libc6/libgcc2/kiivi

[Wouter Verhelst]

Hi,

(not 100% sure that this hasn't already been mentioned here; I vaguely
seem to remember that -- but better say it once too often rather than
not at all).

I thought to update kiivi. Since it was proving troublesome, I thought
it best to remove build-essential 'temporarily' so that I could update,
and then reinstall it.

Turns out libc6 requires libgcc2, which is no longer available (unless
one installs stable). Since upgrading gcc to a fairly recent version now
conflicts against libgcc2, too, upgrading would require throwing libc6
off the system, which obviously won't happen.

This means we have to do a binNMU of libc6, which is going to be fun.

How do we proceed?

(NB: the above obviously also means that kiivi won't be building until
I get a working compiler on it again. Hints are welcome).

--
The biometric identification system at the gates of the CIA headquarters
works because there's a guard with a large gun making sure no one is
trying to fool the system.
  http://www.schneier.com/blog/archives/2009/01/biometrics.html


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Stephen R Marenka]

On Mon, Jun 22, 2009 at 12:10:32AM +0200, Wouter Verhelst wrote:

> Turns out libc6 requires libgcc2, which is no longer available (unless
> one installs stable). Since upgrading gcc to a fairly recent version now
> conflicts against libgcc2, too, upgrading would require throwing libc6
> off the system, which obviously won't happen.

In sid, libgcc2 is provided by gcc-4.4, which FTBFS. That's rumored to
be due to a bad debian patch, but I haven't tried to isolate it yet.

> This means we have to do a binNMU of libc6, which is going to be fun.

libc6 2.5-11 is the latest that builds for m68k. After that came TLS
and major breakage. Anyway, gcc-4.4 is the answer to libgcc2. That's
also what kills build-essential.

> How do we proceed?

Fixing gcc-4.4 is pretty much mandatory. I've heard there are patches
for TLS making their way into upstream somewhere, sometime.

> (NB: the above obviously also means that kiivi won't be building until
> I get a working compiler on it again. Hints are welcome).

I need to do the same thing on zlin2, something kicked build-essential
out somehow. dports has weekly snapshots at [0]. I think judicious
selection from there should work, but I haven't tried yet.

[0] http://ftp.debian-ports.org/debian-snapshot/

Peace,

Stephen

--
Stephen R. Marenka     If life's not fun, you're not doing it right!
<stephen@...>

[Maxim Kuvyrkov-2]

Stephen R Marenka wrote:

...

> Fixing gcc-4.4 is pretty much mandatory. I've heard there are patches
> for TLS making their way into upstream somewhere, sometime.

FYI, TLS patches are in gcc-trunk (what will become gcc-4.5).

--
Maxim K.
CodeSourcery


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Brad Boyer]

On Sat, Jun 27, 2009 at 10:25:53PM +0400, Maxim Kuvyrkov wrote:
> Stephen R Marenka wrote:
> >Fixing gcc-4.4 is pretty much mandatory. I've heard there are patches
> >for TLS making their way into upstream somewhere, sometime.
>
> FYI, TLS patches are in gcc-trunk (what will become gcc-4.5).

That's good to hear. Thank you for passing along the status. Any word on
the kernel changes that I presume are required to use the new TLS code?

        Brad Boyer
        flar@...


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Stephen R Marenka]

On Sat, Jun 27, 2009 at 10:25:53PM +0400, Maxim Kuvyrkov wrote:

>> Fixing gcc-4.4 is pretty much mandatory. I've heard there are patches
>> for TLS making their way into upstream somewhere, sometime.
>
> FYI, TLS patches are in gcc-trunk (what will become gcc-4.5).

Any chance they are easily backportable to gcc-4.4?

Thanks for the update!

Stephen

--
Stephen R. Marenka     If life's not fun, you're not doing it right!
<stephen@...>

[Maxim Kuvyrkov-2]

Brad Boyer wrote:
> On Sat, Jun 27, 2009 at 10:25:53PM +0400, Maxim Kuvyrkov wrote:
>> Stephen R Marenka wrote:
>>> Fixing gcc-4.4 is pretty much mandatory. I've heard there are patches
>>> for TLS making their way into upstream somewhere, sometime.
>> FYI, TLS patches are in gcc-trunk (what will become gcc-4.5).
>
> That's good to hear. Thank you for passing along the status. Any word on
> the kernel changes that I presume are required to use the new TLS code?

Kernel and GLIBC patches are in early debug stage now; keep in mind,
though, that these are for ColdFire architecture, so generic m68k may
need further work.

--
Maxim K.
CodeSourcery


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Maxim Kuvyrkov-2]

Stephen R Marenka wrote:
> On Sat, Jun 27, 2009 at 10:25:53PM +0400, Maxim Kuvyrkov wrote:
>
>>> Fixing gcc-4.4 is pretty much mandatory. I've heard there are patches
>>> for TLS making their way into upstream somewhere, sometime.
>> FYI, TLS patches are in gcc-trunk (what will become gcc-4.5).
>
> Any chance they are easily backportable to gcc-4.4?

