Re: ext/debuggers.hg

[Mukesh Rathor <mukesh.rathor@oracle.com>]

Alexey Tumanov wrote:
> Hi Mukesh,
> 
> I've just discovered your ongoing work on ext/debuggers.hg and had a
> few quick questions. First of all, though, thanks a lot for doing
> this. I personally, can't work without an ability to step through the
> code with a debugger, esp. when learning a new part of the system.

  Same here.

> q1: has your work been merged to xen-unstable or not? I see you've
> kept a separate tree under ext/debuggers.hg and kept it relatively
> in-sync with xen-unstable up until c/set 19569.

  Not yet, hopefully soon tho.

> q2: have you considered maintaining a mercurial patch queue on top of
> xen-unstable as opposed to a separate tree? That would make it easier
> to look at the diff you've contributed

Limited on time. Most people preferred external tree.
Creating a patch is easy tho.

   - hg clone ext/debugger.hg
   - hg clone xen-unstable -r 19569 or whatever debugger is synced to.
   - diff the two
   - apply the patch to any c/s of unstable

> q3: (related to q2 & q1) I'm not entirely sure whether I can get
> built-in kdb support when I build vanilla xen-unstable. Basically, can
> I get some sort of debugging support for the hypervisor/dom0 from
> xen-unstable.hg and what does it take (make debug=y ?)

apply the diff from q2 to unstable, for the most part it should apply
cleanly. Please note, debug=y is not supported. Please start with kdb/README.

> q4: is there a differerence between the gdb debugging support in your
> tree compared to xen-unstable tools/debugger/gdb?

i'm not sure what state that is, but kdb is targeted for xen and dom0
debug. you may experience hang with dom0 single step, i've fixed it.
i'll upload it soon as soon as i get over this stupid flu.
Email me if you want the patch.

> 
> Thanks a lot for reading this, and keep up the great work you've done
> easing the learning curve for those who are just starting out with Xen
> dev!!
> 
> Alex.


take care,
Mukesh