unicorn/qemu/target/arm/translate.h

303 lines
10 KiB
C
Raw Normal View History

2015-08-21 07:04:50 +00:00
#ifndef TARGET_ARM_TRANSLATE_H
#define TARGET_ARM_TRANSLATE_H
#include "exec/translator.h"
#include "internals.h"
2015-08-21 07:04:50 +00:00
/* internal defines */
typedef struct DisasContext {
DisasContextBase base;
const ARMISARegisters *isar;
/* The address of the current instruction being translated. */
target_ulong pc_curr;
target_ulong page_start;
2015-08-21 07:04:50 +00:00
uint32_t insn;
/* Nonzero if this instruction has been conditionally skipped. */
int condjmp;
/* The label that will be jumped to when the instruction is skipped. */
TCGLabel *condlabel;
2015-08-21 07:04:50 +00:00
/* Thumb-2 conditional execution bits. */
int condexec_mask;
int condexec_cond;
int thumb;
int sctlr_b;
TCGMemOp be_data;
2015-08-21 07:04:50 +00:00
#if !defined(CONFIG_USER_ONLY)
int user;
#endif
ARMMMUIdx mmu_idx; /* MMU index to use for normal loads/stores */
uint8_t tbii; /* TBI1|TBI0 for insns */
uint8_t tbid; /* TBI1|TBI0 for data */
bool ns; /* Use non-secure CPREG bank on access */
int fp_excp_el; /* FP exception EL or 0 if enabled */
int sve_excp_el; /* SVE exception EL or 0 if enabled */
int sve_len; /* SVE vector length in bytes */
/* Flag indicating that exceptions from secure mode are routed to EL3. */
bool secure_routed_to_el3;
2015-08-21 07:04:50 +00:00
bool vfp_enabled; /* FP enabled via FPSCR.EN */
int vec_len;
int vec_stride;
bool v7m_handler_mode;
bool v8m_secure; /* true if v8M and we're in Secure mode */
bool v8m_stackcheck; /* true if we need to perform v8M stack limit checks */
bool v8m_fpccr_s_wrong; /* true if v8M FPCCR.S != v8m_secure */
bool v7m_new_fp_ctxt_needed; /* ASPEN set but no active FP context */
bool v7m_lspact; /* FPCCR.LSPACT set */
2015-08-21 07:04:50 +00:00
/* Immediate value in AArch32 SVC insn; must be set if is_jmp == DISAS_SWI
* so that top level loop can generate correct syndrome information.
*/
uint32_t svc_imm;
int aarch64;
int current_el;
/* Debug target exception level for single-step exceptions */
int debug_target_el;
2015-08-21 07:04:50 +00:00
GHashTable *cp_regs;
uint64_t features; /* CPU features bits */
/* Because unallocated encodings generate different exception syndrome
* information from traps due to FP being disabled, we can't do a single
* "is fp access disabled" check at a high level in the decode tree.
* To help in catching bugs where the access check was forgotten in some
* code path, we set this flag when the access check is done, and assert
* that it is set at the point where we actually touch the FP regs.
*/
bool fp_access_checked;
/* ARMv8 single-step state (this is distinct from the QEMU gdbstub
* single-step support).
*/
bool ss_active;
bool pstate_ss;
/* True if the insn just emitted was a load-exclusive instruction
* (necessary for syndrome information for single step exceptions),
* ie A64 LDX*, LDAX*, A32/T32 LDREX*, LDAEX*.
*/
bool is_ldex;
/* True if v8.3-PAuth is active. */
bool pauth_active;
2015-08-21 07:04:50 +00:00
/* Bottom two bits of XScale c15_cpar coprocessor access control reg */
int c15_cpar;
/* True with v8.5-BTI and SCTLR_ELx.BT* set. */
bool bt;
/*
* >= 0, a copy of PSTATE.BTYPE, which will be 0 without v8.5-BTI.
* < 0, set by the current instruction.
*/
int8_t btype;
/* True if this page is guarded. */
bool guarded_page;
/* TCG op of the current insn_start. */
TCGOp *insn_start;
2015-08-21 07:04:50 +00:00
#define TMP_A64_MAX 16
int tmp_a64_count;
TCGv_i64 tmp_a64[TMP_A64_MAX];
2018-10-06 07:31:27 +00:00
// Unicorn: Moved here to avoid global state.
