[Whitebox-ia64] Adaptec 2015S RAID daughter card (Tyan S2882UG3NR motherboard)

Robert Heller heller@cs.umass.edu
Mon, 04 Oct 2004 11:53:13 -0400



In message <200409291524.i8TFOuKf030789@athos.cs.umass.edu>, Robert Heller writ
es:
>OK, the machine is up sans the I2O RAID daughter card.  I am trying to
>build the kernel module, but I am having trouble even building a kernel
>and/or building kernel modules on this system.  Is there some 'trick' to
>building a kernel and/or kernel modules?  I am using the config's that
>were install in the /boot/ directory -- even the vanila configs don't
>seem to work.

After a bit of further mucking with atempting to rebuilt kernel modules
I am finding some weirdnesses WRT gcc.

The system has had verious upgrades (security, bug fixes, etc.), so the
version of gcc is

gcc-3.2.3-42

and the version on the installer CDs is

gcc-3.2.3-34

After poking in the modules that fail to compile (including the dpt_i2o
module), I find that the code is #ifdef'ing on the compiler define __ia64__,
which does not seem to be defined by gcc-3.2.3-42!  Other things are
being defined, like __x86_64__ and such like.

Question:

1) Is gcc-3.2.3-42 broken in NOT defining __ia64__?

2) Or, are (some) the kernel sources broken in not testing for
__x86_64__?

3) Or both?

How have people done *successfull* kernel builds?  Since the binaries
exist, the builds have been done somehow.  What is the magical
incantation?  Or os gcc-3.2.3-42, just plain bad news?

                                     \/
Robert Heller                        ||InterNet:   heller@cs.umass.edu
http://vis-www.cs.umass.edu/~heller  ||            heller@deepsoft.com
http://www.deepsoft.com              /\FidoNet:    1:321/153