[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] General questions on IA64 porting

> 1) The IA64 porting in Xen seems to be stopped or sleeping.
> Q: What is the status of Xen-IA64 development ?

It's still alive.  Xen/IA64 is Dan Magenheimer's baby and he's been mostly 
working on it alone so far.  I believe some people from Intel were looking to 
contribute too.  I've been able to contribute a little time but I have other 
things demanding my attention also.

Xen/IA64 currently boots into a single dom0.  No serious optimisation has been 
done yet.

> 2) In the Xen Users' manual, it is sayed "a port to the IA64 architecture
> is approching completion". It's seems to be optimistic.
> Q: When IA64 support will be officially available ?

That'll probably depend on how many people end up working on it.  It'll be 
months, at least, I'd speculate.

> 3) Several limitations exist on Xen-IA32 (memory <= 4GB, no TLS support,
> problem with AGP graphic cards, etc...).
> Q: Is these limitations also apply on Xen-IA64 ?

I think Dan had his AGP graphics card working OK.

The memory limitation on Xen/IA32 is purely because nobody has wanted to add 
PAE support (we think that IA32 machines with >4gig are relatively rare and 
that supporting x86_64 is a more rewarding route to large memory systems).  
On IA64 this issue doesn't apply.

The TLS support problem on IA32 is an artifact of the strange way TLS is 
implemented on that architecture.  Again, I don't expect that it's an issue 
on IA64.

> 4) Xen-IA32 supports SMP guest OSes.
> Q: Is it true also on IA64 ?

Not yet.  I'm not sure if it'll be worth the effort to support this to start 
with, although it will be a goal to support this eventually.

> 5) The Xen architecture is planned to be integrated in Linux sources.
> Q: Is this integration also applies to Xen-IA64 and at the same time
> than Xen-IA32 or later ?

Definitely later.  The IA32 code is ready to be merged pretty soon and Andrew 
Morton has been making very encouraging noises about it.  The IA64 code won't 
be ready by then.

The IA64 modifications should be much smaller, so they may be possible to 
integrate into the arch/ia64 tree itself.

> 6) The Linux distributions for IA64 (ie: RHEL-3) are based on Linux 2.4
> kernel.
> Q: So why IA64 porting is based on Linux 2.6 kernel ? What is the reason ?

I can't speak for Dan on this but I imagine because it's the kernel of the 
future - RHEL 4 will be based on the 2.6 kernel.  If Xen/IA64 gets integrated 
into a future RHEL, it'll be based on 2.6 anyhow.

> 7) The RHEL (Red Hat Entreprise Linux) license doesn't permits to modify
> or rebuilt the Linux kernel without losing the Red Hat support.
> Q: So in case of problems on a customer site, who is responsible ?
> Is Xen will provide a direct or indirect support in this case ?

Well...  I guess future releases of RHEL may integrate Xen support (things 
seems to be headed that way for IA32), in which case you won't need to 
replace your kernel to get it.  Xen/IA64 isn't ready for production use yet, 

There'd be some support available on this mailing list and XenSource might be 
able to provide some additional support guarantees (regarding the Xen side of 
things, anyhow).  I'm doubtful that having paid for a RHEL license you'd want 
to invalidate the Red Hat support contract, however.

> 8) How can I help you in the IA64 porting and/or testing ?

Dan can probably provide more useful direction here.  The first step would be 
to get IA64 booting on your machine (the mailing list archives contain 
various useful discussions and links regarding this).


This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl
Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.