Messages in this thread | | | From | Ian Campbell <> | Date | Tue, 04 Mar 2008 15:12:57 +0000 | Subject | Re: [RFC] use realmode code to reserve end-of-conventional-memory to 1MB |
| |
On Tue, 2008-03-04 at 15:33 +0100, Ingo Molnar wrote: > * Mark McLoughlin <markmc@redhat.com> wrote: > > > > The patch below seems like the right thing to do. It certainly boots > > > in a domU without the DMI problem (without any of the other related > > > patches such as Alexander's). > > > > Yep, this patch on its own fixes 2.6.25-rc3 DomU boot for me here. > > thanks - i picked it up for v2.6.25 merging.
Thanks, I'd been planning to resend once I got the ddcprobe thing sorted but I've been a bit snowed under (deadlines :-|). This patch is independently correct though and turns the ddcprobe failure from a domU kernel crash into a userspace hang.
ddcprobe hangs because the page at pfn 0 is comprised entirely 0xc2 bytes (the Xen memory scrub poison value -- I'm running a debug hypervisor) which causes lrmi to try and jump to 0xc2c2*16+0xc2c2 to implement int $0xXX. gdb won't let me actually examine that address for some reason so I don't actually know what it's hitting to cause the loop/hang but no good can come of jumping to that address in any case...
Sanest solution seems to me to be to explicitly zero PFN 0 -- I have a patch I just haven't had a chance to clean it up for submission yet.
> i'm wondering, what triggered this bug, and why didnt we have these > problems in the past?
Did someone make the boot allocator start at the bottom instead of the top of memory or something?
Ian.
-- Ian Campbell Current Noise: Nile - Even The Gods Must Die
There is a natural hootchy-kootchy to a goldfish. -- Walt Disney
| |