From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:58848) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cb0tL-0005lV-60 for qemu-devel@nongnu.org; Tue, 07 Feb 2017 03:12:47 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1cb0tG-0002Jr-7o for qemu-devel@nongnu.org; Tue, 07 Feb 2017 03:12:47 -0500 Received: from mx1.redhat.com ([209.132.183.28]:37738) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1cb0tG-0002IE-2s for qemu-devel@nongnu.org; Tue, 07 Feb 2017 03:12:42 -0500 Date: Tue, 7 Feb 2017 16:12:33 +0800 From: Peter Xu Message-ID: <20170207081233.GZ5151@pxdev.xzpeter.org> References: <1486110164-13797-1-git-send-email-peterx@redhat.com> <1486110164-13797-17-git-send-email-peterx@redhat.com> <7d0891fb-62f2-28f3-0f04-99c136d3bed8@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <7d0891fb-62f2-28f3-0f04-99c136d3bed8@redhat.com> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v6 16/18] intel_iommu: do replay when context invalidate List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jason Wang Cc: qemu-devel@nongnu.org, tianyu.lan@intel.com, kevin.tian@intel.com, mst@redhat.com, jan.kiszka@siemens.com, David Gibson , alex.williamson@redhat.com, bd.aviv@gmail.com On Tue, Feb 07, 2017 at 02:16:15PM +0800, Jason Wang wrote: >=20 >=20 > On 2017=E5=B9=B402=E6=9C=8803=E6=97=A5 16:22, Peter Xu wrote: > >Before this one we only invalidate context cache when we receive conte= xt > >entry invalidations. However it's possible that the invalidation also > >contains a domain switch (only if cache-mode is enabled for vIOMMU). I= n > >that case we need to notify all the registered components about the ne= w > >mapping. > > > >Signed-off-by: Peter Xu > >--- > > hw/i386/intel_iommu.c | 10 ++++++++++ > > 1 file changed, 10 insertions(+) > > > >diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c > >index 0a42c01..363227d 100644 > >--- a/hw/i386/intel_iommu.c > >+++ b/hw/i386/intel_iommu.c > >@@ -1115,6 +1115,16 @@ static void vtd_context_device_invalidate(Intel= IOMMUState *s, > > trace_vtd_inv_desc_cc_device(bus_n, VTD_PCI_SLOT(dev= fn_it), > > VTD_PCI_FUNC(devfn_it))= ; > > vtd_as->context_cache_entry.context_cache_gen =3D 0; > >+ /* > >+ * So a device is moving out of (or moving into) a > >+ * domain, a replay() suites here to notify all the > >+ * IOMMU_NOTIFIER_MAP registers about this change. > >+ * This won't bring bad even if we have no such > >+ * notifier registered - the IOMMU notification > >+ * framework will skip MAP notifications if that > >+ * happened. > >+ */ > >+ memory_region_iommu_replay_all(&vtd_as->iommu); > > } > > } > > } >=20 > Reviewed-by: Jason Wang >=20 > We may consider to squash this into patch 18. I can do it. Thanks, -- peterx