[lfs-dev] r10560 - in trunk/BOOK: . appendices bootscripts bootscripts/lfs/init.d chapter01 chapter03 chapter05 chapter06

Bruce Dubbs bruce.dubbs at gmail.com
Sun May 11 10:51:34 PDT 2014


Pierre Labastie wrote:
> Le 11/05/2014 19:07, Bruce Dubbs a écrit :
>> Pierre Labastie wrote:
>>> Le 11/05/2014 18:03, Bruce Dubbs a écrit :
>>>> Fernando de Oliveira wrote:
>>>>> Em 11-05-2014 12:02, bdubbs at higgs.linuxfromscratch.org escreveu:
>>>>>> Author: bdubbs
>>>>>> Date: Sun May 11 08:02:47 2014
>>>>>> New Revision: 10560
>>>>
>>>>>> +    <para>For x86 machines, the limited number of registers is a bottleneck
>>>>>> +    for the system.  Free one up by not using a fram pointer that is not
>>>>>> +    needed:</para>
>>>>>
>>>>> Typo:
>>>>>
>>>>> s/fram pointer/frame pointer/
>>>>
>>>> Fixed in my sandbox.  I'm working on/testing the LTO patch for binutils right
>>>> now and will fix the above with my next commit.
>>
>>> Maybe I was not clear enough. -fomit-frame-pointer has been the default for 32
>>> bit (as well à 64 bit) since gcc-4.6. I guess that part should be totally
>>> suppressed.
>>
>> I haven't built x86 in a while.  I can remove that part in both Chapters 5 and
>> 6 if you think it doesn't make any difference.
>>
>>    -- Bruce
>>
>>
> I did not test thoroughly, but if I believe the gcc docs and logs, yes, it
> should make no difference... Actually, there is a switch in gcc's configure
> (not the top configure) to enable -fno-omit-frame-pointer (notice the "no"),
> which means, the "no" is not the default (sorry for the double negation).

OK.  I'll comment out those sections in the book and wait and see if we 
get any feedback.

   - -Bruce





More information about the lfs-dev mailing list