mirrors/qemu
0
0
Fork 0
mirror of https://github.com/qemu/qemu.git synced 2024-12-12 05:03:42 +08:00
Code Issues Packages Projects Releases Wiki Activity

* i386: fix issue with cache topology passthrough

Browse Source 
* scsi-disk: migrate emulated requests
 * i386/sev: fix Coverity issues
 * i386/tcg: more conversions to new decoder
 -----BEGIN PGP SIGNATURE-----
 
 iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmZv6kMUHHBib256aW5p
 QHJlZGhhdC5jb20ACgkQv/vSX3jHroOn4Af/evnpsae1fm8may1NQmmezKiks/4X
 cR0GaQ7w75Oas05jKsG7Xnrq3Vn6p5wllf3Wf00p7F1iJX18azY9rQgIsUVUgVem
 /EIZk1eM6+mDxuIG0taPxc5Aw3cfIBWAjUmzsXrSr55e/wyiIxZCeUo2zk8Il+iL
 Z4ceNzY5PZzc2Fl10D3cGs/+ynfiDM53ucwe3ve2T6NrxEVfKQPp5jkIUkBUba6z
 zM5O4Q5KTEZYVth1gbDTB/uUJLUFjQ12kCQfRCNX+bEPDHwARr0UWr/Oxtz0jZSd
 FvXohz7tI+v+ph0xHyE4tEFqryvLCII1td2ohTAYZZXNGkjK6XZildngBw==
 =m4BE
 -----END PGP SIGNATURE-----

Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging

* i386: fix issue with cache topology passthrough
* scsi-disk: migrate emulated requests
* i386/sev: fix Coverity issues
* i386/tcg: more conversions to new decoder

# -----BEGIN PGP SIGNATURE-----
#
# iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmZv6kMUHHBib256aW5p
# QHJlZGhhdC5jb20ACgkQv/vSX3jHroOn4Af/evnpsae1fm8may1NQmmezKiks/4X
# cR0GaQ7w75Oas05jKsG7Xnrq3Vn6p5wllf3Wf00p7F1iJX18azY9rQgIsUVUgVem
# /EIZk1eM6+mDxuIG0taPxc5Aw3cfIBWAjUmzsXrSr55e/wyiIxZCeUo2zk8Il+iL
# Z4ceNzY5PZzc2Fl10D3cGs/+ynfiDM53ucwe3ve2T6NrxEVfKQPp5jkIUkBUba6z
# zM5O4Q5KTEZYVth1gbDTB/uUJLUFjQ12kCQfRCNX+bEPDHwARr0UWr/Oxtz0jZSd
# FvXohz7tI+v+ph0xHyE4tEFqryvLCII1td2ohTAYZZXNGkjK6XZildngBw==
# =m4BE
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 17 Jun 2024 12:48:19 AM PDT
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]

* tag 'for-upstream' of https://gitlab.com/bonzini/qemu: (25 commits)
  target/i386: SEV: do not assume machine->cgs is SEV
  target/i386: convert CMPXCHG to new decoder
  target/i386: convert XADD to new decoder
  target/i386: convert LZCNT/TZCNT/BSF/BSR/POPCNT to new decoder
  target/i386: convert SHLD/SHRD to new decoder
  target/i386: adapt gen_shift_count for SHLD/SHRD
  target/i386: pull load/writeback out of gen_shiftd_rm_T1
  target/i386: convert non-grouped, helper-based 2-byte opcodes
  target/i386: split X86_CHECK_prot into PE and VM86 checks
  target/i386: finish converting 0F AE to the new decoder
  target/i386: fix bad sorting of entries in the 0F table
  target/i386: replace read_crN helper with read_cr8
  target/i386: convert MOV from/to CR and DR to new decoder
  target/i386: fix processing of intercept 0 (read CR0)
  target/i386: replace NoSeg special with NoLoadEA
  target/i386: change X86_ENTRYwr to use T0, use it for moves
  target/i386: change X86_ENTRYr to use T0
  target/i386: put BLS* input in T1, use generic flag writeback
  target/i386: rewrite flags writeback for ADCX/ADOX
  target/i386: remove CPUX86State argument from generator functions
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This commit is contained in:
Richard Henderson 2024-06-17 11:38:44 -07:00
commit 85743f54fa
12 changed files with 1173 additions and 1219 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
hw/core/machine.c
View File

@ -36,6 +36,7 @@
GlobalProperty hw_compat_9_0[] = {
{"arm-cpu", "backcompat-cntfrq", "true" },
{"scsi-disk-base", "migrate-emulated-scsi-request", "false" },
{"vfio-pci", "skip-vsc-check", "false" },
};
const size_t hw_compat_9_0_len = G_N_ELEMENTS(hw_compat_9_0);

24
hw/scsi/scsi-disk.c
View File

@ -114,6 +114,7 @@ struct SCSIDiskState {
* 0xffff - reserved
*/
uint16_t rotation_rate;
bool migrate_emulated_scsi_request;
};
static void scsi_free_request(SCSIRequest *req)
@ -162,6 +163,15 @@ static void scsi_disk_save_request(QEMUFile *f, SCSIRequest *req)
}
}
static void scsi_disk_emulate_save_request(QEMUFile *f, SCSIRequest *req)
{
SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, req->dev);
if (s->migrate_emulated_scsi_request) {
scsi_disk_save_request(f, req);
}
}
static void scsi_disk_load_request(QEMUFile *f, SCSIRequest *req)
{
SCSIDiskReq *r = DO_UPCAST(SCSIDiskReq, req, req);
@ -185,6 +195,15 @@ static void scsi_disk_load_request(QEMUFile *f, SCSIRequest *req)
qemu_iovec_init_external(&r->qiov, &r->iov, 1);
}
static void scsi_disk_emulate_load_request(QEMUFile *f, SCSIRequest *req)
{
SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, req->dev);
if (s->migrate_emulated_scsi_request) {
scsi_disk_load_request(f, req);
}
}
/*
* scsi_handle_rw_error has two return values. False means that the error
* must be ignored, true means that the error has been processed and the
@ -2606,6 +2625,8 @@ static const SCSIReqOps scsi_disk_emulate_reqops = {
.read_data = scsi_disk_emulate_read_data,
.write_data = scsi_disk_emulate_write_data,
.get_buf = scsi_get_buf,
.load_request = scsi_disk_emulate_load_request,
.save_request = scsi_disk_emulate_save_request,
};
static const SCSIReqOps scsi_disk_dma_reqops = {
@ -3114,7 +3135,8 @@ static const TypeInfo scsi_disk_base_info = {
DEFINE_PROP_STRING("serial", SCSIDiskState, serial), \
DEFINE_PROP_STRING("vendor", SCSIDiskState, vendor), \
DEFINE_PROP_STRING("product", SCSIDiskState, product), \
DEFINE_PROP_STRING("device_id", SCSIDiskState, device_id)
DEFINE_PROP_STRING("device_id", SCSIDiskState, device_id), \
DEFINE_PROP_BOOL("migrate-emulated-scsi-request", SCSIDiskState, migrate_emulated_scsi_request, true)
static Property scsi_hd_properties[] = {

6
target/i386/cpu.c
View File

@ -6455,10 +6455,8 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
if (*eax & 31) {
int host_vcpus_per_cache = 1 + ((*eax & 0x3FFC000) >> 14);
if (cores_per_pkg > 1) {
*eax &= ~0xFC000000;
*eax |= max_core_ids_in_package(&topo_info) << 26;
}
*eax &= ~0xFC000000;
*eax |= max_core_ids_in_package(&topo_info) << 26;
if (host_vcpus_per_cache > threads_per_pkg) {
*eax &= ~0x3FFC000;

9
target/i386/cpu.h
View File

@ -1260,6 +1260,8 @@ uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
/* Use a clearer name for this. */
#define CPU_INTERRUPT_INIT CPU_INTERRUPT_RESET
#define CC_OP_HAS_EFLAGS(op) ((op) >= CC_OP_EFLAGS && (op) <= CC_OP_ADCOX)
/* Instead of computing the condition codes after each x86 instruction,
* QEMU just stores one operand (called CC_SRC), the result
* (called CC_DST) and the type of operation (called CC_OP). When the
@ -1270,6 +1272,9 @@ uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
typedef enum {
CC_OP_DYNAMIC, /* must use dynamic code to get cc_op */
CC_OP_EFLAGS, /* all cc are explicitly computed, CC_SRC = flags */
CC_OP_ADCX, /* CC_DST = C, CC_SRC = rest. */
CC_OP_ADOX, /* CC_SRC2 = O, CC_SRC = rest. */
CC_OP_ADCOX, /* CC_DST = C, CC_SRC2 = O, CC_SRC = rest. */
CC_OP_MULB, /* modify all flags, C, O = (CC_SRC != 0) */
CC_OP_MULW,
@ -1326,10 +1331,6 @@ typedef enum {
CC_OP_BMILGL,
CC_OP_BMILGQ,
CC_OP_ADCX, /* CC_DST = C, CC_SRC = rest. */
CC_OP_ADOX, /* CC_DST = O, CC_SRC = rest. */
CC_OP_ADCOX, /* CC_DST = C, CC_SRC2 = O, CC_SRC = rest. */
CC_OP_CLR, /* Z set, all other flags clear. */
CC_OP_POPCNT, /* Z via CC_SRC, all other flags clear. */

2
target/i386/helper.h
View File

@ -95,7 +95,7 @@ DEF_HELPER_FLAGS_2(monitor, TCG_CALL_NO_WG, void, env, tl)
DEF_HELPER_FLAGS_2(mwait, TCG_CALL_NO_WG, noreturn, env, int)
DEF_HELPER_1(rdmsr, void, env)
DEF_HELPER_1(wrmsr, void, env)
DEF_HELPER_FLAGS_2(read_crN, TCG_CALL_NO_RWG, tl, env, int)
DEF_HELPER_FLAGS_1(read_cr8, TCG_CALL_NO_RWG, tl, env)
DEF_HELPER_FLAGS_3(write_crN, TCG_CALL_NO_RWG, void, env, int, tl)
#endif /* !CONFIG_USER_ONLY */

11
target/i386/sev.c
View File

@ -587,6 +587,7 @@ static SevCapability *sev_get_capabilities(Error **errp)
sev_common = SEV_COMMON(MACHINE(qdev_get_machine())->cgs);
if (!sev_common) {
error_setg(errp, "SEV is not configured");
return NULL;
}
sev_device = object_property_get_str(OBJECT(sev_common), "sev-device",
@ -1529,11 +1530,12 @@ int
sev_encrypt_flash(hwaddr gpa, uint8_t *ptr, uint64_t len, Error **errp)
{
SevCommonState *sev_common = SEV_COMMON(MACHINE(qdev_get_machine())->cgs);
SevCommonStateClass *klass = SEV_COMMON_GET_CLASS(sev_common);
SevCommonStateClass *klass;
if (!sev_common) {
return 0;
}
klass = SEV_COMMON_GET_CLASS(sev_common);
/* if SEV is in update state then encrypt the data else do nothing */
if (sev_check_state(sev_common, SEV_STATE_LAUNCH_UPDATE)) {
@ -1710,7 +1712,9 @@ void sev_es_set_reset_vector(CPUState *cpu)
{
X86CPU *x86;
CPUX86State *env;
SevCommonState *sev_common = SEV_COMMON(MACHINE(qdev_get_machine())->cgs);
ConfidentialGuestSupport *cgs = MACHINE(qdev_get_machine())->cgs;
SevCommonState *sev_common = SEV_COMMON(
object_dynamic_cast(OBJECT(cgs), TYPE_SEV_COMMON));
/* Only update if we have valid reset information */
if (!sev_common || !sev_common->reset_data_valid) {
@ -2165,6 +2169,7 @@ sev_snp_guest_set_id_block(Object *obj, const char *value, Error **errp)
struct kvm_sev_snp_launch_finish *finish = &sev_snp_guest->kvm_finish_conf;
gsize len;
finish->id_block_en = 0;
g_free(sev_snp_guest->id_block);
g_free((guchar *)finish->id_block_uaddr);
@ -2184,7 +2189,7 @@ sev_snp_guest_set_id_block(Object *obj, const char *value, Error **errp)
return;
}
finish->id_block_en = (len) ? 1 : 0;
finish->id_block_en = 1;
}
static char *

