Re: iBCS w/1.3.71

William M. Perkins (bill@grnwood.richmond.us.net)
Thu, 7 Mar 1996 23:10:25 -0500 (EST)

> On Tue, 5 Mar 1996, William M. Perkins wrote:
> > > On Tue, 5 Mar 1996, Baba Z Buehler wrote:
> > > > After rebuilding iBCS for 1.3.71, this is what I get when I try to insmod it:
> > > >
> > > > root@tigger[3]: insmod iBCS
> > > > create_tables undefined
> > > > Failed to load module! The symbols from kernel 1.3.71 don't match 1.3.71
> > >
> > > A quick fix is to use the insmod-1.3.69-HACKER_TOOL.tar.gz, available
> > > from:
> > > ftp://tsx-11.mit.edu/pub/linux/ALPHA/dosemu/insmod-1.3.69-HACKER_TOOL.tar.gz
> > >
> > > It allows lookup of undefined symbols in modules using /System.map or any
> > > other specified map file.
> > >
> > > -james
> >
> > I am having the same problem that Baba is experiencing. I checked
> > the System.map file created for 1.3.71 and discovered that the
> > 'create_tables' entry did not exist. It did exist in the 1.3.70
> > System.map file. The System.map for the 1.3.71 did have the
> > following entry: create_aout_tables. It looks like the entry's
> > name has been changed, but changing the 'create_tables' name
> > to 'create_aout_tables' in the iBCS sources and re-compiling does
> > not work either. Loading iBCS in that case says that
> > 'create_aout_tables' is undefined. :-)
> >
> And if you see below on System.map, then you can find create_elf_tables
> label. I think this label not help you (and me :-). Now I returned to
> 1.3.70.
>
> Regards,
>
> Yuri.

Hello Yuri,

I was able to get the iBCS code to compile and load as a module with the
1.3.71 kernel from information included from a linux-kernel posting from
Tom Grigg <st588@bayou.uh.edu> (Subject: patches for 1.3.71 w/iBCS). His
posting contained a patch to the 1.3.71 kernel and the 1 February 1996
version of iBCS (ibcs-1.3-960201.tar.gz). The patch installed okey, but
I had trouble compiling the iBCSemul/iBCS code due to the lack of a
/usr/include/linux/modversions.h file. I had to delete the line in the
iBCSemul/Makefile that referenced this file, but it did compile and load
okey. The x286emul code has never compiled for me, but I do not expect
that I will ever have the need to execute any Xenix programs on my Linix
system. :-)

Regarding a seperate posting from Linus Torvalds in reference to Tom
Grigg's patch and the create_aout_tables() function in the kernel, I
think that we shall very soon see some additional patches for iBCS.
I do not think that Linus will include the iBCS kernel portion of Tom's
patch in future versions of the kernel because of the issue dealing
with including code in the kernel that would be better left in user
space. You may reach a different conclusion after you have read Linus'
and Tom's postings in linux-kernel about iBCS, but I think I am correct.

It will be nice to have a new and stable kernel again, hopefully soon!

Cheers!

Bill

-- 
William M. Perkins                       Internet - wperkins@us.net
The Greenwood                               or    - bill@cais.com
Commodore is dead.  Long lives the Amiga!  (AmigaOS/Linux/NetBSD)