![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to linuxelf.h CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [GCC 2.x] / gcc / config / i386|
Return to linuxelf.h CVS log | Up to [GCC 2.x] / gcc / config / i386 |
1.1 root 1: /* Definitions for Intel 386 running Linux with ELF format
2: Written by Eric Youngdale. */
3:
4: /* A lie, I guess, but the general idea behind linux/ELF is that we are
5: supposed to be outputting something that will assemble under SVr4.
6: This gets us pretty close. */
7: #include "i386/i386.h" /* Base i386 target machine definitions */
8: #include "i386/att.h" /* Use the i386 AT&T assembler syntax */
9: #include "svr4.h"
10: #undef MD_EXEC_PREFIX
11: #undef MD_STARTFILE_PREFIX
12: #undef TARGET_VERSION
13: #define TARGET_VERSION fprintf (stderr, " (i386 Linux/ELF)");
14: /* Output at beginning of assembler file. */
15: /* The .file command should always begin the output. */
16: #undef ASM_FILE_START
17: #define ASM_FILE_START(FILE) \
18: do { \
19: output_file_directive (FILE, main_input_filename); \
20: fprintf (FILE, "\t.version\t\"01.01\"\n"); \
21: } while (0)
22: /* The svr4 ABI for the i386 says that records and unions are returned
23: in memory. */
1.1.1.2 ! root 24: #undef DEFAULT_PCC_STRUCT_RETURN
! 25: #define DEFAULT_PCC_STRUCT_RETURN 1
1.1 root 26: /* This is how to output an element of a case-vector that is relative.
27: This is only used for PIC code. See comments by the `casesi' insn in
28: i386.md for an explanation of the expression this outputs. */
29: #undef ASM_OUTPUT_ADDR_DIFF_ELT
30: #define ASM_OUTPUT_ADDR_DIFF_ELT(FILE, VALUE, REL) \
31: fprintf (FILE, "\t.long _GLOBAL_OFFSET_TABLE_+[.-%s%d]\n", LPREFIX, VALUE)
32: /* Indicate that jump tables go in the text section. This is
33: necessary when compiling PIC code. */
34: #define JUMP_TABLES_IN_TEXT_SECTION
35: /* Copy this from the svr4 specifications... */
36: /* Define the register numbers to be used in Dwarf debugging information.
37: The SVR4 reference port C compiler uses the following register numbers
38: in its Dwarf output code:
39: 0 for %eax (gnu regno = 0)
40: 1 for %ecx (gnu regno = 2)
41: 2 for %edx (gnu regno = 1)
42: 3 for %ebx (gnu regno = 3)
43: 4 for %esp (gnu regno = 7)
44: 5 for %ebp (gnu regno = 6)
45: 6 for %esi (gnu regno = 4)
46: 7 for %edi (gnu regno = 5)
47: The following three DWARF register numbers are never generated by
48: the SVR4 C compiler or by the GNU compilers, but SDB on x86/svr4
49: believes these numbers have these meanings.
50: 8 for %eip (no gnu equivalent)
51: 9 for %eflags (no gnu equivalent)
52: 10 for %trapno (no gnu equivalent)
53: It is not at all clear how we should number the FP stack registers
54: for the x86 architecture. If the version of SDB on x86/svr4 were
55: a bit less brain dead with respect to floating-point then we would
56: have a precedent to follow with respect to DWARF register numbers
57: for x86 FP registers, but the SDB on x86/svr4 is so completely
58: broken with respect to FP registers that it is hardly worth thinking
59: of it as something to strive for compatibility with.
60: The verison of x86/svr4 SDB I have at the moment does (partially)
61: seem to believe that DWARF register number 11 is associated with
62: the x86 register %st(0), but that's about all. Higher DWARF
63: register numbers don't seem to be associated with anything in
64: particular, and even for DWARF regno 11, SDB only seems to under-
65: stand that it should say that a variable lives in %st(0) (when
66: asked via an `=' command) if we said it was in DWARF regno 11,
67: but SDB still prints garbage when asked for the value of the
68: variable in question (via a `/' command).
69: (Also note that the labels SDB prints for various FP stack regs
70: when doing an `x' command are all wrong.)
71: Note that these problems generally don't affect the native SVR4
72: C compiler because it doesn't allow the use of -O with -g and
73: because when it is *not* optimizing, it allocates a memory
74: location for each floating-point variable, and the memory
75: location is what gets described in the DWARF AT_location
76: attribute for the variable in question.
77: Regardless of the severe mental illness of the x86/svr4 SDB, we
78: do something sensible here and we use the following DWARF
79: register numbers. Note that these are all stack-top-relative
80: numbers.
81: 11 for %st(0) (gnu regno = 8)
82: 12 for %st(1) (gnu regno = 9)
83: 13 for %st(2) (gnu regno = 10)
84: 14 for %st(3) (gnu regno = 11)
85: 15 for %st(4) (gnu regno = 12)
86: 16 for %st(5) (gnu regno = 13)
87: 17 for %st(6) (gnu regno = 14)
88: 18 for %st(7) (gnu regno = 15)
89: */
90: #undef DBX_REGISTER_NUMBER
91: #define DBX_REGISTER_NUMBER(n) \
92: ((n) == 0 ? 0 \
93: : (n) == 1 ? 2 \
94: : (n) == 2 ? 1 \
95: : (n) == 3 ? 3 \
96: : (n) == 4 ? 6 \
97: : (n) == 5 ? 7 \
98: : (n) == 6 ? 5 \
99: : (n) == 7 ? 4 \
100: : ((n) >= FIRST_STACK_REG && (n) <= LAST_STACK_REG) ? (n)+3 \
101: : (-1))
102: #define LINUX_ELF
103: #include "i386/linux.h"
104:
105: #undef YES_UNDERSCORE
106: #undef DBX_DEBUGGING_INFO
107: #define DWARF_DEBUGGING_INFO
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.