"Greg A. Woods" <woods%planix.ca@localhost> writes:
> At Fri, 05 Mar 2021 15:41:01 -0500, Greg Troxel <gdt%lexort.com@localhost> wrote:
> Subject: XEN3_DOM0 rationalization
>>
>> 4) hardware that's very crufty and maybe we don't (cardbus/pcmcia) but
>> maybe we do
>
> I would say you could get rid of much of that kind of stuff too....
>
> (but then again I only run Xen on semi-modern servers... though it might
> be fun to try using on an old MacBook Pro, etc. ala XenClient)
Where I'm coming from is that I see two styles of using Xen. One I'd
call serious use, where you find a box that is adequate to run some
domUs with a level of CPU/RAM/disk where you would actually want them to
be doing something useful. My box is towards the low end of that today:
2 E5700 CPUs, 8G RAM, 1T SSD -- basically a very nice computer from 2010
with a disk upgrade.
The other I'd call "trying it out", where someone has a computer that
will boot GENERIC amd64 that can run Xen, which means at least one CPU,
and probably at least 1G of RAM. On such a box it should be fairly easy
to install xen{kernel,tools}413, adjust boot.cfg and reboot. As part of
making the Xen learning curve less painful, I have a notion that there
shouldn't be hardware support regresssions from GENERIC to XEN3_DOM0. I
have a MBP from 2008 that I might have tried this on (Core 2 Duo, 4GB).
I guess there's a larger question of whether GENERIC should still
support cardbus/pcmcia, or whether one should neeed RETROGENERIC for
that. But I don't want to push that question. And in the end I don't
see a good reason for XEN3_DOM0 to decline to support things that work
on GENERIC.
Attachment:
signature.asc
Description: PGP signature