t df, int64_t arg1, int64_t arg2) { return SIGNED_ODD(arg1, df) - SIGNED_EVEN(arg2, df); @@ -3176,6 +3719,398 @@ void helper_msa_hsub_u_d(CPUMIPSState *env, } +static inline int64_t msa_subs_s_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + int64_t max_int = DF_MAX_INT(df); + int64_t min_int = DF_MIN_INT(df); + if (arg2 > 0) { + return (min_int + arg2 < arg1) ? arg1 - arg2 : min_int; + } else { + return (arg1 < max_int + arg2) ? arg1 - arg2 : max_int; + } +} + +void helper_msa_subs_s_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_subs_s_df(DF_BYTE, pws->b[0], pwt->b[0]); + pwd->b[1] = msa_subs_s_df(DF_BYTE, pws->b[1], pwt->b[1]); + pwd->b[2] = msa_subs_s_df(DF_BYTE, pws->b[2], pwt->b[2]); + pwd->b[3] = msa_subs_s_df(DF_BYTE, pws->b[3], pwt->b[3]); + pwd->b[4] = msa_subs_s_df(DF_BYTE, pws->b[4], pwt->b[4]); + pwd->b[5] = msa_subs_s_df(DF_BYTE, pws->b[5], pwt->b[5]); + pwd->b[6] = msa_subs_s_df(DF_BYTE, pws->b[6], pwt->b[6]); + pwd->b[7] = msa_subs_s_df(DF_BYTE, pws->b[7], pwt->b[7]); + pwd->b[8] = msa_subs_s_df(DF_BYTE, pws->b[8], pwt->b[8]); + pwd->b[9] = msa_subs_s_df(DF_BYTE, pws->b[9], pwt->b[9]); + pwd->b[10] = msa_subs_s_df(DF_BYTE, pws->b[10], pwt->b[10]); + pwd->b[11] = msa_subs_s_df(DF_BYTE, pws->b[11], pwt->b[11]); + pwd->b[12] = msa_subs_s_df(DF_BYTE, pws->b[12], pwt->b[12]); + pwd->b[13] = msa_subs_s_df(DF_BYTE, pws->b[13], pwt->b[13]); + pwd->b[14] = msa_subs_s_df(DF_BYTE, pws->b[14], pwt->b[14]); + pwd->b[15] = msa_subs_s_df(DF_BYTE, pws->b[15], pwt->b[15]); +} + +void helper_msa_subs_s_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_subs_s_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_subs_s_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_subs_s_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_subs_s_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_subs_s_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_subs_s_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_subs_s_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_subs_s_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_subs_s_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_subs_s_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_subs_s_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_subs_s_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_subs_s_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_subs_s_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_subs_s_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_subs_s_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_subs_u_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + uint64_t u_arg1 = UNSIGNED(arg1, df); + uint64_t u_arg2 = UNSIGNED(arg2, df); + return (u_arg1 > u_arg2) ? u_arg1 - u_arg2 : 0; +} + +void helper_msa_subs_u_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_subs_u_df(DF_BYTE, pws->b[0], pwt->b[0]); + pwd->b[1] = msa_subs_u_df(DF_BYTE, pws->b[1], pwt->b[1]); + pwd->b[2] = msa_subs_u_df(DF_BYTE, pws->b[2], pwt->b[2]); + pwd->b[3] = msa_subs_u_df(DF_BYTE, pws->b[3], pwt->b[3]); + pwd->b[4] = msa_subs_u_df(DF_BYTE, pws->b[4], pwt->b[4]); + pwd->b[5] = msa_subs_u_df(DF_BYTE, pws->b[5], pwt->b[5]); + pwd->b[6] = msa_subs_u_df(DF_BYTE, pws->b[6], pwt->b[6]); + pwd->b[7] = msa_subs_u_df(DF_BYTE, pws->b[7], pwt->b[7]); + pwd->b[8] = msa_subs_u_df(DF_BYTE, pws->b[8], pwt->b[8]); + pwd->b[9] = msa_subs_u_df(DF_BYTE, pws->b[9], pwt->b[9]); + pwd->b[10] = msa_subs_u_df(DF_BYTE, pws->b[10], pwt->b[10]); + pwd->b[11] = msa_subs_u_df(DF_BYTE, pws->b[11], pwt->b[11]); + pwd->b[12] = msa_subs_u_df(DF_BYTE, pws->b[12], pwt->b[12]); + pwd->b[13] = msa_subs_u_df(DF_BYTE, pws->b[13], pwt->b[13]); + pwd->b[14] = msa_subs_u_df(DF_BYTE, pws->b[14], pwt->b[14]); + pwd->b[15] = msa_subs_u_df(DF_BYTE, pws->b[15], pwt->b[15]); +} + +void