TCG code generator

git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@3943 c046a42c-6fe2-441c-8c8c-71466251a162
This commit is contained in:
bellard 2008-02-01 10:05:41 +00:00
parent 56abbcfff3
commit c896fe29d6
13 changed files with 7026 additions and 0 deletions

3
tcg/LICENSE Normal file
View File

@ -0,0 +1,3 @@
All the files in this directory and subdirectories are released under
a BSD like license (see header in each file). No other license is
accepted.

420
tcg/README Normal file
View File

@ -0,0 +1,420 @@
Tiny Code Generator - Fabrice Bellard.
1) Introduction
TCG (Tiny Code Generator) began as a generic backend for a C
compiler. It was simplified to be used in QEMU. It also has its roots
in the QOP code generator written by Paul Brook.
2) Definitions
The TCG "target" is the architecture for which we generate the
code. It is of course not the same as the "target" of QEMU which is
the emulated architecture. As TCG started as a generic C backend used
for cross compiling, it is assumed that the TCG target is different
from the host, although it is never the case for QEMU.
A TCG "function" corresponds to a QEMU Translated Block (TB).
A TCG "temporary" is a variable only live in a given
function. Temporaries are allocated explicitely in each function.
A TCG "global" is a variable which is live in all the functions. They
are defined before the functions defined. A TCG global can be a memory
location (e.g. a QEMU CPU register), a fixed host register (e.g. the
QEMU CPU state pointer) or a memory location which is stored in a
register outside QEMU TBs (not implemented yet).
A TCG "basic block" corresponds to a list of instructions terminated
by a branch instruction.
3) Intermediate representation
3.1) Introduction
TCG instructions operate on variables which are temporaries or
globals. TCG instructions and variables are strongly typed. Two types
are supported: 32 bit integers and 64 bit integers. Pointers are
defined as an alias to 32 bit or 64 bit integers depending on the TCG
target word size.
Each instruction has a fixed number of output variable operands, input
variable operands and always constant operands.
The notable exception is the call instruction which has a variable
number of outputs and inputs.
In the textual form, output operands come first, followed by input
operands, followed by constant operands. The output type is included
in the instruction name. Constants are prefixed with a '$'.
add_i32 t0, t1, t2 (t0 <- t1 + t2)
sub_i64 t2, t3, $4 (t2 <- t3 - 4)
3.2) Assumptions
* Basic blocks
- Basic blocks end after branches (e.g. brcond_i32 instruction),
goto_tb and exit_tb instructions.
- Basic blocks end before legacy dyngen operations.
- Basic blocks start after the end of a previous basic block, at a
set_label instruction or after a legacy dyngen operation.
After the end of a basic block, temporaries at destroyed and globals
are stored at their initial storage (register or memory place
depending on their declarations).
* Floating point types are not supported yet
* Pointers: depending on the TCG target, pointer size is 32 bit or 64
bit. The type TCG_TYPE_PTR is an alias to TCG_TYPE_I32 or
TCG_TYPE_I64.
* Helpers:
Using the tcg_gen_helper_x_y it is possible to call any function
taking i32, i64 or pointer types types. Before calling an helper, all
globals are stored at their canonical location and it is assumed that
the function can modify them. In the future, function modifiers will
be allowed to tell that the helper does not read or write some globals.
On some TCG targets (e.g. x86), several calling conventions are
supported.
* Branches:
Use the instruction 'br' to jump to a label. Use 'jmp' to jump to an
explicit address. Conditional branches can only jump to labels.
3.3) Code Optimizations
When generating instructions, you can count on at least the following
optimizations:
- Single instructions are simplified, e.g.
and_i32 t0, t0, $0xffffffff
is suppressed.
- A liveness analysis is done at the basic block level. The
information is used to suppress moves from a dead temporary to
another one. It is also used to remove instructions which compute
dead results. The later is especially useful for condition code
optimisation in QEMU.
In the following example:
add_i32 t0, t1, t2
add_i32 t0, t0, $1
mov_i32 t0, $1
only the last instruction is kept.
- A macro system is supported (may get closer to function inlining
some day). It is useful if the liveness analysis is likely to prove
that some results of a computation are indeed not useful. With the
macro system, the user can provide several alternative
implementations which are used depending on the used results. It is
especially useful for condition code optimisation in QEMU.
Here is an example:
macro_2 t0, t1, $1
mov_i32 t0, $0x1234
The macro identified by the ID "$1" normally returns the values t0
and t1. Suppose its implementation is:
macro_start
brcond_i32 t2, $0, $TCG_COND_EQ, $1
mov_i32 t0, $2
br $2
set_label $1
mov_i32 t0, $3
set_label $2
add_i32 t1, t3, t4
macro_end
If t0 is not used after the macro, the user can provide a simpler
implementation:
macro_start
add_i32 t1, t2, t4
macro_end
TCG automatically chooses the right implementation depending on
which macro outputs are used after it.
Note that if TCG did more expensive optimizations, macros would be
less useful. In the previous example a macro is useful because the
liveness analysis is done on each basic block separately. Hence TCG
cannot remove the code computing 't0' even if it is not used after
the first macro implementation.
3.4) Instruction Reference
********* Function call
* call <ret> <params> ptr
call function 'ptr' (pointer type)
<ret> optional 32 bit or 64 bit return value
<params> optional 32 bit or 64 bit parameters
********* Jumps/Labels
* jmp t0
Absolute jump to address t0 (pointer type).
* set_label $label
Define label 'label' at the current program point.
* br $label
Jump to label.
* brcond_i32/i64 cond, t0, t1, label
Conditional jump if t0 cond t1 is true. cond can be:
TCG_COND_EQ