389
target/i386/tcg/decode-new.c.inc
View File

@ -151,6 +151,8 @@
X86_OP_GROUP3(op, op0, s0, 2op, s0, op1, s1, ## __VA_ARGS__)
#define X86_OP_GROUPw(op, op0, s0, ...) \
X86_OP_GROUP3(op, op0, s0, None, None, None, None, ## __VA_ARGS__)
#define X86_OP_GROUPwr(op, op0, s0, op1, s1, ...) \
X86_OP_GROUP3(op, op0, s0, op1, s1, None, None, ## __VA_ARGS__)
#define X86_OP_GROUP0(op, ...) \
X86_OP_GROUP3(op, None, None, None, None, None, None, ## __VA_ARGS__)
@ -180,20 +182,20 @@
#define X86_OP_ENTRYrr(op, op0, s0, op1, s1, ...) \
X86_OP_ENTRY3(op, None, None, op0, s0, op1, s1, ## __VA_ARGS__)
#define X86_OP_ENTRYwr(op, op0, s0, op1, s1, ...) \
X86_OP_ENTRY3(op, op0, s0, None, None, op1, s1, ## __VA_ARGS__)
X86_OP_ENTRY3(op, op0, s0, op1, s1, None, None, ## __VA_ARGS__)
#define X86_OP_ENTRY2(op, op0, s0, op1, s1, ...) \
X86_OP_ENTRY3(op, op0, s0, 2op, s0, op1, s1, ## __VA_ARGS__)
#define X86_OP_ENTRYw(op, op0, s0, ...) \
X86_OP_ENTRY3(op, op0, s0, None, None, None, None, ## __VA_ARGS__)
#define X86_OP_ENTRYr(op, op0, s0, ...) \
X86_OP_ENTRY3(op, None, None, None, None, op0, s0, ## __VA_ARGS__)
X86_OP_ENTRY3(op, None, None, op0, s0, None, None, ## __VA_ARGS__)
#define X86_OP_ENTRY1(op, op0, s0, ...) \
X86_OP_ENTRY3(op, op0, s0, 2op, s0, None, None, ## __VA_ARGS__)
#define X86_OP_ENTRY0(op, ...) \
X86_OP_ENTRY3(op, None, None, None, None, None, None, ## __VA_ARGS__)
#define cpuid(feat) .cpuid = X86_FEAT_##feat,
#define noseg .special = X86_SPECIAL_NoSeg,
#define nolea .special = X86_SPECIAL_NoLoadEA,
#define xchg .special = X86_SPECIAL_Locked,
#define lock .special = X86_SPECIAL_HasLock,
#define mmx .special = X86_SPECIAL_MMX,
@ -221,7 +223,9 @@
#define vex13 .vex_class = 13,
#define chk(a) .check = X86_CHECK_##a,
#define svm(a) .intercept = SVM_EXIT_##a,
#define chk2(a, b) .check = X86_CHECK_##a | X86_CHECK_##b,
#define chk3(a, b, c) .check = X86_CHECK_##a | X86_CHECK_##b | X86_CHECK_##c,
#define svm(a) .intercept = SVM_EXIT_##a, .has_intercept = true,
#define avx2_256 .vex_special = X86_VEX_AVX2_256,
@ -267,20 +271,41 @@ static inline const X86OpEntry *decode_by_prefix(DisasContext *s, const X86OpEnt
static void decode_group15(DisasContext *s, CPUX86State *env, X86OpEntry *entry, uint8_t *b)
{
/* only includes ldmxcsr and stmxcsr, because they have AVX variants. */
static const X86OpEntry group15_reg[8] = {
[0] = X86_OP_ENTRYw(RDxxBASE, R,y, cpuid(FSGSBASE) chk(o64) p_f3),
[1] = X86_OP_ENTRYw(RDxxBASE, R,y, cpuid(FSGSBASE) chk(o64) p_f3),
[2] = X86_OP_ENTRYr(WRxxBASE, R,y, cpuid(FSGSBASE) chk(o64) p_f3 zextT0),
[3] = X86_OP_ENTRYr(WRxxBASE, R,y, cpuid(FSGSBASE) chk(o64) p_f3 zextT0),
[5] = X86_OP_ENTRY0(LFENCE, cpuid(SSE2) p_00),
[6] = X86_OP_ENTRY0(MFENCE, cpuid(SSE2) p_00),
[7] = X86_OP_ENTRY0(SFENCE, cpuid(SSE2) p_00),
};
static const X86OpEntry group15_mem[8] = {
[2] = X86_OP_ENTRYr(LDMXCSR, E,d, vex5 chk(VEX128)),
[3] = X86_OP_ENTRYw(STMXCSR, E,d, vex5 chk(VEX128)),
[0] = X86_OP_ENTRYw(FXSAVE, M,y, cpuid(FXSR) p_00),
[1] = X86_OP_ENTRYr(FXRSTOR, M,y, cpuid(FXSR) p_00),
[2] = X86_OP_ENTRYr(LDMXCSR, E,d, vex5 chk(VEX128) p_00),
[3] = X86_OP_ENTRYw(STMXCSR, E,d, vex5 chk(VEX128) p_00),
[4] = X86_OP_ENTRYw(XSAVE, M,y, cpuid(XSAVE) p_00),
[5] = X86_OP_ENTRYr(XRSTOR, M,y, cpuid(XSAVE) p_00),
[6] = X86_OP_ENTRYw(XSAVEOPT, M,b, cpuid(XSAVEOPT) p_00),
[7] = X86_OP_ENTRYw(NOP, M,b, cpuid(CLFLUSH) p_00),
};
static const X86OpEntry group15_mem_66[8] = {
[6] = X86_OP_ENTRYw(NOP, M,b, cpuid(CLWB)),
[7] = X86_OP_ENTRYw(NOP, M,b, cpuid(CLFLUSHOPT)),
};
uint8_t modrm = get_modrm(s, env);
int op = (modrm >> 3) & 7;
if ((modrm >> 6) == 3) {
*entry = group15_reg[(modrm >> 3) & 7];
*entry = group15_reg[op];
} else if (s->prefix & PREFIX_DATA) {
*entry = group15_mem_66[op];
} else {
*entry = group15_mem[(modrm >> 3) & 7];
*entry = group15_mem[op];
}
}
@ -425,6 +450,50 @@ static void decode_0F7F(DisasContext *s, CPUX86State *env, X86OpEntry *entry, ui
*entry = *decode_by_prefix(s, opcodes_0F7F);
}
static void decode_0FB8(DisasContext *s, CPUX86State *env, X86OpEntry *entry, uint8_t *b)
{
static const X86OpEntry popcnt =
X86_OP_ENTRYwr(POPCNT, G,v, E,v, cpuid(POPCNT) zextT0);
if (s->prefix & PREFIX_REPZ) {
*entry = popcnt;
} else {
memset(entry, 0, sizeof(*entry));
}
}
static void decode_0FBC(DisasContext *s, CPUX86State *env, X86OpEntry *entry, uint8_t *b)
{
/* For BSF, pass 2op as the third operand so that we can use zextT0 */
static const X86OpEntry opcodes_0FBC[4] = {
X86_OP_ENTRY3(BSF, G,v, E,v, 2op,v, zextT0),
X86_OP_ENTRY3(BSF, G,v, E,v, 2op,v, zextT0), /* 0x66 */
X86_OP_ENTRYwr(TZCNT, G,v, E,v, zextT0), /* 0xf3 */
X86_OP_ENTRY3(BSF, G,v, E,v, 2op,v, zextT0), /* 0xf2 */
};
if (!(s->cpuid_ext3_features & CPUID_EXT3_ABM)) {
*entry = opcodes_0FBC[0];
} else {
*entry = *decode_by_prefix(s, opcodes_0FBC);
}
}
static void decode_0FBD(DisasContext *s, CPUX86State *env, X86OpEntry *entry, uint8_t *b)
{
/* For BSR, pass 2op as the third operand so that we can use zextT0 */
static const X86OpEntry opcodes_0FBD[4] = {
X86_OP_ENTRY3(BSR, G,v, E,v, 2op,v, zextT0),
X86_OP_ENTRY3(BSR, G,v, E,v, 2op,v, zextT0), /* 0x66 */
X86_OP_ENTRYwr(LZCNT, G,v, E,v, zextT0), /* 0xf3 */
X86_OP_ENTRY3(BSR, G,v, E,v, 2op,v, zextT0), /* 0xf2 */
};
if (!(s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_BMI1)) {
*entry = opcodes_0FBD[0];
} else {
*entry = *decode_by_prefix(s, opcodes_0FBD);
}
}
static void decode_0FD6(DisasContext *s, CPUX86State *env, X86OpEntry *entry, uint8_t *b)
{
static const X86OpEntry movq[4] = {
@ -612,15 +681,15 @@ static const X86OpEntry opcodes_0F38_00toEF[240] = {
/* five rows for no prefix, 66, F3, F2, 66+F2 */
static const X86OpEntry opcodes_0F38_F0toFF[16][5] = {
[0] = {
X86_OP_ENTRY3(MOVBE, G,y, M,y, None,None, cpuid(MOVBE)),
X86_OP_ENTRY3(MOVBE, G,w, M,w, None,None, cpuid(MOVBE)),
X86_OP_ENTRYwr(MOVBE, G,y, M,y, cpuid(MOVBE)),
X86_OP_ENTRYwr(MOVBE, G,w, M,w, cpuid(MOVBE)),
{},
X86_OP_ENTRY2(CRC32, G,d, E,b, cpuid(SSE42)),
X86_OP_ENTRY2(CRC32, G,d, E,b, cpuid(SSE42)),
},
[1] = {
X86_OP_ENTRY3(MOVBE, M,y, G,y, None,None, cpuid(MOVBE)),
X86_OP_ENTRY3(MOVBE, M,w, G,w, None,None, cpuid(MOVBE)),
X86_OP_ENTRYwr(MOVBE, M,y, G,y, cpuid(MOVBE)),
X86_OP_ENTRYwr(MOVBE, M,w, G,w, cpuid(MOVBE)),
{},
X86_OP_ENTRY2(CRC32, G,d, E,y, cpuid(SSE42)),
X86_OP_ENTRY2(CRC32, G,d, E,w, cpuid(SSE42)),
@ -633,7 +702,7 @@ static const X86OpEntry opcodes_0F38_F0toFF[16][5] = {
{},
},
[3] = {
X86_OP_GROUP3(group17, B,y, E,y, None,None, vex13 cpuid(BMI1)),
X86_OP_GROUP3(group17, B,y, None,None, E,y, vex13 cpuid(BMI1)),
{},
{},
{},
@ -985,14 +1054,30 @@ static void decode_0FE6(DisasContext *s, CPUX86State *env, X86OpEntry *entry, ui
*entry = *decode_by_prefix(s, opcodes_0FE6);
}
static const X86OpEntry opcodes_0F[256] = {
[0x0E] = X86_OP_ENTRY0(EMMS, cpuid(3DNOW)), /* femms */
/*
* These ignore the mod bits (assume (modrm&0xc0)==0xc0), so group the
* pre-decode tweak here for all MOVs from/to CR and DR.
*
* AMD documentation (24594.pdf) and testing of Intel 386 and 486
* processors all show that the mod bits are assumed to be 1's,
* regardless of actual values.
*/
static void decode_MOV_CR_DR(DisasContext *s, CPUX86State *env, X86OpEntry *entry, uint8_t *b)
{
/*
* 3DNow!'s opcode byte comes *after* modrm and displacements, making it
* more like an Ib operand. Dispatch to the right helper in a single gen_*
* function.
*/
[0x0F] = X86_OP_ENTRY3(3dnow, P,q, Q,q, I,b, cpuid(3DNOW)),
get_modrm(s, env);
s->modrm |= 0xC0;
entry->gen = gen_MOV;
}
static const X86OpEntry opcodes_0F[256] = {
[0x02] = X86_OP_ENTRYwr(LAR, G,v, E,w, chk(prot)),
[0x03] = X86_OP_ENTRYwr(LSL, G,v, E,w, chk(prot)),
[0x05] = X86_OP_ENTRY0(SYSCALL, chk(o64_intel)),
[0x06] = X86_OP_ENTRY0(CLTS, chk(cpl0) svm(WRITE_CR0)),
[0x07] = X86_OP_ENTRY0(SYSRET, chk3(o64_intel, prot, cpl0)),
[0x10] = X86_OP_GROUP0(0F10),
[0x11] = X86_OP_GROUP0(0F11),
@ -1004,6 +1089,22 @@ static const X86OpEntry opcodes_0F[256] = {
/* Incorrectly listed as Mq,Vq in the manual */
[0x17] = X86_OP_ENTRY3(VMOVHPx_st, M,q, None,None, V,dq, vex5 p_00_66),
/*
* Incorrectly listed as using "d" operand type in the manual. In reality
* there's no 16-bit version (like y) and it does not use REX.W (like d64).
*/
[0x20] = X86_OP_GROUPwr(MOV_CR_DR, R,y_d64, C,y_d64, chk(cpl0) svm(READ_CR0)),
[0x21] = X86_OP_GROUPwr(MOV_CR_DR, R,y_d64, D,y_d64, chk(cpl0) svm(READ_DR0)),
[0x22] = X86_OP_GROUPwr(MOV_CR_DR, C,y_d64, R,y_d64, zextT0 chk(cpl0) svm(WRITE_CR0)),
[0x23] = X86_OP_GROUPwr(MOV_CR_DR, D,y_d64, R,y_d64, zextT0 chk(cpl0) svm(WRITE_DR0)),
[0x30] = X86_OP_ENTRY0(WRMSR, chk(cpl0)),
[0x31] = X86_OP_ENTRY0(RDTSC),
[0x32] = X86_OP_ENTRY0(RDMSR, chk(cpl0)),
[0x33] = X86_OP_ENTRY0(RDPMC),
[0x34] = X86_OP_ENTRY0(SYSENTER, chk2(i64_amd, prot_or_vm86)),
[0x35] = X86_OP_ENTRY0(SYSEXIT, chk3(i64_amd, prot, cpl0)),
[0x40] = X86_OP_ENTRY2(CMOVcc, G,v, E,v, cpuid(CMOV)),
[0x41] = X86_OP_ENTRY2(CMOVcc, G,v, E,v, cpuid(CMOV)),
[0x42] = X86_OP_ENTRY2(CMOVcc, G,v, E,v, cpuid(CMOV)),
