Question about rtx format for insn, call_insn and junp_insn
kito
npickito@gmail.com
Thu May 20 03:57:00 GMT 2010
sorry I get incorrect gcc version info
the correct version which doing dump is gcc 4.1.2
I dump rtl with the gcc 4.4.1 have the line info
thanks for you reply
2010/5/20 Ian Lance Taylor <iant@google.com>:
> kito <npickito@gmail.com> writes:
>
>> Hello every body
>> I have read the rtl.h & rtl.c,
>> but I don't realize the format for insn, call_insn and junp_insn
>>
>> it's define in rtl.def
>>
>> DEF_RTL_EXPR(JUMP_INSN, "jump_insn", "iuuBieie0", RTX_INSN)
>>
>> and it's dump by some real program
>>
>> (jump_insn 14 /* i */
>> 13 /* u */
>> 15 /* u */
>> 1 /* B */
>> (set (pc)
>> (if_then_else (le (reg:CCGC 17 flags)
>> (const_int 0 [0x0]))
>> (label_ref 24)
>> (pc))) /* e?? */
>> -1 /* i? */
>> (nil) /* e? */
>> (nil) /* 0? */ )
>>
>> so my question is where is the i after B ??
>
> Which program is doing the dump? The generator programs (genattrtab,
> etc.) produce different dumps from the compilers.
>
> The 'i' after the 'B' represents an insn locator, a value of type
> source_location aka location_t aka unsigned int. A generator program
> will never print that value. A compiler will print it only if it has
> a value--if the instruction has a known location. It will be printed
> as filename:lineno when available.
>
> The corresponding code in print-rtl.c is:
>
> case 'i':
> if (i == 4 && INSN_P (in_rtx))
> {
> #ifndef GENERATOR_FILE
> /* Pretty-print insn locators. Ignore scoping as it is mostly
> redundant with line number information and do not print anything
> when there is no location information available. */
> if (INSN_LOCATOR (in_rtx) && insn_file (in_rtx))
> fprintf(outfile, " %s:%i", insn_file (in_rtx), insn_line (in_rtx));
> #endif
>
> Ian
>
More information about the Gcc
mailing list