From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:34486) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XlvbT-0004FZ-FU for qemu-devel@nongnu.org; Wed, 05 Nov 2014 03:06:14 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XlvbL-00020m-Vx for qemu-devel@nongnu.org; Wed, 05 Nov 2014 03:06:07 -0500 Message-ID: <5459DA66.6070602@suse.de> Date: Wed, 05 Nov 2014 09:05:58 +0100 From: Alexander Graf MIME-Version: 1.0 References: <1415168221-2324-1-git-send-email-sam.mj@au1.ibm.com> <1415168221-2324-3-git-send-email-sam.mj@au1.ibm.com> In-Reply-To: <1415168221-2324-3-git-send-email-sam.mj@au1.ibm.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [Qemu-ppc] [PATCH 2/2] spapr: Fix stale HTAB during live migration (TCG) List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Samuel Mendoza-Jonas , qemu-ppc@nongnu.org, qemu-devel@nongnu.org On 05.11.14 07:17, Samuel Mendoza-Jonas wrote: > If a TCG guest reboots during a running migration HTAB entries are not > marked dirty, and the destination boots with an invalid HTAB. >=20 > When a reboot occurs reset the state of HTAB migration, and explicitly > inform the destination of invalid entries. >=20 > Signed-off-by: Samuel Mendoza-Jonas > --- > hw/ppc/spapr.c | 59 +++++++++++++++++++++++++++++++++++-------= -------- > include/hw/ppc/spapr.h | 1 + > 2 files changed, 42 insertions(+), 18 deletions(-) >=20 > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > index 1610c28..9f419e8 100644 > --- a/hw/ppc/spapr.c > +++ b/hw/ppc/spapr.c > @@ -829,26 +829,30 @@ static void spapr_reset_htab(sPAPREnvironment *sp= apr) > =20 > shift =3D kvmppc_reset_htab(spapr->htab_shift); > =20 > + pthread_mutex_lock(&spapr->htab_mutex); > if (shift > 0) { > /* Kernel handles htab, we don't need to allocate one */ > spapr->htab_shift =3D shift; > kvmppc_kern_htab =3D true; > =20 > /* Tell readers to update their file descriptor */ > - pthread_mutex_lock(&spapr->htab_mutex); > if (spapr->htab_fd > 0) { > spapr->htab_fd_stale =3D true; > } > - pthread_mutex_unlock(&spapr->htab_mutex); > } else { > if (!spapr->htab) { > /* Allocate an htab if we don't yet have one */ > spapr->htab =3D qemu_memalign(HTAB_SIZE(spapr), HTAB_SIZE(= spapr)); > + } else { > + spapr->htab_mig_full =3D true; > + spapr->htab_first_pass =3D true; > + spapr->htab_save_index =3D 0; You could just set the dirty bitmap to "all dirty" here, no? Then you don't need all the changes belong I presume? > } > =20 > /* And clear it */ > memset(spapr->htab, 0, HTAB_SIZE(spapr)); ... so instead of memset(0)ing it, you could just ppc_hash64_store_hpte(env, i, HPTE64_V_HPTE_DIRTY, 0); the HTAB in a loop. Alex