[WBEL-devel] WBEL4 RC1 on VIA C3 CPU

Steve Phillips stevep at lnxrules.com
Thu Apr 28 21:50:16 CDT 2005


On Thu, 2005-04-28 at 21:01 -0500, John Morris wrote:
> On Wed, 2005-04-27 at 12:26, Johnny Hughes wrote:
> 
> > The RHEL4 kernel doesn't support i586 ... I don't think that John modified
> > the WBEL4-RC1 boot kernel to support i586 either. (I could be wrong).
> > 
> > If WBEL4 is not going to support i586, you might try CentOS-4, as the boot
> > kernel for anaconda is i586 and there is i586 support.
> 
> Oh thanks for the reminder!  I had noticed that issue while spinning an
> internal test release on Monday and then went off to give a talk about
> wireless in libraries and would have forgot it.  Heck, had forgot about
> it when replying to a message just a couple of minutes ago.
> 
> Boot kernels.  I want WBEL4 to be able to use vendor released driver
> discs for RHEL4.  When it builds the boot media it looks for
> kernel-BOOT, then i386, i586 and finally will take i686 if it is the
> only thing available.  For RHEL4 that is going to be i686 so any driver
> discs are going to be built for i686.
> 
> So we have a problem.  Looks like it is going to be possible to have
> Pentium support OR RHEL4 driver compatibility at install time but not
> both.  Or is there some way out of this paradox I'm just not seeing?
> 
> > Before anyone complains about CentOS trying to steal WBEL users ... I am
> > not.  If WBEL-4 supports i586 and your are a WBEL user, please use WBEL-4
> > :).  If not, I just wanted to point out that CentOS-4 might allow you to
> > install on that platform.
> 
> Actually it might be THE option for i586 support.  Since CentOS does
> exist and supports Pentium class hardware it might make more sense for
> WBEL to go for driver compatibility.  Migration between rebuilds is
> pretty simple so nobody gets orphaned without an upgrade path.
> 
> Time for a fast debate folks.  Survivor time, who gets voted off the
> island, Epias or Enterprise hardware? The rebuild process is finally at
> a stage where I think all packages (on both arches) are going to build
> flawlessly on the pass currently running so it is this decision and
> installing Oracle to get tora right that are the only remaining holds.
> 
> Assuming of course the compile doesn't hit yet another snag.  And of
> course I probably should go ahead and roll in the OO.o errata from this
> week to avoid THAT hog of a download after every install.
> 
Hi John,
Don't really see the need for a debate... 

>From http://www.whiteboxlinux.org/faq.html -

items 2 and 3:

"Why didn't you build all of the packages for the i686/athlon/etc. 

Short answer: Because RedHat didn't. WBEL aims to be 100% binary
compatible with RHEL which means no tweaking.

<SNIP>

  * Why didn't/won't you include package foo 

Short answer: Because RedHat didn't include it.

<SNIP>

Those are the primary reasons I chose to stick with Whitebox. So far
every issue I've ever had was reproducible on RHEL. This means I can
have a discussion about a project, toss WBEL on a server for proof of
concept, and be pretty sure that if the project proceeds to production
on RHEL - it'll behave as expected. Saves me a lot of frustration as
well as a 45-60 day purchase cycle up front. I'm sure there are many
others in the same boat. The only reason that I know of to vary off the
stated path is if it doesn't meet the needs of the library, and as you
pointed out, Centos may fill the need to support additional
hardware/features.

Thanks for all the hard work!
stevep

BTW: Any ETA for WBEL4 (as in days, week, months)?


More information about the Whitebox-devel mailing list