FreeBSD and RTEMS

Joel Sherrill joel.sherrill at OARcorp.com
Sat Jun 10 21:34:53 UTC 2000


James Housley wrote:
> 
> Joel Sherrill wrote:
> >
> > This is fabulous news!!  It must have gone pretty smoothly since it
> > did not take you much time.
> Not too long.  A faster disk sub-system would help, but DUAL CPUs help a
> lot.

The RPMs are built on an old Pentium Pro running Redhat 5.2 for
for glibc compatibility reasons.  It has a pretty fast disk subsystem
and the builds are not THAT bad.  

OAR's Solaris machine is an SS5.. It took 43 hours to build JUST the
gcc/newlib tools on it.

> > Not being familiar with FreeBSD packaging, I have a few questions:
> >
> > + Is there any revision number like on the RPMs to indicate which
> >   patches went with them?
> I hove no knowledge of RPM so you have to tell me more first.  But
> normally there isn't an adational version attached besides that of the
> sources.

RPMs normally associate a -REVISION on the RPM to indicate patch
level.  So you end up with something like this:

sh-rtems-gdb-4.18-4.i386.rpm

indicating that this is the fourth round of RPMS for sh-rtems
on gdb 4.18.  For gdb 5.0, we would reset to revision 1.

> > + Do your gcc .tgz's have all languages or just C/C++?  The other
> >   hosts are packaged to avoid making people download and install
> >   things you don't need.
> Right now just C/C++.  And I am re-building them all and verifing them
> now.  I was up working till mid-night, also cause my daughter was sick.

I am sorry to hear that.  I have four kids.  When they are sick it 
really disrupts the normal routine of life.

> > + How large did these end up being?  The other binaries are about
> >   165-200 Mbytes for all binaries in a single host format for all
> >   targets and languages.
> They look like 2-5Meg with docs for each.  binutils are one and
> gcc+newlib are the other.  Compiled with -O2 and no debug.

This seems too small... what is in them?  I have a size report for
i386-rtems RPMs here:

File sizes in bytes:

5039708 i386-rtems-binutils-2.10-1.i386.rpm
1745208 i386-rtems-chill-gcc2.95.2newlib1.8.2-7.i386.rpm
2026825 i386-rtems-g77-gcc2.95.2newlib1.8.2-7.i386.rpm
4963383 i386-rtems-gcc-gcc2.95.2newlib1.8.2-7.i386.rpm
1671126 i386-rtems-gcj-gcc2.95.2newlib1.8.2-7.i386.rpm
1376081 i386-rtems-gdb-4.18-4.i386.rpm
1712817 i386-rtems-objc-gcc2.95.2newlib1.8.2-7.i386.rpm
1297828 rtems-base-binutils-2.10-1.i386.rpm
  14290 rtems-base-chill-gcc2.95.2newlib1.8.2-7.i386.rpm
 318913 rtems-base-g77-gcc2.95.2newlib1.8.2-7.i386.rpm
 841482 rtems-base-gcc-gcc2.95.2newlib1.8.2-7.i386.rpm
 119147 rtems-base-gcj-gcc2.95.2newlib1.8.2-7.i386.rpm
 181182 rtems-base-gdb-4.18-4.i386.rpm

Install Size Summary for i386-rtems (in kilobytes)
  Binutils:             27221.31
  GCC (C/C++ Only):     21218.11
  GDB:                   4605.16
  Chill:                 5321.60
  G77:                   5862.54
  GCJ:                   4526.03
  Objective-C:           4771.90

  Total C/C++ Toolset without GDB:                 48439.42
  Total C/C++ Toolset with GDB:                    53044.58
  Total Complete Toolset (All Languages and GDB):  73526.65

Out of curiousity, is this a good way to present the installed
sizes?

> 
> >
> > + Is gdb proving hard to build or have you just not gotten to it yet?
> Haven't tried yet.

I am curious how it goes.  I think I worked the issues out building on
the 3 OSes I did.

> > > Jim
> > > --
> > > Nothing is fool proof, because fools are too ingenious.
> >
> > --
> > 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
> 
> --
> Nothing is fool proof, because fools are too ingenious.

-- 
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



More information about the users mailing list