@ -1060,9 +1161,64 @@ static const X86OpEntry opcodes_0F[256] = {
[0xa0] = X86_OP_ENTRYr(PUSH, FS, w),
[0xa1] = X86_OP_ENTRYw(POP, FS, w),
[0xa2] = X86_OP_ENTRY0(CPUID),
[0xa4] = X86_OP_ENTRY4(SHLD, E,v, 2op,v, G,v),
[0xa5] = X86_OP_ENTRY3(SHLD, E,v, 2op,v, G,v),
[0xb0] = X86_OP_ENTRY2(CMPXCHG,E,b, G,b, lock),
[0xb1] = X86_OP_ENTRY2(CMPXCHG,E,v, G,v, lock),
[0xb2] = X86_OP_ENTRY3(LSS, G,v, EM,p, None, None),
[0xb4] = X86_OP_ENTRY3(LFS, G,v, EM,p, None, None),
[0xb5] = X86_OP_ENTRY3(LGS, G,v, EM,p, None, None),
[0xb6] = X86_OP_ENTRY3(MOV, G,v, E,b, None, None, zextT0), /* MOVZX */
[0xb7] = X86_OP_ENTRY3(MOV, G,v, E,w, None, None, zextT0), /* MOVZX */
[0xc0] = X86_OP_ENTRY2(XADD, E,b, G,b, lock),
[0xc1] = X86_OP_ENTRY2(XADD, E,v, G,v, lock),
[0xc2] = X86_OP_ENTRY4(VCMP, V,x, H,x, W,x, vex2_rep3 p_00_66_f3_f2),
[0xc3] = X86_OP_ENTRY3(MOV, EM,y,G,y, None,None, cpuid(SSE2)), /* MOVNTI */
[0xc4] = X86_OP_ENTRY4(PINSRW, V,dq,H,dq,E,w, vex5 mmx p_00_66),
[0xc5] = X86_OP_ENTRY3(PEXTRW, G,d, U,dq,I,b, vex5 mmx p_00_66),
[0xc6] = X86_OP_ENTRY4(VSHUF, V,x, H,x, W,x, vex4 p_00_66),
[0xd0] = X86_OP_ENTRY3(VADDSUB, V,x, H,x, W,x, vex2 cpuid(SSE3) p_66_f2),
[0xd1] = X86_OP_ENTRY3(PSRLW_r, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd2] = X86_OP_ENTRY3(PSRLD_r, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd3] = X86_OP_ENTRY3(PSRLQ_r, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd4] = X86_OP_ENTRY3(PADDQ, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd5] = X86_OP_ENTRY3(PMULLW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd6] = X86_OP_GROUP0(0FD6),
[0xd7] = X86_OP_ENTRY3(PMOVMSKB, G,d, None,None, U,x, vex7 mmx avx2_256 p_00_66),
[0xe0] = X86_OP_ENTRY3(PAVGB, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe1] = X86_OP_ENTRY3(PSRAW_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xe2] = X86_OP_ENTRY3(PSRAD_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xe3] = X86_OP_ENTRY3(PAVGW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe4] = X86_OP_ENTRY3(PMULHUW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe5] = X86_OP_ENTRY3(PMULHW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe6] = X86_OP_GROUP0(0FE6),
[0xe7] = X86_OP_ENTRY3(MOVDQ, W,x, None,None, V,x, vex1 mmx p_00_66), /* MOVNTQ/MOVNTDQ */
[0xf0] = X86_OP_ENTRY3(MOVDQ, V,x, None,None, WM,x, vex4_unal cpuid(SSE3) p_f2), /* LDDQU */
[0xf1] = X86_OP_ENTRY3(PSLLW_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xf2] = X86_OP_ENTRY3(PSLLD_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xf3] = X86_OP_ENTRY3(PSLLQ_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xf4] = X86_OP_ENTRY3(PMULUDQ, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xf5] = X86_OP_ENTRY3(PMADDWD, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xf6] = X86_OP_ENTRY3(PSADBW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xf7] = X86_OP_ENTRY3(MASKMOV, None,None, V,dq, U,dq, vex4_unal avx2_256 mmx p_00_66),
[0x08] = X86_OP_ENTRY0(NOP, svm(INVD)),
[0x09] = X86_OP_ENTRY0(NOP, svm(WBINVD)),
[0x0b] = X86_OP_ENTRY0(UD), /* UD2 */
[0x0d] = X86_OP_ENTRY1(NOP, M,v), /* 3DNow! prefetch */
[0x0e] = X86_OP_ENTRY0(EMMS, cpuid(3DNOW)), /* femms */
/*
* 3DNow!'s opcode byte comes *after* modrm and displacements, making it
* more like an Ib operand. Dispatch to the right helper in a single gen_*
* function.
*/
[0x0f] = X86_OP_ENTRY3(3dnow, P,q, Q,q, I,b, cpuid(3DNOW)),
[0x18] = X86_OP_ENTRY1(NOP, nop,v), /* prefetch/reserved NOP */
[0x19] = X86_OP_ENTRY1(NOP, nop,v), /* reserved NOP */
@ -1137,6 +1293,9 @@ static const X86OpEntry opcodes_0F[256] = {
[0xa8] = X86_OP_ENTRYr(PUSH, GS, w),
[0xa9] = X86_OP_ENTRYw(POP, GS, w),
[0xaa] = X86_OP_ENTRY0(RSM, chk(smm) svm(RSM)),
[0xac] = X86_OP_ENTRY4(SHRD, E,v, 2op,v, G,v),
[0xad] = X86_OP_ENTRY3(SHRD, E,v, 2op,v, G,v),
[0xae] = X86_OP_GROUP0(group15),
/*
* It's slightly more efficient to put Ev operand in T0 and allow gen_IMUL3
@ -1144,23 +1303,14 @@ static const X86OpEntry opcodes_0F[256] = {
*/
[0xaf] = X86_OP_ENTRY3(IMUL3, G,v, E,v, 2op,v, sextT0),
[0xb2] = X86_OP_ENTRY3(LSS, G,v, EM,p, None, None),
[0xb4] = X86_OP_ENTRY3(LFS, G,v, EM,p, None, None),
[0xb5] = X86_OP_ENTRY3(LGS, G,v, EM,p, None, None),
[0xb6] = X86_OP_ENTRY3(MOV, G,v, E,b, None, None, zextT0), /* MOVZX */
[0xb7] = X86_OP_ENTRY3(MOV, G,v, E,w, None, None, zextT0), /* MOVZX */
[0xb8] = X86_OP_GROUP0(0FB8),
/* decoded as modrm, which is visible as a difference between page fault and #UD */
[0xb9] = X86_OP_ENTRYr(UD, nop,v), /* UD1 */
[0xbc] = X86_OP_GROUP0(0FBC),
[0xbd] = X86_OP_GROUP0(0FBD),
[0xbe] = X86_OP_ENTRY3(MOV, G,v, E,b, None, None, sextT0), /* MOVSX */
[0xbf] = X86_OP_ENTRY3(MOV, G,v, E,w, None, None, sextT0), /* MOVSX */
[0xc2] = X86_OP_ENTRY4(VCMP, V,x, H,x, W,x, vex2_rep3 p_00_66_f3_f2),
[0xc3] = X86_OP_ENTRY3(MOV, EM,y,G,y, None,None, cpuid(SSE2)), /* MOVNTI */
[0xc4] = X86_OP_ENTRY4(PINSRW, V,dq,H,dq,E,w, vex5 mmx p_00_66),
[0xc5] = X86_OP_ENTRY3(PEXTRW, G,d, U,dq,I,b, vex5 mmx p_00_66),
[0xc6] = X86_OP_ENTRY4(VSHUF, V,x, H,x, W,x, vex4 p_00_66),
[0xc8] = X86_OP_ENTRY1(BSWAP, LoBits,y),
[0xc9] = X86_OP_ENTRY1(BSWAP, LoBits,y),
[0xca] = X86_OP_ENTRY1(BSWAP, LoBits,y),
@ -1170,33 +1320,6 @@ static const X86OpEntry opcodes_0F[256] = {
[0xce] = X86_OP_ENTRY1(BSWAP, LoBits,y),
[0xcf] = X86_OP_ENTRY1(BSWAP, LoBits,y),
[0xd0] = X86_OP_ENTRY3(VADDSUB, V,x, H,x, W,x, vex2 cpuid(SSE3) p_66_f2),
[0xd1] = X86_OP_ENTRY3(PSRLW_r, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd2] = X86_OP_ENTRY3(PSRLD_r, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd3] = X86_OP_ENTRY3(PSRLQ_r, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd4] = X86_OP_ENTRY3(PADDQ, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd5] = X86_OP_ENTRY3(PMULLW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd6] = X86_OP_GROUP0(0FD6),
[0xd7] = X86_OP_ENTRY3(PMOVMSKB, G,d, None,None, U,x, vex7 mmx avx2_256 p_00_66),
[0xe0] = X86_OP_ENTRY3(PAVGB, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe1] = X86_OP_ENTRY3(PSRAW_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xe2] = X86_OP_ENTRY3(PSRAD_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xe3] = X86_OP_ENTRY3(PAVGW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe4] = X86_OP_ENTRY3(PMULHUW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe5] = X86_OP_ENTRY3(PMULHW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xe6] = X86_OP_GROUP0(0FE6),
[0xe7] = X86_OP_ENTRY3(MOVDQ, W,x, None,None, V,x, vex1 mmx p_00_66), /* MOVNTQ/MOVNTDQ */
[0xf0] = X86_OP_ENTRY3(MOVDQ, V,x, None,None, WM,x, vex4_unal cpuid(SSE3) p_f2), /* LDDQU */
[0xf1] = X86_OP_ENTRY3(PSLLW_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xf2] = X86_OP_ENTRY3(PSLLD_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xf3] = X86_OP_ENTRY3(PSLLQ_r, V,x, H,x, W,x, vex7 mmx avx2_256 p_00_66),
[0xf4] = X86_OP_ENTRY3(PMULUDQ, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xf5] = X86_OP_ENTRY3(PMADDWD, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xf6] = X86_OP_ENTRY3(PSADBW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xf7] = X86_OP_ENTRY3(MASKMOV, None,None, V,dq, U,dq, vex4_unal avx2_256 mmx p_00_66),
/* Incorrectly missing from 2-17 */
[0xd8] = X86_OP_ENTRY3(PSUBUSB, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
[0xd9] = X86_OP_ENTRY3(PSUBUSW, V,x, H,x, W,x, vex4 mmx avx2_256 p_00_66),
@ -1335,9 +1458,9 @@ static void decode_group4_5(DisasContext *s, CPUX86State *env, X86OpEntry *entry
/* 0xff */
[0x08] = X86_OP_ENTRY1(INC, E,v, lock),
[0x09] = X86_OP_ENTRY1(DEC, E,v, lock),
[0x0a] = X86_OP_ENTRY3(CALL_m, None, None, E,f64, None, None, zextT0),
[0x0a] = X86_OP_ENTRYr(CALL_m, E,f64, zextT0),
[0x0b] = X86_OP_ENTRYr(CALLF_m, M,p),
[0x0c] = X86_OP_ENTRY3(JMP_m, None, None, E,f64, None, None, zextT0),
[0x0c] = X86_OP_ENTRYr(JMP_m, E,f64, zextT0),
[0x0d] = X86_OP_ENTRYr(JMPF_m, M,p),
[0x0e] = X86_OP_ENTRYr(PUSH, E,f64),
};
@ -1586,18 +1709,18 @@ static const X86OpEntry opcodes_root[256] = {
[0x7E] = X86_OP_ENTRYr(Jcc, J,b),
[0x7F] = X86_OP_ENTRYr(Jcc, J,b),
[0x88] = X86_OP_ENTRY3(MOV, E,b, G,b, None, None),
[0x89] = X86_OP_ENTRY3(MOV, E,v, G,v, None, None),
[0x8A] = X86_OP_ENTRY3(MOV, G,b, E,b, None, None),
[0x8B] = X86_OP_ENTRY3(MOV, G,v, E,v, None, None),
/* Missing in Table A-2: memory destination is always 16-bit. */
[0x8C] = X86_OP_ENTRY3(MOV, E,v, S,w, None, None, op0_Mw),
[0x8D] = X86_OP_ENTRY3(LEA, G,v, M,v, None, None, noseg),
[0x8E] = X86_OP_ENTRY3(MOV, S,w, E,w, None, None),
[0x88] = X86_OP_ENTRYwr(MOV, E,b, G,b),
[0x89] = X86_OP_ENTRYwr(MOV, E,v, G,v),
[0x8A] = X86_OP_ENTRYwr(MOV, G,b, E,b),
[0x8B] = X86_OP_ENTRYwr(MOV, G,v, E,v),
/* Missing in Table A-2: memory destination is always 16-bit. */
[0x8C] = X86_OP_ENTRYwr(MOV, E,v, S,w, op0_Mw),
[0x8D] = X86_OP_ENTRYwr(LEA, G,v, M,v, nolea),
[0x8E] = X86_OP_ENTRYwr(MOV, S,w, E,w),
[0x8F] = X86_OP_GROUPw(group1A, E,v),
[0x98] = X86_OP_ENTRY1(CBW, 0,v), /* rAX */
[0x99] = X86_OP_ENTRY3(CWD, 2,v, 0,v, None, None), /* rDX, rAX */
