blob: 6ad81d221b701fc633d2e83089c59a2d4998136a [file] [log] [blame]
/*---------------------------------------------------------------*/
/*--- begin guest_amd64_defs.h ---*/
/*---------------------------------------------------------------*/
/*
This file is part of Valgrind, a dynamic binary instrumentation
framework.
Copyright (C) 2004-2012 OpenWorks LLP
info@open-works.net
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License as
published by the Free Software Foundation; either version 2 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301, USA.
The GNU General Public License is contained in the file COPYING.
Neither the names of the U.S. Department of Energy nor the
University of California nor the names of its contributors may be
used to endorse or promote products derived from this software
without prior written permission.
*/
/* Only to be used within the guest-amd64 directory. */
#ifndef __VEX_GUEST_AMD64_DEFS_H
#define __VEX_GUEST_AMD64_DEFS_H
/*---------------------------------------------------------*/
/*--- amd64 to IR conversion ---*/
/*---------------------------------------------------------*/
/* Convert one amd64 insn to IR. See the type DisOneInstrFn in
bb_to_IR.h. */
extern
DisResult disInstr_AMD64 ( IRSB* irbb,
Bool (*resteerOkFn) ( void*, Addr64 ),
Bool resteerCisOk,
void* callback_opaque,
UChar* guest_code,
Long delta,
Addr64 guest_IP,
VexArch guest_arch,
VexArchInfo* archinfo,
VexAbiInfo* abiinfo,
Bool host_bigendian );
/* Used by the optimiser to specialise calls to helpers. */
extern
IRExpr* guest_amd64_spechelper ( HChar* function_name,
IRExpr** args,
IRStmt** precedingStmts,
Int n_precedingStmts );
/* Describes to the optimiser which part of the guest state require
precise memory exceptions. This is logically part of the guest
state description. */
extern
Bool guest_amd64_state_requires_precise_mem_exns ( Int, Int );
extern
VexGuestLayout amd64guest_layout;
/*---------------------------------------------------------*/
/*--- amd64 guest helpers ---*/
/*---------------------------------------------------------*/
/* --- CLEAN HELPERS --- */
extern ULong amd64g_calculate_rflags_all (
ULong cc_op,
ULong cc_dep1, ULong cc_dep2, ULong cc_ndep
);
extern ULong amd64g_calculate_rflags_c (
ULong cc_op,
ULong cc_dep1, ULong cc_dep2, ULong cc_ndep
);
extern ULong amd64g_calculate_condition (
ULong/*AMD64Condcode*/ cond,
ULong cc_op,
ULong cc_dep1, ULong cc_dep2, ULong cc_ndep
);
extern ULong amd64g_calculate_FXAM ( ULong tag, ULong dbl );
extern ULong amd64g_calculate_RCR (
ULong arg, ULong rot_amt, ULong rflags_in, Long sz
);
extern ULong amd64g_calculate_RCL (
ULong arg, ULong rot_amt, ULong rflags_in, Long sz
);
extern ULong amd64g_calculate_pclmul(ULong s1, ULong s2, ULong which);
extern ULong amd64g_check_fldcw ( ULong fpucw );
extern ULong amd64g_create_fpucw ( ULong fpround );
extern ULong amd64g_check_ldmxcsr ( ULong mxcsr );
extern ULong amd64g_create_mxcsr ( ULong sseround );
extern VexEmNote amd64g_dirtyhelper_FLDENV ( VexGuestAMD64State*, HWord );
extern VexEmNote amd64g_dirtyhelper_FRSTOR ( VexGuestAMD64State*, HWord );
extern VexEmNote amd64g_dirtyhelper_FRSTORS ( VexGuestAMD64State*, HWord );
extern void amd64g_dirtyhelper_FSTENV ( VexGuestAMD64State*, HWord );
extern void amd64g_dirtyhelper_FNSAVE ( VexGuestAMD64State*, HWord );
extern void amd64g_dirtyhelper_FNSAVES ( VexGuestAMD64State*, HWord );
/* Translate a guest virtual_addr into a guest linear address by
consulting the supplied LDT/GDT structures. Their representation
must be as specified in pub/libvex_guest_amd64.h. To indicate a
translation failure, 1<<32 is returned. On success, the lower 32
bits of the returned result indicate the linear address.
