problem building memtest86+ on x86_64

[Dennis Veatch]

On 12/02/2011 06:56 PM, Richard B. Pyne wrote:
> Also, after trying to fix it by making the edits to the Makefile to 
> specify 64 bit architecture, the build fails on trying to apply 8-byte 
> relocation to a 4 byte field.
>
> --Richard
>

And that's a similar make failure when I was poking around with 
VirtualBoxOSE on our x86_64. Sure I could circumvent it wanting the 
32bit clients but it would fail on some of the vmm's.

> On 12/2/2011 4:04 PM, Richard B. Pyne wrote:
>> I see the conflict, but the options are as set by memtest's Makefile. 
>> lunar build either needs to detect and modify the provided Makefile 
>> or there needs to be a 64 bit specific module that modifies the 
>> Makefile.
>>
>> --Richard


Not gonna happen on lunar x86_64. It is non-multilib and our x86_64 does 
not support 32bit. That's why it cannot find stubs-32. Sure 
/usr/include/gnu/stubs.h (glibc) will check for a wordsize of 32 but it 
doesn't exist and why there is an elif;

#include <bits/wordsize.h>

#if __WORDSIZE == 32
# include <gnu/stubs-32.h>
#elif __WORDSIZE == 64
# include <gnu/stubs-64.h>
#else
# error "unexpected value for __WORDSIZE macro"
#endif

I suspect their v2.01 enhancement of; Patched Makefile to compile on 
x86_64... is referring to a multi-lib setup.


I could be wrong but don't think so.

>> On 12/2/2011 3:56 PM, Terry Chan wrote:
>>> Seriously? Your email says you are building on x86_64, but your gcc 
>>> options are -march=i486 -m32
>>> You don't see a conflict there??
>>>
>>> Terry Chan
>>> ----------------------
>>> On 12/2/2011 4:51 PM, Richard B. Pyne wrote:
>>>> gcc -c -Wall -march=i486 -m32 -O2 -fomit-frame-pointer -fno-builtin 
>>>> -ffreestanding -fPIC -fno-stack-protector -fno-strict-aliasing reloc.c
>>>> In file included from /usr/include/features.h:382:0,
>>>>                  from /usr/include/string.h:27,
>>>>                  from reloc.c:4:
>>>> /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such 
>>>> file or directory
>>>> compilation terminated.
>>>> make: *** [reloc.o] Error 1
>>>>
>


-- 
Dennis Veatch
Lunar Linux Developer
http://www.lunar-linux.org/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://foo-projects.org/pipermail/lunar/attachments/20111202/3a59021f/attachment-0001.html>