[0x99] = X86_OP_ENTRYwr(CWD, 2,v, 0,v), /* rDX, rAX */
[0x9A] = X86_OP_ENTRYrr(CALLF, I_unsigned,p, I_unsigned,w, chk(i64)),
[0x9B] = X86_OP_ENTRY0(WAIT),
[0x9C] = X86_OP_ENTRY0(PUSHF, chk(vm86_iopl) svm(PUSHF)),
@ -1607,22 +1730,22 @@ static const X86OpEntry opcodes_root[256] = {
[0xA8] = X86_OP_ENTRYrr(AND, 0,b, I,b), /* AL, Ib */
[0xA9] = X86_OP_ENTRYrr(AND, 0,v, I,z), /* rAX, Iz */
[0xAA] = X86_OP_ENTRY3(STOS, Y,b, 0,b, None, None),
[0xAB] = X86_OP_ENTRY3(STOS, Y,v, 0,v, None, None),
[0xAA] = X86_OP_ENTRYwr(STOS, Y,b, 0,b),
[0xAB] = X86_OP_ENTRYwr(STOS, Y,v, 0,v),
/* Manual writeback because REP LODS (!) has to write EAX/RAX after every LODS. */
[0xAC] = X86_OP_ENTRYr(LODS, X,b),
[0xAD] = X86_OP_ENTRYr(LODS, X,v),
[0xAE] = X86_OP_ENTRYrr(SCAS, 0,b, Y,b),
[0xAF] = X86_OP_ENTRYrr(SCAS, 0,v, Y,v),
[0xB8] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xB9] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xBA] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xBB] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xBC] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xBD] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xBE] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xBF] = X86_OP_ENTRY3(MOV, LoBits,v, I,v, None, None),
[0xB8] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xB9] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xBA] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xBB] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xBC] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xBD] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xBE] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xBF] = X86_OP_ENTRYwr(MOV, LoBits,v, I,v),
[0xC8] = X86_OP_ENTRYrr(ENTER, I,w, I,b),
[0xC9] = X86_OP_ENTRY1(LEAVE, A,d64),
@ -1725,6 +1848,10 @@ static bool decode_op_size(DisasContext *s, X86OpEntry *e, X86OpSize size, MemOp
*ot = s->dflag == MO_16 ? MO_32 : s->dflag;
return true;
case X86_SIZE_y_d64: /* Full (not 16-bit) register access */
*ot = CODE64(s) ? MO_64 : MO_32;
return true;
case X86_SIZE_z: /* 16-bit for 16-bit operand size, else 32-bit */
*ot = s->dflag == MO_16 ? MO_16 : MO_32;
return true;
@ -1802,11 +1929,34 @@ static bool decode_op(DisasContext *s, CPUX86State *env, X86DecodedInsn *decode,
case X86_TYPE_C: /* REG in the modrm byte selects a control register */
op->unit = X86_OP_CR;
goto get_reg;
op->n = ((get_modrm(s, env) >> 3) & 7) | REX_R(s);
if (op->n == 0 && (s->prefix & PREFIX_LOCK) &&
(s->cpuid_ext3_features & CPUID_EXT3_CR8LEG)) {
op->n = 8;
s->prefix &= ~PREFIX_LOCK;
}
if (op->n != 0 && op->n != 2 && op->n != 3 && op->n != 4 && op->n != 8) {
return false;
}
if (decode->e.intercept) {
decode->e.intercept += op->n;
}
break;
case X86_TYPE_D: /* REG in the modrm byte selects a debug register */
op->unit = X86_OP_DR;
goto get_reg;
op->n = ((get_modrm(s, env) >> 3) & 7) | REX_R(s);
if (op->n >= 8) {
/*
* illegal opcode. The DR4 and DR5 case is checked in the generated
* code instead, to save on hflags bits.
*/
return false;
}
if (decode->e.intercept) {
decode->e.intercept += op->n;
}
break;
case X86_TYPE_G: /* REG in the modrm byte selects a GPR */
op->unit = X86_OP_INT;
@ -2047,6 +2197,10 @@ static bool has_cpuid_feature(DisasContext *s, X86CPUIDFeature cpuid)
return true;
case X86_FEAT_CMOV:
return (s->cpuid_features & CPUID_CMOV);
case X86_FEAT_CLFLUSH:
return (s->cpuid_features & CPUID_CLFLUSH);
case X86_FEAT_FXSR:
return (s->cpuid_features & CPUID_FXSR);
case X86_FEAT_F16C:
return (s->cpuid_ext_features & CPUID_EXT_F16C);
case X86_FEAT_FMA:
@ -2055,6 +2209,8 @@ static bool has_cpuid_feature(DisasContext *s, X86CPUIDFeature cpuid)
return (s->cpuid_ext_features & CPUID_EXT_MOVBE);
case X86_FEAT_PCLMULQDQ:
return (s->cpuid_ext_features & CPUID_EXT_PCLMULQDQ);
case X86_FEAT_POPCNT:
return (s->cpuid_ext_features & CPUID_EXT_POPCNT);
case X86_FEAT_SSE:
return (s->cpuid_features & CPUID_SSE);
case X86_FEAT_SSE2:
@ -2080,6 +2236,8 @@ static bool has_cpuid_feature(DisasContext *s, X86CPUIDFeature cpuid)
case X86_FEAT_AVX:
return (s->cpuid_ext_features & CPUID_EXT_AVX);
case X86_FEAT_XSAVE:
return (s->cpuid_ext_features & CPUID_EXT_XSAVE);
case X86_FEAT_3DNOW:
return (s->cpuid_ext2_features & CPUID_EXT2_3DNOW);
@ -2094,11 +2252,20 @@ static bool has_cpuid_feature(DisasContext *s, X86CPUIDFeature cpuid)
return (s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_BMI2);
case X86_FEAT_AVX2:
return (s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_AVX2);
case X86_FEAT_CLFLUSHOPT:
return (s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_CLFLUSHOPT);
case X86_FEAT_CLWB:
return (s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_CLWB);
case X86_FEAT_FSGSBASE:
return (s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_FSGSBASE);
case X86_FEAT_SHA_NI:
return (s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_SHA_NI);
case X86_FEAT_CMPCCXADD:
return (s->cpuid_7_1_eax_features & CPUID_7_1_EAX_CMPCCXADD);
case X86_FEAT_XSAVEOPT:
return (s->cpuid_xsave_features & CPUID_XSAVE_XSAVEOPT);
}
g_assert_not_reached();
}
@ -2428,18 +2595,12 @@ static void disas_insn(DisasContext *s, CPUState *cpu)
if (b == 0x0f) {
b = x86_ldub_code(env, s);
switch (b) {
case 0x00 ... 0x03: /* mostly privileged instructions */
case 0x05 ... 0x09:
case 0x00 ... 0x01: /* mostly privileged instructions */
case 0x1a ... 0x1b: /* MPX */
case 0x20 ... 0x23: /* mov from/to CR and DR */
case 0x30 ... 0x35: /* more privileged instructions */
case 0xa2 ... 0xa5: /* CPUID, BT, SHLD */
case 0xaa ... 0xae: /* RSM, SHRD, grp15 */
case 0xb0 ... 0xb1: /* cmpxchg */
case 0xa3: /* bt */
case 0xab: /* bts */
case 0xb3: /* btr */
case 0xb8: /* integer ops */
case 0xba ... 0xbd: /* integer ops */
case 0xc0 ... 0xc1: /* xadd */
case 0xba ... 0xbb: /* grp8, btc */
case 0xc7: /* grp9 */
disas_insn_old(s, cpu, b + 0x100);
return;
@ -2466,18 +2627,28 @@ static void disas_insn(DisasContext *s, CPUState *cpu)
/* Checks that result in #UD come first. */
if (decode.e.check) {
if (decode.e.check & X86_CHECK_i64) {
if (CODE64(s)) {
if (CODE64(s)) {
if (decode.e.check & X86_CHECK_i64) {
goto illegal_op;
}
if ((decode.e.check & X86_CHECK_i64_amd) && env->cpuid_vendor1 != CPUID_VENDOR_INTEL_1) {
goto illegal_op;
}
} else {
if (decode.e.check & X86_CHECK_o64) {
goto illegal_op;
}
if ((decode.e.check & X86_CHECK_o64_intel) && env->cpuid_vendor1 == CPUID_VENDOR_INTEL_1) {
goto illegal_op;
}
}
if (decode.e.check & X86_CHECK_o64) {
if (!CODE64(s)) {
if (decode.e.check & X86_CHECK_prot_or_vm86) {
if (!PE(s)) {
goto illegal_op;
}
}
if (decode.e.check & X86_CHECK_prot) {
if (!PE(s) || VM86(s)) {
if (decode.e.check & X86_CHECK_no_vm86) {
if (VM86(s)) {
goto illegal_op;
}
}
@ -2524,11 +2695,6 @@ static void disas_insn(DisasContext *s, CPUState *cpu)
assert(decode.op[1].unit == X86_OP_INT);
break;
case X86_SPECIAL_NoSeg:
decode.mem.def_seg = -1;
s->override = -1;
break;
case X86_SPECIAL_Op0_Mw:
assert(decode.op[0].unit == X86_OP_INT);
if (decode.op[0].has_ea) {
@ -2556,19 +2722,21 @@ static void disas_insn(DisasContext *s, CPUState *cpu)
* exceptions if there is no memory operand). Exceptions are
* vm86 checks (INTn, IRET, PUSHF/POPF), RSM and XSETBV (!).
*
* RSM and XSETBV will be handled in the gen_* functions
* instead of using chk().
* XSETBV will check for CPL0 in the gen_* function instead of using chk().
*/
if (decode.e.check & X86_CHECK_cpl0) {
if (CPL(s) != 0) {
goto gp_fault;
}
}
if (decode.e.intercept && unlikely(GUEST(s))) {
if (decode.e.has_intercept && unlikely(GUEST(s))) {
gen_helper_svm_check_intercept(tcg_env,
tcg_constant_i32(decode.e.intercept));
}
if (decode.e.check) {
if ((decode.e.check & X86_CHECK_smm) && !(s->flags & HF_SMM_MASK)) {
goto illegal_op;
}
if ((decode.e.check & X86_CHECK_vm86_iopl) && VM86(s)) {
if (IOPL(s) < 3) {
goto gp_fault;
@ -2585,12 +2753,13 @@ static void disas_insn(DisasContext *s, CPUState *cpu)
gen_helper_enter_mmx(tcg_env);
}
if (decode.op[0].has_ea || decode.op[1].has_ea || decode.op[2].has_ea) {
if (decode.e.special != X86_SPECIAL_NoLoadEA &&
(decode.op[0].has_ea || decode.op[1].has_ea || decode.op[2].has_ea)) {
gen_load_ea(s, &decode.mem, decode.e.vex_class == 12);
}
if (s->prefix & PREFIX_LOCK) {
gen_load(s, &decode, 2, s->T1);
decode.e.gen(s, env, &decode);
decode.e.gen(s, &decode);
} else {
if (decode.op[0].unit == X86_OP_MMX) {
compute_mmx_offset(&decode.op[0]);
@ -2599,12 +2768,12 @@ static void disas_insn(DisasContext *s, CPUState *cpu)
}
gen_load(s, &decode, 1, s->T0);
gen_load(s, &decode, 2, s->T1);
decode.e.gen(s, env, &decode);
decode.e.gen(s, &decode);
gen_writeback(s, &decode, 0, s->T0);
}
/*
* Write back flags after last memory access. Some newer ALU instructions, as
* Write back flags after last memory access. Some older ALU instructions, as
* well as SSE instructions, write flags in the gen_* function, but that can
* cause incorrect tracking of CC_OP for instructions that write to both memory
* and flags.