*/
//extern
//ULong amd64g_use_seg_selector ( HWord ldt, HWord gdt,
// UInt seg_selector, UInt virtual_addr );
extern ULong amd64g_calculate_mmx_pmaddwd ( ULong, ULong );
extern ULong amd64g_calculate_mmx_psadbw ( ULong, ULong );
extern ULong amd64g_calculate_mmx_pmovmskb ( ULong );
extern ULong amd64g_calculate_sse_pmovmskb ( ULong w64hi, ULong w64lo );
extern ULong amd64g_calculate_sse_phminposuw ( ULong sLo, ULong sHi );
extern ULong amd64g_calc_crc32b ( ULong crcIn, ULong b );
extern ULong amd64g_calc_crc32w ( ULong crcIn, ULong w );
extern ULong amd64g_calc_crc32l ( ULong crcIn, ULong l );
extern ULong amd64g_calc_crc32q ( ULong crcIn, ULong q );
extern ULong amd64g_calc_mpsadbw ( ULong sHi, ULong sLo,
ULong dHi, ULong dLo,
ULong imm_and_return_control_bit );
/* --- DIRTY HELPERS --- */
extern ULong amd64g_dirtyhelper_loadF80le ( ULong/*addr*/ );
extern void amd64g_dirtyhelper_storeF80le ( ULong/*addr*/, ULong/*data*/ );
extern void amd64g_dirtyhelper_CPUID_baseline ( VexGuestAMD64State* st );
extern void amd64g_dirtyhelper_CPUID_sse3_and_cx16 ( VexGuestAMD64State* st );
extern void amd64g_dirtyhelper_CPUID_sse42_and_cx16 ( VexGuestAMD64State* st );
extern void amd64g_dirtyhelper_CPUID_avx_and_cx16 ( VexGuestAMD64State* st );
extern void amd64g_dirtyhelper_FINIT ( VexGuestAMD64State* );
extern void amd64g_dirtyhelper_FXSAVE ( VexGuestAMD64State*, HWord );
extern VexEmNote amd64g_dirtyhelper_FXRSTOR ( VexGuestAMD64State*, HWord );
extern ULong amd64g_dirtyhelper_RDTSC ( void );
extern ULong amd64g_dirtyhelper_IN ( ULong portno, ULong sz/*1,2 or 4*/ );
extern void amd64g_dirtyhelper_OUT ( ULong portno, ULong data,
ULong sz/*1,2 or 4*/ );
extern void amd64g_dirtyhelper_SxDT ( void* address,
ULong op /* 0 or 1 */ );
/* Helps with PCMP{I,E}STR{I,M}.
CALLED FROM GENERATED CODE: DIRTY HELPER(s). (But not really,
actually it could be a clean helper, but for the fact that we can't
pass by value 2 x V128 to a clean helper, nor have one returned.)
Reads guest state, writes to guest state for the xSTRM cases, no
accesses of memory, is a pure function.
opc_and_imm contains (4th byte of opcode << 8) | the-imm8-byte so
the callee knows which I/E and I/M variant it is dealing with and
what the specific operation is. 4th byte of opcode is in the range
0x60 to 0x63:
istri 66 0F 3A 63
istrm 66 0F 3A 62
estri 66 0F 3A 61
estrm 66 0F 3A 60
gstOffL and gstOffR are the guest state offsets for the two XMM
register inputs. We never have to deal with the memory case since
that is handled by pre-loading the relevant value into the fake
XMM16 register.
For ESTRx variants, edxIN and eaxIN hold the values of those two
registers.
In all cases, the bottom 16 bits of the result contain the new
OSZACP %rflags values. For xSTRI variants, bits[31:16] of the
result hold the new %ecx value. For xSTRM variants, the helper
writes the result directly to the guest XMM0.
Declarable side effects: in all cases, reads guest state at
[gstOffL, +16) and [gstOffR, +16). For xSTRM variants, also writes
guest_XMM0.
Is expected to be called with opc_and_imm combinations which have
actually been validated, and will assert if otherwise. The front
end should ensure we're only called with verified values.
*/
extern ULong amd64g_dirtyhelper_PCMPxSTRx (
VexGuestAMD64State*,
HWord opc4_and_imm,
HWord gstOffL, HWord gstOffR,
HWord edxIN, HWord eaxIN
);
/* Implementation of intel AES instructions as described in
Intel Advanced Vector Extensions
Programming Reference
MARCH 2008
319433-002.
CALLED FROM GENERATED CODE: DIRTY HELPER(s). (But not really,
actually it could be a clean helper, but for the fact that we can't
pass by value 2 x V128 to a clean helper, nor have one returned.)