2018-10-06 07:26:52 +00:00
TCGv_i64 V0;
2018-10-06 07:28:42 +00:00
TCGv_i64 V1;
2018-10-06 07:31:27 +00:00
TCGv_i64 M0;
2015-08-21 07:04:50 +00:00
// Unicorn engine
struct uc_struct *uc;
} DisasContext;
typedef struct DisasCompare {
TCGCond cond;
TCGv_i32 value;
bool value_global;
} DisasCompare;
2015-08-21 07:04:50 +00:00
void unallocated_encoding(DisasContext *s);
2015-08-21 07:04:50 +00:00
static inline int arm_dc_feature(DisasContext *dc, int feature)
{
return (dc->features & (1ULL << feature)) != 0;
}
static inline int get_mem_index(DisasContext *s)
{
return arm_to_core_mmu_idx(s->mmu_idx);
2015-08-21 07:04:50 +00:00
}
/* Function used to determine the target exception EL when otherwise not known
* or default.
*/
static inline int default_exception_el(DisasContext *s)
{
/* If we are coming from secure EL0 in a system with a 32-bit EL3, then
* there is no secure EL1, so we route exceptions to EL3. Otherwise,
* exceptions can only be routed to ELs above 1, so we target the higher of
* 1 or the current EL.
*/
return (s->mmu_idx == ARMMMUIdx_S1SE0 && s->secure_routed_to_el3)
? 3 : MAX(1, s->current_el);
}
static inline void disas_set_insn_syndrome(DisasContext *s, uint32_t syn)
{
/* We don't need to save all of the syndrome so we mask and shift
* out unneeded bits to help the sleb128 encoder do a better job.
*/
syn &= ARM_INSN_START_WORD2_MASK;
syn >>= ARM_INSN_START_WORD2_SHIFT;
/* We check and clear insn_start_idx to catch multiple updates. */
assert(s->insn_start != NULL);
tcg_set_insn_start_param(s->insn_start, 2, syn);
s->insn_start = NULL;
}
2015-08-21 07:04:50 +00:00
/* target-specific extra values for is_jmp */
/* is_jmp field values */
#define DISAS_JUMP DISAS_TARGET_0 /* only pc was modified dynamically */
#define DISAS_UPDATE DISAS_TARGET_1 /* cpu state was modified dynamically */
2015-08-21 07:04:50 +00:00
/* These instructions trap after executing, so the A32/T32 decoder must
* defer them until after the conditional execution state has been updated.
* WFI also needs special handling when single-stepping.
*/
#define DISAS_WFI DISAS_TARGET_2
#define DISAS_SWI DISAS_TARGET_3
2015-08-21 07:04:50 +00:00
/* WFE */
#define DISAS_WFE DISAS_TARGET_4
#define DISAS_HVC DISAS_TARGET_5
#define DISAS_SMC DISAS_TARGET_6
#define DISAS_YIELD DISAS_TARGET_7
arm: Implement M profile exception return properly On M profile, return from exceptions happen when code in Handler mode executes one of the following function call return instructions: * POP or LDM which loads the PC * LDR to PC * BX register and the new PC value is 0xFFxxxxxx. QEMU tries to implement this by not treating the instruction specially but then catching the attempt to execute from the magic address value. This is not ideal, because: * there are guest visible differences from the architecturally specified behaviour (for instance jumping to 0xFFxxxxxx via a different instruction should not cause an exception return but it will in the QEMU implementation) * we have to account for it in various places (like refusing to take an interrupt if the PC is at a magic value, and making sure that the MPU doesn't deny execution at the magic value addresses) Drop these hacks, and instead implement exception return the way the architecture specifies -- by having the relevant instructions check for the magic value and raise the 'do an exception return' QEMU internal exception immediately. The effect on the generated code is minor: bx lr, old code (and new code for Thread mode): TCG: mov_i32 tmp5,r14 movi_i32 tmp6,$0xfffffffffffffffe and_i32 pc,tmp5,tmp6 movi_i32 tmp6,$0x1 and_i32 tmp5,tmp5,tmp6 st_i32 tmp5,env,$0x218 exit_tb $0x0 set_label $L0 exit_tb $0x7f2aabd61993 x86_64 generated code: 0x7f2aabe87019: mov %ebx,%ebp 0x7f2aabe8701b: and $0xfffffffffffffffe,%ebp 0x7f2aabe8701e: mov %ebp,0x3c(%r14) 0x7f2aabe87022: and $0x1,%ebx 0x7f2aabe87025: mov %ebx,0x218(%r14) 0x7f2aabe8702c: xor %eax,%eax 0x7f2aabe8702e: jmpq 0x7f2aabe7c016 bx lr, new code when in Handler mode: TCG: mov_i32 tmp5,r14 movi_i32 tmp6,$0xfffffffffffffffe and_i32 pc,tmp5,tmp6 movi_i32 tmp6,$0x1 and_i32 tmp5,tmp5,tmp6 st_i32 tmp5,env,$0x218 movi_i32 tmp5,$0xffffffffff000000 brcond_i32 pc,tmp5,geu,$L1 exit_tb $0x0 set_label $L1 movi_i32 tmp5,$0x8 call exception_internal,$0x0,$0,env,tmp5 x86_64 generated code: 0x7fe8fa1264e3: mov %ebp,%ebx 0x7fe8fa1264e5: and $0xfffffffffffffffe,%ebx 0x7fe8fa1264e8: mov %ebx,0x3c(%r14) 0x7fe8fa1264ec: and $0x1,%ebp 0x7fe8fa1264ef: mov %ebp,0x218(%r14) 0x7fe8fa1264f6: cmp $0xff000000,%ebx 0x7fe8fa1264fc: jae 0x7fe8fa126509 0x7fe8fa126502: xor %eax,%eax 0x7fe8fa126504: jmpq 0x7fe8fa122016 0x7fe8fa126509: mov %r14,%rdi 0x7fe8fa12650c: mov $0x8,%esi 0x7fe8fa126511: mov $0x56095dbeccf5,%r10 0x7fe8fa12651b: callq *%r10 which is a difference of one cmp/branch-not-taken. This will be lost in the noise of having to exit generated code and look up the next TB anyway. Backports commit 3bb8a96f5348913ee130169504f3642f501b113e from qemu
2018-03-02 19:56:27 +00:00
/* M profile branch which might be an exception return (and so needs
* custom end-of-TB code)
*/
#define DISAS_BX_EXCRET DISAS_TARGET_8
target/arm: optimize indirect branches Speed up indirect branches by jumping to the target if it is valid. Softmmu measurements (see later commit for user-mode results): Note: baseline (i.e. speedup == 1x) is QEMU v2.9.0. - Impact on Boot time | setup | ARM debian jessie boot+shutdown time | stddev | |--------+--------------------------------------+--------| | v2.9.0 | 8.84 | 0.07 | | +cross | 8.85 | 0.03 | | +jr | 8.83 | 0.06 | - NBench, arm-softmmu (debian jessie guest). Host: Intel i7-4790K @ 4.00GHz 1.3x +-+-------------------------------------------------------------------------------------------------------------+-+ | | | cross #### | 1.25x +cross+jr..........................................................#++#.........................................+-+ | #### # # | | +++# # # # | | +++ **** # # # | 1.2x +-+...................................####............*..*..#......#..#.........................................+-+ | **** # * * # # # #### | | * * # * * # # # # # | 1.15x +-+................................*..*..#............*..*..#......#..#.....#..#................................+-+ | * * # * * # # # # # | | * * # #### * * # # # # # | | * * # # # * * # # # # # #### | 1.1x +-+................................*..*..#......#..#..*..*..#......#..#.....#..#.........................#..#...+-+ | * * # # # * * # # # # # # # | | * * # # # * * # # # # # # # | 1.05x +-+..........................####..*..*..#......#..#..*..*..#......#..#.....#..#......+++............*****..#...+-+ | ***** # * * # # # * * # ***** # # # +++ | ****### * * # | | *+++* # * * # # # * * # *+++* # **** # *****### * * # * * # | | *****### +++#### * * # * * # ***** # * * # * * # * * # * | *++# * * # * * # | 1x +-++-+*+++*-+#++****++#++*+-+*++#+-*++*++#-+*+++*-+#++*++*++#++*+-+*++#+-*++*++#-+*+++*-+#++*++*++#++*+-+*++#+-++-+ | * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # | | * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # | 0.95x +-+---*****###--****###--*****###--****###--*****###--****###--*****###--****###--*****###--****###--*****###---+-+ ASSIGNMENT BITFIELD FOURFP EMULATION HUFFMAN LU DECOMPOSITIONEURAL NNUMERIC SOSTRING SORT hmean png: http://imgur.com/eOLmZNR NB. 'cross' represents the previous commit. Backports commit 8a6b28c7b5104263344508df0f4bce97f22cfcaf from qemu
2018-03-03 02:16:56 +00:00
/* For instructions which want an immediate exit to the main loop,
* as opposed to attempting to use lookup_and_goto_ptr. Unlike
* DISAS_UPDATE this doesn't write the PC on exiting the translation
* loop so you need to ensure something (gen_a64_set_pc_im or runtime
* helper) has done so before we reach return from cpu_tb_exec.
