From b07716a4d28eb98f03ca0aae782a97b1c8427735 Mon Sep 17 00:00:00 2001 From: Michael Snyder Date: Fri, 9 Jan 2004 19:44:50 +0000 Subject: [PATCH] 2004-01-07 Michael Snyder * gencode.c: Replace 'Hitachi' with 'Renesas'. (op tab): Add new instructions for sh4a, DBR, SBR. (expand_opcode): Add handling for new movxy combinations. (gensym_caselist): Ditto. (expand_ppi_movxy): Remove movx/movy expansions, now handled in expand_opcode. (gensym): Add some helpful macros. (expand_ppi_code): Flatten loop for simplicity, tweak for 12-bit instead of 8-bit table (some insns are ambiguous to 8 bits). (ppi_gensim, main): Generate 12-bit instead of 8-bit ppi table. * interp.c: Replace 'Hitachi' with 'Renesas'. (union saved_state_type): Add dbr, sgr, ldst. (get_loop_bounds_ext): New function. (init_dsp): Add bfd_mach_sh4al_dsp. (sim_resume): Handle extended loop bounds. --- sim/sh/ChangeLog | 19 ++ sim/sh/gencode.c | 657 ++++++++++++++++++++++++++++++++++++++++++++----------- sim/sh/interp.c | 49 ++++- 3 files changed, 588 insertions(+), 137 deletions(-) diff --git a/sim/sh/ChangeLog b/sim/sh/ChangeLog index 5955cfa4fc..e8afeb66c7 100644 --- a/sim/sh/ChangeLog +++ b/sim/sh/ChangeLog @@ -1,3 +1,22 @@ +2004-01-07 Michael Snyder + + * gencode.c: Replace 'Hitachi' with 'Renesas'. + (op tab): Add new instructions for sh4a, DBR, SBR. + (expand_opcode): Add handling for new movxy combinations. + (gensym_caselist): Ditto. + (expand_ppi_movxy): Remove movx/movy expansions, + now handled in expand_opcode. + (gensym): Add some helpful macros. + (expand_ppi_code): Flatten loop for simplicity, tweak for 12-bit + instead of 8-bit table (some insns are ambiguous to 8 bits). + (ppi_gensim, main): Generate 12-bit instead of 8-bit ppi table. + + * interp.c: Replace 'Hitachi' with 'Renesas'. + (union saved_state_type): Add dbr, sgr, ldst. + (get_loop_bounds_ext): New function. + (init_dsp): Add bfd_mach_sh4al_dsp. + (sim_resume): Handle extended loop bounds. + 2003-12-18 Michael Snyder * gencode.c (expand_opcode): Simplify and reorganize. diff --git a/sim/sh/gencode.c b/sim/sh/gencode.c index 26b461e818..9a139da9cf 100644 --- a/sim/sh/gencode.c +++ b/sim/sh/gencode.c @@ -1,4 +1,5 @@ -/* Simulator/Opcode generator for the Hitachi Super-H architecture. +/* Simulator/Opcode generator for the Renesas + (formerly Hitachi) / SuperH Inc. Super-H architecture. Written by Steve Chamberlain of Cygnus Support. sac@cygnus.com @@ -151,6 +152,11 @@ op tab[] = "SET_SR_T (0);", }, + /* sh4a */ + { "", "", "clrdmxy", "0000000010001000", + "saved_state.asregs.cregs.named.sr &= ~(SR_MASK_DMX | SR_MASK_DMY);" + }, + { "", "0", "cmp/eq #,R0", "10001000i8*1....", "SET_SR_T (R0 == SEXT (i));", }, @@ -195,7 +201,7 @@ op tab[] = "SET_SR_T (0);", }, - { "", "nm", "div1 ,", "0011nnnnmmmm0100", /* ? MVS */ + { "", "nm", "div1 ,", "0011nnnnmmmm0100", "div1 (R, m, n/*, T*/);", }, @@ -285,10 +291,19 @@ op tab[] = }, /* sh4 */ - { "", "", "fipr ,", "1111nnmm11101101", - "/* FIXME: not implemented */", - "RAISE_EXCEPTION (SIGILL);", - "/* FIXME: check for DP and (n & 1) == 0? */", + { "", "", "fipr ,", "1111vvVV11101101", + "if (FPSCR_PR)", + " RAISE_EXCEPTION (SIGILL);", + "else", + "{", + " double fsum = 0;", + " /* FIXME: check for nans and infinities. */", + " fsum += FR (v1+0) * FR (v2+0);", + " fsum += FR (v1+1) * FR (v2+1);", + " fsum += FR (v1+2) * FR (v2+2);", + " fsum += FR (v1+3) * FR (v2+3);", + " SET_FR (v1+3, fsum);", + "}", }, /* sh2e */ @@ -440,6 +455,11 @@ op tab[] = "FP_UNARY(n, -);", }, + /* sh4a */ + { "", "", "fpchg", "1111011111111101", + "SET_FPSCR (GET_FPSCR() ^ FPSCR_MASK_PR);", + }, + /* sh4 */ { "", "", "frchg", "1111101111111101", "if (FPSCR_PR)", @@ -498,6 +518,17 @@ op tab[] = " FPUL = (int)FR(n);", }, + /* sh4 */ + { "", "", "ftrv ", "1111vv0111111101", + "if (FPSCR_PR)", + " RAISE_EXCEPTION (SIGILL);", + "else", + "{", + " /* FIXME not implemented. */", + " printf (\"ftrv