Re: [Crash-utility] crash and libvirt, and more
Richard W.M. Jones wrote:
On Tue, Aug 19, 2008 at 11:06:49AM +0100, Bryn M. Reeves wrote:
> Sounds like another possible use for the "debuginfo-server" idea that
> comes up periodically. There are a couple of things that could benefit
> from a centralised way of managing a collection of kernel (or other
> package) debug data.
Sounds interesting.
What I was actually trying (not successfully yet) was to see if I
could make a fake binary which just contains debug info for the few
kernel structures I care about. Starting with task_struct as an
example. Then it should be a simple matter of loading those symbols
using 'symbol-file'.
These "fake binaries" should be quite a lot smaller than an entire
vmlinux.
As I said I haven't actually made this work yet.
Also working on a virDomainMemoryPeek patch for crash.
Rich.
Out of curiousity, any reason why a libvirt interface couldn't be
created that accesses guest pseudo-physical addresses? And does
the existing interface accept vmalloc addresses, or only unity-mapped
kernel virtual addresses?
Dave