This is the mail archive of the crossgcc@sources.redhat.com mailing list for the crossgcc project.

See the CrossGCC FAQ for lots more information.


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

Re: How to build glibc for ppc variant w/o floating point?


I found some weird/buggy things when looking more at the Linux/PPC-config
files nearer :

Kai Ruottu wrote:
> 
> Peter Barada wrote:
> >
> > I'm trying to put together glibc-2.2.2 for a ppc variant w/o floating
> > point, and I'm not getting very far...
>
> The 't-ppc' must anyhow be edited to be something like:
> 
> -------------------------- clip ------------------------------------
> # PowerPC support without gas

 Of course everyone else but the GCC-developers expect Linux to use gas
as default :-(  It seems that the '--with-gnu-as' selects the 't-ppcos'
as the tmake_file instead of the 't-ppc' and the '-fPIC' option will be
used to build the 'pic' subdir too as default.

 Hmmm what is this mystical default PPC-assembler which is not 'gas'?
Ok, the AIX- and Mac OS X machines may have another assembler, but which
assembler the native Linux/PPC-machines then use, when gas cannot be the
default... Or is this simply a nasty bug?

 Anyway when I didn't see a 't-linux' in the 'gcc/config/rs6000', the
default 't-ppc' and the things it does, told that I had used it when
building... One trap more for all non-paranoid persons ;-)...

>  Using the '-msoft-float' or having it as default so enables one to
> choose the soft-float-package, trying the default fp-bit/dp-bit coming
> with GCC first has its reason, but if more speed is needed, there may
> be alternatives...

 The handling of the fp-bit/dp-bit-stuff was recently changed and the
routines in them can be built 'fine-grained' by defining:

  FP-BIT = fp-bit.c
  DP-BIT = dp-bit.c

instead of:

  LIB2FUNCS_EXTRA = fp-bit.c dp-bit.c ....

in the tmake_file. The old way results to building the fp-bit.o and
dp-bit.o as two big objects while the new way results to building them
as separate small objects for the functions in them. The reduction in the
size of the linked executable can then be 20 kbytes or so...

 The gcc-2.95.3 sources don't use the new way for any PPC-target although
it would be possible !  Not even the embedded 'ppc-eabi' had this 'fine-
grained soft-float routines' enabled... The gcc-3.0 snapshots (and the
release) however has this feature in use...

Cheers, Kai


------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com


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