Re: Modernizing a Macintosh LC III

2013-11-16 Thread Patrick McCarthy

On 11/12/2013 12:20 PM, Thorsten Glaser wrote:

Geert Uytterhoeven dixit:


Oops, another one bitten by that Debian kernel which has
CONFIG_M68KFPU_EMU disabled...


Yesyesyes… but you must admit the Kconfig message did look scary.

http://www.freewrt.org/~tg/f/linux-image-3.12-rc7-m68k_3.12~rc7-1~exp1+m68k.1_m68k.deb

It’ll be in the next src:linux upload in Debian itself, but
the maintainers are currently blocked by the Alioth catastrophic
disc failure-caused outage.

In the meantime, here’s the vmlinuz from it and the initrd
from my ARAnyM system (i.e. “I can boot it but no guarantee
you can”):

http://www.freewrt.org/~tg/f/vmlinuz-3.12-rc7-m68k
http://www.freewrt.org/~tg/f/initrd.img-3.12-rc7-m68k

a193808d4d8f687ca981051f7a485d09  initrd.img-3.12-rc7-m68k
7e19b242a6779df61fac62151e617c3d  
linux-image-3.12-rc7-m68k_3.12~rc7-1~exp1+m68k.1_m68k.deb
fc7cdec3d73098449f3315ed4e54e243  vmlinuz-3.12-rc7-m68k

bye,
//mirabilos

PS: I’ll be dead for a few days due to dentist.


The ARAnyM kernel and initrd you posted unfortunately didn't work, I put 
them in Penguin in 32bit mode and it hung at 'MacLinux ABCDEFGHIJK'.


Is the one in the deb more Mac-centric? I assume it's best to manually 
unpack into the newly untarred modern fs?  Thanks again.



--
To UNSUBSCRIBE, email to debian-68k-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/52878920.1050...@gmail.com



Re: Modernizing a Macintosh LC III

2013-11-16 Thread Finn Thain

On Tue, 12 Nov 2013, Scott Holder wrote:

> On 11/12/2013 12:12 PM, Thorsten Glaser wrote:
> > 
> > In the meantime, here's the vmlinuz from it and the initrd from my 
> > ARAnyM system (i.e. "I can boot it but no guarantee you can"):
> >
> > http://www.freewrt.org/~tg/f/vmlinuz-3.12-rc7-m68k 
> > http://www.freewrt.org/~tg/f/initrd.img-3.12-rc7-m68k
> 
> I can report that these do not work for me on the LC 475. I don't even 
> get an initial boot console, it just freezes in Penguin after gunzipping 
> the kernel and beginning the process and doing something for the slot 
> interrupts.

This regression, also reported by Patrick (hang at "K"), looks like an 
unhandled interrupt. If so, you can probably get slightly different 
results (e.g. hang a bit later) by simply repeating the process.

I would also check what happens if you first boot MacOS with extensions 
off (so that the MacOS NIC driver doesn't load).

Also, I'm curious to know whether the Penguin settings for the serial 
ports affect the failure mode (you don't want Penguin to initialise them).

You might also try disconnecting any serial devices.

Finn

> 
> Scott
> 
> 
> 


-- 
To UNSUBSCRIBE, email to debian-68k-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/alpine.LNX.2.00.1311170913380.6636@nippy.intranet



Re: Modernizing a Macintosh LC III

2013-11-16 Thread Thorsten Glaser
Patrick McCarthy dixit:

>> PS: I’ll be dead for a few days due to dentist.

It was a horror, I tell you. And will be for some more days.

> Is the one in the deb more Mac-centric?

This is the very same as in the .deb just “installed”.
We do not do subarchitecture-specific kernels for the
generic distro any more.

> I assume it's best to manually unpack
> into the newly untarred modern fs?  Thanks again.

No, you cannot do this any more, since we switched to initrd.
(You can, however, do this with monolithic kernel images you
build yourself and tailor to your computer.)

bye,
//mirabilos
-- 
„Also irgendwie hast du IMMER recht. Hier zuckelte gerade ein Triebwagen mit
der Aufschrift "Ostdeutsche Eisenbahn" durch Wuppertal. Ich glaubs machmal
nicht…“ -- Natureshadow, per SMS
„Hilf mir mal grad beim Denken“ -- Natureshadow, IRL, 2x


--
To UNSUBSCRIBE, email to debian-68k-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/pine.bsm.4.64l.1311162253250.11...@herc.mirbsd.org



gcc-4.8: [m68k] patch set 2013-11 (was Re: Bug#711558: PR52306 (was Re: Bug#711558: gcc-4.8: [m68k] patch set 2))

2013-11-16 Thread Thorsten Glaser
Matthias Klose dixit:

>Am 23.08.2013 00:26, schrieb Thorsten Glaser:

>> --- a/src/gcc/common.opt
>> +++ b/src/gcc/common.opt
>> @@ -858,7 +858,7 @@ Common Report Var(flag_asynchronous_unwi
>>  Generate unwind tables that are exact at each instruction boundary
>>  
>>  fauto-inc-dec
>> -Common Report Var(flag_auto_inc_dec) Init(1)
>> +Common Report Var(flag_auto_inc_dec) Init(0)
>>  Generate auto-inc/dec instructions
>>  
>>  ; -fcheck-bounds causes gcc to generate array bounds checks.

