From mboxrd@z Thu Jan  1 00:00:00 1970
From: Hollis Blanchard <hollisb@us.ibm.com>
Date: Mon, 09 Jun 2008 14:57:26 +0000
Subject: Re: [PATCH 2 of 2] Add gdb break point support to PowerPC kvm
Message-Id: <1213023446.3358.5.camel@localhost.localdomain>
List-Id: <kvm-ppc.vger.kernel.org>
References: <d917195fbef25227967d.1212771276@thinkpadL>
In-Reply-To: <d917195fbef25227967d.1212771276@thinkpadL>
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
To: kvm-ppc@vger.kernel.org

On Fri, 2008-06-06 at 15:23 -0500, Jerone Young wrote:
>=20
> > > > > +       case BOOKE_INTERRUPT_DEBUG: {
> > > > > +               u32 dbsr;
> > > > > +               struct kvm_guest_debug *dbg > &vcpu->guest_debug;
> > > > > +
> > > > > +               vcpu->arch.pc =3D mfspr(SPRN_CSRR0);
> > > > > +
> > > > > +               /* clear events in DBSR register */
> > > > > +               dbsr =3D mfspr(SPRN_DBSR);
> > > > > +
> > > > > +               if (vcpu->arch.pc =3D dbg->bp[0])
> > > > > +                       dbsr |=3D 1<<23; /* clear IAC1 event */
> > > > > +               if (vcpu->arch.pc =3D dbg->bp[1])
> > > > > +                       dbsr |=3D 1<<22; /* clear IAC2 event */
> > > > > +               if (vcpu->arch.pc =3D dbg->bp[2])
> > > > > +                       dbsr |=3D 1<<21; /* clear IAC3 event */
> > > > > +               if (vcpu->arch.pc =3D dbg->bp[3])
> > > > > +                       dbsr |=3D 1<<20; /* clear IAC4 event */
> > > >=20
> > > > This doesn't seem like the right thing to do here. What if the
> > > > breakpoint isn't due to any of the IACs?
> > >=20
> > > If it is an IAC register that triggered the event then you have to
> clear
> > > it. The exact event in the DBSR.
> >=20
> > That's my point: why are you using dbg->bp[] instead of DBSR bits?
> If
> > anything, I would expect an error message if unexpected DBSR bits
> were
> > set, and in any case:
> >       dbsr =3D mfspr(=EF=BB=BFSPRN_DBSR);
> >       mtspr(SPRN_DBSR, dbsr);
>=20
> Oh the issue is if 2 debug interrupts got off from two different
> registers. At least that was my train of the thought. Though the fact
> if
> that does happen that would mean they both had the exact same
> breakpoint. So we would only wat to process one of them.=20

That may be your train of thought, but it's not mine. Please reread my
comments and let me know if you still have questions.

--=20
Hollis Blanchard
IBM Linux Technology Center