Subject: Re: Softfloat and -current
To: Bruce ONeel <edoneel@sdf.lonestar.org>
From: Bruce ONeel <edoneel@sdf.lonestar.org>
List: port-mac68k
Date: 11/12/2004 16:30:58
Also, while we are on the subject, where in the source tree might we be
when one gets the line out about

root on xxx
and then one gets the date.

Is this init starting?  Any help will be great since I don't know
where to look.

cheers

bruce

Bruce ONeel <edoneel@sdf.lonestar.org> wrote:
> Date: Fri, 12 Nov 2004 14:01:17 +0200
> From: Bruce ONeel <edoneel@sdf.lonestar.org>
> Subject: Re: Softfloat and -current
> To: "Julio M. Merino Vidal" <jmmv@menta.net>
> Cc: port-mac68k@NetBSD.org
> 
> Hi,
> 
> Sorry this took so long.  Has anyone tried current with a full up
> 040 recently?  Does it work?
> 
> Good news and bad news. 
> 
> The good news is that the patch looks almost correct, but,
> the changes to the libm makefile didn't get copied.  The changes 
> are below.
> 
> The bad news is that while the 2.99.10 kernel seems to work fine,
> install kernel does not work for me as well.
> 
> I'm doing a clean build just in case the install kernel was using 
> libm.
> 
> cheers
> 
> bruce
> 
> 
> [edoneel@lain]/usr/edoneel/current/src/lib/libm[1141]: diff Makefile
> ~/current/Makefile
> 80a81
> > .if (${MKSOFTFLOAT} != "yes")
> 86a88
> > .endif
> [edoneel@lain]/usr/edoneel/current/src/lib/libm[1142]:
> -- 
> 
> .if (${MKSOFTFLOAT} != "yes")
> .PATH:  ${.CURDIR}/arch/mc68881 ${.CURDIR}/arch/m68k
> ARCH_SRCS = e_acos.S e_asin.S e_atanh.S e_cosh.S e_exp.S e_fmod.S
> e_log.S \
>             e_log10.S e_remainder.S e_scalb.S e_sinh.S e_sqrt.S s_atan.S
> \
>             s_ceil.S s_copysign.S s_cos.S s_expm1.S s_finite.S s_floor.S
> \
>             s_log1p.S s_logb.S s_rint.S s_scalbn.S s_sin.S s_tan.S
> s_tanh.S
> .endif
> 
> "Julio M. Merino Vidal" <jmmv@menta.net> wrote:
> > Date: Sat, 9 Oct 2004 14:51:54 +0200
> > From: "Julio M. Merino Vidal" <jmmv@menta.net>
> > Subject: Softfloat and -current
> > To: port-mac68k@NetBSD.org
> > content-length: 1488
> > 
> > Hi all,
> > 
> > I can't get -current kernels (as of today's sources) to work on my Performa 630
> > (68040LC chip).  If I boot the installation kernel (netbsd-INSTALL.gz), I get
> > messages like:
> > 
> > [...]
> > md0: internal 2560 KB image area
> > Kernel Illegal Instruction trap.
> > trap type 2, code = 0x0, v = 0x0
> > kernel program counter = 0x3fd9b8
> > pid = 0, lid = 1, pc = 003FD9B8, ps = 2600, sfc = 1, dfc = 1
> > [...registers and stack dump...]
> > panic: Illegal instruction
> > 
> > Booting my customized kernel generally results in a hang while probing adb0
> > (i.e., I see 'adb0:' and it gets stalled there); no panic nor anything else,
> > just a freeze (strange...).  One of these boots, however, resulted in a panic
> > similar to the above one, just before it tried to mount the root devie (IIRC).
> > 
> > Note that I'm building everything with softfloat enabled (-V MKSOFTFLOAT=yes
> > on build.sh's command line).
> > 
> > However, a netbsd-INSTALL.gz image from 2.0_RC3 built by Bruce works properly.
> > 
> > Maybe there is a problem in the patches I merged in the code (though I tried
> > them a while ago and worked flawlessy).  And, AFAIK the kernel should not
> > include FPU code, right?  If so, I don't see why it panics (though the
> > messages look very much like FPU problems).
> > 
> > Is anybody running current on such machines?  If not, can you try to build a
> > kernel and run it on them?  Any idea about what can be happening?
> > 
> > Thanks.
> > 
> > -- 
> > Julio M. Merino Vidal <jmmv@menta.net>
> > http://www.livejournal.com/users/jmmv/
> > The NetBSD Project - http://www.NetBSD.org/