This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: -malign on Pentiums


> > [p2align] is picked up at configure time unless gas.h is explictly included.
> > Look in the gcc/configure.in.  When I do a build on my RH5 box, it
> > definitely gets detected.
> 
> Then you have better luck than I do.  Using an up-to-date CVS version, I tried
> the following:

Perhaps I just had lower standards than you do. ;-) It's definitely
turned on it auto-host.h, but I think I agree.  It is not used in any
way that I can see.


> sds10> strings final.o | grep p2align

I don't see it there, either.

> I cannot see anything about gas.h.  Am I looking at the wrong place?

Bernd, do you think that this:

/* A C statement to output to the stdio stream FILE an assembler
   command to advance the location counter to a multiple of 1<<LOG
   bytes if it is within MAX_SKIP bytes.

   This is used to align code labels according to Intel recommendations.  */

#ifdef HAVE_GAS_MAX_SKIP_P2ALIGN
#  define ASM_OUTPUT_MAX_SKIP_ALIGN(FILE,LOG,MAX_SKIP) \
     if ((LOG)!=0) \
       if ((MAX_SKIP)==0) fprintf ((FILE), "\t.p2align %d\n", (LOG)); \
       else fprintf ((FILE), "\t.p2align %d,,%d\n", (LOG), (MAX_SKIP))
#endif

should be wrenched out of i386/gas.h and slammed into i386.h?  This
would get the code used in targets that had this facility but that didn't
include gas.h.   This would include, say, Linux.

> Maybe you were just commenting on the autoconf detection of the .p2align
> statement.  In that case I agree with you that it is indeed detected.  The
> problem is that it is not used, at least not in final.c.

Agreed.   The list of targets that would pluck the code out of gas.h
seems to be pretty inconsistent.

RJL


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]