Messages in this thread | | | Date | Sun, 01 Mar 2009 17:16:39 -0800 | From | Jeremy Fitzhardinge <> | Subject | Re: brk patches.. |
| |
H. Peter Anvin wrote: > If the brk is bounded-size on the same order or smaller as the kernel, > we should just mark it as an unallocated (bss) section in the ELF image > and be done with it... there really is no point in trying to be smarter > (we'd be subject to failures to load the kernel proper.) If the brk is > significantly bigger, then yes, we need to be smarter. However, that is > not my current understanding of the requirements. >
Yes, right. And in my case I actually need it to generate an appropriate e820 table, so adding a dependency on e820 would be circular...
(To be specific: I reshape the guest e820 table so that it doesn't have memory in any forbidden areas of the host e820 table. That may require moving the pseudo-physical address of pages into a new overflow e820 entry, which would also require allocating pages for the p2m radix tree.)
J
| |