diff options
Diffstat (limited to 'target/riscv/translate.c')
| -rw-r--r-- | target/riscv/translate.c | 54 |
1 files changed, 35 insertions, 19 deletions
diff --git a/target/riscv/translate.c b/target/riscv/translate.c index cef61b5..9ddef2d 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -20,7 +20,6 @@ #include "qemu/log.h" #include "cpu.h" #include "tcg/tcg-op.h" -#include "exec/exec-all.h" #include "exec/helper-proto.h" #include "exec/helper-gen.h" #include "exec/target_page.h" @@ -1210,11 +1209,6 @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc) /* The specification allows for longer insns, but not supported by qemu. */ #define MAX_INSN_LEN 4 -static inline int insn_len(uint16_t first_word) -{ - return (first_word & 3) == 3 ? 4 : 2; -} - const RISCVDecoder decoder_table[] = { { always_true_p, decode_insn32 }, { has_xthead_p, decode_xthead}, @@ -1223,13 +1217,35 @@ const RISCVDecoder decoder_table[] = { const size_t decoder_table_size = ARRAY_SIZE(decoder_table); -static void decode_opc(CPURISCVState *env, DisasContext *ctx, uint16_t opcode) +static void decode_opc(CPURISCVState *env, DisasContext *ctx) { + uint32_t opcode; + bool pc_is_4byte_align = ((ctx->base.pc_next % 4) == 0); + ctx->virt_inst_excp = false; - ctx->cur_insn_len = insn_len(opcode); + if (pc_is_4byte_align) { + /* + * Load 4 bytes at once to make instruction fetch atomically. + * + * Note: When pc is 4-byte aligned, 4-byte instruction wouldn't be + * across pages. We could preload 4 bytes instruction no matter + * real one is 2 or 4 bytes. Instruction preload wouldn't trigger + * additional page fault. + */ + opcode = translator_ldl(env, &ctx->base, ctx->base.pc_next); + } else { + /* + * For unaligned pc, instruction preload may trigger additional + * page fault so we only load 2 bytes here. + */ + opcode = (uint32_t) translator_lduw(env, &ctx->base, ctx->base.pc_next); + } + ctx->ol = ctx->xl; + + ctx->cur_insn_len = insn_len((uint16_t)opcode); /* Check for compressed insn */ if (ctx->cur_insn_len == 2) { - ctx->opcode = opcode; + ctx->opcode = (uint16_t)opcode; /* * The Zca extension is added as way to refer to instructions in the C * extension that do not include the floating-point loads and stores @@ -1239,15 +1255,17 @@ static void decode_opc(CPURISCVState *env, DisasContext *ctx, uint16_t opcode) return; } } else { - uint32_t opcode32 = opcode; - opcode32 = deposit32(opcode32, 16, 16, - translator_lduw(env, &ctx->base, - ctx->base.pc_next + 2)); - ctx->opcode = opcode32; + if (!pc_is_4byte_align) { + /* Load last 2 bytes of instruction here */ + opcode = deposit32(opcode, 16, 16, + translator_lduw(env, &ctx->base, + ctx->base.pc_next + 2)); + } + ctx->opcode = opcode; for (guint i = 0; i < ctx->decoders->len; ++i) { riscv_cpu_decode_fn func = g_ptr_array_index(ctx->decoders, i); - if (func(ctx, opcode32)) { + if (func(ctx, opcode)) { return; } } @@ -1282,7 +1300,7 @@ static void riscv_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) ctx->cfg_vta_all_1s = cpu->cfg.rvv_ta_all_1s; ctx->vstart_eq_zero = FIELD_EX32(tb_flags, TB_FLAGS, VSTART_EQ_ZERO); ctx->vl_eq_vlmax = FIELD_EX32(tb_flags, TB_FLAGS, VL_EQ_VLMAX); - ctx->misa_mxl_max = mcc->misa_mxl_max; + ctx->misa_mxl_max = mcc->def->misa_mxl_max; ctx->xl = FIELD_EX32(tb_flags, TB_FLAGS, XL); ctx->address_xl = FIELD_EX32(tb_flags, TB_FLAGS, AXL); ctx->cs = cs; @@ -1325,10 +1343,8 @@ static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu) { DisasContext *ctx = container_of(dcbase, DisasContext, base); CPURISCVState *env = cpu_env(cpu); - uint16_t opcode16 = translator_lduw(env, &ctx->base, ctx->base.pc_next); - ctx->ol = ctx->xl; - decode_opc(env, ctx, opcode16); + decode_opc(env, ctx); ctx->base.pc_next += ctx->cur_insn_len; /* |