xmtrx, FV%d\\n\", v1);", + "}", + }, + /* sh2e */ { "", "", "fsts FPUL,", "1111nnnn00001101", " union", @@ -535,12 +566,18 @@ op tab[] = { "", "n", "ldc ,MOD", "0100nnnn01011110", "SET_MOD (R[n]);", }, -#if 0 { "", "n", "ldc ,DBR", "0100nnnn11111010", - "DBR = R[n];", - "/* FIXME: user mode */", + "if (SR_MD)", + " DBR = R[n]; /* priv mode */", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", + }, + { "", "n", "ldc ,SGR", "0100nnnn00111010", + "if (SR_MD)", + " SGR = R[n]; /* priv mode */", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", }, -#endif { "n", "n", "ldc.l @+,", "0100nnnnmmmm0111", "MA (1);", "CREG (m) = RLAT (R[n]);", @@ -558,14 +595,26 @@ op tab[] = "SET_MOD (RLAT (R[n]));", "R[n] += 4;", }, -#if 0 { "n", "n", "ldc.l @+,DBR", "0100nnnn11110110", - "MA (1);", - "DBR = RLAT (R[n]);", - "R[n] += 4;", - "/* FIXME: user mode */", + "if (SR_MD)", + "{ /* priv mode */", + " MA (1);", + " DBR = RLAT (R[n]);", + " R[n] += 4;", + "}", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", + }, + { "n", "n", "ldc.l @+,SGR", "0100nnnn00110110", + "if (SR_MD)", + "{ /* priv mode */", + " MA (1);", + " SGR = RLAT (R[n]);", + " R[n] += 4;", + "}", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", }, -#endif /* sh-dsp */ { "", "", "ldre @(,PC)", "10001110i8p1....", @@ -575,6 +624,20 @@ op tab[] = "RS = SEXT (i) * 2 + 4 + PH2T (PC);", }, + /* sh4a */ + { "", "n", "ldrc ", "0100nnnn00110100", + "SET_RC (R[n]);", + "loop = get_loop_bounds_ext (RS, RE, memory, mem_end, maskw, endianw);", + "CHECK_INSN_PTR (insn_ptr);", + "RE |= 1;", + }, + { "", "", "ldrc #", "10001010i8*1....", + "SET_RC (i);", + "loop = get_loop_bounds_ext (RS, RE, memory, mem_end, maskw, endianw);", + "CHECK_INSN_PTR (insn_ptr);", + "RE |= 1;", + }, + { "", "n", "lds ,", "0100nnnnssss1010", "SREG (m) = R[n];", }, @@ -777,14 +840,49 @@ op tab[] = "WLAT (R[n], R[0]);", }, + { "n", "0", "movco.l R0, @", "0000nnnn01110011", + "/* LDST -> T */", + "SET_SR_T (LDST);", + "/* if (T) R0 -> (Rn) */", + "if (T)", + " WLAT (R[n], R[0]);", + "/* 0 -> LDST */", + "SET_LDST (0);", + }, + + { "0", "n", "movli.l @, R0", "0000nnnn01100011", + "/* 1 -> LDST */", + "SET_LDST (1);", + "/* (Rn) -> R0 */", + "R[0] = RLAT (R[n]);", + "/* if (interrupt/exception) 0 -> LDST */", + "/* (we don't simulate asynchronous interrupts/exceptions) */", + }, + { "n", "", "movt ", "0000nnnn00101001", "R[n] = T;", }, + { "0", "n", "movua.l @,R0", "0100nnnn10101001", + "int regn = R[n];", + "MA (1);", + "R[0] = (RBAT (regn) << 24) + (RBAT (regn + 1) << 16) + ", + " (RBAT (regn + 2) << 8) + RBAT (regn + 3);", + "L (n);", + }, + { "0n", "n", "movua.l @+,R0", "0100nnnn11101001", + "int regn = R[n];", + "MA (1);", + "R[0] = (RBAT (regn) << 24) + (RBAT (regn + 1) << 16) + ", + " (RBAT (regn + 2) << 8) + RBAT (regn + 3);", + "R[n] += 4;", + "L (n);", + }, { "", "mn", "mul.l ,", "0000nnnnmmmm0111", - "MACL = ((int)R[n]) * ((int)R[m]);", + "MACL = ((int) R[n]) * ((int) R[m]);", }, -#if 0 +#if 0 /* FIXME: The above cast to int is not really portable. + It should be replaced by a SEXT32 macro. */ { "", "nm", "mul.l ,", "0000nnnnmmmm0111", "MACL = R[n] * R[m];", }, @@ -820,6 +918,12 @@ op tab[] = "R[n] = ~R[m];", }, + /* sh4a */ + { "", "n", "icbi @", "0000nnnn11100011", + "/* Except for the effect on the cache - which is not simulated -", + " this is like a nop. */", + }, + { "", "n", "ocbi @", "0000nnnn10010011", "RSBAT (R[n]); /* Take exceptions like byte load, otherwise noop. */", "/* FIXME: Cache not implemented */", @@ -851,6 +955,18 @@ op tab[] = " this is like a nop. */", }, + /* sh4a */ + { "", "n", "prefi @", "0000nnnn11010011", + "/* Except for the effect on the cache - which is not simulated -", + " this is like a nop. */", + }, + + /* sh4a */ + { "", "", "synco", "0000000010101011", + "/* Except for the effect on the pipeline - which is not simulated -", + " this is like a nop. */", + }, + { "n", "n", "rotcl ", "0100nnnn00100100", "ult = R[n] < 0;", "R[n] = (R[n] << 1) | T;", @@ -898,6 +1014,18 @@ op tab[] = "Delay_Slot (PC + 2);", }, + /* sh4a */ + { "", "", "setdmx", "0000000010011000", + "saved_state.asregs.cregs.named.sr |= SR_MASK_DMX;" + "saved_state.asregs.cregs.named.sr &= ~SR_MASK_DMY;" + }, + + /* sh4a */ + { "", "", "setdmy", "0000000011001000", + "saved_state.asregs.cregs.named.sr |= SR_MASK_DMY;" + "saved_state.asregs.cregs.named.sr &= ~SR_MASK_DMX;" + }, + /* sh-dsp */ { "", "n", "setrc ", "0100nnnn00010100", "SET_RC (R[n]);", @@ -980,31 +1108,43 @@ op tab[] = "R[n] = CREG (m);", }, -#if 0 { "n", "", "stc SGR,", "0000nnnn00111010", - "R[n] = SGR;", + "if (SR_MD)", + " R[n] = SGR; /* priv mode */", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", }, { "n", "", "stc DBR,", "0000nnnn11111010", - "R[n] = DBR;", + "if (SR_MD)", + " R[n] = DBR; /* priv mode */", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", }, -#endif { "n", "n", "stc.l ,@-", "0100nnnnmmmm0011", "MA (1);", "R[n] -= 4;", "WLAT (R[n], CREG (m));", }, -#if 0 { "n", "n", "stc.l SGR,@-", "0100nnnn00110010", - "MA (1);", - "R[n] -= 4;", - "WLAT (R[n], SGR);", + "if (SR_MD)", + "{ /* priv mode */", + " MA (1);", + " R[n] -= 4;", + " WLAT (R[n], SGR);", + "}", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", }, { "n", "n", "stc.l DBR,@-", "0100nnnn11110010", - "MA (1);", - "R[n] -= 4;", - "WLAT (R[n], DBR);", + "if (SR_MD)", + "{ /* priv mode */", + " MA (1);", + " R[n] -= 4;", + " WLAT (R[n], DBR);", + "}", + "else", + " RAISE_EXCEPTION (SIGILL); /* user mode */", }, -#endif { "n", "", "sts ,", "0000nnnnssss1010", "R[n] = SREG (m);", @@ -1258,56 +1398,118 @@ op movsxy_tab[] = "WLAT (R[n], SEXT (DSP_R (m)));", "R[n] += R[8];", }, - { "", "n", "movx.w @,", "111100xxXX000100", + { "", "n", "movx.w @,", "111100xyXY0001??", "DSP_R (m) = RSWAT (R[n]) << 16;", - "iword &= 0xfd53; goto top;", + "if (iword & 3)", + " {", + " iword &= 0xfd53; goto top;", + " }", + }, + { "", "n", "movx.l @,", "111100xyXY010100", + "DSP_R (m) = RLAT (R[n]);", }, - { "n", "n", "movx.w @+,", "111100xxXX001000", + { "n", "n", "movx.w @+,", "111100xyXY0010??", "DSP_R (m) = RSWAT (R[n]) << 16;", "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : 2;", - "iword &= 0xfd53; goto top;", + "if (iword & 3)", + " {", + " iword &= 0xfd53; goto top;", + " }", + }, + { "n", "n", "movx.l @+,", "111100xyXY011000", + "DSP_R (m) = RLAT (R[n]);", + "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : 4;", }, - { "n", "n8","movx.w @+REG_8,", "111100xxXX001100", + { "n", "n8","movx.w @+REG_8,", "111100xyXY0011??", "DSP_R (m) = RSWAT (R[n]) << 16;", "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : R[8];", - "iword &= 0xfd53; goto top;", + "if (iword & 3)", + " {", + " iword &= 0xfd53; goto top;", + " }", + }, + { "n", "n8","movx.l @+REG_8,", "111100xyXY011100", + "DSP_R (m) = RLAT (R[n]);", + "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : R[8];", }, - { "", "n", "movx.w ,@", "111100xxaa100100", + { "", "n", "movx.w ,@", "111100xyax1001??", "WWAT (R[n], DSP_R (m) >> 16);", - "iword &= 0xfd53; goto top;", + "if (iword & 3)", + " {", + " iword &= 0xfd53; goto top;", + " }", + }, + { "", "n", "movx.l ,@", "111100xyax110100", + "WLAT (R[n], DSP_R (m));", }, - { "n", "n", "movx.w ,@+", "111100xxaa101000", + { "n", "n", "movx.w ,@+", "111100xyax1010??", "WWAT (R[n], DSP_R (m) >> 16);", "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : 2;", - "iword &= 0xfd53; goto top;", + "if (iword & 3)", + " {", + " iword &= 0xfd53; goto top;", + " }", }, - { "n", "n8","movx.w ,@+REG_8","111100xxaa101100", + { "n", "n", "movx.l ,@+", "111100xyax111000", + "WLAT (R[n], DSP_R (m));", + "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : 4;", + }, + { "n", "n8","movx.w ,@+REG_8","111100xyax1011??", "WWAT (R[n], DSP_R (m) >> 16);", "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : R[8];", - "iword &= 0xfd53; goto top;", + "if (iword & 