Karmic/Lucid i386 virtual flavour
John Johansen
john.johansen at canonical.com
Tue Feb 9 18:44:17 UTC 2010
Andy Whitcroft wrote:
> There has been some discussion about XEN not working with the 32bit kernels
> in both Karmic and Lucid it seems that when we moved from having a server
> flavour to having a generic-pae flavour in 32bit we moved from M686 to
> M586 for that PAE based kernel. XEN dom-U now does not work because
> it is dependant on TSC support in the processor. Note that the Jaunty
> server flavour config was CONFIG_M686 which includes TSC automatically.
> It does appear from Wikipedia that PAE support comes after TSC support in
> the Pentium timeline and therefore we may well be able to enable TSC and
> XEN by switching the generic-pae config over to CONFIG_M586TSC mode, as
> to have PAE capabilities you must also have a TSC. M586TSC is basically
> just M586 (what we have now) plus TSC.
>
> For TSC:
> The Time Stamp Counter is a 64-bit register present on all x86
> processors since the Pentium. It counts the number of ticks
> since reset. Instruction RDTSC returns the TSC in EDX:EAX. Its
> opcode is 0F 31.[1] Pentium competitors such as the Cyrix 6x86
> did not always have a TSC and may consider RDTSC an illegal
> instruction. Cyrix included a Time Stamp Counter in their MII
>
> For PAE:
> In computing, Physical Address Extension (PAE) is a feature of
> some x86 and x86-64 processors that enables the use of more than
> 4 gigabytes[1] of physical memory to be used in 32-bit systems
> - given appropriate operating system support. PAE is provided
> by Intel Pentium Pro (and above) CPUs - including all later
> Pentium-series processors except the 400 MHz bus versions of the
> Pentium M, as well as by other processors such as the AMD Athlon
> and later AMD processor models with similar or more advanced
> versions of the same architecture.
>
> I think this should be safe for Lucid, Karmic may be a little more
> difficult. I propose we make the change in Lucid and get some miles on
> it in the Alphas before we decide on Karmic.
>
> Comments?
>
Trying this again has been on my todo list, as I would love to kill the
EC2 topic branch. I will just note that this was what was initially tried
with Karmic but ended up failing in 1 out of 4 zones on EC2.
More information about the kernel-team
mailing list