Hi Dave,
> Have you abandoned the idea of adding an ELF note (or some
other
> mechanism) for passing the "pfn_to_mfn_frame_list_list" mfn value
> from the shared, per-domain, "arch_shared_info" structure(s)?
I talked with our kdump patch managers. They don't want to change
kernel level codes.. I'm sorry for I didn't tell this to you. I'll
talk with them again.
> As we proved back in June with the xen vmcore that you sent me,
> there would be no need for dom0cut since the xen vmcore could be
> used unmodified.
Now we think about partial (not whole-machine, but only a specific
domain or xen hypervisor) dump for large memory machines (mainly to
shorten dump taking time). dom0cut will read /proc/vmcore and write
an image file.
Furthermore -- if you could pass the
"pfn_to_mfn_frame_list_list" mfn
values for all existing domains, it would be possible to bring up
crash sessions for dom0, as well as for any of the domU's -- all
with a single unmodified vmcore.
It sounds good.
Also, I'm curious, why doesn't the unmodified vmcore work
with
gdb?
vmcore image's virtual address space is different from xen's one.
Because kdump kernel uses different address space from xen, and it
doesn't know anything about xen.
We know kdump kernel with some tricks can make gdb-enabled vmcore
image, but smaller patch is better to be applied.
regards,
--
Kazuo Moriwaka <moriwaka(a)valinux.co.jp>