helper_msa_subs_u_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_subs_u_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_subs_u_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_subs_u_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_subs_u_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_subs_u_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_subs_u_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_subs_u_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_subs_u_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_subs_u_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_subs_u_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_subs_u_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_subs_u_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_subs_u_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_subs_u_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_subs_u_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_subs_u_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_subsus_u_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + uint64_t u_arg1 = UNSIGNED(arg1, df); + uint64_t max_uint = DF_MAX_UINT(df); + if (arg2 >= 0) { + uint64_t u_arg2 = (uint64_t)arg2; + return (u_arg1 > u_arg2) ? + (int64_t)(u_arg1 - u_arg2) : + 0; + } else { + uint64_t u_arg2 = (uint64_t)(-arg2); + return (u_arg1 < max_uint - u_arg2) ? + (int64_t)(u_arg1 + u_arg2) : + (int64_t)max_uint; + } +} + +void helper_msa_subsus_u_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_subsus_u_df(DF_BYTE, pws->b[0], pwt->b[0]); + pwd->b[1] = msa_subsus_u_df(DF_BYTE, pws->b[1], pwt->b[1]); + pwd->b[2] = msa_subsus_u_df(DF_BYTE, pws->b[2], pwt->b[2]); + pwd->b[3] = msa_subsus_u_df(DF_BYTE, pws->b[3], pwt->b[3]); + pwd->b[4] = msa_subsus_u_df(DF_BYTE, pws->b[4], pwt->b[4]); + pwd->b[5] = msa_subsus_u_df(DF_BYTE, pws->b[5], pwt->b[5]); + pwd->b[6] = msa_subsus_u_df(DF_BYTE, pws->b[6], pwt->b[6]); + pwd->b[7] = msa_subsus_u_df(DF_BYTE, pws->b[7], pwt->b[7]); + pwd->b[8] = msa_subsus_u_df(DF_BYTE, pws->b[8], pwt->b[8]); + pwd->b[9] = msa_subsus_u_df(DF_BYTE, pws->b[9], pwt->b[9]); + pwd->b[10] = msa_subsus_u_df(DF_BYTE, pws->b[10], pwt->b[10]); + pwd->b[11] = msa_subsus_u_df(DF_BYTE, pws->b[11], pwt->b[11]); + pwd->b[12] = msa_subsus_u_df(DF_BYTE, pws->b[12], pwt->b[12]); + pwd->b[13] = msa_subsus_u_df(DF_BYTE, pws->b[13], pwt->b[13]); + pwd->b[14] = msa_subsus_u_df(DF_BYTE, pws->b[14], pwt->b[14]); + pwd->b[15] = msa_subsus_u_df(DF_BYTE, pws->b[15], pwt->b[15]); +} + +void helper_msa_subsus_u_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_subsus_u_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_subsus_u_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_subsus_u_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_subsus_u_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_subsus_u_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_subsus_u_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_subsus_u_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_subsus_u_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_subsus_u_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_subsus_u_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_subsus_u_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_subsus_u_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_subsus_u_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_subsus_u_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_subsus_u_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_subsus_u_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_subsuu_s_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + