[uClibc]setjmp/longjmp and others

Erik Andersen andersen at codepoet.org
Tue Jan 8 09:28:19 UTC 2002


On Mon Jan 07, 2002 at 08:23:16PM -0800, David Schleef wrote:
> On Mon, Jan 07, 2002 at 08:52:38PM -0700, Erik Andersen wrote:
> > On Mon Jan 07, 2002 at 03:30:26PM -0800, David Schleef wrote:
> > > 
> > > Note that adding '-falign-functions=0' to the compile flags for
> > > uClibc can decrease it's size by several percent.
> > 
> > I have that now for x86.  It wasn't certain that it was
> > safe as a general case optimization.  Are you certain that
> > would be safe for ARM, sparc, etc?
> 
> (I _meant_ -malign-functions=0)
> 
> -malign-functions is only available on i386, AFAIK, and only for
> gcc-2.95.  gcc-3.0 has -falign-functions, which replaces the
> -malign-functions option.  HOWEVER, as an added bonus to those
> who aren't paying attention, -falign-functions is specified in
> bytes, not powers-of-2, as used by -malign-functions.  So the
> correct option would be -falign-functions=1.  I just checked this
> on PowerPC, and it works correctly.
> 
> gcc-3.0 looks pretty nice.  Haven't looked at code size, however.

I just compiled uClibc up with gcc 3.0.3 and everything
seems to work just fine.  Binary sizes look pretty comparable 
on x86.  Havn't tried on ARM yet,

 -Erik

--
Erik B. Andersen             http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--



More information about the uClibc mailing list