TCG_COND_NE
TCG_COND_LT /* signed */
TCG_COND_GE /* signed */
TCG_COND_LE /* signed */
TCG_COND_GT /* signed */
TCG_COND_LTU /* unsigned */
TCG_COND_GEU /* unsigned */
TCG_COND_LEU /* unsigned */
TCG_COND_GTU /* unsigned */
********* Arithmetic
* add_i32/i64 t0, t1, t2
t0=t1+t2
* sub_i32/i64 t0, t1, t2
t0=t1-t2
* mul_i32/i64 t0, t1, t2
t0=t1*t2
* div_i32/i64 t0, t1, t2
t0=t1/t2 (signed). Undefined behavior if division by zero or overflow.
* divu_i32/i64 t0, t1, t2
t0=t1/t2 (unsigned). Undefined behavior if division by zero.
* rem_i32/i64 t0, t1, t2
t0=t1%t2 (signed). Undefined behavior if division by zero or overflow.
* remu_i32/i64 t0, t1, t2
t0=t1%t2 (unsigned). Undefined behavior if division by zero.
* and_i32/i64 t0, t1, t2
********* Logical
t0=t1&t2
* or_i32/i64 t0, t1, t2
t0=t1|t2
* xor_i32/i64 t0, t1, t2
t0=t1^t2
* shl_i32/i64 t0, t1, t2
********* Shifts
* shl_i32/i64 t0, t1, t2
t0=t1 << t2. Undefined behavior if t2 < 0 or t2 >= 32 (resp 64)
* shr_i32/i64 t0, t1, t2
t0=t1 >> t2 (unsigned). Undefined behavior if t2 < 0 or t2 >= 32 (resp 64)
* sar_i32/i64 t0, t1, t2
t0=t1 >> t2 (signed). Undefined behavior if t2 < 0 or t2 >= 32 (resp 64)
********* Misc
* mov_i32/i64 t0, t1
t0 = t1
Move t1 to t0 (both operands must have the same type).
* ext8s_i32/i64 t0, t1
ext16s_i32/i64 t0, t1
ext32s_i64 t0, t1
8, 16 or 32 bit sign extension (both operands must have the same type)
* bswap16_i32 t0, t1
16 bit byte swap on a 32 bit value. The two high order bytes must be set
to zero.
* bswap_i32 t0, t1
32 bit byte swap
* bswap_i64 t0, t1
64 bit byte swap
********* Type conversions
* ext_i32_i64 t0, t1
Convert t1 (32 bit) to t0 (64 bit) and does sign extension
* extu_i32_i64 t0, t1
Convert t1 (32 bit) to t0 (64 bit) and does zero extension
* trunc_i64_i32 t0, t1
Truncate t1 (64 bit) to t0 (32 bit)
********* Load/Store
* ld_i32/i64 t0, t1, offset
ld8s_i32/i64 t0, t1, offset
ld8u_i32/i64 t0, t1, offset
ld16s_i32/i64 t0, t1, offset
ld16u_i32/i64 t0, t1, offset
ld32s_i64 t0, t1, offset
ld32u_i64 t0, t1, offset
t0 = read(t1 + offset)
Load 8, 16, 32 or 64 bits with or without sign extension from host memory.
offset must be a constant.
* st_i32/i64 t0, t1, offset
st8_i32/i64 t0, t1, offset
st16_i32/i64 t0, t1, offset
st32_i64 t0, t1, offset
write(t0, t1 + offset)
Write 8, 16, 32 or 64 bits to host memory.
********* QEMU specific operations
* tb_exit t0
Exit the current TB and return the value t0 (word type).
* goto_tb index
Exit the current TB and jump to the TB index 'index' (constant) if the
current TB was linked to this TB. Otherwise execute the next
instructions.
* qemu_ld_i32/i64 t0, t1, flags
qemu_ld8u_i32/i64 t0, t1, flags
qemu_ld8s_i32/i64 t0, t1, flags
qemu_ld16u_i32/i64 t0, t1, flags
qemu_ld16s_i32/i64 t0, t1, flags
qemu_ld32u_i64 t0, t1, flags
qemu_ld32s_i64 t0, t1, flags
Load data at the QEMU CPU address t1 into t0. t1 has the QEMU CPU
address type. 'flags' contains the QEMU memory index (selects user or
kernel access) for example.
* qemu_st_i32/i64 t0, t1, flags
qemu_st8_i32/i64 t0, t1, flags
qemu_st16_i32/i64 t0, t1, flags
qemu_st32_i64 t0, t1, flags
Store the data t0 at the QEMU CPU Address t1. t1 has the QEMU CPU
address type. 'flags' contains the QEMU memory index (selects user or
kernel access) for example.
Note 1: Some shortcuts are defined when the last operand is known to be
a constant (e.g. addi for add, movi for mov).
Note 2: When using TCG, the opcodes must never be generated directly
as some of them may not be available as "real" opcodes. Always use the
function tcg_gen_xxx(args).
4) Backend
tcg-target.h contains the target specific definitions. tcg-target.c
contains the target specific code.
4.1) Assumptions
The target word size (TCG_TARGET_REG_BITS) is expected to be 32 bit or
64 bit. It is expected that the pointer has the same size as the word.
On a 32 bit target, all 64 bit operations are converted to 32 bits. A
few specific operations must be implemented to allow it (see add2_i32,
sub2_i32, brcond2_i32).
Floating point operations are not supported in this version. A
previous incarnation of the code generator had full support of them,
but it is better to concentrate on integer operations first.
On a 64 bit target, no assumption is made in TCG about the storage of
the 32 bit values in 64 bit registers.
4.2) Constraints
GCC like constraints are used to define the constraints of every
instruction. Memory constraints are not supported in this
version. Aliases are specified in the input operands as for GCC.
A target can define specific register or constant constraints. If an
operation uses a constant input constraint which does not allow all
constants, it must also accept registers in order to have a fallback.
The movi_i32 and movi_i64 operations must accept any constants.
The mov_i32 and mov_i64 operations must accept any registers of the
same type.
The ld/st instructions must accept signed 32 bit constant offsets. It
can be implemented by reserving a specific register to compute the
address if the offset is too big.
The ld/st instructions must accept any destination (ld) or source (st)
register.
4.3) Function call assumptions
- The only supported types for parameters and return value are: 32 and
64 bit integers and pointer.
- The stack grows downwards.
- The first N parameters are passed in registers.
- The next parameters are passed on the stack by storing them as words.
- Some registers are clobbered during the call.
- The function can return 0 or 1 value in registers. On a 32 bit
target, functions must be able to return 2 values in registers for
64 bit return type.
5) Migration from dyngen to TCG
TCG is backward compatible with QEMU "dyngen" operations. It means
that TCG instructions can be freely mixed with dyngen operations. It
is expected that QEMU targets will be progressively fully converted to
TCG. Once a target is fully converted to dyngen, it will be possible
to apply more optimizations because more registers will be free for
the generated code.
The exception model is the same as the dyngen one.

