Re: ct-ng 1.16.0: setting glibc 2.8 cause build to fail

[Zvi Vered <veredz72 at gmail dot com>]

Dear Mr. Morin, All,

glibc 2.8 was last modified at 26-Feb-2009.
I must use glibc 2.8 (or earlier) because Intel Compiler does not
support glibc 2.9 (or later).

I have to use kernel 3.2.x (or later) because this kernel contains the
drivers I need for the PC104 chipset.
Kernel 3.2.25 (for example) was released at: 08-Feb-2012

When I using gcc 4.3.2 (09-Mar-2008) I got:

linux-gnu/buildtools/lib/gcc/i686-nptl-linux-gnu/4.3.2/include-fixed
-isystem /home/zvivered/ct-ng-1.16.0/release/i686-nptl-linux-gnu/sysroot/usr/include
-D_LIBC_REENTRANT -include ../include/libc-symbols.h  -DPIC -DSHARED
  -DASSEMBLER  -DGAS_SYNTAX  -Wa,--noexecstack  -o
/home/zvivered/ct-ng-1.16.0/bin/.build/i686-nptl-linux-gnu/build/build-libc-final/math/s_frexp.os
-MD -MP -MF /home/zvivered/ct-ng-1.16.0/bin/.build/i686-nptl-linux-gnu/build/build-libc-final/math/s_frexp.os.dt
-MT /home/zvivered/ct-ng-1.16.0/bin/.build/i686-nptl-linux-gnu/build/build-libc-final/math/s_frexp.os
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S: Assembler messages:
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: invalid identifier
for ".ifdef"
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: junk at end of
line, first unrecognized character is `1'
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: junk at end of
line, first unrecognized character is `1'
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: junk at end of
line, first unrecognized character is `1'
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: junk at end of
line, first unrecognized character is `.'
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: junk at end of
line, first unrecognized character is `1'
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: expected comma
after name `' in .size directive
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: ".endif" without ".if"
[ALL  ]    ../sysdeps/i386/fpu/s_frexp.S:66: Error: junk
`.get_pc_thunk.dx' after expression
[ERROR]    make[3]: ***
[/home/zvivered/ct-ng-1.16.0/bin/.build/i686-nptl-linux-gnu/build/build-libc-final/math/s_frexp.os]
Error 1
[ALL  ]    make[3]: *** Waiting for unfinished jobs....
[ALL  ]    make[3]: Leaving directory
`/home/zvivered/ct-ng-1.16.0/bin/.build/src/glibc-2.8/math'
[ERROR]    make[2]: *** [math/subdir_lib] Error 2
[ALL  ]    make[2]: Leaving directory
`/home/zvivered/ct-ng-1.16.0/bin/.build/src/glibc-2.8'
[ERROR]    make[1]: *** [all] Error 2
[ALL  ]    make[1]: Leaving directory
`/home/zvivered/ct-ng-1.16.0/bin/.build/i686-nptl-linux-gnu/build/build-libc-final'

Do you think it is possible to use kernel 3.x with glib2.8 ?

Thanks,
Zvika

On Tue, Oct 30, 2012 at 2:06 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> Zvi, All,
>
> On Monday 29 October 2012 Zvi Vered wrote:
>> I'm using version 1.16.0 to build a toolchain fo i686-nptl-linux-gnu.
>>
>> I ran:
>> ct-ng i686-nptl-linux-gnu
>> Then, with menuconfig I set glibc to 2.8 (instead of 2.9)
>>
>> The build fails with the following messages:
> [--SNIP--]
>> [ERROR]    configure: error:
>> [CFG  ]    *** These critical programs are missing or too old: as ld
>> [CFG  ]    *** Check the INSTALL file for required versions.
>>
>> What is the reason for this failure ?
>
> glibc-2.8's ./configure has an incorrect check for binutils version. In
> this case glibc-2.8 was not prepared to recognised binutils-2.2x (IIRC),
> and thus its ./configure does not recognise them.
>
> As a rule of thumb, it is highly suggested that one uses versions of the
> same /age/. Mixing 'old' versions of a component with a newer version of
> another component can be problematic.
>
> Advice: either upgrade to a newer glibc, or downgrade the binutils to a
> version that was current at the time glibc-2.8 was released. Ditto for
> the compiler and the kernel headers.
>
> Quote of the day:
>     Friends don't let friends build toolchains
>     with disparate component versions.
>
> Regards,
> Yann E. MORIN.
>
> --
> .-----------------.--------------------.------------------.--------------------.
> |  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
> | +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
> | +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
> | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
> '------------------------------^-------^------------------^--------------------'

--
For unsubscribe information see http://sourceware.org/lists.html#faq