target/arm: optimize indirect branches Speed up indirect branches by jumping to the target if it is valid. Softmmu measurements (see later commit for user-mode results): Note: baseline (i.e. speedup == 1x) is QEMU v2.9.0. - Impact on Boot time | setup | ARM debian jessie boot+shutdown time | stddev | |--------+--------------------------------------+--------| | v2.9.0 | 8.84 | 0.07 | | +cross | 8.85 | 0.03 | | +jr | 8.83 | 0.06 | - NBench, arm-softmmu (debian jessie guest). Host: Intel i7-4790K @ 4.00GHz 1.3x +-+-------------------------------------------------------------------------------------------------------------+-+ | | | cross #### | 1.25x +cross+jr..........................................................#++#.........................................+-+ | #### # # | | +++# # # # | | +++ **** # # # | 1.2x +-+...................................####............*..*..#......#..#.........................................+-+ | **** # * * # # # #### | | * * # * * # # # # # | 1.15x +-+................................*..*..#............*..*..#......#..#.....#..#................................+-+ | * * # * * # # # # # | | * * # #### * * # # # # # | | * * # # # * * # # # # # #### | 1.1x +-+................................*..*..#......#..#..*..*..#......#..#.....#..#.........................#..#...+-+ | * * # # # * * # # # # # # # | | * * # # # * * # # # # # # # | 1.05x +-+..........................####..*..*..#......#..#..*..*..#......#..#.....#..#......+++............*****..#...+-+ | ***** # * * # # # * * # ***** # # # +++ | ****### * * # | | *+++* # * * # # # * * # *+++* # **** # *****### * * # * * # | | *****### +++#### * * # * * # ***** # * * # * * # * * # * | *++# * * # * * # | 1x +-++-+*+++*-+#++****++#++*+-+*++#+-*++*++#-+*+++*-+#++*++*++#++*+-+*++#+-*++*++#-+*+++*-+#++*++*++#++*+-+*++#+-++-+ | * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # | | * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # | 0.95x +-+---*****###--****###--*****###--****###--*****###--****###--*****###--****###--*****###--****###--*****###---+-+ ASSIGNMENT BITFIELD FOURFP EMULATION HUFFMAN LU DECOMPOSITIONEURAL NNUMERIC SOSTRING SORT hmean png: http://imgur.com/eOLmZNR NB. 'cross' represents the previous commit. Backports commit 8a6b28c7b5104263344508df0f4bce97f22cfcaf from qemu
2018-03-03 02:16:56 +00:00
*/
#define DISAS_EXIT DISAS_TARGET_9
2015-08-21 07:04:50 +00:00
#ifdef TARGET_AARCH64
void a64_translate_init(struct uc_struct *uc);
void gen_a64_set_pc_im(DisasContext *s, uint64_t val);
extern const TranslatorOps aarch64_translator_ops;
2015-08-21 07:04:50 +00:00
#else
static inline void a64_translate_init(struct uc_struct *uc)
{
}
static inline void gen_a64_set_pc_im(DisasContext *s, uint64_t val)
2015-08-21 07:04:50 +00:00
{
}
#endif
void arm_test_cc(DisasContext *s, DisasCompare *cmp, int cc);
void arm_free_cc(DisasContext *s, DisasCompare *cmp);
void arm_jump_cc(DisasContext *s, DisasCompare *cmp, TCGLabel *label);
void arm_gen_test_cc(DisasContext *s, int cc, TCGLabel *label);
2015-08-21 07:04:50 +00:00
/* Return state of Alternate Half-precision flag, caller frees result */
static inline TCGv_i32 get_ahp_flag(DisasContext *s)
{
TCGContext *tcg_ctx = s->uc->tcg_ctx;
TCGv_i32 ret = tcg_temp_new_i32(tcg_ctx);