32
tcg/TODO Normal file
View File

@ -0,0 +1,32 @@
- test macro system
- test conditional jumps
- test mul, div, ext8s, ext16s, bswap
- generate a global TB prologue and epilogue to save/restore registers
to/from the CPU state and to reserve a stack frame to optimize
helper calls. Modify cpu-exec.c so that it does not use global
register variables (except maybe for 'env').
- fully convert the x86 target. The minimal amount of work includes:
- add cc_src, cc_dst and cc_op as globals
- disable its eflags optimization (the liveness analysis should
suffice)
- move complicated operations to helpers (in particular FPU, SSE, MMX).
- optimize the x86 target:
- move some or all the registers as globals
- use the TB prologue and epilogue to have QEMU target registers in
pre assigned host registers.
Ideas:
- Move the slow part of the qemu_ld/st ops after the end of the TB.
- Experiment: change instruction storage to simplify macro handling
and to handle dynamic allocation and see if the translation speed is
OK.
- change exception syntax to get closer to QOP system (exception
parameters given with a specific instruction).

1163
tcg/i386/tcg-target.c Normal file

File diff suppressed because it is too large Load Diff

54
tcg/i386/tcg-target.h Normal file
View File

@ -0,0 +1,54 @@
/*
* Tiny Code Generator for QEMU
*
* Copyright (c) 2008 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#define TCG_TARGET_I386 1
#define TCG_TARGET_REG_BITS 32
//#define TCG_TARGET_WORDS_BIGENDIAN
#define TCG_TARGET_NB_REGS 8
enum {
TCG_REG_EAX = 0,
TCG_REG_ECX,
TCG_REG_EDX,
TCG_REG_EBX,
TCG_REG_ESP,
TCG_REG_EBP,
TCG_REG_ESI,
TCG_REG_EDI,
};
/* used for function call generation */
#define TCG_REG_CALL_STACK TCG_REG_ESP
#define TCG_TARGET_STACK_ALIGN 16
/* Note: must be synced with dyngen-exec.h */
#define TCG_AREG0 TCG_REG_EBP
#define TCG_AREG1 TCG_REG_EBX
#define TCG_AREG2 TCG_REG_ESI
#define TCG_AREG3 TCG_REG_EDI
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
}