3)", + " {", + " iword &= 0xfd53; goto top;", + " }", + }, + { "n", "n8","movx.l ,@+REG_8","111100xyax111100", + "WLAT (R[n], DSP_R (m));", + "R[n] += ((R[n] & 0xffff) == MOD_ME) ? MOD_DELTA : R[8];", }, - { "", "n", "movy.w @,", "111100yyYY000001", + { "", "n", "movy.w @,", "111100yxYX000001", "DSP_R (m) = RSWAT (R[n]) << 16;", }, - { "n", "n", "movy.w @+,", "111100yyYY000010", + { "n", "n", "movy.w @+,", "111100yxYX000010", "DSP_R (m) = RSWAT (R[n]) << 16;", "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : 2;", }, - { "n", "n9","movy.w @+REG_9,", "111100yyYY000011", + { "n", "n9","movy.w @+REG_9,", "111100yxYX000011", "DSP_R (m) = RSWAT (R[n]) << 16;", "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : R[9];", }, - { "", "n", "movy.w ,@", "111100yyAA010001", + { "", "n", "movy.w ,@", "111100yxAY010001", "WWAT (R[n], DSP_R (m) >> 16);", }, - { "n", "n", "movy.w ,@+", "111100yyAA010010", + { "n", "n", "movy.w ,@+", "111100yxAY010010", "WWAT (R[n], DSP_R (m) >> 16);", "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : 2;", }, - { "n", "n9", "movy.w ,@+REG_9", "111100yyAA010011", + { "n", "n9", "movy.w ,@+REG_9", "111100yxAY010011", "WWAT (R[n], DSP_R (m) >> 16);", "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : R[9];", }, + { "", "n", "movy.l @,", "111100yxYX100001", + "DSP_R (m) = RLAT (R[n]);", + }, + { "n", "n", "movy.l @+,", "111100yxYX100010", + "DSP_R (m) = RLAT (R[n]);", + "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : 4;", + }, + { "n", "n9","movy.l @+REG_9,", "111100yxYX100011", + "DSP_R (m) = RLAT (R[n]);", + "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : R[9];", + }, + { "", "n", "movy.l ,@", "111100yxAY110001", + "WLAT (R[n], DSP_R (m));", + }, + { "n", "n", "movy.l ,@+", "111100yxAY110010", + "WLAT (R[n], DSP_R (m));", + "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : 4;", + }, + { "n", "n9", "movy.l ,@+REG_9", "111100yxAY110011", + "WLAT (R[n], DSP_R (m));", + "R[n] += ((R[n] | ~0xffff) == MOD_ME) ? MOD_DELTA : R[9];", + }, { "", "", "nopx nopy", "1111000000000000", "/* nop */", }, @@ -1462,7 +1664,7 @@ op ppi_tab[] = "DSR |= carry;\n", "goto assign_z;\n", }, - { "","", "pcmp Sx,Sy", "10000100xxyy....", + { "","", "pcmp Sx,Sy", "10000100xxyyzzzz", "int Sx = DSP_R (x);", "int Sx_grd = GET_DSP_GRD (x);", "int Sy = DSP_R (y);", @@ -1479,6 +1681,24 @@ op ppi_tab[] = }, { "","", "pwad Sx,Sy,Dz", "10110100xxyyzzzz", }, + { "","", "(if cc) pabs Sx,Dz", "100010ccxx01zzzz", + "/* FIXME: duplicate code pabs. */", + "res = DSP_R (x);", + "res_grd = GET_DSP_GRD (x);", + "if (res >= 0)", + " carry = 0;", + "else", + " {", + " res = -res;", + " carry = (res != 0); /* The manual has a bug here. */", + " res_grd = -res_grd - carry;", + " }", + "COMPUTE_OVERFLOW;", + "/* ??? The re-computing of overflow after", + " saturation processing is specific to pabs. */", + "overflow = res_grd != SIGN32 (res) ? DSR_MASK_V : 0;", + "ADD_SUB_GE;", + }, { "","", "pabs Sx,Dz", "10001000xx..zzzz", "res = DSP_R (x);", "res_grd = GET_DSP_GRD (x);", @@ -1496,6 +1716,18 @@ op ppi_tab[] = "overflow = res_grd != SIGN32 (res) ? DSR_MASK_V : 0;", "ADD_SUB_GE;", }, + + { "","", "(if cc) prnd Sx,Dz", "100110ccxx01zzzz", + "/* FIXME: duplicate code prnd. */", + "int Sx = DSP_R (x);", + "int Sx_grd = GET_DSP_GRD (x);", + "", + "res = (Sx + 0x8000) & 0xffff0000;", + "carry = (unsigned) res < (unsigned) Sx;", + "res_grd = Sx_grd + carry;", + "COMPUTE_OVERFLOW;", + "ADD_SUB_GE;", + }, { "","", "prnd Sx,Dz", "10011000xx..zzzz", "int Sx = DSP_R (x);", "int Sx_grd = GET_DSP_GRD (x);", @@ -1506,6 +1738,31 @@ op ppi_tab[] = "COMPUTE_OVERFLOW;", "ADD_SUB_GE;", }, + + { "","", "(if cc) pabs Sy,Dz", "101010cc01yyzzzz", + "/* FIXME: duplicate code pabs. */", + "res = DSP_R (y);", + "res_grd = 0;", + "overflow = 0;", + "greater_equal = DSR_MASK_G;", + "if (res >= 0)", + " carry = 0;", + "else", + " {", + " res = -res;", + " carry = 1;", + " if (res < 