The SVR4 reference port C compiler uses the following register numbers
in its Dwarf output code:
- 0 for %eax (regno = 0)
- 1 for %ecx (regno = 2)
- 2 for %edx (regno = 1)
- 3 for %ebx (regno = 3)
- 4 for %esp (regno = 7)
- 5 for %ebp (regno = 6)
- 6 for %esi (regno = 4)
- 7 for %edi (regno = 5)
-
- 8 for FP_REGS[tos] (regno = 8)
- 9 for FP_REGS[tos-1] (regno = 9)
- 10 for FP_REGS[tos-2] (regno = 10)
- 11 for FP_REGS[tos-3] (regno = 11)
- 12 for FP_REGS[tos-4] (regno = 12)
- 13 for FP_REGS[tos-5] (regno = 13)
- 14 for FP_REGS[tos-6] (regno = 14)
- 15 for FP_REGS[tos-7] (regno = 15)
+ 0 for %eax (gnu regno = 0)
+ 1 for %ecx (gnu regno = 2)
+ 2 for %edx (gnu regno = 1)
+ 3 for %ebx (gnu regno = 3)
+ 4 for %esp (gnu regno = 7)
+ 5 for %ebp (gnu regno = 6)
+ 6 for %esi (gnu regno = 4)
+ 7 for %edi (gnu regno = 5)
+
+ The following three DWARF register numbers are never generated by
+ the SVR4 C compiler or by the GNU compilers, but SDB on x86/svr4
+ believes these numbers have these meanings.
+
+ 8 for %eip (no gnu equivalent)
+ 9 for %eflags (no gnu equivalent)
+ 10 for %trapno (no gnu equivalent)
+
+ It is not at all clear how we should number the FP stack registers
+ for the x86 architecture. If the version of SDB on x86/svr4 were
+ a bit less brain dead with respect to floating-point then we would
+ have a precedent to follow with respect to DWARF register numbers
+ for x86 FP registers, but the SDB on x86/svr4 is so completely
+ broken with respect to FP registers that it is hardly worth thinking
+ of it as something to strive for compatability with.
+
+ The verison of x86/svr4 SDB I have at the moment does (partially)
+ seem to believe that DWARF register number 11 is associated with
+ the x86 register %st(0), but that's about all. Higher DWARF
+ register numbers don't seem to be associated with anything in
+ particular, and even for DWARF regno 11, SDB only seems to under-
+ stand that it should say that a variable lives in %st(0) (when
+ asked via an `=' command) if we said it was in DWARF regno 11,
+ but SDB still prints garbage when asked for the value of the
+ variable in question (via a `/' command).
+
+ (Also note that the labels SDB prints for various FP stack regs
+ when doing an `x' command are all wrong.)
+
+ Note that these problems generally don't affect the native SVR4
+ C compiler because it doesn't allow the use of -O with -g and
+ because when it is *not* optimizing, it allocates a memory
+ location for each floating-point variable, and the memory
+ location is what gets described in the DWARF AT_location
+ attribute for the variable in question.
+
+ Regardless of the severe mental illness of the x86/svr4 SDB, we
+ do something sensible here and we use the following DWARF
+ register numbers. Note that these are all stack-top-relative
+ numbers.
+
+ 11 for %st(0) (gnu regno = 8)
+ 12 for %st(1) (gnu regno = 9)
+ 13 for %st(2) (gnu regno = 10)
+ 14 for %st(3) (gnu regno = 11)
+ 15 for %st(4) (gnu regno = 12)
+ 16 for %st(5) (gnu regno = 13)
+ 17 for %st(6) (gnu regno = 14)
+ 18 for %st(7) (gnu regno = 15)
*/
#undef DBX_REGISTER_NUMBER
: (n) == 5 ? 7 \
: (n) == 6 ? 5 \
: (n) == 7 ? 4 \
- : ((n) >= FIRST_STACK_REG && (n) <= LAST_STACK_REG) ? (n) \
+ : ((n) >= FIRST_STACK_REG && (n) <= LAST_STACK_REG) ? (n)+3 \
: (abort (), 0))
/* The routine used to output sequences of byte values. We use a special
OSDN Git Service