482
tcg/tcg-dyngen.c Normal file
View File

@ -0,0 +1,482 @@
/*
* Tiny Code Generator for QEMU
*
* Copyright (c) 2008 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include <assert.h>
#include <stdarg.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <inttypes.h>
#include "config.h"
#include "osdep.h"
#include "tcg.h"
int __op_param1, __op_param2, __op_param3;
#if defined(__sparc__) || defined(__arm__)
void __op_gen_label1(){}
void __op_gen_label2(){}
void __op_gen_label3(){}
#else
int __op_gen_label1, __op_gen_label2, __op_gen_label3;
#endif
int __op_jmp0, __op_jmp1, __op_jmp2, __op_jmp3;
#if 0
#if defined(__s390__)
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
}
#elif defined(__ia64__)
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
while (start < stop) {
asm volatile ("fc %0" :: "r"(start));
start += 32;
}
asm volatile (";;sync.i;;srlz.i;;");
}
#elif defined(__powerpc__)
#define MIN_CACHE_LINE_SIZE 8 /* conservative value */
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
unsigned long p;
start &= ~(MIN_CACHE_LINE_SIZE - 1);
stop = (stop + MIN_CACHE_LINE_SIZE - 1) & ~(MIN_CACHE_LINE_SIZE - 1);
for (p = start; p < stop; p += MIN_CACHE_LINE_SIZE) {
asm volatile ("dcbst 0,%0" : : "r"(p) : "memory");
}
asm volatile ("sync" : : : "memory");
for (p = start; p < stop; p += MIN_CACHE_LINE_SIZE) {
asm volatile ("icbi 0,%0" : : "r"(p) : "memory");
}
asm volatile ("sync" : : : "memory");
asm volatile ("isync" : : : "memory");
}
#elif defined(__alpha__)
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
asm ("imb");
}
#elif defined(__sparc__)
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
unsigned long p;
p = start & ~(8UL - 1UL);
stop = (stop + (8UL - 1UL)) & ~(8UL - 1UL);
for (; p < stop; p += 8)
__asm__ __volatile__("flush\t%0" : : "r" (p));
}
#elif defined(__arm__)
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
register unsigned long _beg __asm ("a1") = start;
register unsigned long _end __asm ("a2") = stop;
register unsigned long _flg __asm ("a3") = 0;
__asm __volatile__ ("swi 0x9f0002" : : "r" (_beg), "r" (_end), "r" (_flg));
}
#elif defined(__mc68000)
# include <asm/cachectl.h>
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
cacheflush(start,FLUSH_SCOPE_LINE,FLUSH_CACHE_BOTH,stop-start+16);
}
#elif defined(__mips__)
#include <sys/cachectl.h>
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
_flush_cache ((void *)start, stop - start, BCACHE);
}
#else
#error unsupported CPU
#endif
#ifdef __alpha__
register int gp asm("$29");
static inline void immediate_ldah(void *p, int val) {
uint32_t *dest = p;
long high = ((val >> 16) + ((val >> 15) & 1)) & 0xffff;
*dest &= ~0xffff;
*dest |= high;
*dest |= 31 << 16;
}
static inline void immediate_lda(void *dest, int val) {
*(uint16_t *) dest = val;
}
void fix_bsr(void *p, int offset) {
uint32_t *dest = p;
*dest &= ~((1 << 21) - 1);
*dest |= (offset >> 2) & ((1 << 21) - 1);
}
#endif /* __alpha__ */
#ifdef __arm__
#define ARM_LDR_TABLE_SIZE 1024
typedef struct LDREntry {
uint8_t *ptr;
uint32_t *data_ptr;
unsigned type:2;
} LDREntry;
static LDREntry arm_ldr_table[1024];
static uint32_t arm_data_table[ARM_LDR_TABLE_SIZE];
extern char exec_loop;
static inline void arm_reloc_pc24(uint32_t *ptr, uint32_t insn, int val)
{
*ptr = (insn & ~0xffffff) | ((insn + ((val - (int)ptr) >> 2)) & 0xffffff);
}
static uint8_t *arm_flush_ldr(uint8_t *gen_code_ptr,
LDREntry *ldr_start, LDREntry *ldr_end,
uint32_t *data_start, uint32_t *data_end,
int gen_jmp)
{
LDREntry *le;
uint32_t *ptr;
int offset, data_size, target;
uint8_t *data_ptr;
uint32_t insn;
uint32_t mask;
data_size = (data_end - data_start) << 2;
if (gen_jmp) {
/* generate branch to skip the data */
if (data_size == 0)
return gen_code_ptr;
target = (long)gen_code_ptr + data_size + 4;
arm_reloc_pc24((uint32_t *)gen_code_ptr, 0xeafffffe, target);
gen_code_ptr += 4;
}
/* copy the data */
data_ptr = gen_code_ptr;
memcpy(gen_code_ptr, data_start, data_size);
gen_code_ptr += data_size;
/* patch the ldr to point to the data */
for(le = ldr_start; le < ldr_end; le++) {
ptr = (uint32_t *)le->ptr;
offset = ((unsigned long)(le->data_ptr) - (unsigned long)data_start) +
(unsigned long)data_ptr -
(unsigned long)ptr - 8;
if (offset < 0) {
fprintf(stderr, "Negative constant pool offset\n");
tcg_abort();
}
switch (le->type) {
case 0: /* ldr */
mask = ~0x00800fff;
if (offset >= 4096) {
fprintf(stderr, "Bad ldr offset\n");
tcg_abort();
}
break;
case 1: /* ldc */
mask = ~0x008000ff;
if (offset >= 1024 ) {
fprintf(stderr, "Bad ldc offset\n");
tcg_abort();
}
break;
case 2: /* add */
mask = ~0xfff;
if (offset >= 1024 ) {
fprintf(stderr, "Bad add offset\n");
tcg_abort();
}
break;
default:
fprintf(stderr, "Bad pc relative fixup\n");
tcg_abort();
}
insn = *ptr & mask;
switch (le->type) {
case 0: /* ldr */
insn |= offset | 0x00800000;
break;
case 1: /* ldc */
insn |= (offset >> 2) | 0x00800000;
break;
case 2: /* add */
insn |= (offset >> 2) | 0xf00;
break;
}
*ptr = insn;
}
return gen_code_ptr;
}
#endif /* __arm__ */
#ifdef __ia64
/* Patch instruction with "val" where "mask" has 1 bits. */
static inline void ia64_patch (uint64_t insn_addr, uint64_t mask, uint64_t val)
{
uint64_t m0, m1, v0, v1, b0, b1, *b = (uint64_t *) (insn_addr & -16);
# define insn_mask ((1UL << 41) - 1)
unsigned long shift;
b0 = b[0]; b1 = b[1];
shift = 5 + 41 * (insn_addr % 16); /* 5 template, 3 x 41-bit insns */
if (shift >= 64) {
m1 = mask << (shift - 64);
v1 = val << (shift - 64);
} else {
m0 = mask << shift; m1 = mask >> (64 - shift);
v0 = val << shift; v1 = val >> (64 - shift);
b[0] = (b0 & ~m0) | (v0 & m0);
}
b[1] = (b1 & ~m1) | (v1 & m1);
}
static inline void ia64_patch_imm60 (uint64_t insn_addr, uint64_t val)
{
ia64_patch(insn_addr,
0x011ffffe000UL,