0)", + " {", + " if (S)", + " res = 0x7fffffff;", + " else", + " {", + " overflow = DSR_MASK_V;", + " greater_equal = 0;", + " }", + " }", + " }", + }, { "","", "pabs Sy,Dz", "10101000..yyzzzz", "res = DSP_R (y);", "res_grd = 0;", @@ -1529,6 +1786,17 @@ op ppi_tab[] = " }", " }", }, + { "","", "(if cc) prnd Sy,Dz", "101110cc01yyzzzz", + "/* FIXME: duplicate code prnd. */", + "int Sy = DSP_R (y);", + "int Sy_grd = SIGN32 (Sy);", + "", + "res = (Sy + 0x8000) & 0xffff0000;", + "carry = (unsigned) res < (unsigned) Sy;", + "res_grd = Sy_grd + carry;", + "COMPUTE_OVERFLOW;", + "ADD_SUB_GE;", + }, { "","", "prnd Sy,Dz", "10111000..yyzzzz", "int Sy = DSP_R (y);", "int Sy_grd = SIGN32 (Sy);", @@ -1609,6 +1877,18 @@ op ppi_tab[] = "COMPUTE_OVERFLOW;", "ADD_SUB_GE;", }, + { "","", "(if cc) psub Sy,Sx,Dz", "100001ccxxyyzzzz", + "int Sx = DSP_R (x);", + "int Sx_grd = GET_DSP_GRD (x);", + "int Sy = DSP_R (y);", + "int Sy_grd = SIGN32 (Sy);", + "", + "res = Sy - Sx;", + "carry = (unsigned) res > (unsigned) Sy;", + "res_grd = Sy_grd - Sx_grd - carry;", + "COMPUTE_OVERFLOW;", + "ADD_SUB_GE;", + }, { "","", "(if cc) padd Sx,Sy,Dz", "101100ccxxyyzzzz", "int Sx = DSP_R (x);", "int Sx_grd = GET_DSP_GRD (x);", @@ -1698,6 +1978,21 @@ op ppi_tab[] = "overflow = 0;", "greater_equal = 1;", }, + { "","", "pclr Du pmuls Se,Sf,Dg", "0100eeff0001gguu", + "/* Do multiply. */", + "res = (DSP_R (e) >> 16) * (DSP_R (f) >> 16) * 2;", + "if (res == 0x80000000)", + " res = 0x7fffffff;", + "DSP_R (g) = res;", + "DSP_GRD (g) = SIGN32 (res);", + "/* FIXME: update DSR based on results of multiply! */", + "", + "/* Do clr. */", + "z = u;", + "res = 0;", + "res_grd = 0;", + "goto assign_z;", + }, { "","", "(if cc) pdmsb Sx,Dz", "100111ccxx..zzzz", "unsigned Sx = DSP_R (x);", "int Sx_grd = GET_DSP_GRD (x);", @@ -1813,6 +2108,27 @@ op ppi_tab[] = " MACL = DSP_R (z) = res;", "return;", }, + /* sh4a */ + { "","", "(if cc) pswap Sx,Dz", "100111ccxx01zzzz", + "int Sx = DSP_R (x);", + "", + "res = ((Sx & 0xffff) * 65536) + ((Sx >> 16) & 0xffff);", + "res_grd = GET_DSP_GRD (x);", + "carry = 0;", + "overflow = 0;", + "greater_equal = res & 0x80000000 ? 0 : DSR_MASK_G;", + }, + /* sh4a */ + { "","", "(if cc) pswap Sy,Dz", "101111cc01yyzzzz", + "int Sy = DSP_R (y);", + "", + "res = ((Sy & 0xffff) * 65536) + ((Sy >> 16) & 0xffff);", + "res_grd = SIGN32 (Sy);", + "carry = 0;", + "overflow = 0;", + "greater_equal = res & 0x80000000 ? 0 : DSR_MASK_G;", + }, + {0, 0} }; @@ -1976,26 +2292,32 @@ expand_opcode (val, i, s) exit (1); case '0': case '1': - /* Consume an arbitrary number of ones and zeros. */ - do { - j = (j << 1) + (s[m++] - '0'); - } while (s[m] == '0' || s[m] == '1'); - expand_opcode ((val << m) | j, i, s + m); - break; - + /* Consume an arbitrary number of ones and zeros. */ + do { + j = (j << 1) + (s[m++] - '0'); + } while (s[m] == '0' || s[m] == '1'); + expand_opcode ((val << m) | j, i, s + m); + break; case 'N': /* NN -- four-way fork */ for (j = 0; j < 4; j++) expand_opcode ((val << 2) | j, i, s + 2); break; - case 'x': /* xx -- 2-way fork */ - /* Cross-breeding with movy moved to separate function. */ - for (j = 0; j < 4; j += 2) + case 'x': /* xx or xy -- two-way or four-way fork */ + for (j = 0; j < 4; j += (s[1] == 'x' ? 2 : 1)) expand_opcode ((val << 2) | j, i, s + 2); break; - case 'y': /* yy -- two-way fork */ - for (j = 0; j < 2; j++) + case 'y': /* yy or yx -- two-way or four-way fork */ + for (j = 0; j < (s[1] == 'x' ? 4 : 2); j++) expand_opcode ((val << 2) | j, i, s + 2); break; + case '?': /* Seven-way "wildcard" fork for movxy */ + expand_opcode ((val << 2), i, s + 2); + for (j = 1; j < 4; j++) + { + expand_opcode ((val << 2) | j, i, s + 2); + expand_opcode ((val << 2) | (j + 16), i, s + 2); + } + break; case 'i': /* eg. "i8*1" */ case '.': /* "...." is a wildcard */ case 'n': @@ -2033,17 +2355,41 @@ expand_opcode (val, i, s) expand_opcode ((val << 4) | j, i, s + 4); break; case 'X': + /* XX/XY -- 2/4 way fork. */ + for (j = 0; j < 4; j += (s[1] == 'X' ? 