>I think, setting the flag for the option to 0 as the default, and applying this
>for m68k only would be the second best option, provided that you cannot find 
>out
>how to implement Mikael's suggestion.

It turns out that there is no pre-options target hook, only a post
one, so this is what I eventually did.

I also integrated the other patches Mikael suggested, as well as
two more fixes for issues that popped up in the meantime (boost1.54
ICE, and reloction errors in smokeqt and python-qt4 that are new
and appear to be due to the size (-fpic/-fPIC issue)).

The .debdiff is attached, for your convenience to read through and
comment on it already. I’m currently building this, and I’ll report
success (or not) in about a week (since it takes at least five days
for this to compile, from experience).

Changelog:

+  * m68k-ada.diff: Add gcc-4.8.0-m68k-ada-pr48835-2.patch and
+gcc-4.8.0-m68k-ada-pr51483.patch by Mikael Pettersson, to
+fix more CC0-specific and m68k/Ada-specific problems.

Two suggested patches, which I just folded into m68k-ada.diff
for simplicity. Applied on m68k only, as it was previously.

+  * m68k-picflag.diff: New, backport from trunk, by Andreas Schwab,
+to avoid relocation errors when linking big shared objects.
+  * pr58369.diff: New, backport from trunk, by Jeffrey A. Law,
+to fix ICE while building boost 1.54.

The aforementioned two new issues. Always applied: m68k-picflag
doesn’t affect other platforms, and pr58369 is “always applied”
in GCC trunk too, but if you disagree with this, feel free to
apply it m68k only.

+  * pr52306.diff: Supersedes pr52306-retry-hack.diff (in debports
+unreleased), disables -fauto-inc-dec by default on m68k to
+work around ICE when building C++ code (e.g. Qt-related).

As discussed above: changing the default, m68k only.

+  * rules.conf: Do not force libcloog-isl-dev (>= 0.18) on older
+distribution releases that do not have this version yet.

ISTR writing about this already: as I regularily build cross
compilers for wheezy/amd64 which doesn’t have this version
yet, and all other new version dependencies are for sid and
whatever animal *buntu has as development version only (and
the changelog entry adding the 0.18 restriction indicates it
was only done to force the transition in Debian), I disabled
this for “older” releases, just like the mpc version dependency.

bye,
//mirabilos
-- 
„Also irgendwie hast du IMMER recht. Hier zuckelte gerade ein Triebwagen mit
der Aufschrift "Ostdeutsche Eisenbahn" durch Wuppertal. Ich glaubs machmal
nicht…“ -- Natureshadow, per SMS
„Hilf mir mal grad beim Denken“ -- Natureshadow, IRL, 2xdiff -u gcc-4.8-4.8.2/debian/changelog gcc-4.8-4.8.2/debian/changelog
--- gcc-4.8-4.8.2/debian/changelog
+++ gcc-4.8-4.8.2/debian/changelog
@@ -1,3 +1,20 @@
+gcc-4.8 (4.8.2-5+m68k.1) unreleased; urgency=low
+
+  * m68k-ada.diff: Add gcc-4.8.0-m68k-ada-pr48835-2.patch and
+gcc-4.8.0-m68k-ada-pr51483.patch by Mikael Pettersson, to
+fix more CC0-specific and m68k/Ada-specific problems.
+  * m68k-picflag.diff: New, backport from trunk, by Andreas Schwab,
+to avoid relocation errors when linking big shared objects.
+  * pr58369.diff: New, backport from trunk, by Jeffrey A. Law,
+to fix ICE while building boost 1.54.
+  * pr52306.diff: Supersedes pr52306-retry-hack.diff (in debports
+unreleased), disables -fauto-inc-dec by default on m68k to
+work around ICE when building C++ code (e.g. Qt-related).
+  * rules.conf: Do not force libcloog-isl-dev (>= 0.18) on older
+distribution releases that do not have this version yet.
+
+ -- Thorsten Glaser   Sun, 17 Nov 2013 00:41:12 +0100
+
 gcc-4.8 (4.8.2-5) unstable; urgency=low
 
   * Update to SVN 20131115 (r204839) from the gcc-4_8-branch.
diff -u gcc-4.8-4.8.2/debian/patches/m68k-ada.diff 
gcc-4.8-4.8.2/debian/patches/m68k-ada.diff
--- gcc-4.8-4.8.2/debian/patches/m68k-ada.diff
+++ gcc-4.8-4.8.2/debian/patches/m68k-ada.diff
@@ -1,3 +1,11 @@
+gcc/
+
+2013-03-23  Mikael Pettersson  
+
+   PR ada/48835
+   * cse.c (fold_rtx) : If prev_insn_cc0 is NULL
+   return the orginal expression x.
+
 gcc/ada/
 
 2011-10-12  Mikael Pettersson  
@@ -11,6 +19,16 @@
Add Gnat_Malloc -> __gnat_malloc export.
* s-memory.ads: Remove Alloc -> __gnat_malloc export.
 
+2012-01-28  Mikael Pettersson  
+
+   PR ada/51483
+   * gcc-interface/misc.c (enumerate_modes): Pass both precision and
+