On Sat, Sep 22, 2007 at 04:31:32AM -0400, Priya Narasimhan wrote: > Hi, > > I am compiling relatively simple C programs using "gcc -S > -fomit-frame-pointer -mcpu=xscale" on a gumstix (ARM Xscale) processor. I > noticed that the generated assembly code has additional stuff in, beyond > what I would typically see in a standard ARM assembly program (e.g., one > that was generated instead through the ARM Development Suite toolchain). > > For instance, I see the following comments in the GNU-generated assembly > code. What do these signify? > > @ zero_extendqisi2 I think this is just a debugging aid so that you can tell the difference between e.g. ldrb r2, [r5, r3] @ zero_extendqisi2 and ldrb r2, [r5, r3] since without the comment, they'll look just the same. I'm guessing the comment is a remnant from before the -dp option was added to GCC. It's not unusual for more than one GCC instruction pattern to map to the same machine instruction. The -dp option will add a comment to clear the ambiguity. In the example above, in addition to loading 8 bits from memory to a register, GCC expects the first instruction to clear the most significant 24 bits to zero (comment "*arm_zero_extendqisi2" with -dp) while making no such expectation of the second instruction (comment "*arm_movqi_insn" with -dp). > @ args =0, pretend =0 (what does pretend mean?) > @ frame_needed = 0, uses_anonymous_args = 0 (what does uses_anonymous_args > mean?) Also debugging aids, this time with some of the information GCC uses when generation function prologues and epilogues. > Secondly, I notice that without the "-O" option, saving registers to the > stack happens almost on every alternate line of the generated assembly code. Please refer to the manual <URL:http://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html>. > Finally, is there any documentation on what "-O" actually accomplishes for > an ARM/Xscale processor? For instance, does it use the barrel shifter or > conditional execution wherever possible? AFAIK, there's not much documentation about the machine specific optimizations. If you have a particular piece of source code where you want to know how it was optimized, you can look at the dump files. Please see <URL:http://gcc.gnu.org/onlinedocs/gcc/Debugging-Options.html>. The most interesting options when it comes to the machine specific optimizations are: -fdump-rtl-combine (or even better recently, -fdump-rtl-combine-details) -fdump-rtl-ce1 -fdump-rtl-ce2 -fdump-rtl-ce3 -fdump-rtl-peephole2 Frequently, you'll need to compare two dump files using your favourite diff utility to see what a particular optimization pass changed. Then you might as well just specify -fdump-rtl-all(-details) and get all the dump files. -- Rask Ingemann Lambertsen Danish law requires addresses in e-mail to be logged and stored for a year