( ((val & 0x0800000000000000UL) >> 23) /* bit 59 -> 36 */
| ((val & 0x00000000000fffffUL) << 13) /* bit 0 -> 13 */));
ia64_patch(insn_addr - 1, 0x1fffffffffcUL, val >> 18);
}
static inline void ia64_imm64 (void *insn, uint64_t val)
{
/* Ignore the slot number of the relocation; GCC and Intel
toolchains differed for some time on whether IMM64 relocs are
against slot 1 (Intel) or slot 2 (GCC). */
uint64_t insn_addr = (uint64_t) insn & ~3UL;
ia64_patch(insn_addr + 2,
0x01fffefe000UL,
( ((val & 0x8000000000000000UL) >> 27) /* bit 63 -> 36 */
| ((val & 0x0000000000200000UL) << 0) /* bit 21 -> 21 */
| ((val & 0x00000000001f0000UL) << 6) /* bit 16 -> 22 */
| ((val & 0x000000000000ff80UL) << 20) /* bit 7 -> 27 */
| ((val & 0x000000000000007fUL) << 13) /* bit 0 -> 13 */)
);
ia64_patch(insn_addr + 1, 0x1ffffffffffUL, val >> 22);
}
static inline void ia64_imm60b (void *insn, uint64_t val)
{
/* Ignore the slot number of the relocation; GCC and Intel
toolchains differed for some time on whether IMM64 relocs are
against slot 1 (Intel) or slot 2 (GCC). */
uint64_t insn_addr = (uint64_t) insn & ~3UL;
if (val + ((uint64_t) 1 << 59) >= (1UL << 60))
fprintf(stderr, "%s: value %ld out of IMM60 range\n",
__FUNCTION__, (int64_t) val);
ia64_patch_imm60(insn_addr + 2, val);
}
static inline void ia64_imm22 (void *insn, uint64_t val)
{
if (val + (1 << 21) >= (1 << 22))
fprintf(stderr, "%s: value %li out of IMM22 range\n",
__FUNCTION__, (int64_t)val);
ia64_patch((uint64_t) insn, 0x01fffcfe000UL,
( ((val & 0x200000UL) << 15) /* bit 21 -> 36 */
| ((val & 0x1f0000UL) << 6) /* bit 16 -> 22 */
| ((val & 0x00ff80UL) << 20) /* bit 7 -> 27 */
| ((val & 0x00007fUL) << 13) /* bit 0 -> 13 */));
}
/* Like ia64_imm22(), but also clear bits 20-21. For addl, this has
the effect of turning "addl rX=imm22,rY" into "addl
rX=imm22,r0". */
static inline void ia64_imm22_r0 (void *insn, uint64_t val)
{
if (val + (1 << 21) >= (1 << 22))
fprintf(stderr, "%s: value %li out of IMM22 range\n",
__FUNCTION__, (int64_t)val);
ia64_patch((uint64_t) insn, 0x01fffcfe000UL | (0x3UL << 20),
( ((val & 0x200000UL) << 15) /* bit 21 -> 36 */
| ((val & 0x1f0000UL) << 6) /* bit 16 -> 22 */
| ((val & 0x00ff80UL) << 20) /* bit 7 -> 27 */
| ((val & 0x00007fUL) << 13) /* bit 0 -> 13 */));
}
static inline void ia64_imm21b (void *insn, uint64_t val)
{
if (val + (1 << 20) >= (1 << 21))
fprintf(stderr, "%s: value %li out of IMM21b range\n",
__FUNCTION__, (int64_t)val);
ia64_patch((uint64_t) insn, 0x11ffffe000UL,
( ((val & 0x100000UL) << 16) /* bit 20 -> 36 */
| ((val & 0x0fffffUL) << 13) /* bit 0 -> 13 */));
}
static inline void ia64_nop_b (void *insn)
{
ia64_patch((uint64_t) insn, (1UL << 41) - 1, 2UL << 37);
}
static inline void ia64_ldxmov(void *insn, uint64_t val)
{
if (val + (1 << 21) < (1 << 22))
ia64_patch((uint64_t) insn, 0x1fff80fe000UL, 8UL << 37);
}
static inline int ia64_patch_ltoff(void *insn, uint64_t val,
int relaxable)
{
if (relaxable && (val + (1 << 21) < (1 << 22))) {
ia64_imm22_r0(insn, val);
return 0;
}
return 1;
}
struct ia64_fixup {
struct ia64_fixup *next;
void *addr; /* address that needs to be patched */
long value;
};
#define IA64_PLT(insn, plt_index) \
do { \
struct ia64_fixup *fixup = alloca(sizeof(*fixup)); \
fixup->next = plt_fixes; \
plt_fixes = fixup; \
fixup->addr = (insn); \
fixup->value = (plt_index); \
plt_offset[(plt_index)] = 1; \
} while (0)
#define IA64_LTOFF(insn, val, relaxable) \
do { \
if (ia64_patch_ltoff(insn, val, relaxable)) { \
struct ia64_fixup *fixup = alloca(sizeof(*fixup)); \
fixup->next = ltoff_fixes; \
ltoff_fixes = fixup; \
fixup->addr = (insn); \
fixup->value = (val); \
} \
} while (0)
static inline void ia64_apply_fixes (uint8_t **gen_code_pp,
struct ia64_fixup *ltoff_fixes,
uint64_t gp,
struct ia64_fixup *plt_fixes,
int num_plts,
unsigned long *plt_target,
unsigned int *plt_offset)
{
static const uint8_t plt_bundle[] = {
0x04, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, /* nop 0; movl r1=GP */
0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x60,
0x05, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, /* nop 0; brl IP */
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xc0
};
uint8_t *gen_code_ptr = *gen_code_pp, *plt_start, *got_start;
uint64_t *vp;
struct ia64_fixup *fixup;
unsigned int offset = 0;
struct fdesc {
long ip;
long gp;
} *fdesc;
int i;
if (plt_fixes) {
plt_start = gen_code_ptr;
for (i = 0; i < num_plts; ++i) {
if (plt_offset[i]) {
plt_offset[i] = offset;
offset += sizeof(plt_bundle);
fdesc = (struct fdesc *) plt_target[i];
memcpy(gen_code_ptr, plt_bundle, sizeof(plt_bundle));
ia64_imm64 (gen_code_ptr + 0x02, fdesc->gp);
ia64_imm60b(gen_code_ptr + 0x12,
(fdesc->ip - (long) (gen_code_ptr + 0x10)) >> 4);
gen_code_ptr += sizeof(plt_bundle);
}
}
for (fixup = plt_fixes; fixup; fixup = fixup->next)
ia64_imm21b(fixup->addr,
((long) plt_start + plt_offset[fixup->value]
- ((long) fixup->addr & ~0xf)) >> 4);
}
got_start = gen_code_ptr;
/* First, create the GOT: */
for (fixup = ltoff_fixes; fixup; fixup = fixup->next) {
/* first check if we already have this value in the GOT: */
for (vp = (uint64_t *) got_start; vp < (uint64_t *) gen_code_ptr; ++vp)
if (*vp == fixup->value)
break;
if (vp == (uint64_t *) gen_code_ptr) {
/* Nope, we need to put the value in the GOT: */
*vp = fixup->value;
gen_code_ptr += 8;
}
ia64_imm22(fixup->addr, (long) vp - gp);
}
/* Keep code ptr aligned. */
if ((long) gen_code_ptr & 15)
gen_code_ptr += 8;
*gen_code_pp = gen_code_ptr;
}
#endif
#endif
const TCGArg *dyngen_op(TCGContext *s, int opc, const TCGArg *opparam_ptr)
{
uint8_t *gen_code_ptr;
gen_code_ptr = s->code_ptr;
switch(opc) {
/* op.h is dynamically generated by dyngen.c from op.c */
#include "op.h"
default:
tcg_abort();
}
s->code_ptr = gen_code_ptr;
return opparam_ptr;
}

