Pass two TCGReg to tcg_out_tlb_load, rather than idx+args. Move ldst_optimization routines just below tcg_out_tlb_load to avoid the need for forward declarations.
Use TCGReg enum in preference to int where apprpriate. Signed-off-by: Richard Henderson <r...@twiddle.net> --- tcg/i386/tcg-target.c | 457 +++++++++++++++++++++++--------------------------- 1 file changed, 208 insertions(+), 249 deletions(-) diff --git a/tcg/i386/tcg-target.c b/tcg/i386/tcg-target.c index 1b86009..a7ff8a3 100644 --- a/tcg/i386/tcg-target.c +++ b/tcg/i386/tcg-target.c @@ -1043,22 +1043,10 @@ static const void * const qemu_st_helpers[4] = { helper_ret_stq_mmu, }; -static void add_qemu_ldst_label(TCGContext *s, - int is_ld, - int opc, - int data_reg, - int data_reg2, - int addrlo_reg, - int addrhi_reg, - int mem_index, - uint8_t *raddr, - uint8_t **label_ptr); - /* Perform the TLB load and compare. Inputs: - ADDRLO_IDX contains the index into ARGS of the low part of the - address; the high part of the address is at ADDR_LOW_IDX+1. + ADDRLO and ADDRHI contain the low and high part of the address. MEM_INDEX and S_BITS are the memory context and log2 size of the load. @@ -1076,14 +1064,12 @@ static void add_qemu_ldst_label(TCGContext *s, First argument register is clobbered. */ -static inline void tcg_out_tlb_load(TCGContext *s, int addrlo_idx, +static inline void tcg_out_tlb_load(TCGContext *s, TCGReg addrlo, TCGReg addrhi, int mem_index, TCGMemOp s_bits, - const TCGArg *args, uint8_t **label_ptr, int which) { - const int addrlo = args[addrlo_idx]; - const int r0 = TCG_REG_L0; - const int r1 = TCG_REG_L1; + const TCGReg r0 = TCG_REG_L0; + const TCGReg r1 = TCG_REG_L1; TCGType ttype = TCG_TYPE_I32; TCGType htype = TCG_TYPE_I32; int trexw = 0, hrexw = 0; @@ -1132,7 +1118,7 @@ static inline void tcg_out_tlb_load(TCGContext *s, int addrlo_idx, if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) { /* cmp 4(r0), addrhi */ - tcg_out_modrm_offset(s, OPC_CMP_GvEv, args[addrlo_idx+1], r0, 4); + tcg_out_modrm_offset(s, OPC_CMP_GvEv, addrhi, r0, 4); /* jne slow_path */ tcg_out_opc(s, OPC_JCC_long + JCC_JNE, 0, 0, 0); @@ -1146,6 +1132,183 @@ static inline void tcg_out_tlb_load(TCGContext *s, int addrlo_idx, tcg_out_modrm_offset(s, OPC_ADD_GvEv + hrexw, r1, r0, offsetof(CPUTLBEntry, addend) - which); } + +/* + * Record the context of a call to the out of line helper code for the slow path + * for a load or store, so that we can later generate the correct helper code + */ +static void add_qemu_ldst_label(TCGContext *s, int is_ld, TCGMemOp opc, + TCGReg datalo, TCGReg datahi, + TCGReg addrlo, TCGReg addrhi, + int mem_index, uint8_t *raddr, + uint8_t **label_ptr) +{ + TCGLabelQemuLdst *label = new_ldst_label(s); + + label->is_ld = is_ld; + label->opc = opc; + label->datalo_reg = datalo; + label->datahi_reg = datahi; + label->addrlo_reg = addrlo; + label->addrhi_reg = addrhi; + label->mem_index = mem_index; + label->raddr = raddr; + label->label_ptr[0] = label_ptr[0]; + if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) { + label->label_ptr[1] = label_ptr[1]; + } +} + +/* + * Generate code for the slow path for a load at the end of block + */ +static void tcg_out_qemu_ld_slow_path(TCGContext *s, TCGLabelQemuLdst *l) +{ + TCGMemOp opc = l->opc; + TCGMemOp s_bits = opc & MO_SIZE; + TCGReg data_reg; + uint8_t **label_ptr = &l->label_ptr[0]; + + /* resolve label address */ + *(uint32_t *)label_ptr[0] = (uint32_t)(s->code_ptr - label_ptr[0] - 4); + if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) { + *(uint32_t *)label_ptr[1] = (uint32_t)(s->code_ptr - label_ptr[1] - 4); + } + + if (TCG_TARGET_REG_BITS == 32) { + int ofs = 0; + + tcg_out_st(s, TCG_TYPE_PTR, TCG_AREG0, TCG_REG_ESP, ofs); + ofs += 4; + + tcg_out_st(s, TCG_TYPE_I32, l->addrlo_reg, TCG_REG_ESP, ofs); + ofs += 4; + + if (TARGET_LONG_BITS == 64) { + tcg_out_st(s, TCG_TYPE_I32, l->addrhi_reg, TCG_REG_ESP, ofs); + ofs += 4; + } + + tcg_out_sti(s, TCG_TYPE_I32, TCG_REG_ESP, ofs, l->mem_index); + ofs += 4; + + tcg_out_sti(s, TCG_TYPE_I32, TCG_REG_ESP, ofs, (uintptr_t)l->raddr); + } else { + tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0], TCG_AREG0); + /* The second argument is already loaded with addrlo. */ + tcg_out_movi(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2], + l->mem_index); + tcg_out_movi(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[3], + (uintptr_t)l->raddr); + } + + tcg_out_calli(s, (uintptr_t)qemu_ld_helpers[s_bits]); + + data_reg = l->datalo_reg; + switch (opc & MO_SSIZE) { + case MO_SB: + tcg_out_ext8s(s, data_reg, TCG_REG_EAX, P_REXW); + break; + case MO_SW: + tcg_out_ext16s(s, data_reg, TCG_REG_EAX, P_REXW); + break; +#if TCG_TARGET_REG_BITS == 64 + case MO_SL: + tcg_out_ext32s(s, data_reg, TCG_REG_EAX); + break; +#endif + case MO_UB: + case MO_UW: + /* Note that the helpers have zero-extended to tcg_target_long. */ + case MO_UL: + tcg_out_mov(s, TCG_TYPE_I32, data_reg, TCG_REG_EAX); + break; + case MO_Q: + if (TCG_TARGET_REG_BITS == 64) { + tcg_out_mov(s, TCG_TYPE_I64, data_reg, TCG_REG_RAX); + } else if (data_reg == TCG_REG_EDX) { + /* xchg %edx, %eax */ + tcg_out_opc(s, OPC_XCHG_ax_r32 + TCG_REG_EDX, 0, 0, 0); + tcg_out_mov(s, TCG_TYPE_I32, l->datahi_reg, TCG_REG_EAX); + } else { + tcg_out_mov(s, TCG_TYPE_I32, data_reg, TCG_REG_EAX); + tcg_out_mov(s, TCG_TYPE_I32, l->datahi_reg, TCG_REG_EDX); + } + break; + default: + tcg_abort(); + } + + /* Jump to the code corresponding to next IR of qemu_st */ + tcg_out_jmp(s, (uintptr_t)l->raddr); +} + +/* + * Generate code for the slow path for a store at the end of block + */ +static void tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *l) +{ + TCGMemOp opc = l->opc; + TCGMemOp s_bits = opc & MO_SIZE; + uint8_t **label_ptr = &l->label_ptr[0]; + TCGReg retaddr; + + /* resolve label address */ + *(uint32_t *)label_ptr[0] = (uint32_t)(s->code_ptr - label_ptr[0] - 4); + if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) { + *(uint32_t *)label_ptr[1] = (uint32_t)(s->code_ptr - label_ptr[1] - 4); + } + + if (TCG_TARGET_REG_BITS == 32) { + int ofs = 0; + + tcg_out_st(s, TCG_TYPE_PTR, TCG_AREG0, TCG_REG_ESP, ofs); + ofs += 4; + + tcg_out_st(s, TCG_TYPE_I32, l->addrlo_reg, TCG_REG_ESP, ofs); + ofs += 4; + + if (TARGET_LONG_BITS == 64) { + tcg_out_st(s, TCG_TYPE_I32, l->addrhi_reg, TCG_REG_ESP, ofs); + ofs += 4; + } + + tcg_out_st(s, TCG_TYPE_I32, l->datalo_reg, TCG_REG_ESP, ofs); + ofs += 4; + + if (s_bits == MO_64) { + tcg_out_st(s, TCG_TYPE_I32, l->datahi_reg, TCG_REG_ESP, ofs); + ofs += 4; + } + + tcg_out_sti(s, TCG_TYPE_I32, TCG_REG_ESP, ofs, l->mem_index); + ofs += 4; + + retaddr = TCG_REG_EAX; + tcg_out_movi(s, TCG_TYPE_I32, retaddr, (uintptr_t)l->raddr); + tcg_out_st(s, TCG_TYPE_I32, retaddr, TCG_REG_ESP, ofs); + } else { + tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0], TCG_AREG0); + /* The second argument is already loaded with addrlo. */ + tcg_out_mov(s, (s_bits == MO_64 ? TCG_TYPE_I64 : TCG_TYPE_I32), + tcg_target_call_iarg_regs[2], l->datalo_reg); + tcg_out_movi(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3], + l->mem_index); + + if (ARRAY_SIZE(tcg_target_call_iarg_regs) > 4) { + retaddr = tcg_target_call_iarg_regs[4]; + tcg_out_movi(s, TCG_TYPE_PTR, retaddr, (uintptr_t)l->raddr); + } else { + retaddr = TCG_REG_RAX; + tcg_out_movi(s, TCG_TYPE_PTR, retaddr, (uintptr_t)l->raddr); + tcg_out_st(s, TCG_TYPE_PTR, retaddr, TCG_REG_ESP, 0); + } + } + + /* "Tail call" to the helper, with the return address back inline. */ + tcg_out_push(s, retaddr); + tcg_out_jmp(s, (uintptr_t)qemu_st_helpers[s_bits]); +} #elif defined(__x86_64__) && defined(__linux__) # include <asm/prctl.h> # include <sys/prctl.h> @@ -1250,46 +1413,36 @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg datalo, TCGReg datahi, common. */ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, TCGMemOp opc) { - int data_reg, data_reg2 = 0; - int addrlo_idx; + TCGReg datalo, datahi, addrlo; #if defined(CONFIG_SOFTMMU) + TCGReg addrhi; int mem_index; TCGMemOp s_bits; uint8_t *label_ptr[2]; #endif - data_reg = args[0]; - addrlo_idx = 1; - if (TCG_TARGET_REG_BITS == 32 && (opc & MO_SIZE) == MO_64) { - data_reg2 = args[1]; - addrlo_idx = 2; - } + datalo = *args++; + datahi = (TCG_TARGET_REG_BITS == 32 && opc == 3 ? *args++ : 0); + addrlo = *args++; #if defined(CONFIG_SOFTMMU) - mem_index = args[addrlo_idx + 1 + (TARGET_LONG_BITS > TCG_TARGET_REG_BITS)]; + addrhi = (TARGET_LONG_BITS > TCG_TARGET_REG_BITS ? *args++ : 0); + mem_index = *args++; s_bits = opc & MO_SIZE; - tcg_out_tlb_load(s, addrlo_idx, mem_index, s_bits, args, + tcg_out_tlb_load(s, addrlo, addrhi, mem_index, s_bits, label_ptr, offsetof(CPUTLBEntry, addr_read)); /* TLB Hit. */ - tcg_out_qemu_ld_direct(s, data_reg, data_reg2, TCG_REG_L1, 0, 0, opc); + tcg_out_qemu_ld_direct(s, datalo, datahi, TCG_REG_L1, 0, 0, opc); /* Record the current context of a load into ldst label */ - add_qemu_ldst_label(s, - 1, - opc, - data_reg, - data_reg2, - args[addrlo_idx], - args[addrlo_idx + 1], - mem_index, - s->code_ptr, - label_ptr); + add_qemu_ldst_label(s, 1, opc, datalo, datahi, addrlo, addrhi, + mem_index, s->code_ptr, label_ptr); #else { int32_t offset = GUEST_BASE; - int base = args[addrlo_idx]; + TCGReg base = addrlo; int seg = 0; /* ??? We assume all operations have left us with register contents @@ -1307,7 +1460,7 @@ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, TCGMemOp opc) offset = 0; } - tcg_out_qemu_ld_direct(s, data_reg, data_reg2, base, offset, seg, opc); + tcg_out_qemu_ld_direct(s, datalo, datahi, base, offset, seg, opc); } #endif } @@ -1374,46 +1527,36 @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg datalo, TCGReg datahi, static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, TCGMemOp opc) { - int data_reg, data_reg2 = 0; - int addrlo_idx; + TCGReg datalo, datahi, addrlo; #if defined(CONFIG_SOFTMMU) + TCGReg addrhi; int mem_index; TCGMemOp s_bits; uint8_t *label_ptr[2]; #endif - data_reg = args[0]; - addrlo_idx = 1; - if (TCG_TARGET_REG_BITS == 32 && opc == 3) { - data_reg2 = args[1]; - addrlo_idx = 2; - } + datalo = *args++; + datahi = (TCG_TARGET_REG_BITS == 32 && opc == 3 ? *args++ : 0); + addrlo = *args++; #if defined(CONFIG_SOFTMMU) - mem_index = args[addrlo_idx + 1 + (TARGET_LONG_BITS > TCG_TARGET_REG_BITS)]; + addrhi = (TARGET_LONG_BITS > TCG_TARGET_REG_BITS ? *args++ : 0); + mem_index = *args++; s_bits = opc & MO_SIZE; - tcg_out_tlb_load(s, addrlo_idx, mem_index, s_bits, args, + tcg_out_tlb_load(s, addrlo, addrhi, mem_index, s_bits, label_ptr, offsetof(CPUTLBEntry, addr_write)); /* TLB Hit. */ - tcg_out_qemu_st_direct(s, data_reg, data_reg2, TCG_REG_L1, 0, 0, opc); + tcg_out_qemu_st_direct(s, datalo, datahi, TCG_REG_L1, 0, 0, opc); /* Record the current context of a store into ldst label */ - add_qemu_ldst_label(s, - 0, - opc, - data_reg, - data_reg2, - args[addrlo_idx], - args[addrlo_idx + 1], - mem_index, - s->code_ptr, - label_ptr); + add_qemu_ldst_label(s, 0, opc, datalo, datahi, addrlo, addrhi, + mem_index, s->code_ptr, label_ptr); #else { int32_t offset = GUEST_BASE; - int base = args[addrlo_idx]; + TCGReg base = addrlo; int seg = 0; /* ??? We assume all operations have left us with register contents @@ -1431,194 +1574,10 @@ static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, TCGMemOp opc) offset = 0; } - tcg_out_qemu_st_direct(s, data_reg, data_reg2, base, offset, seg, opc); - } -#endif -} - -#if defined(CONFIG_SOFTMMU) -/* - * Record the context of a call to the out of line helper code for the slow path - * for a load or store, so that we can later generate the correct helper code - */ -static void add_qemu_ldst_label(TCGContext *s, - int is_ld, - int opc, - int data_reg, - int data_reg2, - int addrlo_reg, - int addrhi_reg, - int mem_index, - uint8_t *raddr, - uint8_t **label_ptr) -{ - TCGLabelQemuLdst *label = new_ldst_label(s); - - label->is_ld = is_ld; - label->opc = opc; - label->datalo_reg = data_reg; - label->datahi_reg = data_reg2; - label->addrlo_reg = addrlo_reg; - label->addrhi_reg = addrhi_reg; - label->mem_index = mem_index; - label->raddr = raddr; - label->label_ptr[0] = label_ptr[0]; - if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) { - label->label_ptr[1] = label_ptr[1]; - } -} - -/* - * Generate code for the slow path for a load at the end of block - */ -static void tcg_out_qemu_ld_slow_path(TCGContext *s, TCGLabelQemuLdst *l) -{ - TCGMemOp opc = l->opc; - TCGMemOp s_bits = opc & MO_SIZE; - TCGReg data_reg; - uint8_t **label_ptr = &l->label_ptr[0]; - - /* resolve label address */ - *(uint32_t *)label_ptr[0] = (uint32_t)(s->code_ptr - label_ptr[0] - 4); - if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) { - *(uint32_t *)label_ptr[1] = (uint32_t)(s->code_ptr - label_ptr[1] - 4); - } - - if (TCG_TARGET_REG_BITS == 32) { - int ofs = 0; - - tcg_out_st(s, TCG_TYPE_PTR, TCG_AREG0, TCG_REG_ESP, ofs); - ofs += 4; - - tcg_out_st(s, TCG_TYPE_I32, l->addrlo_reg, TCG_REG_ESP, ofs); - ofs += 4; - - if (TARGET_LONG_BITS == 64) { - tcg_out_st(s, TCG_TYPE_I32, l->addrhi_reg, TCG_REG_ESP, ofs); - ofs += 4; - } - - tcg_out_sti(s, TCG_TYPE_I32, TCG_REG_ESP, ofs, l->mem_index); - ofs += 4; - - tcg_out_sti(s, TCG_TYPE_I32, TCG_REG_ESP, ofs, (uintptr_t)l->raddr); - } else { - tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0], TCG_AREG0); - /* The second argument is already loaded with addrlo. */ - tcg_out_movi(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2], - l->mem_index); - tcg_out_movi(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[3], - (uintptr_t)l->raddr); + tcg_out_qemu_st_direct(s, datalo, datahi, base, offset, seg, opc); } - - tcg_out_calli(s, (uintptr_t)qemu_ld_helpers[s_bits]); - - data_reg = l->datalo_reg; - switch (opc & MO_SSIZE) { - case MO_SB: - tcg_out_ext8s(s, data_reg, TCG_REG_EAX, P_REXW); - break; - case MO_SW: - tcg_out_ext16s(s, data_reg, TCG_REG_EAX, P_REXW); - break; -#if TCG_TARGET_REG_BITS == 64 - case MO_SL: - tcg_out_ext32s(s, data_reg, TCG_REG_EAX); - break; #endif - case MO_UB: - case MO_UW: - /* Note that the helpers have zero-extended to tcg_target_long. */ - case MO_UL: - tcg_out_mov(s, TCG_TYPE_I32, data_reg, TCG_REG_EAX); - break; - case MO_Q: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_mov(s, TCG_TYPE_I64, data_reg, TCG_REG_RAX); - } else if (data_reg == TCG_REG_EDX) { - /* xchg %edx, %eax */ - tcg_out_opc(s, OPC_XCHG_ax_r32 + TCG_REG_EDX, 0, 0, 0); - tcg_out_mov(s, TCG_TYPE_I32, l->datahi_reg, TCG_REG_EAX); - } else { - tcg_out_mov(s, TCG_TYPE_I32, data_reg, TCG_REG_EAX); - tcg_out_mov(s, TCG_TYPE_I32, l->datahi_reg, TCG_REG_EDX); - } - break; - default: - tcg_abort(); - } - - /* Jump to the code corresponding to next IR of qemu_st */ - tcg_out_jmp(s, (uintptr_t)l->raddr); -} - -/* - * Generate code for the slow path for a store at the end of block - */ -static void tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *l) -{ - TCGMemOp opc = l->opc; - TCGMemOp s_bits = opc & MO_SIZE; - uint8_t **label_ptr = &l->label_ptr[0]; - TCGReg retaddr; - - /* resolve label address */ - *(uint32_t *)label_ptr[0] = (uint32_t)(s->code_ptr - label_ptr[0] - 4); - if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) { - *(uint32_t *)label_ptr[1] = (uint32_t)(s->code_ptr - label_ptr[1] - 4); - } - - if (TCG_TARGET_REG_BITS == 32) { - int ofs = 0; - - tcg_out_st(s, TCG_TYPE_PTR, TCG_AREG0, TCG_REG_ESP, ofs); - ofs += 4; - - tcg_out_st(s, TCG_TYPE_I32, l->addrlo_reg, TCG_REG_ESP, ofs); - ofs += 4; - - if (TARGET_LONG_BITS == 64) { - tcg_out_st(s, TCG_TYPE_I32, l->addrhi_reg, TCG_REG_ESP, ofs); - ofs += 4; - } - - tcg_out_st(s, TCG_TYPE_I32, l->datalo_reg, TCG_REG_ESP, ofs); - ofs += 4; - - if (s_bits == MO_64) { - tcg_out_st(s, TCG_TYPE_I32, l->datahi_reg, TCG_REG_ESP, ofs); - ofs += 4; - } - - tcg_out_sti(s, TCG_TYPE_I32, TCG_REG_ESP, ofs, l->mem_index); - ofs += 4; - - retaddr = TCG_REG_EAX; - tcg_out_movi(s, TCG_TYPE_I32, retaddr, (uintptr_t)l->raddr); - tcg_out_st(s, TCG_TYPE_I32, retaddr, TCG_REG_ESP, ofs); - } else { - tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0], TCG_AREG0); - /* The second argument is already loaded with addrlo. */ - tcg_out_mov(s, (s_bits == MO_64 ? TCG_TYPE_I64 : TCG_TYPE_I32), - tcg_target_call_iarg_regs[2], l->datalo_reg); - tcg_out_movi(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3], - l->mem_index); - - if (ARRAY_SIZE(tcg_target_call_iarg_regs) > 4) { - retaddr = tcg_target_call_iarg_regs[4]; - tcg_out_movi(s, TCG_TYPE_PTR, retaddr, (uintptr_t)l->raddr); - } else { - retaddr = TCG_REG_RAX; - tcg_out_movi(s, TCG_TYPE_PTR, retaddr, (uintptr_t)l->raddr); - tcg_out_st(s, TCG_TYPE_PTR, retaddr, TCG_REG_ESP, 0); - } - } - - /* "Tail call" to the helper, with the return address back inline. */ - tcg_out_push(s, retaddr); - tcg_out_jmp(s, (uintptr_t)qemu_st_helpers[s_bits]); } -#endif /* CONFIG_SOFTMMU */ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc, const TCGArg *args, const int *const_args) -- 1.8.1.4