32
target/i386/tcg/decode-new.h
View File

@ -90,6 +90,7 @@ typedef enum X86OpSize {
X86_SIZE_w, /* 16-bit */
X86_SIZE_x, /* 128/256-bit, based on operand size */
X86_SIZE_y, /* 32/64-bit, based on operand size */
X86_SIZE_y_d64, /* 32/64-bit, based on 64-bit mode */
X86_SIZE_z, /* 16-bit for 16-bit operand size, else 32-bit */
X86_SIZE_z_f64, /* 32-bit for 32-bit operand size or 64-bit mode, else 16-bit */
@ -108,12 +109,18 @@ typedef enum X86CPUIDFeature {
X86_FEAT_AVX2,
X86_FEAT_BMI1,
X86_FEAT_BMI2,
X86_FEAT_CLFLUSH,
X86_FEAT_CLFLUSHOPT,
X86_FEAT_CLWB,
X86_FEAT_CMOV,
X86_FEAT_CMPCCXADD,
X86_FEAT_F16C,
X86_FEAT_FMA,
X86_FEAT_FSGSBASE,
X86_FEAT_FXSR,
X86_FEAT_MOVBE,
X86_FEAT_PCLMULQDQ,
X86_FEAT_POPCNT,
X86_FEAT_SHA_NI,
X86_FEAT_SSE,
X86_FEAT_SSE2,
@ -122,6 +129,8 @@ typedef enum X86CPUIDFeature {
X86_FEAT_SSE41,
X86_FEAT_SSE42,
X86_FEAT_SSE4A,
X86_FEAT_XSAVE,
X86_FEAT_XSAVEOPT,
} X86CPUIDFeature;
/* Execution flags */
@ -142,8 +151,8 @@ typedef enum X86InsnCheck {
X86_CHECK_i64 = 1,
X86_CHECK_o64 = 2,
/* Fault outside protected mode */
X86_CHECK_prot = 4,
/* Fault in vm86 mode */
X86_CHECK_no_vm86 = 4,
/* Privileged instruction checks */
X86_CHECK_cpl0 = 8,
@ -159,6 +168,17 @@ typedef enum X86InsnCheck {
/* Fault if VEX.W=0 */
X86_CHECK_W1 = 256,
/* Fault outside protected mode, possibly including vm86 mode */
X86_CHECK_prot_or_vm86 = 512,
X86_CHECK_prot = X86_CHECK_prot_or_vm86 | X86_CHECK_no_vm86,
/* Fault outside SMM */
X86_CHECK_smm = 1024,
/* Vendor-specific checks for Intel/AMD differences */
X86_CHECK_i64_amd = 2048,
X86_CHECK_o64_intel = 4096,
} X86InsnCheck;
typedef enum X86InsnSpecial {
@ -170,8 +190,9 @@ typedef enum X86InsnSpecial {
/* Always locked if it has a memory operand (XCHG) */
X86_SPECIAL_Locked,
/* Do not apply segment base to effective address */
X86_SPECIAL_NoSeg,
/* Do not load effective address in s->A0 */
X86_SPECIAL_NoLoadEA,
/*
* Rd/Mb or Rd/Mw in the manual: register operand 0 is treated as 32 bits
* (and writeback zero-extends it to 64 bits if applicable). PREFIX_DATA
@ -245,7 +266,7 @@ typedef struct X86DecodedInsn X86DecodedInsn;
typedef void (*X86DecodeFunc)(DisasContext *s, CPUX86State *env, X86OpEntry *entry, uint8_t *b);
/* Code generation function. */
typedef void (*X86GenFunc)(DisasContext *s, CPUX86State *env, X86DecodedInsn *decode);
typedef void (*X86GenFunc)(DisasContext *s, X86DecodedInsn *decode);
struct X86OpEntry {
/* Based on the is_decode flags. */
@ -271,6 +292,7 @@ struct X86OpEntry {
unsigned valid_prefix:16;
unsigned check:16;
unsigned intercept:8;
bool has_intercept:1;
bool is_decode:1;
};

