pc386 i386ex Rtems Make issues?

[Joel Sherrill]

John Jeffers wrote:
> 
> Hello
> 

> configure i386-rtems  ..... i386ex was issued.  This completed as expected.
> 
> make RTEMS_BSP="pc386 i386ex"

I assume you did --enable-rtemsbsp="pc386 i386ex" on the configure line
also.
 
> resulted in a number of issues (I put stdio to a log file but stderr did
> not go)

Try this :)

make >logfile 2>&1 &
tail -f logfile

> 1) I used the posted i386 Linux RPM's which built out to /opt/rtems on
> RH6.2 and have not tried to fix for 386 soft float issues but I did not
> expect to encounter them on a OS build.  Is this really the issue with the
> following or do I have Other finger problems?
> 
>   i386ex/lib/linkends:115:parse error o-optimize/hello.exe
> 
>   resulting in premature termination: exit error 1

Somehow this typo crept in to the i386ex linkcmds.  This is a specific
problem with this file and this BSP.  It is not the soft FP problem.
That only impacts the ts386ex BSP.

Index: startup/linkcmds
===================================================================
RCS file:
/usr1/CVS/rtems/c/src/lib/libbsp/i386/i386ex/startup/linkcmds,v
retrieving revision 1.10
retrieving revision 1.11
diff -u -r1.10 -r1.11
--- startup/linkcmds	1999/11/17 17:50:46	1.10
+++ startup/linkcmds	2000/06/14 13:08:41	1.11
@@ -9,7 +9,7 @@
  *  found in the file LICENSE in this distribution or at
  *  http://www.OARcorp.com/rtems/license.html.
  *
- *  $Id: linkcmds,v 1.10 1999/11/17 17:50:46 joel Exp $
+ *  $Id: linkcmds,v 1.11 2000/06/14 13:08:41 joel Exp $
  * 
  * Memory layout:
  *
@@ -112,7 +112,7 @@
 	*(.dtors)
 	LONG(0)
 	__DTOR_END__ = .;
-	_rodata_start 5 . ;
+	_rodata_start = . ;
 	*(.rodata)
 	*(.gnu.linkonce.r*)
 	_erodata = ALIGN( 0x10 ) ;

> 
> (As a note the expected termination could be in the "Getting Started with
> RTEMS for C/C++ Users" eg.Results in ....... Use Make all etc. for Config
> result this would result in less head scratching for newbies at the cost of
> ether a line of print each,  alternately terminating each master script or
> makefile with "Success :<filename>" as a comment has worked for me.)

Aaaaa... the normal output of bit_rtems is included but not a 
regular old configure and make.. good suggestion.
 
> 2) Is this expected?
> 
>   rtems-4.5.0-beta3a/c/src/exec/score/cpu/i386/cpu_asm and
> rtems.s  from   asm.h:52
>   warning undefining 'Register_PREFIX"

Expected.  gcc was/is wrong about its value.  __REGISTER_PREFIX__
is set to nothing on this gcc port and it should be "%".

> Which also occurred in:
> 
>   libcpu                cpu_asm.S, cpu_Model.s, start.S, irq_asm.S:14, time_isr.S:34

All assembly files get this because of above. 
 
> Also the following BSP issues
> 
>   ne2000.c      issued          'ne_init_hardware'
>   ne2000.c:446                  'warning assignment from incompaitable pointer type'

Not ideal but not a problem.

>   uipc_mbuf.c:53                        'mbinit static not defined'
>   linkaddr.c:63                 '%ax' instead of %eax due to 'w' suffix'

I don't see either of these in my last build logs. Could you
doublecheck them?
 
> 3) Am I supposed to put the tarred examples in any specific place before
> the Make.  The Getting Started does not seem to indicate this and in
> Section 6 alludes to there not being any at this stage?

No.  

But note that they probably do not define enough tasks since
the minimums were changed.

> Thanks John

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel at OARcorp.com                 On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
   Support Available             (256) 722-9985