Experimental Physics and
Industrial Control System

[email protected] · Mon, 26 Jul 2004 23:32:39 +0900

Hello Kate,

>RTEMS-4.6.0 is better than  RTEMS-4.6.0pre#.   RTEMS-4.6.0 has
>PowerPC patches  for RTEMS-4.6.0pre#.
He tried to recompile RTEMS-4.6.0 kernels and GeSys_2.1beta codes.
All passed before step what for build cexp : and xsyms :

>To build rtems5500-cexp.bin and rtems5500-cexp.sym :
>0) cd RTEMS5500-GeSys_2.1.beta
>1) make

Did you used bspcheck commands at Makefile line 142.
But when i type make commands after other error messages from Makefile line 142..
What is missing ?

root@rtemsdev RTEMS5500-GeSys_2.1.beta]# make
Makefile:142: *** GeSys has not been ported/tested on this BSP (mvme5500) yet.

Makefile: 142 Line :
all: bspcheck libnms ${ARCH} $(SRCS) $(PGMS)

Regards,
Kiman,
PLS control and taskforce team.

> 
> 
> -----Original message-----
> From: "KateFeng" 
> To: "KimanHa" 
> Cc: [email protected],[email protected]
> Date: 07/17/2004(Sat)01:32
> Subject: Re: RTEMSfor MVME5500
> 
> Kiman Ha wrote:
> 
> >  Hello Kate,
> >
> > Anyone have to tried compile RTEMS for MVME5500?
> >
> > Our Control staffs who Mr OH is recently trying to compile RTEM for
> > MVME5500.
> >
> > But he reported to me some compile error messages
> >
> > Used : RTEMS-4.6.0pre1
> 
> RTEMS-4.6.0 is better than  RTEMS-4.6.0pre#.   RTEMS-4.6.0 has
> PowerPC patches  for RTEMS-4.6.0pre#.  However, the patches
> does not seem to be related to the building error you got.
> 
> > But the compile error is occurs this step;
> >
> > To build rtems5500-cexp.bin and rtems5500-cexp.sym :
> > 0) cd RTEMS5500-GeSys_2.1.beta
> > 2) make
> >  ldep/ldep.c:659:2: warning: #warning TODO weak symbols
> >
> > Warning: Symbol without object file??
> >
> > -> substituting symbol file name, using 'startfiles.o'...
> > (o-optimize/startfiles.nm/line 1)
> >
> > Warning: type mismatch between multiply defined symbols
> >
> >          __deregister_frame_info: known as T, is now w
> >
> > Warning: type mismatch between multiply defined symbols
> >
> >          __register_frame_info: known as T, is now w
> >
> > Warning: type mismatch between multiply defined symbols
> >
> >          environ: known as G, is now B
> >
> > Warning: Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /libbfd.a[archive.o]' couldn't be removed; probably it's needed by the
> > application
> >
> > Warning: Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /libbfd.a[archures.o]' couldn't be removed; probably it's needed by the
> > application
> >
> > Warning: Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /libbfd.a[bfd.o]'
> >
> > :
> >
> > :
> >
> > :
> >
> > Warning: Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /libiberty.a[xexit.o]' couldn't be removed; probably it's needed by the
> > application
> >
> > Warning: Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /libiberty.a[xmalloc.o]' couldn't be removed; probably it's needed by the
> > application
> >
> > Warning: Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /libiberty.a[xstrerror.o]' couldn't be removed; probably it's needed by the
> > application
> >
> > Warning: Object 'librtemscpu.a[dummy.o]' couldn't be removed; probably it's
> > needed by the application
> >
> > Warning: Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /libiberty.a[xstrerror.o]' couldn't be removed; probably it's needed by the
> > application
> >
> > Warning: Object 'librtemscpu.a[dummy.o]' couldn't be removed; probably it's
> > needed by the application
> >
> 
> I got all these warning message as well except the directory is
> '/work/rtems/ppc-rtems//lib/libxxx.a'  instead.  Where  was  your  cexp
> library copied to ?    Theoretically, 'make install' should do the job.
> However, sometimes one might have to manually install some or all of
> the libraries to  /opt/rtems/ppc-rtems/lib and the header files to
> /opt/rtems/include, which is mentioned at the end of INSTALL.
> 
> >
> > Object
> > '/work/rtems/rtems-ppc/lib/gcc-lib/ppc-rtems/3.2.1/../../../../ppc-rtems/lib
> > /librtemsbsp.a[ftpd.o]' not found!
> 
> As well, I got  :
> Warning: Object '/opt/rtems3/ppc-rtems//lib/libiberty.a[strncmp.o]'
> couldn't be removed; probably it's needed by the application
> Object 'librtemsbsp.a[ftpd.o]' not found!
> 
> However, the 'make' continued as what  is attched (GeSys.log)
> Is  librtemsbsp.a installed ?   He might have missed
> one step in installing the  BSP libraries.  Thesy should be in
> /opt/rtems/ppc-rtems/mvme5500/lib directory.  The
> mvme5500Lib.tar.gz should be tared  into /opt/rtems/ppc-rtem.
> I updated the README to cover some more details (attached).
> 
> >
> >
> > --- end of error message ---
> >
> > When are you going to release the BSP source code for MVME5500?
> 
> We followed the advice of the legal office to submitt  a  letter for approval
> of releasing the source code, which hopefully  will  be approved soon.  It is
> a standard procedure that we never  did  before.  Thanks for your patience.
> 
> 
> Some VME modules (e.g. Oms58 driver) might require  PCI sync
> command if mvme5500 is used as the SBC.  Please reference the drvOMS58.cc
> file that is posted in synAppRTEMS.  Even so,  I appreciated the runtime
> performance of RTEMS-mvme5500.   It's faster in initilaizing the same amount
> of motors as compared with mv2307.  The  MEDM screen  for the mv5500 IOC
> startups so much faster than  that of the mv2307  when I needed it for
> 32 motors.
> 
> Regards,
> Kate
> 
> 
> 
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System