1166
target/i386/tcg/emit.c.inc
View File

File diff suppressed because it is too large Load Diff

16
target/i386/tcg/seg_helper.c
View File

@ -2265,11 +2265,11 @@ void helper_sysexit(CPUX86State *env, int dflag)
target_ulong helper_lsl(CPUX86State *env, target_ulong selector1)
{
unsigned int limit;
uint32_t e1, e2, eflags, selector;
uint32_t e1, e2, selector;
int rpl, dpl, cpl, type;
selector = selector1 & 0xffff;
eflags = cpu_cc_compute_all(env);
assert(CC_OP == CC_OP_EFLAGS);
if ((selector & 0xfffc) == 0) {
goto fail;
}
@ -2301,22 +2301,22 @@ target_ulong helper_lsl(CPUX86State *env, target_ulong selector1)
}
if (dpl < cpl || dpl < rpl) {
fail:
CC_SRC = eflags & ~CC_Z;
CC_SRC &= ~CC_Z;
return 0;
}
}
limit = get_seg_limit(e1, e2);
CC_SRC = eflags | CC_Z;
CC_SRC |= CC_Z;
return limit;
}
target_ulong helper_lar(CPUX86State *env, target_ulong selector1)
{
uint32_t e1, e2, eflags, selector;
uint32_t e1, e2, selector;
int rpl, dpl, cpl, type;
selector = selector1 & 0xffff;
eflags = cpu_cc_compute_all(env);
assert(CC_OP == CC_OP_EFLAGS);
if ((selector & 0xfffc) == 0) {
goto fail;
}
@ -2351,11 +2351,11 @@ target_ulong helper_lar(CPUX86State *env, target_ulong selector1)
}
if (dpl < cpl || dpl < rpl) {
fail:
CC_SRC = eflags & ~CC_Z;
CC_SRC &= ~CC_Z;
return 0;
}
}
CC_SRC = eflags | CC_Z;
CC_SRC |= CC_Z;
return e2 & 0x00f0ff00;
}

20
target/i386/tcg/sysemu/misc_helper.c
View File

@ -63,23 +63,13 @@ target_ulong helper_inl(CPUX86State *env, uint32_t port)
cpu_get_mem_attrs(env), NULL);
}
target_ulong helper_read_crN(CPUX86State *env, int reg)
target_ulong helper_read_cr8(CPUX86State *env)
{
target_ulong val;
switch (reg) {
default:
val = env->cr[reg];
break;
case 8:
if (!(env->hflags2 & HF2_VINTR_MASK)) {
val = cpu_get_apic_tpr(env_archcpu(env)->apic_state);
} else {
val = env->int_ctl & V_TPR_MASK;
}
break;
if (!(env->hflags2 & HF2_VINTR_MASK)) {
return cpu_get_apic_tpr(env_archcpu(env)->apic_state);
} else {
return env->int_ctl & V_TPR_MASK;
}
return val;
}
void helper_write_crN(CPUX86State *env, int reg, target_ulong t0)

