T_CFLAGS on x86_64 (gcc in 5.10 and 6.15

Bruce Dubbs bruce.dubbs at gmail.com
Fri Dec 4 20:23:46 PST 2009


Chris Staub wrote:
> On 12/04/2009 08:29 PM, Bruce Dubbs wrote:
>> Ken Moffat wrote:
>>> 2009/12/5 Bruce Dubbs<bruce.dubbs at gmail.com>:
>>>
>>>   A *bootstrap* build on x86_64 doesn't use that option, but we
>>> are passing it anyway (so, presumably, our second set of
>>> compilers are built differently from a bootstrap (and from our
>>> final set of compilers).
>> -fomit-frame-pointer in this context is only an optimization to gcc
>> itself, not it's output.  Does this optimization or lack of it really
>> make a difference for our intermediate compiler?

> The issue is simply consistency in the book's instructions. Both GCC 
> Pass 2 and GCC in Chapter 6 have a sed that adds -fomit-frame-pointer. 
> However, in Chapter 6 this command is in a case statement that causes it 
> to be applied for x86 only, but the command in GCC Pass 2 is not. If you 
> are saying however that it is not important for the temporary compiler, 
> then do you mean that the sed should be removed from Chapter 5 entirely?

I don't consider myself a tool-chain expert, but I don't see why not.  I 
suspect it was only there for ICA analysis when the new toolchain was 
introduced.

I don't see it in Greg's DIY site, but it may be in a patch.  I didn't 
look at those.

   -- Bruce



More information about the lfs-dev mailing list