[gPXE] etherboot and WDS

James kojeroo at telus.net
Fri Jan 15 17:12:00 EST 2010



On 15-Jan-10 12:00 Marty Connor wrote:
> James wrote on 1/15/10 12:53 PM:
>> Sorry, I thought this was a straight forward question for a well known
>> (by the developers) problem.
> 
> James,
> 
> Thanks very much for giving the additional information you gave.
> 
> For the last several years we have focused most of our development 
> energy on gPXE as the successor to Etherboot.
> 
> We made this choice based on where we believed we could best apply our 
> development resources.
> 
> gPXE is rapidly being adopted because of its feature-set, stability, and 
> compatibility.
> 
> We have no current plans to make significant updates to Etherboot.  If 
> someone wants to fund us, or someone offers a patch that makes sense, we 
> may apply it, but we really want people to migrate to gPXE.
> 
Understood.  I would be willing to develop a patch if there are not
any fundamental infrastructure hurdles in Etherboot that really
prevents a fix.

> In the last year or two we have seen quite a few projects begin to use 
> gPXE instead of Etherboot, and we expect this trend to accelerate.  I'm 
> sure the good people of Red Hat are moving in that direction.
> 
I'm currently in a sticky situation.  I have problems adopting gPXE
because it doesn't work well with the Bochs BIOS that ships with
qemu 0.9.1.  Updating qemu to 0.12 isn't an option so that leaves me
to either figure out what it will take to patch qemu to use SeaBIOS,
fix Etherboot to work with WDS, or hack gPXE to behave more like
Etherboot when identifying qemu's Bochs BIOS as PnP capable.
Etherboot hooks in fine but gPXE uses int 19h.

> I can imagine this is not really what you wanted to hear, but it is what 
> it is.  We choose to focus our valuable development energy on gPXE, and 
> will make it as attractive an upgrade option as possible.
> 
> Perhaps someone will want to help you hack on this.  FOSS has the beauty 
> of allowing folks who are sufficiently motivated to figure out almost 
> anything.  Perhaps there is another way to attack this problem by 
> chainloading gPXE or gpxelinux.0.  Let's see what people think.
> 
I've considered chainloading but it isn't an attractive option from an
end user point of view.  I should be able to specify qemu -boot n
and just boot from existing WDS servers to install Windows.

> I hope at least that this message gives you a better understanding of 
> how things are from our end.
> 

Thanks,
James



More information about the gPXE mailing list