716
target/i386/tcg/translate.c
View File

@ -246,10 +246,6 @@ STUB_HELPER(mwait, TCGv_env env, TCGv_i32 pc_ofs)
STUB_HELPER(outb, TCGv_env env, TCGv_i32 port, TCGv_i32 val)
STUB_HELPER(outw, TCGv_env env, TCGv_i32 port, TCGv_i32 val)
STUB_HELPER(outl, TCGv_env env, TCGv_i32 port, TCGv_i32 val)
STUB_HELPER(rdmsr, TCGv_env env)
STUB_HELPER(read_crN, TCGv ret, TCGv_env env, TCGv_i32 reg)
STUB_HELPER(get_dr, TCGv ret, TCGv_env env, TCGv_i32 reg)
STUB_HELPER(set_dr, TCGv_env env, TCGv_i32 reg, TCGv val)
STUB_HELPER(stgi, TCGv_env env)
STUB_HELPER(svm_check_intercept, TCGv_env env, TCGv_i32 type)
STUB_HELPER(vmload, TCGv_env env, TCGv_i32 aflag)
@ -257,7 +253,6 @@ STUB_HELPER(vmmcall, TCGv_env env)
STUB_HELPER(vmrun, TCGv_env env, TCGv_i32 aflag, TCGv_i32 pc_ofs)
STUB_HELPER(vmsave, TCGv_env env, TCGv_i32 aflag)
STUB_HELPER(write_crN, TCGv_env env, TCGv_i32 reg, TCGv val)
STUB_HELPER(wrmsr, TCGv_env env)
#endif
static void gen_jmp_rel(DisasContext *s, MemOp ot, int diff, int tb_num);
@ -439,13 +434,6 @@ static inline MemOp mo_stacksize(DisasContext *s)
return CODE64(s) ? MO_64 : SS32(s) ? MO_32 : MO_16;
}
/* Select size 8 if lsb of B is clear, else OT. Used for decoding
byte vs word opcodes. */
static inline MemOp mo_b_d(int b, MemOp ot)
{
return b & 1 ? ot : MO_8;
}
/* Compute the result of writing t0 to the OT-sized register REG.
*
* If DEST is NULL, store the result into the register and return the
@ -540,15 +528,6 @@ static inline void gen_op_st_v(DisasContext *s, int idx, TCGv t0, TCGv a0)
tcg_gen_qemu_st_tl(t0, a0, s->mem_index, idx | MO_LE);
}
static inline void gen_op_st_rm_T0_A0(DisasContext *s, int idx, int d)
{
if (d == OR_TMP0) {
gen_op_st_v(s, idx, s->T0, s->A0);
} else {
gen_op_mov_reg_v(s, idx, d, s->T0);
}
}
static void gen_update_eip_next(DisasContext *s)
{
assert(s->pc_save != -1);
@ -729,11 +708,6 @@ static TCGv gen_ext_tl(TCGv dst, TCGv src, MemOp size, bool sign)
return dst;
}
static void gen_extu(MemOp ot, TCGv reg)
{
gen_ext_tl(reg, reg, ot, false);
}
static void gen_exts(MemOp ot, TCGv reg)
{
gen_ext_tl(reg, reg, ot, true);
@ -837,17 +811,6 @@ static void gen_movs(DisasContext *s, MemOp ot)
gen_op_add_reg(s, s->aflag, R_EDI, dshift);
}
static void gen_op_update1_cc(DisasContext *s)
{
tcg_gen_mov_tl(cpu_cc_dst, s->T0);
}
static void gen_op_update2_cc(DisasContext *s)
{
tcg_gen_mov_tl(cpu_cc_src, s->T1);
tcg_gen_mov_tl(cpu_cc_dst, s->T0);
}
/* compute all eflags to reg */
static void gen_mov_eflags(DisasContext *s, TCGv reg)
{
@ -1448,64 +1411,11 @@ static bool check_cpl0(DisasContext *s)
return false;
}
static void gen_shift_flags(DisasContext *s, MemOp ot, TCGv result,
TCGv shm1, TCGv count, bool is_right)
{
TCGv_i32 z32, s32, oldop;
TCGv z_tl;
/* Store the results into the CC variables. If we know that the
variable must be dead, store unconditionally. Otherwise we'll
need to not disrupt the current contents. */
z_tl = tcg_constant_tl(0);
if (cc_op_live[s->cc_op] & USES_CC_DST) {
tcg_gen_movcond_tl(TCG_COND_NE, cpu_cc_dst, count, z_tl,
result, cpu_cc_dst);
} else {
tcg_gen_mov_tl(cpu_cc_dst, result);
}
if (cc_op_live[s->cc_op] & USES_CC_SRC) {
tcg_gen_movcond_tl(TCG_COND_NE, cpu_cc_src, count, z_tl,
shm1, cpu_cc_src);
} else {
tcg_gen_mov_tl(cpu_cc_src, shm1);
}
/* Get the two potential CC_OP values into temporaries. */
tcg_gen_movi_i32(s->tmp2_i32, (is_right ? CC_OP_SARB : CC_OP_SHLB) + ot);
if (s->cc_op == CC_OP_DYNAMIC) {
oldop = cpu_cc_op;
} else {
tcg_gen_movi_i32(s->tmp3_i32, s->cc_op);
oldop = s->tmp3_i32;
}
/* Conditionally store the CC_OP value. */
z32 = tcg_constant_i32(0);
s32 = tcg_temp_new_i32();
tcg_gen_trunc_tl_i32(s32, count);
tcg_gen_movcond_i32(TCG_COND_NE, cpu_cc_op, s32, z32, s->tmp2_i32, oldop);
/* The CC_OP value is no longer predictable. */
set_cc_op(s, CC_OP_DYNAMIC);
}
/* XXX: add faster immediate case */
static void gen_shiftd_rm_T1(DisasContext *s, MemOp ot, int op1,
bool is_right, TCGv count_in)
static void gen_shiftd_rm_T1(DisasContext *s, MemOp ot,
bool is_right, TCGv count)
{
target_ulong mask = (ot == MO_64 ? 63 : 31);
TCGv count;
/* load */
if (op1 == OR_TMP0) {
gen_op_ld_v(s, ot, s->T0, s->A0);
} else {
gen_op_mov_v_reg(s, ot, s->T0, op1);
}
count = tcg_temp_new();
tcg_gen_andi_tl(count, count_in, mask);
switch (ot) {
case MO_16:
@ -1567,11 +1477,6 @@ static void gen_shiftd_rm_T1(DisasContext *s, MemOp ot, int op1,
tcg_gen_or_tl(s->T0, s->T0, s->T1);
break;
}
/* store */
gen_op_st_rm_T0_A0(s, ot, op1);
gen_shift_flags(s, ot, s->T0, s->tmp0, count, is_right);
}
#define X86_MAX_INSN_LENGTH 15
@ -3081,108 +2986,11 @@ static void disas_insn_old(DisasContext *s, CPUState *cpu, int b)
CPUX86State *env = cpu_env(cpu);
int prefixes = s->prefix;
MemOp dflag = s->dflag;
int shift;
MemOp ot;
int modrm, reg, rm, mod, op, opreg, val;
int modrm, reg, rm, mod, op, val;
/* now check op code */
switch (b) {
/**************************/
/* arith & logic */
case 0x1c0:
case 0x1c1: /* xadd Ev, Gv */
ot = mo_b_d(b, dflag);
modrm = x86_ldub_code(env, s);
reg = ((modrm >> 3) & 7) | REX_R(s);
mod = (modrm >> 6) & 3;
gen_op_mov_v_reg(s, ot, s->T0, reg);
if (mod == 3) {
rm = (modrm & 7) | REX_B(s);
gen_op_mov_v_reg(s, ot, s->T1, rm);
tcg_gen_add_tl(s->T0, s->T0, s->T1);
gen_op_mov_reg_v(s, ot, reg, s->T1);
gen_op_mov_reg_v(s, ot, rm, s->T0);
} else {
gen_lea_modrm(env, s, modrm);
if (s->prefix & PREFIX_LOCK) {
tcg_gen_atomic_fetch_add_tl(s->T1, s->A0, s->T0,
s->mem_index, ot | MO_LE);
tcg_gen_add_tl(s->T0, s->T0, s->T1);
} else {
gen_op_ld_v(s, ot, s->T1, s->A0);
tcg_gen_add_tl(s->T0, s->T0, s->T1);
gen_op_st_v(s, ot, s->T0, s->A0);
}
gen_op_mov_reg_v(s, ot, reg, s->T1);
}
gen_op_update2_cc(s);
set_cc_op(s, CC_OP_ADDB + ot);
break;
case 0x1b0:
case 0x1b1: /* cmpxchg Ev, Gv */
{
TCGv oldv, newv, cmpv, dest;
ot = mo_b_d(b, dflag);
modrm = x86_ldub_code(env, s);
reg = ((modrm >> 3) & 7) | REX_R(s);
mod = (modrm >> 6) & 3;
oldv = tcg_temp_new();
newv = tcg_temp_new();
cmpv = tcg_temp_new();
gen_op_mov_v_reg(s, ot, newv, reg);
tcg_gen_mov_tl(cmpv, cpu_regs[R_EAX]);
gen_extu(ot, cmpv);
if (s->prefix & PREFIX_LOCK) {
if (mod == 3) {
goto illegal_op;
}
gen_lea_modrm(env, s, modrm);
tcg_gen_atomic_cmpxchg_tl(oldv, s->A0, cmpv, newv,
s->mem_index, ot | MO_LE);
} else {
if (mod == 3) {
rm = (modrm & 7) | REX_B(s);
gen_op_mov_v_reg(s, ot, oldv, rm);
gen_extu(ot, oldv);
/*
* Unlike the memory case, where "the destination operand receives
* a write cycle without regard to the result of the comparison",
* rm must not be touched altogether if the write fails, including
* not zero-extending it on 64-bit processors. So, precompute
* the result of a successful writeback and perform the movcond
* directly on cpu_regs. Also need to write accumulator first, in
* case rm is part of RAX too.
*/
dest = gen_op_deposit_reg_v(s, ot, rm, newv, newv);
tcg_gen_movcond_tl(TCG_COND_EQ, dest, oldv, cmpv, newv, dest);
} else {
gen_lea_modrm(env, s, modrm);
gen_op_ld_v(s, ot, oldv, s->A0);
/*
* Perform an unconditional store cycle like physical cpu;
* must be before changing accumulator to ensure
* idempotency if the store faults and the instruction
* is restarted
*/
tcg_gen_movcond_tl(TCG_COND_EQ, newv, oldv, cmpv, newv, oldv);
gen_op_st_v(s, ot, newv, s->A0);
}
}
/*
* Write EAX only if the cmpxchg fails; reuse newv as the destination,
* since it's dead here.
*/
dest = gen_op_deposit_reg_v(s, ot, R_EAX, newv, oldv);
tcg_gen_movcond_tl(TCG_COND_EQ, dest, oldv, cmpv, dest, newv);
tcg_gen_mov_tl(cpu_cc_src, oldv);
tcg_gen_mov_tl(s->cc_srcT, cmpv);
tcg_gen_sub_tl(cpu_cc_dst, cmpv, oldv);
set_cc_op(s, CC_OP_SUBB + ot);
}
break;
case 0x1c7: /* cmpxchg8b */
modrm = x86_ldub_code(env, s);
mod = (modrm >> 6) & 3;
@ -3245,45 +3053,6 @@ static void disas_insn_old(DisasContext *s, CPUState *cpu, int b)
}
break;
/**************************/
/* shifts */
case 0x1a4: /* shld imm */
op = 0;
shift = 1;
goto do_shiftd;
case 0x1a5: /* shld cl */
op = 0;
shift = 0;
goto do_shiftd;
case 0x1ac: /* shrd imm */
op = 1;
shift = 1;
goto do_shiftd;
case 0x1ad: /* shrd cl */
op = 1;
shift = 0;
do_shiftd:
ot = dflag;
modrm = x86_ldub_code(env, s);
mod = (modrm >> 6) & 3;
rm = (modrm & 7) | REX_B(s);
reg = ((modrm >> 3) & 7) | REX_R(s);
if (mod != 3) {
gen_lea_modrm(env, s, modrm);
opreg = OR_TMP0;
} else {
opreg = rm;
}
gen_op_mov_v_reg(s, ot, s->T1, reg);
if (shift) {
TCGv imm = tcg_constant_tl(x86_ldub_code(env, s));
gen_shiftd_rm_T1(s, ot, opreg, op, imm);
} else {
gen_shiftd_rm_T1(s, ot, opreg, op, cpu_regs[R_ECX]);
}
break;
/************************/
/* bit operations */
case 0x1ba: /* bt/bts/btr/btc Gv, im */
@ -3423,147 +3192,6 @@ static void disas_insn_old(DisasContext *s, CPUState *cpu, int b)
break;
}
break;
case 0x1bc: /* bsf / tzcnt */
case 0x1bd: /* bsr / lzcnt */
ot = dflag;
modrm = x86_ldub_code(env, s);
reg = ((modrm >> 3) & 7) | REX_R(s);
gen_ld_modrm(env, s, modrm, ot);
gen_extu(ot, s->T0);
/* Note that lzcnt and tzcnt are in different extensions. */
if ((prefixes & PREFIX_REPZ)
&& (b & 1
? s->cpuid_ext3_features & CPUID_EXT3_ABM
: s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_BMI1)) {
int size = 8 << ot;
/* For lzcnt/tzcnt, C bit is defined related to the input. */
tcg_gen_mov_tl(cpu_cc_src, s->T0);
if (b & 1) {
/* For lzcnt, reduce the target_ulong result by the
number of zeros that we expect to find at the top. */
tcg_gen_clzi_tl(s->T0, s->T0, TARGET_LONG_BITS);
tcg_gen_subi_tl(s->T0, s->T0, TARGET_LONG_BITS - size);
} else {
/* For tzcnt, a zero input must return the operand size. */
tcg_gen_ctzi_tl(s->T0, s->T0, size);
}
/* For lzcnt/tzcnt, Z bit is defined related to the result. */
gen_op_update1_cc(s);
set_cc_op(s, CC_OP_BMILGB + ot);
} else {
/* For bsr/bsf, only the Z bit is defined and it is related
to the input and not the result. */
tcg_gen_mov_tl(cpu_cc_dst, s->T0);
set_cc_op(s, CC_OP_LOGICB + ot);
/* ??? The manual says that the output is undefined when the
input is zero, but real hardware leaves it unchanged, and
real programs appear to depend on that. Accomplish this
by passing the output as the value to return upon zero. */
if (b & 1) {
/* For bsr, return the bit index of the first 1 bit,
not the count of leading zeros. */
tcg_gen_xori_tl(s->T1, cpu_regs[reg], TARGET_LONG_BITS - 1);
tcg_gen_clz_tl(s->T0, s->T0, s->T1);
tcg_gen_xori_tl(s->T0, s->T0, TARGET_LONG_BITS - 1);
} else {
tcg_gen_ctz_tl(s->T0, s->T0, cpu_regs[reg]);
}
}
gen_op_mov_reg_v(s, ot, reg, s->T0);
break;
case 0x130: /* wrmsr */
case 0x132: /* rdmsr */
if (check_cpl0(s)) {
gen_update_cc_op(s);
gen_update_eip_cur(s);
if (b & 2) {
gen_helper_rdmsr(tcg_env);
} else {
gen_helper_wrmsr(tcg_env);
s->base.is_jmp = DISAS_EOB_NEXT;
}
}
break;
case 0x131: /* rdtsc */
gen_update_cc_op(s);
gen_update_eip_cur(s);
translator_io_start(&s->base);
gen_helper_rdtsc(tcg_env);
break;
