[Xen-devel] [PATCH 0/2] extract vmcoreinfo from /proc/vmcore for Xen

Keir Fraser keir.fraser at eu.citrix.com
Thu Jun 12 10:36:33 EDT 2008


I'm a bit confused as to what this is for. If you know where Xen is located
in physical or virtual address space, and you have access to Xen symbols,
why would you need the few symbol addresses provided by vmcoreinfo? I can
hardly see what symbols dom_xen and dom_io would be particularly useful for.
Similarly, why would you record structure sizes when this is statically
determinable given the Xen sources?

 -- Keir

On 12/6/08 09:22, "Itsuro ODA" <oda at valinux.co.jp> wrote:

> Hi,
> 
> The following patchset is to support extracting vmcoreinfo from
> /proc/vmcore for Xen.
> 
> This is the up-port of the patchset I made for 3.2.0 and posted
> the end of March.
> 
> [PATCH 1/2] patch for xen hypervisor (xen-unstable.hg)
> [PATCH 2/2] patch for linux (linux-2.6.18-xen.hg)
> 
> --- background ----------------------------------------------------
> * what the makedumpfile is:
>   To shorten the size of the dumpfile and the time of creating the
>   dumpfile, makedumpfile copies only the necessary pages for analysis
>   to the dumpfile from /proc/vmcore. You can specify the kind of
>   unnecessary pages with dump_level. If you want to shorten the size
>   further, enable the compression of the page data.
> * Xen extraction 
>   makedumpfile also has a feature that extracting the part of
>   xen hypervisor and domain-0 from /proc/vmcore taken under xen
>   environment. In most cases under xen environment, it is enough to
>   analyze a cause of system crash.
> * extract vmcoreinfo from /proc/vmcore
>   makedumpfile 1.2.0 supports to extract the vmcoreinfo data from
>   /proc/vmcore and uses it for dump filtering. (for a vmcore taken
>   under the native environment.)
> ------------------------------------------------------------------
> 
> Thanks.





More information about the kexec mailing list