tcg_gen_ld_i32(tcg_ctx, ret, tcg_ctx->cpu_env,
offsetof(CPUARMState, vfp.xregs[ARM_VFP_FPSCR]));
tcg_gen_extract_i32(tcg_ctx, ret, ret, 26, 1);
return ret;
}
/* Set bits within PSTATE. */
static inline void set_pstate_bits(DisasContext *s, uint32_t bits)
{
TCGContext *tcg_ctx = s->uc->tcg_ctx;
TCGv_i32 p = tcg_temp_new_i32(tcg_ctx);
tcg_debug_assert(!(bits & CACHED_PSTATE_BITS));
tcg_gen_ld_i32(tcg_ctx, p, tcg_ctx->cpu_env, offsetof(CPUARMState, pstate));
tcg_gen_ori_i32(tcg_ctx, p, p, bits);
tcg_gen_st_i32(tcg_ctx, p, tcg_ctx->cpu_env, offsetof(CPUARMState, pstate));
tcg_temp_free_i32(tcg_ctx, p);
}
/* Clear bits within PSTATE. */
static inline void clear_pstate_bits(DisasContext *s, uint32_t bits)
{
TCGContext *tcg_ctx = s->uc->tcg_ctx;
TCGv_i32 p = tcg_temp_new_i32(tcg_ctx);
tcg_debug_assert(!(bits & CACHED_PSTATE_BITS));
tcg_gen_ld_i32(tcg_ctx, p, tcg_ctx->cpu_env, offsetof(CPUARMState, pstate));
tcg_gen_andi_i32(tcg_ctx, p, p, ~bits);
tcg_gen_st_i32(tcg_ctx, p, tcg_ctx->cpu_env, offsetof(CPUARMState, pstate));
tcg_temp_free_i32(tcg_ctx, p);
}
/* If the singlestep state is Active-not-pending, advance to Active-pending. */
static inline void gen_ss_advance(DisasContext *s)
{
if (s->ss_active) {
s->pstate_ss = 0;
clear_pstate_bits(s, PSTATE_SS);
}
}
static inline void gen_exception(DisasContext *s, int excp, uint32_t syndrome,
uint32_t target_el)
{
TCGContext *tcg_ctx = s->uc->tcg_ctx;
TCGv_i32 tcg_excp = tcg_const_i32(tcg_ctx, excp);
TCGv_i32 tcg_syn = tcg_const_i32(tcg_ctx, syndrome);
TCGv_i32 tcg_el = tcg_const_i32(tcg_ctx, target_el);
gen_helper_exception_with_syndrome(tcg_ctx, tcg_ctx->cpu_env, tcg_excp,
tcg_syn, tcg_el);
tcg_temp_free_i32(tcg_ctx, tcg_el);
tcg_temp_free_i32(tcg_ctx, tcg_syn);
tcg_temp_free_i32(tcg_ctx, tcg_excp);
}
/* Generate an architectural singlestep exception */
static inline void gen_swstep_exception(DisasContext *s, int isv, int ex)
{
bool same_el = (s->debug_target_el == s->current_el);
/*
* If singlestep is targeting a lower EL than the current one,
* then s->ss_active must be false and we can never get here.
*/
assert(s->debug_target_el >= s->current_el);
gen_exception(s, EXCP_UDEF, syn_swstep(same_el, isv, ex), s->debug_target_el);
}
/*
* Given a VFP floating point constant encoded into an 8 bit immediate in an
* instruction, expand it to the actual constant value of the specified
* size, as per the VFPExpandImm() pseudocode in the Arm ARM.
*/
uint64_t vfp_expand_imm(int size, uint8_t imm8);
/* Vector operations shared between ARM and AArch64. */
extern const GVecGen3 cmtst_op[4];
extern const GVecGen3 mla_op[4];
extern const GVecGen3 mls_op[4];
2018-11-10 15:32:29 +00:00
extern const GVecGen2i ssra_op[4];
extern const GVecGen2i usra_op[4];
extern const GVecGen2i sri_op[4];
extern const GVecGen2i sli_op[4];
extern const GVecGen4 uqadd_op[4];
extern const GVecGen4 sqadd_op[4];
extern const GVecGen4 uqsub_op[4];
extern const GVecGen4 sqsub_op[4];
void gen_cmtst_i64(TCGContext* tcg_ctx, TCGv_i64 d, TCGv_i64 a, TCGv_i64 b);
/*
* Forward to the isar_feature_* tests given a DisasContext pointer.
*/
#define dc_isar_feature(name, ctx) \
({ DisasContext *ctx_ = (ctx); isar_feature_##name(ctx_->isar); })
2015-08-21 07:04:50 +00:00
#endif /* TARGET_ARM_TRANSLATE_H */