I think the patch should apply without conflicts to 4.4 and require no
or little backporting.

--
Maxim


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Stephen R Marenka]

On Sun, Jun 28, 2009 at 11:39:17AM +0400, Maxim Kuvyrkov wrote:

> Stephen R Marenka wrote:
>> On Sat, Jun 27, 2009 at 10:25:53PM +0400, Maxim Kuvyrkov wrote:
>>
>>>> Fixing gcc-4.4 is pretty much mandatory. I've heard there are patches
>>>> for TLS making their way into upstream somewhere, sometime.
>>> FYI, TLS patches are in gcc-trunk (what will become gcc-4.5).
>>
>> Any chance they are easily backportable to gcc-4.4?
>
> I think the patch should apply without conflicts to 4.4 and require no  
> or little backporting.

binutils and gcc patches are at [0]. Let the fun begin. :)

[0] http://people.debian.org/~smarenka/m68k/tls/

Peace,

Stephen

--
Stephen R. Marenka     If life's not fun, you're not doing it right!
<stephen@...>

[Brad Boyer]

On Sun, Jun 28, 2009 at 03:50:07PM -0500, Stephen R Marenka wrote:
> binutils and gcc patches are at [0]. Let the fun begin. :)
>
> [0] http://people.debian.org/~smarenka/m68k/tls/

Bit of a problem:

Forbidden

You don't have permission to access /~smarenka/m68k/tls/binutils_patch1 on this server.

        Brad Boyer
        flar@...


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Stephen R Marenka]

On Sun, Jun 28, 2009 at 02:13:48PM -0700, Brad Boyer wrote:

> On Sun, Jun 28, 2009 at 03:50:07PM -0500, Stephen R Marenka wrote:
> > binutils and gcc patches are at [0]. Let the fun begin. :)
> >
> > [0] http://people.debian.org/~smarenka/m68k/tls/
>
> Bit of a problem:
>
> Forbidden
>
> You don't have permission to access /~smarenka/m68k/tls/binutils_patch1 on this server.

Arrggggh!!!! fixed.

Thanks!

Stephen

--
Stephen R. Marenka     If life's not fun, you're not doing it right!
<stephen@...>

[Kolbjørn Barmen-5]

On Mon, 29 Jun 2009, Stephen R Marenka wrote:

> On Sun, Jun 28, 2009 at 02:13:48PM -0700, Brad Boyer wrote:
> > On Sun, Jun 28, 2009 at 03:50:07PM -0500, Stephen R Marenka wrote:
> > > binutils and gcc patches are at [0]. Let the fun begin. :)
> > >
> > > [0] http://people.debian.org/~smarenka/m68k/tls/
> >
> > Bit of a problem:
> >
> > Forbidden
> >
> > You don't have permission to access /~smarenka/m68k/tls/binutils_patch1 on this server.
>
> Arrggggh!!!! fixed.

Looking at the binutils patch...

2000
2001    BFD_ASSERT (!elf_m68k_hash_table (info)->allow_multigot_p
2002 -        || ((big->rel_8o_n_entries
2003 -        <= ELF_M68K_REL_8O_MAX_N_ENTRIES_IN_GOT (info))
2004 -       && (big->rel_8o_16o_n_entries
2005 */
2006 -      arg->rel_16o_offset = (arg->got_offset
2007 -              + (arg->got_offset - arg->rel_16o_offset));
2008 -  }

patch: **** malformed patch at line 2005: */

What's up with line 2005 there?

So I remove line 2005, and move on...

2030 -           /* We've assigned a negative offset to this entry,
2031 -              next entry should get (+abs(offset) + 0).  */
2032 -           arg->rel_32o_offset = (arg->got_offset
2033 -                                  + (arg->got_offset - arg->rel_32o_offset));
2034 -       }
2035 -      else
2036 -       /* Next entry will simply get next offset.  */
2037 -       arg->rel_32o_offset += 4;
2038 -
2039 -      break;
2040 -

patch: **** malformed patch at line 2037: -     arg->rel_32o_offset += 4;

Huh - this one i dont grasp :)

-- kolla


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Maxim Kuvyrkov-2]

Kolbjørn Barmen wrote:

> On Mon, 29 Jun 2009, Stephen R Marenka wrote:
>
>> On Sun, Jun 28, 2009 at 02:13:48PM -0700, Brad Boyer wrote:
>>> On Sun, Jun 28, 2009 at 03:50:07PM -0500, Stephen R Marenka wrote:
>>>> binutils and gcc patches are at [0]. Let the fun begin. :)
>>>>
>>>> [0] http://people.debian.org/~smarenka/m68k/tls/
>>> Bit of a problem:
>>>
>>> Forbidden
>>>
>>> You don't have permission to access /~smarenka/m68k/tls/binutils_patch1 on this server.
>> Arrggggh!!!! fixed.
>
> Looking at the binutils patch...

I suggest you use binutils mainline checkout as that contains a couple
of bug fixes to the TLS support.  Same applies to GCC.  Both mainline
m68k binutils and GCC are in good shape for development.

--
Maxim


--
To UNSUBSCRIBE, email to debian-68k-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...