Reads guest state, writes to guest state, no
accesses of memory, is a pure function.
opc4 contains the 4th byte of opcode. Front-end should only
give opcode corresponding to AESENC/AESENCLAST/AESDEC/AESDECLAST/AESIMC.
(will assert otherwise).
gstOffL and gstOffR are the guest state offsets for the two XMM
register inputs, gstOffD is the guest state offset for the XMM register
output. We never have to deal with the memory case since that is handled
by pre-loading the relevant value into the fake XMM16 register.
*/
extern void amd64g_dirtyhelper_AES (
VexGuestAMD64State* gst,
HWord opc4, HWord gstOffD,
HWord gstOffL, HWord gstOffR
);
/* Implementation of AESKEYGENASSIST.
CALLED FROM GENERATED CODE: DIRTY HELPER(s). (But not really,
actually it could be a clean helper, but for the fact that we can't
pass by value 1 x V128 to a clean helper, nor have one returned.)
Reads guest state, writes to guest state, no
accesses of memory, is a pure function.
imm8 is the Round Key constant.
gstOffL and gstOffR are the guest state offsets for the two XMM
register input and output. We never have to deal with the memory case since
that is handled by pre-loading the relevant value into the fake
XMM16 register.
*/
extern void amd64g_dirtyhelper_AESKEYGENASSIST (
VexGuestAMD64State* gst,
HWord imm8,
HWord gstOffL, HWord gstOffR
);
//extern void amd64g_dirtyhelper_CPUID_sse0 ( VexGuestAMD64State* );
//extern void amd64g_dirtyhelper_CPUID_sse1 ( VexGuestAMD64State* );
//extern void amd64g_dirtyhelper_CPUID_sse2 ( VexGuestAMD64State* );
//extern void amd64g_dirtyhelper_FSAVE ( VexGuestAMD64State*, HWord );
//extern VexEmNote
// amd64g_dirtyhelper_FRSTOR ( VexGuestAMD64State*, HWord );
//extern void amd64g_dirtyhelper_FSTENV ( VexGuestAMD64State*, HWord );
//extern VexEmNote
// amd64g_dirtyhelper_FLDENV ( VexGuestAMD64State*, HWord );
/*---------------------------------------------------------*/
/*--- Condition code stuff ---*/
/*---------------------------------------------------------*/
/* rflags masks */
#define AMD64G_CC_SHIFT_O 11
#define AMD64G_CC_SHIFT_S 7
#define AMD64G_CC_SHIFT_Z 6
#define AMD64G_CC_SHIFT_A 4
#define AMD64G_CC_SHIFT_C 0
#define AMD64G_CC_SHIFT_P 2
#define AMD64G_CC_MASK_O (1ULL << AMD64G_CC_SHIFT_O)
#define AMD64G_CC_MASK_S (1ULL << AMD64G_CC_SHIFT_S)
#define AMD64G_CC_MASK_Z (1ULL << AMD64G_CC_SHIFT_Z)
#define AMD64G_CC_MASK_A (1ULL << AMD64G_CC_SHIFT_A)
#define AMD64G_CC_MASK_C (1ULL << AMD64G_CC_SHIFT_C)
#define AMD64G_CC_MASK_P (1ULL << AMD64G_CC_SHIFT_P)
/* FPU flag masks */
#define AMD64G_FC_SHIFT_C3 14
#define AMD64G_FC_SHIFT_C2 10
#define AMD64G_FC_SHIFT_C1 9
#define AMD64G_FC_SHIFT_C0 8
#define AMD64G_FC_MASK_C3 (1ULL << AMD64G_FC_SHIFT_C3)
#define AMD64G_FC_MASK_C2 (1ULL << AMD64G_FC_SHIFT_C2)
#define AMD64G_FC_MASK_C1 (1ULL << AMD64G_FC_SHIFT_C1)
#define AMD64G_FC_MASK_C0 (1ULL << AMD64G_FC_SHIFT_C0)
/* %RFLAGS thunk descriptors. A four-word thunk is used to record
details of the most recent flag-setting operation, so the flags can
be computed later if needed. It is possible to do this a little
more efficiently using a 3-word thunk, but that makes it impossible
to describe the flag data dependencies sufficiently accurately for
Memcheck. Hence 4 words are used, with minimal loss of efficiency.
The four words are:
CC_OP, which describes the operation.
CC_DEP1 and CC_DEP2. These are arguments to the operation.
We want Memcheck to believe that the resulting flags are
data-dependent on both CC_DEP1 and CC_DEP2, hence the
name DEP.
