[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-devel] HVM guest debugging

  • To: "Keir Fraser" <Keir.Fraser@xxxxxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Guy Zana" <guy@xxxxxxxxxxxx>
  • Date: Sun, 11 Nov 2007 11:12:45 -0500
  • Delivery-date: Sun, 11 Nov 2007 08:17:23 -0800
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AcgkSmGrcubXbaPbQQCM+8zmWtqsXAAMEFdmAAApJOA=
  • Thread-topic: [Xen-devel] HVM guest debugging

Oh, I see that there is a ptrace() implementation that uses the setdebugging hypercall.
Does it really support HVMs? Is that OS dependant? What happens in Windows?
Btw, what's that condition doing in the vmexit handler? Why not reflect the exception back to the guest instead of crashing the domain?

From: Keir Fraser [mailto:Keir.Fraser@xxxxxxxxxxxx]
Sent: Sunday, November 11, 2007 5:51 PM
To: Guy Zana; xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Xen-devel] HVM guest debugging

See XEN_DOMCTL_setdebugging.

 -- Keir

On 11/11/07 10:11, "Guy Zana" <guy@xxxxxxxxxxxx> wrote:

In vmx.c, in the vmexit handler (reason = EXIT_REASON_EXCEPTION_NMI)
I found this code:

       switch ( vector )
        case TRAP_debug:
        case TRAP_int3:
            if ( !v->domain->debugger_attached )
                goto exit_and_crash;

Now, I don't understand who updates debugger_attached for a HVM guest.
(I worked with windbg in order to debug a windows HVM, and I guess EXIT_REASON_EXCEPTION_NMI is the exit reason for breakpoints I placed)

Can anyone clear this for me?


Xen-devel mailing list

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.