uint64_t u_arg1 = UNSIGNED(arg1, df); + uint64_t u_arg2 = UNSIGNED(arg2, df); + int64_t max_int = DF_MAX_INT(df); + int64_t min_int = DF_MIN_INT(df); + if (u_arg1 > u_arg2) { + return u_arg1 - u_arg2 < (uint64_t)max_int ? + (int64_t)(u_arg1 - u_arg2) : + max_int; + } else { + return u_arg2 - u_arg1 < (uint64_t)(-min_int) ? + (int64_t)(u_arg1 - u_arg2) : + min_int; + } +} + +void helper_msa_subsuu_s_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_subsuu_s_df(DF_BYTE, pws->b[0], pwt->b[0]); + pwd->b[1] = msa_subsuu_s_df(DF_BYTE, pws->b[1], pwt->b[1]); + pwd->b[2] = msa_subsuu_s_df(DF_BYTE, pws->b[2], pwt->b[2]); + pwd->b[3] = msa_subsuu_s_df(DF_BYTE, pws->b[3], pwt->b[3]); + pwd->b[4] = msa_subsuu_s_df(DF_BYTE, pws->b[4], pwt->b[4]); + pwd->b[5] = msa_subsuu_s_df(DF_BYTE, pws->b[5], pwt->b[5]); + pwd->b[6] = msa_subsuu_s_df(DF_BYTE, pws->b[6], pwt->b[6]); + pwd->b[7] = msa_subsuu_s_df(DF_BYTE, pws->b[7], pwt->b[7]); + pwd->b[8] = msa_subsuu_s_df(DF_BYTE, pws->b[8], pwt->b[8]); + pwd->b[9] = msa_subsuu_s_df(DF_BYTE, pws->b[9], pwt->b[9]); + pwd->b[10] = msa_subsuu_s_df(DF_BYTE, pws->b[10], pwt->b[10]); + pwd->b[11] = msa_subsuu_s_df(DF_BYTE, pws->b[11], pwt->b[11]); + pwd->b[12] = msa_subsuu_s_df(DF_BYTE, pws->b[12], pwt->b[12]); + pwd->b[13] = msa_subsuu_s_df(DF_BYTE, pws->b[13], pwt->b[13]); + pwd->b[14] = msa_subsuu_s_df(DF_BYTE, pws->b[14], pwt->b[14]); + pwd->b[15] = msa_subsuu_s_df(DF_BYTE, pws->b[15], pwt->b[15]); +} + +void helper_msa_subsuu_s_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_subsuu_s_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_subsuu_s_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_subsuu_s_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_subsuu_s_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_subsuu_s_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_subsuu_s_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_subsuu_s_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_subsuu_s_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_subsuu_s_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_subsuu_s_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_subsuu_s_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_subsuu_s_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_subsuu_s_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_subsuu_s_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_subsuu_s_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_subsuu_s_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} + + +static inline int64_t msa_subv_df(uint32_t df, int64_t arg1, int64_t arg2) +{ + return arg1 - arg2; +} + +void helper_msa_subv_b(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->b[0] = msa_subv_df(DF_BYTE, pws->b[0], pwt->b[0]); + pwd->b[1] = msa_subv_df(DF_BYTE, pws->b[1], pwt->b[1]); + pwd->b[2] = msa_subv_df(DF_BYTE, pws->b[2], pwt->b[2]); + pwd->b[3] = msa_subv_df(DF_BYTE, pws->b[3], pwt->b[3]); + pwd->b[4] = msa_subv_df(DF_BYTE, pws->b[4], pwt->b[4]); + pwd->b[5] = msa_subv_df(DF_BYTE, pws->b[5], pwt->b[5]); + pwd->b[6] = msa_subv_df(DF_BYTE, pws->b[6], pwt->b[6]); + pwd->b[7] = msa_subv_df(DF_BYTE, pws->b[7], pwt->b[7]); + pwd->b[8] = msa_subv_df(DF_BYTE, pws->b[8], pwt->b[8]); + pwd->b[9] = msa_subv_df(DF_BYTE, pws->b[9], pwt->b[9]); + pwd->b[10] = msa_subv_df(DF_BYTE, pws->b[10], pwt->b[10]); + pwd->b[11] = msa_subv_df(DF_BYTE, pws->b[11], pwt->b[11]); + pwd->b[12] = msa_subv_df(DF_BYTE, pws->b[12], pwt->b[12]); + pwd->b[13] = msa_subv_df(DF_BYTE, pws->b[13], pwt->b[13]); + pwd->b[14] = msa_subv_df(DF_BYTE, pws->b[14], pwt->b[14]); + pwd->b[15] = msa_subv_df(DF_BYTE, pws->b[15], pwt->b[15]); +} + +void