Kernel using LZMA compression

Bryn M. Reeves bmr at redhat.com
Wed Nov 4 18:24:17 UTC 2009


On 11/04/2009 06:18 PM, Ikem Krueger wrote:
>> The executive summary is: Xen does not let a kernel boot itself,
>> because mimicking bare hardware is too tedious (and pointless.)
>> Instead, Xen instantiates an instance of a kernel into the Xen
>> environment.  To do this instantiation, Xen does its own
>> decompression, so Xen must know everything about the compression.
>
> I know you're right. But that sound stupid to me: The kernel itself
> has routines built-in for decompression. Why isn't it enough to let
> Xen use the same routines for decompression as the kernel?
>

I am reading between the lines here (I have never looked at this stuff 
in Xen) but I would assume it's for the reason given above. The kernel's 
own decompression routines must run very early on in the boot process - 
well before the first line of C code runs and while the CPU (on x86) is 
still running in legacy real addressing mode (right after the handover 
from the bootloader and relocation of the kernel image).

It's emulating this early-boot environment that is tedious and pointless 
and being able to use the in-kernel decompresser is not sufficient 
motivation to go down that route.

Regards,
Bryn.




More information about the devel mailing list