1175
tcg/tcg-op.h Normal file

File diff suppressed because it is too large Load Diff

228
tcg/tcg-opc.h Normal file
View File

@ -0,0 +1,228 @@
/*
* Tiny Code Generator for QEMU
*
* Copyright (c) 2008 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "dyngen-opc.h"
#ifndef DEF2
#define DEF2(name, oargs, iargs, cargs, flags) DEF(name, oargs + iargs + cargs, 0)
#endif
/* predefined ops */
DEF2(end, 0, 0, 0, 0) /* must be kept first */
DEF2(nop, 0, 0, 0, 0)
DEF2(nop1, 0, 0, 1, 0)
DEF2(nop2, 0, 0, 2, 0)
DEF2(nop3, 0, 0, 3, 0)
DEF2(nopn, 0, 0, 1, 0) /* variable number of parameters */
/* macro handling */
DEF2(macro_2, 2, 0, 1, 0)
DEF2(macro_start, 0, 0, 2, 0)
DEF2(macro_end, 0, 0, 2, 0)
DEF2(macro_goto, 0, 0, 3, 0)
DEF2(set_label, 0, 0, 1, 0)
DEF2(call, 0, 1, 2, 0) /* variable number of parameters */
DEF2(jmp, 0, 1, 0, TCG_OPF_BB_END)
DEF2(br, 0, 0, 1, TCG_OPF_BB_END)
DEF2(mov_i32, 1, 1, 0, 0)
DEF2(movi_i32, 1, 0, 1, 0)
/* load/store */
DEF2(ld8u_i32, 1, 1, 1, 0)
DEF2(ld8s_i32, 1, 1, 1, 0)
DEF2(ld16u_i32, 1, 1, 1, 0)
DEF2(ld16s_i32, 1, 1, 1, 0)
DEF2(ld_i32, 1, 1, 1, 0)
DEF2(st8_i32, 0, 2, 1, 0)
DEF2(st16_i32, 0, 2, 1, 0)
DEF2(st_i32, 0, 2, 1, 0)
/* arith */
DEF2(add_i32, 1, 2, 0, 0)
DEF2(sub_i32, 1, 2, 0, 0)
DEF2(mul_i32, 1, 2, 0, 0)
#ifdef TCG_TARGET_HAS_div_i32
DEF2(div_i32, 1, 2, 0, 0)
DEF2(divu_i32, 1, 2, 0, 0)
DEF2(rem_i32, 1, 2, 0, 0)
DEF2(remu_i32, 1, 2, 0, 0)
#else
DEF2(div2_i32, 2, 3, 0, 0)
DEF2(divu2_i32, 2, 3, 0, 0)
#endif
DEF2(and_i32, 1, 2, 0, 0)
DEF2(or_i32, 1, 2, 0, 0)
DEF2(xor_i32, 1, 2, 0, 0)
/* shifts */
DEF2(shl_i32, 1, 2, 0, 0)
DEF2(shr_i32, 1, 2, 0, 0)
DEF2(sar_i32, 1, 2, 0, 0)
DEF2(brcond_i32, 0, 2, 2, TCG_OPF_BB_END)
#if TCG_TARGET_REG_BITS == 32
DEF2(add2_i32, 2, 4, 0, 0)
DEF2(sub2_i32, 2, 4, 0, 0)
DEF2(brcond2_i32, 0, 4, 2, TCG_OPF_BB_END)
DEF2(mulu2_i32, 2, 2, 0, 0)
#endif
#ifdef TCG_TARGET_HAS_ext8s_i32
DEF2(ext8s_i32, 1, 1, 0, 0)
#endif
#ifdef TCG_TARGET_HAS_ext16s_i32
DEF2(ext16s_i32, 1, 1, 0, 0)
#endif
#ifdef TCG_TARGET_HAS_bswap_i32
DEF2(bswap_i32, 1, 1, 0, 0)
#endif
#if TCG_TARGET_REG_BITS == 64
DEF2(mov_i64, 1, 1, 0, 0)
DEF2(movi_i64, 1, 0, 1, 0)
/* load/store */
DEF2(ld8u_i64, 1, 1, 1, 0)
DEF2(ld8s_i64, 1, 1, 1, 0)
DEF2(ld16u_i64, 1, 1, 1, 0)
DEF2(ld16s_i64, 1, 1, 1, 0)
DEF2(ld32u_i64, 1, 1, 1, 0)
DEF2(ld32s_i64, 1, 1, 1, 0)
DEF2(ld_i64, 1, 1, 1, 0)
DEF2(st8_i64, 0, 2, 1, 0)
DEF2(st16_i64, 0, 2, 1, 0)
DEF2(st32_i64, 0, 2, 1, 0)
DEF2(st_i64, 0, 2, 1, 0)
/* arith */
DEF2(add_i64, 1, 2, 0, 0)
DEF2(sub_i64, 1, 2, 0, 0)
DEF2(mul_i64, 1, 2, 0, 0)
#ifdef TCG_TARGET_HAS_div_i64
DEF2(div_i64, 1, 2, 0, 0)
DEF2(divu_i64, 1, 2, 0, 0)
DEF2(rem_i64, 1, 2, 0, 0)
DEF2(remu_i64, 1, 2, 0, 0)
#else
DEF2(div2_i64, 2, 3, 0, 0)
DEF2(divu2_i64, 2, 3, 0, 0)
#endif
DEF2(and_i64, 1, 2, 0, 0)
DEF2(or_i64, 1, 2, 0, 0)
DEF2(xor_i64, 1, 2, 0, 0)
/* shifts */
DEF2(shl_i64, 1, 2, 0, 0)
DEF2(shr_i64, 1, 2, 0, 0)
DEF2(sar_i64, 1, 2, 0, 0)
DEF2(brcond_i64, 0, 2, 2, TCG_OPF_BB_END)
#ifdef TCG_TARGET_HAS_ext8s_i64
DEF2(ext8s_i64, 1, 1, 0, 0)
#endif
#ifdef TCG_TARGET_HAS_ext16s_i64
DEF2(ext16s_i64, 1, 1, 0, 0)
#endif
#ifdef TCG_TARGET_HAS_ext32s_i64
DEF2(ext32s_i64, 1, 1, 0, 0)
#endif
#ifdef TCG_TARGET_HAS_bswap_i64
DEF2(bswap_i64, 1, 1, 0, 0)
#endif
#endif
/* QEMU specific */
DEF2(exit_tb, 0, 0, 1, TCG_OPF_BB_END)
DEF2(goto_tb, 0, 0, 1, TCG_OPF_BB_END)
/* Note: even if TARGET_LONG_BITS is not defined, the INDEX_op
constants must be defined */
#if TCG_TARGET_REG_BITS == 32
#if TARGET_LONG_BITS == 32
DEF2(qemu_ld8u, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_ld8u, 1, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_ld8s, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_ld8s, 1, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_ld16u, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_ld16u, 1, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_ld16s, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_ld16s, 1, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_ld32u, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_ld32u, 1, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_ld32s, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_ld32s, 1, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_ld64, 2, 1, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_ld64, 2, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_st8, 0, 2, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_st8, 0, 3, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_st16, 0, 2, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_st16, 0, 3, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_st32, 0, 2, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_st32, 0, 3, 1, TCG_OPF_CALL_CLOBBER)
#endif
#if TARGET_LONG_BITS == 32
DEF2(qemu_st64, 0, 3, 1, TCG_OPF_CALL_CLOBBER)
#else
DEF2(qemu_st64, 0, 4, 1, TCG_OPF_CALL_CLOBBER)
#endif
#else /* TCG_TARGET_REG_BITS == 32 */
DEF2(qemu_ld8u, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_ld8s, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_ld16u, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_ld16s, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_ld32u, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_ld32s, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_ld64, 1, 1, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_st8, 0, 2, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_st16, 0, 2, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_st32, 0, 2, 1, TCG_OPF_CALL_CLOBBER)
DEF2(qemu_st64, 0, 2, 1, TCG_OPF_CALL_CLOBBER)
#endif /* TCG_TARGET_REG_BITS != 32 */
#undef DEF2

68
tcg/tcg-runtime.c Normal file
View File

@ -0,0 +1,68 @@
/*
* Tiny Code Generator for QEMU
*
* Copyright (c) 2008 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include <stdarg.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <inttypes.h>
#include "config.h"
#include "osdep.h"
#include "tcg.h"
int64_t tcg_helper_shl_i64(int64_t arg1, int64_t arg2)
{
return arg1 << arg2;
}
int64_t tcg_helper_shr_i64(int64_t arg1, int64_t arg2)
{
return (uint64_t)arg1 >> arg2;
}
int64_t tcg_helper_sar_i64(int64_t arg1, int64_t arg2)
{
return arg1 >> arg2;
}
int64_t tcg_helper_div_i64(int64_t arg1, int64_t arg2)
{
return arg1 / arg2;
}
int64_t tcg_helper_rem_i64(int64_t arg1, int64_t arg2)
{
return arg1 / arg2;
}
uint64_t tcg_helper_divu_i64(uint64_t arg1, uint64_t arg2)
{
return arg1 / arg2;
}
uint64_t tcg_helper_remu_i64(uint64_t arg1, uint64_t arg2)
{
return arg1 / arg2;
}

