[jornada] kernel crash depends on ELF entry point

Hi, all

According as,

  http://www.handhelds.org/pipermail/jornada/2002-June/000190.html

I could succeed in using a rxvt on X window in Jornada 720. This is
pretty cool except kernel crash just after starting some commands,
such as ifconfig, ldconfig, etc.

I also compiled the kernel 2.4.18-rmk3-hh10 by myself after changing
some points indicated at

  http://www.handhelds.org/pipermail/jornada/2002-June/000191.html

But situation is same. Before this crash, nothing reported on serial
console, just crashing. However when jornada LCD screen turns out
couple of minute after crash (but backlight is still on), this message
is only appeared on serial console.

  e1356fb_blank: blank=1 info->modename=

By using /usr/bin/readelf on ipaq H3600 familiar/intimate
distribution, I tried to figure out the differences between two types
of command that cause or not cause kernel crash. I can realize that a
command causes crash has a ELF entry point 0x00008000's, and other
commands with 0x02000000's can run perfectly.

In intimate or Debian distributions, almost all command is located at
0x00008000's entry point.

This means that all command I need have to be recompiled as virtual
address above 0x0200000 from their source code. This makes difficult
to use Debian on jornada 720.

On my ipaq, both ELF entry points have no problem to run. What is the
difference between ipaq and jornada 720?.

Can someone indicate the way to fix it?

---
Takato