From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kiszka Subject: Re: [stable,master] [PATCH] qemu-kvm: Don't break device assignment INTx Date: Thu, 24 Mar 2011 08:13:21 +0100 Message-ID: <4D8AEF11.8060401@web.de> References: <20110323173440.19806.63813.stgit@s20.home> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="------------enig9399153B59D99430ABCCE06A" Cc: kvm@vger.kernel.org, franxisco1988@gmail.com, extmaillist@linuxbox.cz, yongjie.ren@intel.com To: Alex Williamson Return-path: Received: from fmmailgate01.web.de ([217.72.192.221]:56874 "EHLO fmmailgate01.web.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933034Ab1CXHNz (ORCPT ); Thu, 24 Mar 2011 03:13:55 -0400 In-Reply-To: <20110323173440.19806.63813.stgit@s20.home> Sender: kvm-owner@vger.kernel.org List-ID: This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enig9399153B59D99430ABCCE06A Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On 2011-03-23 18:38, Alex Williamson wrote: > Commit 1a836445 moved pci.o from a target object to a generic hardware > object, which drops CONFIG_KVM_DEVICE_ASSIGNMENT. This results in > the device assignment kludge to update INTx vectors on interrupt > routing changes never getting called, which means device assignment > level triggered interrupts don't work. This mostly reverts that > commit, but makes pci.o conditional on CONFIG_PCI, which hopefully > addresses the original intent of the change. Fixes: What a mess - but a still required one. >=20 > https://bugs.launchpad.net/qemu/+bug/730441 >=20 > Reported-by: Yongjie Ren > Signed-off-by: Alex Williamson > --- >=20 > Makefile.objs | 2 +- > Makefile.target | 1 + > hw/pci.c | 2 +- > 3 files changed, 3 insertions(+), 2 deletions(-) >=20 > diff --git a/Makefile.objs b/Makefile.objs > index d1a2264..6080db0 100644 > --- a/Makefile.objs > +++ b/Makefile.objs > @@ -172,7 +172,7 @@ hw-obj-y =3D > hw-obj-y +=3D loader.o > hw-obj-$(CONFIG_VIRTIO) +=3D virtio.o virtio-console.o > hw-obj-y +=3D fw_cfg.o > -hw-obj-$(CONFIG_PCI) +=3D pci.o pci_bridge.o > +hw-obj-$(CONFIG_PCI) +=3D pci_bridge.o > hw-obj-$(CONFIG_PCI) +=3D msix.o msi.o > hw-obj-$(CONFIG_PCI) +=3D pci_host.o pcie_host.o > hw-obj-$(CONFIG_PCI) +=3D ioh3420.o xio3130_upstream.o xio3130_downstr= eam.o > diff --git a/Makefile.target b/Makefile.target > index fdbdc6c..95597e7 100644 > --- a/Makefile.target > +++ b/Makefile.target > @@ -199,6 +199,7 @@ obj-y =3D arch_init.o cpus.o monitor.o machine.o gd= bstub.o vl.o balloon.o > # virtio has to be here due to weird dependency between PCI and virtio= -net. > # need to fix this properly > obj-$(CONFIG_NO_PCI) +=3D pci-stub.o > +obj-$(CONFIG_PCI) +=3D pci.o > obj-$(CONFIG_VIRTIO) +=3D virtio-blk.o virtio-balloon.o virtio-net.o v= irtio-serial-bus.o > obj-$(CONFIG_VIRTIO_PCI) +=3D virtio-pci.o > obj-y +=3D vhost_net.o > diff --git a/hw/pci.c b/hw/pci.c > index 730df5f..51bf98d 100644 > --- a/hw/pci.c > +++ b/hw/pci.c > @@ -29,8 +29,8 @@ > #include "net.h" > #include "sysemu.h" > #include "loader.h" > +#include "qemu-kvm.h" > #include "hw/pc.h" > -#include "kvm.h" > #include "device-assignment.h" > #include "qemu-objects.h" > #include "range.h" However, this hunk remains obsolete (kvm.h includes qemu-kvm.h), just drop it. Jan --------------enig9399153B59D99430ABCCE06A Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.15 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAk2K7xUACgkQitSsb3rl5xSHYQCfVZQXp/G+Een7+PVvitVucYsS zwoAoNki3RZumVZvDwTUUp4xdcxfyI64 =gmmw -----END PGP SIGNATURE----- --------------enig9399153B59D99430ABCCE06A--