1781
tcg/tcg.c Normal file

File diff suppressed because it is too large Load Diff

324
tcg/tcg.h Normal file
View File

@ -0,0 +1,324 @@
/*
* Tiny Code Generator for QEMU
*
* Copyright (c) 2008 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "tcg-target.h"
#if TCG_TARGET_REG_BITS == 32
typedef int32_t tcg_target_long;
typedef uint32_t tcg_target_ulong;
#define TCG_PRIlx PRIx32
#define TCG_PRIld PRId32
#elif TCG_TARGET_REG_BITS == 64
typedef int64_t tcg_target_long;
typedef uint64_t tcg_target_ulong;
#define TCG_PRIlx PRIx64
#define TCG_PRIld PRId64
#else
#error unsupported
#endif
#if TCG_TARGET_NB_REGS <= 32
typedef uint32_t TCGRegSet;
#elif TCG_TARGET_NB_REGS <= 64
typedef uint64_t TCGRegSet;
#else
#error unsupported
#endif
enum {
#define DEF(s, n, copy_size) INDEX_op_ ## s,
#include "tcg-opc.h"
#undef DEF
NB_OPS,
};
#define tcg_regset_clear(d) (d) = 0
#define tcg_regset_set(d, s) (d) = (s)
#define tcg_regset_set32(d, reg, val32) (d) |= (val32) << (reg)
#define tcg_regset_set_reg(d, r) (d) |= 1 << (r)
#define tcg_regset_reset_reg(d, r) (d) &= ~(1 << (r))
#define tcg_regset_test_reg(d, r) (((d) >> (r)) & 1)
#define tcg_regset_or(d, a, b) (d) = (a) | (b)
#define tcg_regset_and(d, a, b) (d) = (a) & (b)
#define tcg_regset_andnot(d, a, b) (d) = (a) & ~(b)
#define tcg_regset_not(d, a) (d) = ~(a)
typedef struct TCGRelocation {
struct TCGRelocation *next;
int type;
uint8_t *ptr;
tcg_target_long addend;
} TCGRelocation;
typedef struct TCGLabel {
int has_value;
union {
tcg_target_ulong value;
TCGRelocation *first_reloc;
} u;
} TCGLabel;
typedef struct TCGPool {
struct TCGPool *next;
int size;
uint8_t data[0];
} TCGPool;
#define TCG_POOL_CHUNK_SIZE 32768
#define TCG_MAX_LABELS 512
#define TCG_MAX_TEMPS 256
typedef int TCGType;
#define TCG_TYPE_I32 0
#define TCG_TYPE_I64 1
#if TCG_TARGET_REG_BITS == 32
#define TCG_TYPE_PTR TCG_TYPE_I32
#else
#define TCG_TYPE_PTR TCG_TYPE_I64
#endif
typedef tcg_target_ulong TCGArg;
/* call flags */
#define TCG_CALL_TYPE_MASK 0x000f
#define TCG_CALL_TYPE_STD 0x0000 /* standard C call */
#define TCG_CALL_TYPE_REGPARM_1 0x0001 /* i386 style regparm call (1 reg) */
#define TCG_CALL_TYPE_REGPARM_2 0x0002 /* i386 style regparm call (2 regs) */
#define TCG_CALL_TYPE_REGPARM 0x0003 /* i386 style regparm call (3 regs) */
typedef enum {
TCG_COND_EQ,
TCG_COND_NE,
TCG_COND_LT,
TCG_COND_GE,
TCG_COND_LE,
TCG_COND_GT,
/* unsigned */
TCG_COND_LTU,
TCG_COND_GEU,
TCG_COND_LEU,
TCG_COND_GTU,
} TCGCond;
#define TEMP_VAL_DEAD 0
#define TEMP_VAL_REG 1
#define TEMP_VAL_MEM 2
#define TEMP_VAL_CONST 3
/* XXX: optimize memory layout */
typedef struct TCGTemp {
TCGType base_type;
TCGType type;
int val_type;
int reg;
tcg_target_long val;
int mem_reg;
tcg_target_long mem_offset;
unsigned int fixed_reg:1;
unsigned int mem_coherent:1;
unsigned int mem_allocated:1;
const char *name;
} TCGTemp;
typedef struct TCGHelperInfo {
void *func;
const char *name;
} TCGHelperInfo;
typedef struct TCGContext TCGContext;
typedef void TCGMacroFunc(TCGContext *s, int macro_id, const int *dead_args);
struct TCGContext {
uint8_t *pool_cur, *pool_end;
TCGPool *pool_first, *pool_current;
TCGLabel *labels;
int nb_labels;
TCGTemp *temps; /* globals first, temps after */
int nb_globals;
int nb_temps;
/* constant indexes (end of temp array) */
int const_start;
int const_end;
/* goto_tb support */
uint8_t *code_buf;
unsigned long *tb_next;
uint16_t *tb_next_offset;
uint16_t *tb_jmp_offset; /* != NULL if USE_DIRECT_JUMP */
uint16_t *op_dead_iargs; /* for each operation, each bit tells if the
corresponding input argument is dead */
/* tells in which temporary a given register is. It does not take
into account fixed registers */
int reg_to_temp[TCG_TARGET_NB_REGS];
TCGRegSet reserved_regs;
tcg_target_long current_frame_offset;
tcg_target_long frame_start;
tcg_target_long frame_end;
int frame_reg;
uint8_t *code_ptr;
TCGTemp static_temps[TCG_MAX_TEMPS];
TCGMacroFunc *macro_func;
TCGHelperInfo *helpers;
int nb_helpers;
int allocated_helpers;
};
extern TCGContext tcg_ctx;
extern uint16_t *gen_opc_ptr;
extern TCGArg *gen_opparam_ptr;
extern uint16_t gen_opc_buf[];
extern TCGArg gen_opparam_buf[];
/* pool based memory allocation */
void *tcg_malloc_internal(TCGContext *s, int size);
void tcg_pool_reset(TCGContext *s);
void tcg_pool_delete(TCGContext *s);
static inline void *tcg_malloc(int size)
{
TCGContext *s = &tcg_ctx;
uint8_t *ptr, *ptr_end;
size = (size + sizeof(long) - 1) & ~(sizeof(long) - 1);
ptr = s->pool_cur;
ptr_end = ptr + size;
if (unlikely(ptr_end > s->pool_end)) {
return tcg_malloc_internal(&tcg_ctx, size);
} else {
s->pool_cur = ptr_end;
return ptr;
}
}
void tcg_context_init(TCGContext *s);
void tcg_func_start(TCGContext *s);
int dyngen_code(TCGContext *s, uint8_t *gen_code_buf);
int dyngen_code_search_pc(TCGContext *s, uint8_t *gen_code_buf,
const uint8_t *searched_pc);
void tcg_set_frame(TCGContext *s, int reg,
tcg_target_long start, tcg_target_long size);
void tcg_set_macro_func(TCGContext *s, TCGMacroFunc *func);
int tcg_global_reg_new(TCGType type, int reg, const char *name);
int tcg_global_mem_new(TCGType type, int reg, tcg_target_long offset,
const char *name);
int tcg_temp_new(TCGType type);
char *tcg_get_arg_str(TCGContext *s, char *buf, int buf_size, TCGArg arg);
#define TCG_CT_ALIAS 0x80
#define TCG_CT_IALIAS 0x40
#define TCG_CT_REG 0x01
#define TCG_CT_CONST 0x02 /* any constant of register size */
typedef struct TCGArgConstraint {
uint32_t ct;
union {
TCGRegSet regs;
} u;
} TCGArgConstraint;
#define TCG_MAX_OP_ARGS 16
#define TCG_OPF_BB_END 0x01 /* instruction defines the end of a basic
block */
#define TCG_OPF_CALL_CLOBBER 0x02 /* instruction clobbers call registers */
typedef struct TCGOpDef {
const char *name;
uint8_t nb_oargs, nb_iargs, nb_cargs, nb_args;
uint8_t flags;
uint16_t copy_size;
TCGArgConstraint *args_ct;
int *sorted_args;
} TCGOpDef;
typedef struct TCGTargetOpDef {
int op;
const char *args_ct_str[TCG_MAX_OP_ARGS];
} TCGTargetOpDef;
extern TCGOpDef tcg_op_defs[];
void tcg_target_init(TCGContext *s);
#define tcg_abort() \
do {\
fprintf(stderr, "%s:%d: tcg fatal error\n", __FILE__, __LINE__);\
abort();\
} while (0)
void tcg_add_target_add_op_defs(const TCGTargetOpDef *tdefs);
void tcg_gen_call(TCGContext *s, TCGArg func, unsigned int flags,
unsigned int nb_rets, const TCGArg *rets,
unsigned int nb_params, const TCGArg *args1);
void tcg_gen_shifti_i64(TCGArg ret, TCGArg arg1,
int c, int right, int arith);
/* only used for debugging purposes */
void tcg_register_helper(void *func, const char *name);
#define TCG_HELPER(func) tcg_register_helper(func, #func)
const char *tcg_helper_get_name(TCGContext *s, void *func);
void tcg_dump_ops(TCGContext *s, FILE *outfile);
void dump_ops(const uint16_t *opc_buf, const TCGArg *opparam_buf);
int tcg_const_i32(int32_t val);
int tcg_const_i64(int64_t val);
#if TCG_TARGET_REG_BITS == 32
#define tcg_const_ptr tcg_const_i32
#define tcg_add_ptr tcg_add_i32
#define tcg_sub_ptr tcg_sub_i32
#else
#define tcg_const_ptr tcg_const_i64
#define tcg_add_ptr tcg_add_i64
#define tcg_sub_ptr tcg_sub_i64
#endif
void tcg_out_reloc(TCGContext *s, uint8_t *code_ptr, int type,
int label_index, long addend);
void tcg_reg_alloc_start(TCGContext *s);
void tcg_reg_alloc_bb_end(TCGContext *s);
void tcg_liveness_analysis(TCGContext *s);
const TCGArg *tcg_gen_code_op(TCGContext *s, int opc, const TCGArg *args1,
unsigned int dead_iargs);
const TCGArg *dyngen_op(TCGContext *s, int opc, const TCGArg *opparam_ptr);
/* tcg-runtime.c */
int64_t tcg_helper_shl_i64(int64_t arg1, int64_t arg2);
int64_t tcg_helper_shr_i64(int64_t arg1, int64_t arg2);
int64_t tcg_helper_sar_i64(int64_t arg1, int64_t arg2);
int64_t tcg_helper_div_i64(int64_t arg1, int64_t arg2);
int64_t tcg_helper_rem_i64(int64_t arg1, int64_t arg2);
uint64_t tcg_helper_divu_i64(uint64_t arg1, uint64_t arg2);
uint64_t tcg_helper_remu_i64(uint64_t arg1, uint64_t arg2);