CC_NDEP. This is a 3rd argument to the operation which is
sometimes needed. We arrange things so that Memcheck does
not believe the resulting flags are data-dependent on CC_NDEP
("not dependent").
To make Memcheck believe that (the definedness of) the encoded
flags depends only on (the definedness of) CC_DEP1 and CC_DEP2
requires two things:
(1) In the guest state layout info (amd64guest_layout), CC_OP and
CC_NDEP are marked as always defined.
(2) When passing the thunk components to an evaluation function
(calculate_condition, calculate_eflags, calculate_eflags_c) the
IRCallee's mcx_mask must be set so as to exclude from
consideration all passed args except CC_DEP1 and CC_DEP2.
Strictly speaking only (2) is necessary for correctness. However,
(1) helps efficiency in that since (2) means we never ask about the
definedness of CC_OP or CC_NDEP, we may as well not even bother to
track their definedness.
When building the thunk, it is always necessary to write words into
CC_DEP1 and CC_DEP2, even if those args are not used given the
CC_OP field (eg, CC_DEP2 is not used if CC_OP is CC_LOGIC1/2/4).
This is important because otherwise Memcheck could give false
positives as it does not understand the relationship between the
CC_OP field and CC_DEP1 and CC_DEP2, and so believes that the
definedness of the stored flags always depends on both CC_DEP1 and
CC_DEP2.
However, it is only necessary to set CC_NDEP when the CC_OP value
requires it, because Memcheck ignores CC_NDEP, and the evaluation
functions do understand the CC_OP fields and will only examine
CC_NDEP for suitable values of CC_OP.
A summary of the field usages is:
Operation DEP1 DEP2 NDEP
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
add/sub/mul first arg second arg unused
adc/sbb first arg (second arg)
XOR old_carry old_carry
and/or/xor result zero unused
inc/dec result zero old_carry
shl/shr/sar result subshifted- unused
result
rol/ror result zero old_flags
copy old_flags zero unused.
Therefore Memcheck will believe the following:
* add/sub/mul -- definedness of result flags depends on definedness
of both args.
* adc/sbb -- definedness of result flags depends on definedness of
both args and definedness of the old C flag. Because only two
DEP fields are available, the old C flag is XOR'd into the second
arg so that Memcheck sees the data dependency on it. That means
the NDEP field must contain a second copy of the old C flag
so that the evaluation functions can correctly recover the second
arg.
* and/or/xor are straightforward -- definedness of result flags
depends on definedness of result value.
* inc/dec -- definedness of result flags depends only on
definedness of result. This isn't really true -- it also depends
on the old C flag. However, we don't want Memcheck to see that,
and so the old C flag must be passed in NDEP and not in DEP2.
It's inconceivable that a compiler would generate code that puts
the C flag in an undefined state, then does an inc/dec, which
leaves C unchanged, and then makes a conditional jump/move based
on C. So our fiction seems a good approximation.
* shl/shr/sar -- straightforward, again, definedness of result
flags depends on definedness of result value. The subshifted
value (value shifted one less) is also needed, but its
definedness is the same as the definedness of the shifted value.
* rol/ror -- these only set O and C, and leave A Z C P alone.
However it seems prudent (as per inc/dec) to say the definedness
of all resulting flags depends on the definedness of the result,
hence the old flags must go in as NDEP and not DEP2.
* rcl/rcr are too difficult to do in-line, and so are done by a
helper function. They are not part of this scheme. The helper
function takes the value to be rotated, the rotate amount and the
old flags, and returns the new flags and the rotated value.
Since the helper's mcx_mask does not have any set bits, Memcheck
will lazily propagate undefinedness from any of the 3 args into
both results (flags and actual value).