case 0x133: /* rdpmc */
gen_update_cc_op(s);
gen_update_eip_cur(s);
gen_helper_rdpmc(tcg_env);
s->base.is_jmp = DISAS_NORETURN;
break;
case 0x134: /* sysenter */
/* For AMD SYSENTER is not valid in long mode */
if (LMA(s) && env->cpuid_vendor1 != CPUID_VENDOR_INTEL_1) {
goto illegal_op;
}
if (!PE(s)) {
gen_exception_gpf(s);
} else {
gen_helper_sysenter(tcg_env);
s->base.is_jmp = DISAS_EOB_ONLY;
}
break;
case 0x135: /* sysexit */
/* For AMD SYSEXIT is not valid in long mode */
if (LMA(s) && env->cpuid_vendor1 != CPUID_VENDOR_INTEL_1) {
goto illegal_op;
}
if (!PE(s) || CPL(s) != 0) {
gen_exception_gpf(s);
} else {
gen_helper_sysexit(tcg_env, tcg_constant_i32(dflag - 1));
s->base.is_jmp = DISAS_EOB_ONLY;
}
break;
case 0x105: /* syscall */
/* For Intel SYSCALL is only valid in long mode */
if (!LMA(s) && env->cpuid_vendor1 == CPUID_VENDOR_INTEL_1) {
goto illegal_op;
}
gen_update_cc_op(s);
gen_update_eip_cur(s);
gen_helper_syscall(tcg_env, cur_insn_len_i32(s));
/* condition codes are modified only in long mode */
if (LMA(s)) {
assume_cc_op(s, CC_OP_EFLAGS);
}
/* TF handling for the syscall insn is different. The TF bit is checked
after the syscall insn completes. This allows #DB to not be
generated after one has entered CPL0 if TF is set in FMASK. */
s->base.is_jmp = DISAS_EOB_RECHECK_TF;
break;
case 0x107: /* sysret */
/* For Intel SYSRET is only valid in long mode */
if (!LMA(s) && env->cpuid_vendor1 == CPUID_VENDOR_INTEL_1) {
goto illegal_op;
}
if (!PE(s) || CPL(s) != 0) {
gen_exception_gpf(s);
} else {
gen_helper_sysret(tcg_env, tcg_constant_i32(dflag - 1));
/* condition codes are modified only in long mode */
if (LMA(s)) {
assume_cc_op(s, CC_OP_EFLAGS);
}
/* TF handling for the sysret insn is different. The TF bit is
checked after the sysret insn completes. This allows #DB to be
generated "as if" the syscall insn in userspace has just
completed. */
s->base.is_jmp = DISAS_EOB_RECHECK_TF;
}
break;
case 0x1a2: /* cpuid */
gen_update_cc_op(s);
gen_update_eip_cur(s);
gen_helper_cpuid(tcg_env);
break;
case 0x100:
modrm = x86_ldub_code(env, s);
mod = (modrm >> 6) & 3;
@ -3967,39 +3595,6 @@ static void disas_insn_old(DisasContext *s, CPUState *cpu, int b)
}
break;
case 0x108: /* invd */
case 0x109: /* wbinvd; wbnoinvd with REPZ prefix */
if (check_cpl0(s)) {
gen_svm_check_intercept(s, (b & 1) ? SVM_EXIT_WBINVD : SVM_EXIT_INVD);
/* nothing to do */
}
break;
case 0x102: /* lar */
case 0x103: /* lsl */
{
TCGLabel *label1;
TCGv t0;
if (!PE(s) || VM86(s))
goto illegal_op;
ot = dflag != MO_16 ? MO_32 : MO_16;
modrm = x86_ldub_code(env, s);
reg = ((modrm >> 3) & 7) | REX_R(s);
gen_ld_modrm(env, s, modrm, MO_16);
t0 = tcg_temp_new();
gen_update_cc_op(s);
if (b == 0x102) {
gen_helper_lar(t0, tcg_env, s->T0);
} else {
gen_helper_lsl(t0, tcg_env, s->T0);
}
tcg_gen_andi_tl(s->tmp0, cpu_cc_src, CC_Z);
label1 = gen_new_label();
tcg_gen_brcondi_tl(TCG_COND_EQ, s->tmp0, 0, label1);
gen_op_mov_reg_v(s, ot, reg, t0);
gen_set_label(label1);
set_cc_op(s, CC_OP_EFLAGS);
}
break;
case 0x11a:
modrm = x86_ldub_code(env, s);
if (s->flags & HF_MPX_EN_MASK) {
@ -4191,311 +3786,6 @@ static void disas_insn_old(DisasContext *s, CPUState *cpu, int b)
}
gen_nop_modrm(env, s, modrm);
break;
case 0x120: /* mov reg, crN */
case 0x122: /* mov crN, reg */
if (!check_cpl0(s)) {
break;
}
modrm = x86_ldub_code(env, s);
/*
* Ignore the mod bits (assume (modrm&0xc0)==0xc0).
* AMD documentation (24594.pdf) and testing of Intel 386 and 486
* processors all show that the mod bits are assumed to be 1's,
* regardless of actual values.
*/
rm = (modrm & 7) | REX_B(s);
reg = ((modrm >> 3) & 7) | REX_R(s);
switch (reg) {
case 0:
if ((prefixes & PREFIX_LOCK) &&
(s->cpuid_ext3_features & CPUID_EXT3_CR8LEG)) {
reg = 8;
}
break;
case 2:
case 3:
case 4:
case 8:
break;
default:
goto unknown_op;
}
ot = (CODE64(s) ? MO_64 : MO_32);
translator_io_start(&s->base);
if (b & 2) {
gen_svm_check_intercept(s, SVM_EXIT_WRITE_CR0 + reg);
gen_op_mov_v_reg(s, ot, s->T0, rm);
gen_helper_write_crN(tcg_env, tcg_constant_i32(reg), s->T0);
s->base.is_jmp = DISAS_EOB_NEXT;
} else {
gen_svm_check_intercept(s, SVM_EXIT_READ_CR0 + reg);
gen_helper_read_crN(s->T0, tcg_env, tcg_constant_i32(reg));
gen_op_mov_reg_v(s, ot, rm, s->T0);
}
break;
case 0x121: /* mov reg, drN */
case 0x123: /* mov drN, reg */
if (check_cpl0(s)) {
modrm = x86_ldub_code(env, s);
/* Ignore the mod bits (assume (modrm&0xc0)==0xc0).
* AMD documentation (24594.pdf) and testing of
* intel 386 and 486 processors all show that the mod bits
* are assumed to be 1's, regardless of actual values.
*/
rm = (modrm & 7) | REX_B(s);
reg = ((modrm >> 3) & 7) | REX_R(s);
if (CODE64(s))
ot = MO_64;
else
ot = MO_32;
if (reg >= 8) {
goto illegal_op;
}
if (b & 2) {
gen_svm_check_intercept(s, SVM_EXIT_WRITE_DR0 + reg);
gen_op_mov_v_reg(s, ot, s->T0, rm);
tcg_gen_movi_i32(s->tmp2_i32, reg);
gen_helper_set_dr(tcg_env, s->tmp2_i32, s->T0);
s->base.is_jmp = DISAS_EOB_NEXT;
} else {
gen_svm_check_intercept(s, SVM_EXIT_READ_DR0 + reg);
tcg_gen_movi_i32(s->tmp2_i32, reg);
gen_helper_get_dr(s->T0, tcg_env, s->tmp2_i32);
gen_op_mov_reg_v(s, ot, rm, s->T0);
}
}
break;
case 0x106: /* clts */
if (check_cpl0(s)) {
gen_svm_check_intercept(s, SVM_EXIT_WRITE_CR0);
gen_helper_clts(tcg_env);
/* abort block because static cpu state changed */
s->base.is_jmp = DISAS_EOB_NEXT;
}
break;
/* MMX/3DNow!/SSE/SSE2/SSE3/SSSE3/SSE4 support */
case 0x1ae:
modrm = x86_ldub_code(env, s);
switch (modrm) {
CASE_MODRM_MEM_OP(0): /* fxsave */
if (!(s->cpuid_features & CPUID_FXSR)
|| (prefixes & PREFIX_LOCK)) {
goto illegal_op;
}
if ((s->flags & HF_EM_MASK) || (s->flags & HF_TS_MASK)) {
gen_exception(s, EXCP07_PREX);
break;
}
gen_lea_modrm(env, s, modrm);
gen_helper_fxsave(tcg_env, s->A0);
break;
CASE_MODRM_MEM_OP(1): /* fxrstor */
if (!(s->cpuid_features & CPUID_FXSR)
|| (prefixes & PREFIX_LOCK)) {
goto illegal_op;
}
if ((s->flags & HF_EM_MASK) || (s->flags & HF_TS_MASK)) {
gen_exception(s, EXCP07_PREX);
break;
}
gen_lea_modrm(env, s, modrm);
gen_helper_fxrstor(tcg_env, s->A0);
break;
CASE_MODRM_MEM_OP(2): /* ldmxcsr */
if ((s->flags & HF_EM_MASK) || !(s->flags & HF_OSFXSR_MASK)) {
goto illegal_op;
}
if (s->flags & HF_TS_MASK) {
gen_exception(s, EXCP07_PREX);
break;
}
gen_lea_modrm(env, s, modrm);
tcg_gen_qemu_ld_i32(s->tmp2_i32, s->A0, s->mem_index, MO_LEUL);
gen_helper_ldmxcsr(tcg_env, s->tmp2_i32);
break;
CASE_MODRM_MEM_OP(3): /* stmxcsr */
if ((s->flags & HF_EM_MASK) || !(s->flags & HF_OSFXSR_MASK)) {
goto illegal_op;
}
if (s->flags & HF_TS_MASK) {
gen_exception(s, EXCP07_PREX);
break;
}
gen_helper_update_mxcsr(tcg_env);
gen_lea_modrm(env, s, modrm);
tcg_gen_ld32u_tl(s->T0, tcg_env, offsetof(CPUX86State, mxcsr));
gen_op_st_v(s, MO_32, s->T0, s->A0);
break;
CASE_MODRM_MEM_OP(4): /* xsave */
if ((s->cpuid_ext_features & CPUID_EXT_XSAVE) == 0
|| (prefixes & (PREFIX_LOCK | PREFIX_DATA
| PREFIX_REPZ | PREFIX_REPNZ))) {
goto illegal_op;
}
gen_lea_modrm(env, s, modrm);
tcg_gen_concat_tl_i64(s->tmp1_i64, cpu_regs[R_EAX],
cpu_regs[R_EDX]);
gen_helper_xsave(tcg_env, s->A0, s->tmp1_i64);
break;
CASE_MODRM_MEM_OP(5): /* xrstor */
if ((s->cpuid_ext_features & CPUID_EXT_XSAVE) == 0
|| (prefixes & (PREFIX_LOCK | PREFIX_DATA
| PREFIX_REPZ | PREFIX_REPNZ))) {
goto illegal_op;
}
gen_lea_modrm(env, s, modrm);
tcg_gen_concat_tl_i64(s->tmp1_i64, cpu_regs[R_EAX],
cpu_regs[R_EDX]);
gen_helper_xrstor(tcg_env, s->A0, s->tmp1_i64);
/* XRSTOR is how MPX is enabled, which changes how
we translate. Thus we need to end the TB. */
s->base.is_jmp = DISAS_EOB_NEXT;
break;
CASE_MODRM_MEM_OP(6): /* xsaveopt / clwb */
if (prefixes & PREFIX_LOCK) {
goto illegal_op;
}
if (prefixes & PREFIX_DATA) {
/* clwb */
if (!(s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_CLWB)) {
goto illegal_op;
}
gen_nop_modrm(env, s, modrm);
} else {
/* xsaveopt */
if ((s->cpuid_ext_features & CPUID_EXT_XSAVE) == 0
|| (s->cpuid_xsave_features & CPUID_XSAVE_XSAVEOPT) == 0
|| (prefixes & (PREFIX_REPZ | PREFIX_REPNZ))) {
goto illegal_op;
}
gen_lea_modrm(env, s, modrm);
tcg_gen_concat_tl_i64(s->tmp1_i64, cpu_regs[R_EAX],
cpu_regs[R_EDX]);
gen_helper_xsaveopt(tcg_env, s->A0, s->tmp1_i64);
}
break;
CASE_MODRM_MEM_OP(7): /* clflush / clflushopt */
if (prefixes & PREFIX_LOCK) {
goto illegal_op;
}
if (prefixes & PREFIX_DATA) {
/* clflushopt */
if (!(s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_CLFLUSHOPT)) {
goto illegal_op;
}
} else {
/* clflush */
if ((s->prefix & (PREFIX_REPZ | PREFIX_REPNZ))
|| !(s->cpuid_features & CPUID_CLFLUSH)) {
goto illegal_op;
}
}
gen_nop_modrm(env, s, modrm);
break;
case 0xc0 ... 0xc7: /* rdfsbase (f3 0f ae /0) */
case 0xc8 ... 0xcf: /* rdgsbase (f3 0f ae /1) */
case 0xd0 ... 0xd7: /* wrfsbase (f3 0f ae /2) */
case 0xd8 ... 0xdf: /* wrgsbase (f3 0f ae /3) */
if (CODE64(s)
&& (prefixes & PREFIX_REPZ)
&& !(prefixes & PREFIX_LOCK)
&& (s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_FSGSBASE)) {
TCGv base, treg, src, dst;
/* Preserve hflags bits by testing CR4 at runtime. */
tcg_gen_movi_i32(s->tmp2_i32, CR4_FSGSBASE_MASK);
gen_helper_cr4_testbit(tcg_env, s->tmp2_i32);
base = cpu_seg_base[modrm & 8 ? R_GS : R_FS];
treg = cpu_regs[(modrm & 7) | REX_B(s)];
if (modrm & 0x10) {
/* wr*base */
dst = base, src = treg;
} else {
/* rd*base */
dst = treg, src = base;
}
if (s->dflag == MO_32) {
tcg_gen_ext32u_tl(dst, src);
} else {
tcg_gen_mov_tl(dst, src);
}
break;
}
goto unknown_op;
case 0xf8 ... 0xff: /* sfence */
if (!(s->cpuid_features & CPUID_SSE)
|| (prefixes & PREFIX_LOCK)) {
goto illegal_op;
}
tcg_gen_mb(TCG_MO_ST_ST | TCG_BAR_SC);
break;
case 0xe8 ... 0xef: /* lfence */
if (!(s->cpuid_features & CPUID_SSE)
|| (prefixes & PREFIX_LOCK)) {
goto illegal_op;
}
tcg_gen_mb(TCG_MO_LD_LD | TCG_BAR_SC);
break;
case 0xf0 ... 0xf7: /* mfence */
if (!(s->cpuid_features & CPUID_SSE2)
|| (prefixes & PREFIX_LOCK)) {
goto illegal_op;
}
tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
break;
default:
goto unknown_op;
}
break;
case 0x1aa: /* rsm */
gen_svm_check_intercept(s, SVM_EXIT_RSM);
if (!(s->flags & HF_SMM_MASK))
goto illegal_op;
#ifdef CONFIG_USER_ONLY
/* we should not be in SMM mode */
g_assert_not_reached();
#else
gen_helper_rsm(tcg_env);
assume_cc_op(s, CC_OP_EFLAGS);
#endif /* CONFIG_USER_ONLY */
s->base.is_jmp = DISAS_EOB_ONLY;
break;
case 0x1b8: /* SSE4.2 popcnt */
if ((prefixes & (PREFIX_REPZ | PREFIX_LOCK | PREFIX_REPNZ)) !=
PREFIX_REPZ)
goto illegal_op;
if (!(s->cpuid_ext_features & CPUID_EXT_POPCNT))
goto illegal_op;
modrm = x86_ldub_code(env, s);
reg = ((modrm >> 3) & 7) | REX_R(s);
ot = dflag;
gen_ld_modrm(env, s, modrm, ot);
gen_extu(ot, s->T0);
tcg_gen_mov_tl(cpu_cc_src, s->T0);
tcg_gen_ctpop_tl(s->T0, s->T0);
gen_op_mov_reg_v(s, ot, reg, s->T0);
set_cc_op(s, CC_OP_POPCNT);
break;
default:
g_assert_not_reached();
}