1227
tcg/x86_64/tcg-target.c Normal file

File diff suppressed because it is too large Load Diff

69
tcg/x86_64/tcg-target.h Normal file
View File

@ -0,0 +1,69 @@
/*
* Tiny Code Generator for QEMU
*
* Copyright (c) 2008 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#define TCG_TARGET_X86_64 1
#define TCG_TARGET_REG_BITS 64
//#define TCG_TARGET_WORDS_BIGENDIAN
#define TCG_TARGET_NB_REGS 16
enum {
TCG_REG_RAX = 0,
TCG_REG_RCX,
TCG_REG_RDX,
TCG_REG_RBX,
TCG_REG_RSP,
TCG_REG_RBP,
TCG_REG_RSI,
TCG_REG_RDI,
TCG_REG_R8,
TCG_REG_R9,
TCG_REG_R10,
TCG_REG_R11,
TCG_REG_R12,
TCG_REG_R13,
TCG_REG_R14,
TCG_REG_R15,
};
#define TCG_CT_CONST_S32 0x100
#define TCG_CT_CONST_U32 0x200
/* used for function call generation */
#define TCG_REG_CALL_STACK TCG_REG_RSP
#define TCG_TARGET_STACK_ALIGN 16
/* optional instructions */
#define TCG_TARGET_HAS_bswap_i32
#define TCG_TARGET_HAS_bswap_i64
/* Note: must be synced with dyngen-exec.h */
#define TCG_AREG0 TCG_REG_R14
#define TCG_AREG1 TCG_REG_R15
#define TCG_AREG2 TCG_REG_R12
#define TCG_AREG3 TCG_REG_R13
static inline void flush_icache_range(unsigned long start, unsigned long stop)
{
}