*/
enum {
AMD64G_CC_OP_COPY=0, /* DEP1 = current flags, DEP2 = 0, NDEP = unused */
/* just copy DEP1 to output */
AMD64G_CC_OP_ADDB, /* 1 */
AMD64G_CC_OP_ADDW, /* 2 DEP1 = argL, DEP2 = argR, NDEP = unused */
AMD64G_CC_OP_ADDL, /* 3 */
AMD64G_CC_OP_ADDQ, /* 4 */
AMD64G_CC_OP_SUBB, /* 5 */
AMD64G_CC_OP_SUBW, /* 6 DEP1 = argL, DEP2 = argR, NDEP = unused */
AMD64G_CC_OP_SUBL, /* 7 */
AMD64G_CC_OP_SUBQ, /* 8 */
AMD64G_CC_OP_ADCB, /* 9 */
AMD64G_CC_OP_ADCW, /* 10 DEP1 = argL, DEP2 = argR ^ oldCarry, NDEP = oldCarry */
AMD64G_CC_OP_ADCL, /* 11 */
AMD64G_CC_OP_ADCQ, /* 12 */
AMD64G_CC_OP_SBBB, /* 13 */
AMD64G_CC_OP_SBBW, /* 14 DEP1 = argL, DEP2 = argR ^ oldCarry, NDEP = oldCarry */
AMD64G_CC_OP_SBBL, /* 15 */
AMD64G_CC_OP_SBBQ, /* 16 */
AMD64G_CC_OP_LOGICB, /* 17 */
AMD64G_CC_OP_LOGICW, /* 18 DEP1 = result, DEP2 = 0, NDEP = unused */
AMD64G_CC_OP_LOGICL, /* 19 */
AMD64G_CC_OP_LOGICQ, /* 20 */
AMD64G_CC_OP_INCB, /* 21 */
AMD64G_CC_OP_INCW, /* 22 DEP1 = result, DEP2 = 0, NDEP = oldCarry (0 or 1) */
AMD64G_CC_OP_INCL, /* 23 */
AMD64G_CC_OP_INCQ, /* 24 */
AMD64G_CC_OP_DECB, /* 25 */
AMD64G_CC_OP_DECW, /* 26 DEP1 = result, DEP2 = 0, NDEP = oldCarry (0 or 1) */
AMD64G_CC_OP_DECL, /* 27 */
AMD64G_CC_OP_DECQ, /* 28 */
AMD64G_CC_OP_SHLB, /* 29 DEP1 = res, DEP2 = res', NDEP = unused */
AMD64G_CC_OP_SHLW, /* 30 where res' is like res but shifted one bit less */
AMD64G_CC_OP_SHLL, /* 31 */
AMD64G_CC_OP_SHLQ, /* 32 */
AMD64G_CC_OP_SHRB, /* 33 DEP1 = res, DEP2 = res', NDEP = unused */
AMD64G_CC_OP_SHRW, /* 34 where res' is like res but shifted one bit less */
AMD64G_CC_OP_SHRL, /* 35 */
AMD64G_CC_OP_SHRQ, /* 36 */
AMD64G_CC_OP_ROLB, /* 37 */
AMD64G_CC_OP_ROLW, /* 38 DEP1 = res, DEP2 = 0, NDEP = old flags */
AMD64G_CC_OP_ROLL, /* 39 */
AMD64G_CC_OP_ROLQ, /* 40 */
AMD64G_CC_OP_RORB, /* 41 */
AMD64G_CC_OP_RORW, /* 42 DEP1 = res, DEP2 = 0, NDEP = old flags */
AMD64G_CC_OP_RORL, /* 43 */
AMD64G_CC_OP_RORQ, /* 44 */
AMD64G_CC_OP_UMULB, /* 45 */
AMD64G_CC_OP_UMULW, /* 46 DEP1 = argL, DEP2 = argR, NDEP = unused */
AMD64G_CC_OP_UMULL, /* 47 */
AMD64G_CC_OP_UMULQ, /* 48 */
AMD64G_CC_OP_SMULB, /* 49 */
AMD64G_CC_OP_SMULW, /* 50 DEP1 = argL, DEP2 = argR, NDEP = unused */
AMD64G_CC_OP_SMULL, /* 51 */
AMD64G_CC_OP_SMULQ, /* 52 */
AMD64G_CC_OP_NUMBER
};
typedef
enum {
AMD64CondO = 0, /* overflow */
AMD64CondNO = 1, /* no overflow */
AMD64CondB = 2, /* below */
AMD64CondNB = 3, /* not below */
AMD64CondZ = 4, /* zero */
AMD64CondNZ = 5, /* not zero */
AMD64CondBE = 6, /* below or equal */
AMD64CondNBE = 7, /* not below or equal */
AMD64CondS = 8, /* negative */
AMD64CondNS = 9, /* not negative */
AMD64CondP = 10, /* parity even */
AMD64CondNP = 11, /* not parity even */
AMD64CondL = 12, /* jump less */
AMD64CondNL = 13, /* not less */
AMD64CondLE = 14, /* less or equal */
AMD64CondNLE = 15, /* not less or equal */
AMD64CondAlways = 16 /* HACK */
}
AMD64Condcode;
#endif /* ndef __VEX_GUEST_AMD64_DEFS_H */
/*---------------------------------------------------------------*/
/*--- end guest_amd64_defs.h ---*/
/*---------------------------------------------------------------*/