2 : 1)) + expand_opcode ((val << 2) | j, i, s + 2); + break; case 'a': - /* XX, aa -- two-way fork */ - for (j = 0; j < 4; j += 2) + /* aa/ax -- 2/4 way fork. */ + for (j = 0; j < 4; j += (s[1] == 'a' ? 2 : 1)) expand_opcode ((val << 2) | j, i, s + 2); break; case 'Y': + /* YY/YX -- 2/4 way fork. */ + for (j = 0; j < (s[1] == 'Y' ? 2 : 4); j += 1) + expand_opcode ((val << 2) | j, i, s + 2); + break; case 'A': - /* YY, AA -- two-way fork */ - for (j = 0; j < 2; j++) + /* AA/AY: 2/4 way fork. */ + for (j = 0; j < (s[1] == 'A' ? 2 : 4); j += 1) expand_opcode ((val << 2) | j, i, s + 2); break; + case 'v': + /* vv(VV) -- 4(16) way fork. */ + /* Vector register fv0/4/8/12. */ + if (s[2] == 'V') + { + /* 2 vector registers. */ + for (j = 0; j < 15; j++) + expand_opcode ((val << 4) | j, i, s + 4); + } + else + { + /* 1 vector register. */ + for (j = 0; j < 4; j += 1) + expand_opcode ((val << 2) | j, i, s + 2); + } + break; } } } @@ -2113,24 +2459,6 @@ expand_ppi_movxy () int i; for (i = 0xf000; i < 0xf400; i++) - if ((i & 3) == 0 && (i & 12) != 0 && table[i] != 0) - { - /* A movx insn, which needs to be filled out with the - corresponding movy insns. This used to be done in - expand_opcode. */ - int m, mv; - - for (m = 0; m < 32; m++) - { - /* Ignore illegal nopy */ - if ((m & 7) == 0 && m != 0) - continue; - mv = m & 3 | (m & 4) << 2 | (m & 8) << 3 | (m & 16) << 4; - table [i | mv] = table [i]; - } - } - - for (i = 0xf000; i < 0xf400; i++) if (table[i]) table[i + 0x800] = table[0xf800]; } @@ -2161,6 +2489,18 @@ gensim_caselist (p) *s); exit (1); break; + case '?': + /* Wildcard expansion, nothing to do here. */ + s += 2; + break; + case 'v': + printf (" int v1 = ((iword >> 10) & 3) * 4;\n"); + s += 2; + break; + case 'V': + printf (" int v2 = ((iword >> 8) & 3) * 4;\n"); + s += 2; + break; case '0': case '1': s += 2; @@ -2179,12 +2519,26 @@ gensim_caselist (p) s += 2; break; case 'x': - printf (" int n = ((iword >> 9) & 1) + 4;\n"); + if (s[1] == 'y') /* xy */ + { + printf (" int n = (iword & 3) ? \n"); + printf (" ((iword >> 9) & 1) + 4 : \n"); + printf (" REG_xy ((iword >> 8) & 3);\n"); + } + else + printf (" int n = ((iword >> 9) & 1) + 4;\n"); needn = 1; s += 2; break; case 'y': - printf (" int n = ((iword >> 8) & 1) + 6;\n"); + if (s[1] == 'x') /* yx */ + { + printf (" int n = (iword & 0xc) ? \n"); + printf (" ((iword >> 8) & 1) + 6 : \n"); + printf (" REG_yx ((iword >> 8) & 3);\n"); + } + else + printf (" int n = ((iword >> 8) & 1) + 6;\n"); needn = 1; s += 2; break; @@ -2197,19 +2551,47 @@ gensim_caselist (p) s += 4; break; case 'X': - printf (" int m = ((iword >> 7) & 1) + 8;\n"); + if (s[1] == 'Y') /* XY */ + { + printf (" int m = (iword & 3) ? \n"); + printf (" ((iword >> 7) & 1) + 8 : \n"); + printf (" DSP_xy ((iword >> 6) & 3);\n"); + } + else + printf (" int m = ((iword >> 7) & 1) + 8;\n"); s += 2; break; case 'a': - printf (" int m = 7 - ((iword >> 6) & 2);\n"); + if (s[1] == 'x') /* ax */ + { + printf (" int m = (iword & 3) ? \n"); + printf (" 7 - ((iword >> 6) & 2) : \n"); + printf (" DSP_ax ((iword >> 6) & 3);\n"); + } + else + printf (" int m = 7 - ((iword >> 6) & 2);\n"); s += 2; break; case 'Y': - printf (" int m = ((iword >> 6) & 1) + 10;\n"); + if (s[1] == 'X') /* YX */ + { + printf (" int m = (iword & 0xc) ? \n"); + printf (" ((iword >> 6) & 1) + 10 : \n"); + printf (" DSP_yx ((iword >> 6) & 3);\n"); + } + else + printf (" int m = ((iword >> 6) & 1) + 10;\n"); s += 2; break; case 'A': - printf (" int m = 7 - ((iword >> 5) & 2);\n"); + if (s[1] == 'Y') /* AY */ + { + printf (" int m = (iword & 0xc) ? \n"); + printf (" 7 - ((iword >> 5) & 2) : \n"); + printf (" DSP_ay ((iword >> 6) & 3);\n"); + } + else + printf (" int m = 7 - ((iword >> 5) & 2);\n"); s += 2; break; @@ -2303,6 +2685,18 @@ static void gensim () { printf ("{\n"); + printf ("/* REG_xy = [r4, r5, r0, r1]. */\n"); + printf ("#define REG_xy(R) ((R)==0 ? 