Re: [RFC][PATCH] modular kgdb-light (was: Re: [git pull] kgdb-light -v10)

[Andi Kleen <andi@firstfloor.org>]

On Fri, Feb 15, 2008 at 01:35:36PM +0100, Jan Kiszka wrote:
> Andi Kleen wrote:
> >> This includes things like having "breakpoint reservations" (discussed 
> >> earlier) and just generally trying to add lots of infrastructure to make 
> >> kgdb "fit in" to the kernel.
> > 
> > I think that part is actually mostly ok now (old kgdb stubs were
> > much worse in this regard) 
> > 
> > I still think the ultimative proof for this would be working
> > "modprobe kgdb" though.
> 
> To pick up this idea again I did the experimental patch below. It
> applies against Jason's latest kgdb-light patch queue:

Very nice! If it's that simple then the kgdb integration is really
clean.

> 
> http://git.kernel.org/?p=linux/kernel/git/jwessel/linux-2.6-kgdb.git;a=shortlog;h=for_ingo
> 
> The patch nicely demonstrates what deeper dependencies on kernel
> services currently exist in kgdb-light. The following symbols were
> unresolvable:
> 

Some comments on the specific patch:

>  o genapic - for send_IPI_allbutself, ie. CPU roundup

I would rather export some generic wrapper for that than the full genapic
structure.

>  o machine_emergency_restart - for implementing "R0" gdb packet

Hmm, might be a bit dangerous to call this directly -- there are various
quirks with e.g. not rebooting on CPU #0 and not resetting APIC 
state. But ok [this is not directly related to the fact that it's
exported now, just mentioning this in general]

-Andi