helper_msa_subv_h(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->h[0] = msa_subv_df(DF_HALF, pws->h[0], pwt->h[0]); + pwd->h[1] = msa_subv_df(DF_HALF, pws->h[1], pwt->h[1]); + pwd->h[2] = msa_subv_df(DF_HALF, pws->h[2], pwt->h[2]); + pwd->h[3] = msa_subv_df(DF_HALF, pws->h[3], pwt->h[3]); + pwd->h[4] = msa_subv_df(DF_HALF, pws->h[4], pwt->h[4]); + pwd->h[5] = msa_subv_df(DF_HALF, pws->h[5], pwt->h[5]); + pwd->h[6] = msa_subv_df(DF_HALF, pws->h[6], pwt->h[6]); + pwd->h[7] = msa_subv_df(DF_HALF, pws->h[7], pwt->h[7]); +} + +void helper_msa_subv_w(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->w[0] = msa_subv_df(DF_WORD, pws->w[0], pwt->w[0]); + pwd->w[1] = msa_subv_df(DF_WORD, pws->w[1], pwt->w[1]); + pwd->w[2] = msa_subv_df(DF_WORD, pws->w[2], pwt->w[2]); + pwd->w[3] = msa_subv_df(DF_WORD, pws->w[3], pwt->w[3]); +} + +void helper_msa_subv_d(CPUMIPSState *env, + uint32_t wd, uint32_t ws, uint32_t wt) +{ + wr_t *pwd = &(env->active_fpu.fpr[wd].wr); + wr_t *pws = &(env->active_fpu.fpr[ws].wr); + wr_t *pwt = &(env->active_fpu.fpr[wt].wr); + + pwd->d[0] = msa_subv_df(DF_DOUBLE, pws->d[0], pwt->d[0]); + pwd->d[1] = msa_subv_df(DF_DOUBLE, pws->d[1], pwt->d[1]); +} + + /* * Interleave * ---------- @@ -4400,11 +5335,6 @@ void helper_msa_shf_df(CPUMIPSState *env, uint32_t df, uint32_t wd, msa_move_v(pwd, pwx); } -static inline int64_t msa_subv_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - return arg1 - arg2; -} - #define MSA_BINOP_IMM_DF(helper, func) \ void helper_msa_ ## helper ## _df(CPUMIPSState *env, uint32_t df, \ uint32_t wd, uint32_t ws, int32_t u5) \ @@ -4586,97 +5516,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl) MSA_TEROP_IMMU_DF(binsri, binsr) #undef MSA_TEROP_IMMU_DF -static inline int64_t msa_subs_s_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - int64_t max_int = DF_MAX_INT(df); - int64_t min_int = DF_MIN_INT(df); - if (arg2 > 0) { - return (min_int + arg2 < arg1) ? arg1 - arg2 : min_int; - } else { - return (arg1 < max_int + arg2) ? arg1 - arg2 : max_int; - } -} - -static inline int64_t msa_subs_u_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - uint64_t u_arg1 = UNSIGNED(arg1, df); - uint64_t u_arg2 = UNSIGNED(arg2, df); - return (u_arg1 > u_arg2) ? u_arg1 - u_arg2 : 0; -} - -static inline int64_t msa_subsus_u_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - uint64_t u_arg1 = UNSIGNED(arg1, df); - uint64_t max_uint = DF_MAX_UINT(df); - if (arg2 >= 0) { - uint64_t u_arg2 = (uint64_t)arg2; - return (u_arg1 > u_arg2) ? - (int64_t)(u_arg1 - u_arg2) : - 0; - } else { - uint64_t u_arg2 = (uint64_t)(-arg2); - return (u_arg1 < max_uint - u_arg2) ? - (int64_t)(u_arg1 + u_arg2) : - (int64_t)max_uint; - } -} - -static inline int64_t msa_subsuu_s_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - uint64_t u_arg1 = UNSIGNED(arg1, df); - uint64_t u_arg2 = UNSIGNED(arg2, df); - int64_t max_int = DF_MAX_INT(df); - int64_t min_int = DF_MIN_INT(df); - if (u_arg1 > u_arg2) { - return u_arg1 - u_arg2 < (uint64_t)max_int ? - (int64_t)(u_arg1 - u_arg2) : - max_int; - } else { - return u_arg2 - u_arg1 < (uint64_t)(-min_int) ? - (int64_t)(u_arg1 - u_arg2) : - min_int; - } -} - -static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - return arg1 * arg2; -} - -#define SIGNED_EXTRACT(e, o, a, df) \ - do { \ - e = SIGNED_EVEN(a, df); \ - o = SIGNED_ODD(a, df); \ - } while (0) - -#define UNSIGNED_EXTRACT(e, o, a, df) \ - do { \ - e = UNSIGNED_EVEN(a, df); \ - o = UNSIGNED_ODD(a, df); \ - } while (0) - -static inline int64_t msa_dotp_s_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - int64_t even_arg1; - int64_t even_arg2; - int64_t odd_arg1; - int64_t odd_arg2; - SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); - SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); - return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); -} - -static inline int64_t msa_dotp_u_df(uint32_t df, int64_t arg1, int64_t arg2) -{ - int64_t even_arg1; - int64_t even_arg2; - int64_t odd_arg1; - int64_t odd_arg2; - UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); - UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); - return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); -} - #define CONCATENATE_AND_SLIDE(s, k) \ do { \ for (i = 0; i < s; i++) { \ @@ -4794,15 +5633,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, \ } \ } -MSA_BINOP_DF(subv) -MSA_BINOP_DF(subs_s) -MSA_BINOP_DF(subs_u) -MSA_BINOP_DF(subsus_u) -MSA_BINOP_DF(subsuu_s) -MSA_BINOP_DF(mulv) -MSA_BINOP_DF(dotp_s) -MSA_BINOP_DF(dotp_u) - MSA_BINOP_DF(mul_q) MSA_BINOP_DF(mulr_q) #undef MSA_BINOP_DF @@ -4816,66 +5646,6 @@ void helper_msa_sld_df(CPUMIPSState *env, uint32_t df, uint32_t wd, msa_sld_df(df, pwd, pws, env->active_tc.gpr[rt]); } -static inline int64_t msa_maddv_df(uint32_t df, int64_t dest, int64_t arg1, - int64_t arg2) -{ - return dest + arg1 * arg2; -} - -static inline int64_t msa_msubv_df(uint32_t df, int64_t dest, int64_t arg1, - int64_t arg2) -{ - return dest - arg1 * arg2; -} - -static inline int64_t msa_dpadd_s_df(uint32_t df, int64_t dest, int64_t arg1, - int64_t arg2) -{ - int64_t even_arg1; - int64_t even_arg2; - int64_t odd_arg1; - int64_t odd_arg2; - SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); - SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); - return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); -} - -static inline int64_t msa_dpadd_u_df(uint32_t df, int64_t dest, int64_t arg1, - int64_t arg2) -{ - int64_t even_arg1; - int64_t even_arg2; - int64_t odd_arg1; - int64_t odd_arg2; - UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); - UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); - return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2); -} - -static inline int64_t msa_dpsub_s_df(uint32_t df, int64_t dest, int64_t arg1, - int64_t arg2) -{ - int64_t even_arg1; - int64_t even_arg2; - int64_t odd_arg1; - int64_t odd_arg2; - SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); - SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); - return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2)); -} - -static inline int64_t msa_dpsub_u_df(uint32_t df, int64_t dest, int64_t arg1, - int64_t arg2) -{ - int64_t even_arg1; - int64_t even_arg2; - int64_t odd_arg1; - int64_t odd_arg2; - UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df); - UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df); - return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2)); -} - static inline int64_t msa_madd_q_df(uint32_t df, int64_t dest, int64_t arg1, int64_t arg2) { @@ -5002,12 +5772,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, uint32_t wd, \ } \ } -MSA_TEROP_DF(maddv) -MSA_TEROP_DF(msubv) -MSA_TEROP_DF(dpadd_s) -MSA_TEROP_DF(dpadd_u) -MSA_TEROP_DF(dpsub_s) -MSA_TEROP_DF(dpsub_u) MSA_TEROP_DF(binsl) MSA_TEROP_DF(binsr) MSA_TEROP_DF(madd_q) |