4 : (R)==2 ? 5 : (R)==1 ? 0 : 1)\n"); + printf ("/* REG_yx = [r6, r7, r2, r3]. */\n"); + printf ("#define REG_yx(R) ((R)==0 ? 6 : (R)==1 ? 7 : (R)==2 ? 2 : 3)\n"); + printf ("/* DSP_ax = [a0, a1, x0, x1]. */\n"); + printf ("#define DSP_ax(R) ((R)==0 ? 7 : (R)==2 ? 5 : (R)==1 ? 8 : 9)\n"); + printf ("/* DSP_ay = [a0, a1, y0, y1]. */\n"); + printf ("#define DSP_ay(R) ((R)==0 ? 7 : (R)==1 ? 5 : (R)==2 ? 10 : 11)\n"); + printf ("/* DSP_xy = [x0, x1, y0, y1]. */\n"); + printf ("#define DSP_xy(R) ((R)==0 ? 8 : (R)==2 ? 9 : (R)==1 ? 10 : 11)\n"); + printf ("/* DSP_yx = [y0, y1, x0, x1]. */\n"); + printf ("#define DSP_yx(R) ((R)==0 ? 10 : (R)==1 ? 11 : (R)==2 ? 8 : 9)\n"); printf (" switch (jump_table[iword]) {\n"); gensim_caselist (tab); @@ -2341,7 +2735,8 @@ gendefines () static int ppi_index; /* Take a ppi code, expand all varying fields in it and fill all the - right entries in 'table' with the opcode index. */ + right entries in 'table' with the opcode index. + NOTE: tail recursion optimization removed for simplicity. */ static void expand_ppi_code (val, i, s) @@ -2351,46 +2746,40 @@ expand_ppi_code (val, i, s) { int j; - for (;;) + switch (s[0]) { - switch (s[0]) - { - default: - fprintf (stderr, "gencode/expand_ppi_code: Illegal char '%c'\n", - s[0]); - exit (2); - break; - /* The last eight bits are disregarded for the switch table. */ - case 'm': - case 'x': - case '.': - table[val] = i; - return; - case '0': - val += val; - s++; - break; - case '1': - val += val + 1; - s++; - break; - case 'i': - case 'e': case 'f': - val += val; - s++; - expand_ppi_code (val, i, s); - val++; - break; - case 'c': - val <<= 2; - s += 2; - val++; - expand_ppi_code (val, ppi_index++, s); - val++; - expand_ppi_code (val, i, s); - val++; - break; - } + default: + fprintf (stderr, "gencode/expand_ppi_code: Illegal char '%c'\n", s[0]); + exit (2); + break; + case 'g': + case 'z': + /* The last four bits are disregarded for the switch table. */ + table[val] = i; + return; + case 'm': + /* Four-bit expansion. */ + for (j = 0; j < 16; j++) + expand_ppi_code ((val << 4) + j, i, s + 4); + break; + case '.': + case '0': + expand_ppi_code ((val << 1), i, s + 1); + break; + case '1': + expand_ppi_code ((val << 1) + 1, i, s + 1); + break; + case 'i': + case 'e': case 'f': + case 'x': case 'y': + expand_ppi_code ((val << 1), i, s + 1); + expand_ppi_code ((val << 1) + 1, i, s + 1); + break; + case 'c': + expand_ppi_code ((val << 2) + 1, ppi_index++, s + 2); + expand_ppi_code ((val << 2) + 2, i, s + 2); + expand_ppi_code ((val << 2) + 3, i, s + 2); + break; } } @@ -2442,18 +2831,24 @@ ppi_gensim () printf ("ppi_insn (iword)\n"); printf (" int iword;\n"); printf ("{\n"); + printf (" /* 'ee' = [x0, x1, y0, a1] */\n"); printf (" static char e_tab[] = { 8, 9, 10, 5};\n"); + printf (" /* 'ff' = [y0, y1, x0, a1] */\n"); printf (" static char f_tab[] = {10, 11, 8, 5};\n"); + printf (" /* 'xx' = [x0, x1, a0, a1] */\n"); printf (" static char x_tab[] = { 8, 9, 7, 5};\n"); + printf (" /* 'yy' = [y0, y1, m0, m1] */\n"); printf (" static char y_tab[] = {10, 11, 12, 14};\n"); + printf (" /* 'gg' = [m0, m1, a0, a1] */\n"); printf (" static char g_tab[] = {12, 14, 7, 5};\n"); + printf (" /* 'uu' = [x0, y0, a0, a1] */\n"); printf (" static char u_tab[] = { 8, 10, 7, 5};\n"); printf ("\n"); printf (" int z;\n"); printf (" int res, res_grd;\n"); printf (" int carry, overflow, greater_equal;\n"); printf ("\n"); - printf (" switch (ppi_table[iword >> 8]) {\n"); + printf (" switch (ppi_table[iword >> 4]) {\n"); for (; p->name; p++) { @@ -2614,7 +3009,7 @@ main (ac, av) memset (table, 0, sizeof table); ppi_filltable (); - dumptable ("ppi_table", 1 << 8, 0); + dumptable ("ppi_table", 1 << 12, 0); } else if (strcmp (av[1], "-x") == 0) { diff --git a/sim/sh/interp.c b/sim/sh/interp.c index 5ca8e61a0a..2bbb690ba4 100644 --- a/sim/sh/interp.c +++ b/sim/sh/interp.c @@ -1,4 +1,4 @@ -/* Simulator for the Hitachi SH architecture. +/* Simulator for the Renesas (formerly Hitachi) / SuperH Inc. SH architecture. Written by Steve Chamberlain of Cygnus Support. sac@cygnus.com @@ -120,6 +120,9 @@ typedef union int re; /* sh3 */ int bank[8]; + int dbr; /* debug base register */ + int sgr; /* saved gr15 */ + int ldst; /* load/store flag (boolean) */ } named; int i[16]; } cregs; @@ -167,10 +170,8 @@ static int target_dsp; static int host_little_endian; static char **prog_argv; -#if 1 static int maskw = 0; static int maskl = 0; -#endif static SIM_OPEN_KIND sim_kind; static char *myname; @@ -189,8 +190,10 @@ static char *myname; #define CREG(n) (saved_state.asregs.cregs.i[(n)]) #define GBR saved_state.asregs.cregs.named.gbr #define VBR saved_state.asregs.cregs.named.vbr +#define DBR saved_state.asregs.cregs.named.dbr #define SSR saved_state.asregs.cregs.named.ssr #define SPC saved_state.asregs.cregs.named.spc +#define SGR saved_state.asregs.cregs.named.sgr #define SREG(n) (saved_state.asregs.sregs.i[(n)]) #define MACH saved_state.asregs.sregs.named.mach #define MACL saved_state.asregs.sregs.named.macl @@ -228,6 +231,7 @@ static char *myname; #define Q ((saved_state.asregs.cregs.named.sr & SR_MASK_Q) != 0) #define S ((saved_state.asregs.cregs.named.sr & SR_MASK_S) != 0) #define T ((saved_state.asregs.cregs.named.sr & SR_MASK_T) != 0) +#define LDST ((saved_state.asregs.cregs.named.ldst) != 0) #define SR_BL ((saved_state.asregs.cregs.named.sr & SR_MASK_BL) != 0) #define SR_RB ((saved_state.asregs.cregs.named.sr & SR_MASK_RB) != 0) @@ -249,6 +253,7 @@ do { \ #define SET_SR_Q(EXP) SET_SR_BIT ((EXP), SR_MASK_Q) #define SET_SR_S(EXP) SET_SR_BIT ((EXP), SR_MASK_S) #define SET_SR_T(EXP) SET_SR_BIT ((EXP), SR_MASK_T) +#define SET_LDST(EXP) (saved_state.asregs.cregs.named.ldst = ((EXP) != 0)) /* stc currently relies on being able to read SR without modifications. */ #define GET_SR() (saved_state.asregs.cregs.named.sr - 0) @@ -1411,6 +1416,32 @@ macl (regs, memory, n, m) MACH = mach; } + +/* GET_LOOP_BOUNDS {EXTENDED} + These two functions compute the actual starting and ending point + of the repeat loop, based on the RS and RE registers (repeat start, + repeat stop). The extended version is called for LDRC, and the + regular version is called for SETRC. The difference is that for + LDRC, the loop start and end instructions are literally the ones + pointed to by RS and RE -- for SETRC, they're not (see docs). */ + +static struct loop_bounds +get_loop_bounds_ext (rs, re, memory, mem_end, maskw, endianw) + int rs, re; + unsigned char *memory, *mem_end; + int maskw, endianw; +{ + struct loop_bounds loop; + + /* FIXME: should I verify RS < RE? */ + loop.start = PT2H (RS); /* FIXME not using the params? */ + loop.end = PT2H (RE & ~1); /* Ignore bit 0 of RE. */ + SKIP_INSN (loop.end); + if (loop.end >= mem_end) + loop.end = PT2H (0); + return loop; +} + float fsca_s (int in, double (*f) (double)) { @@ -1541,7 +1572,9 @@ init_dsp (abfd) int was_dsp = target_dsp; unsigned long mach = bfd_get_mach (abfd); - if (mach == bfd_mach_sh_dsp || mach == bfd_mach_sh3_dsp) + if (mach == bfd_mach_sh_dsp || + mach == bfd_mach_sh4al_dsp || + mach == bfd_mach_sh3_dsp) { int ram_area_size, xram_start, yram_start; int new_select; @@ -1556,7 +1589,7 @@ init_dsp (abfd) xram_start = 0x0800f000; ram_area_size = 0x1000; } - if (mach == bfd_mach_sh3_dsp) + if (mach == bfd_mach_sh3_dsp || mach == bfd_mach_sh4al_dsp) { /* SH7612: 8KB each for X & Y memory; @@ -1740,7 +1773,11 @@ sim_resume (sd, step, siggnal) memory = saved_state.asregs.memory; mem_end = memory + saved_state.asregs.msize; - loop = get_loop_bounds (RS, RE, memory, mem_end, maskw, endianw); + if (RE & 1) + loop = get_loop_bounds_ext (RS, RE, memory, mem_end, maskw, endianw); + else + loop = get_loop_bounds (RS, RE, memory, mem_end, maskw, endianw); + insn_ptr = PT2H (saved_state.asregs.pc); CHECK_INSN_PTR (insn_ptr); -- 2.11.0