Received: from eggs.gnu.org ([2001:4830:134:3::10]:54205) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZL6gg-0003cg-H8 for qemu-devel@nongnu.org; Fri, 31 Jul 2015 05:33:14 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZL6gc-0005hE-Fp for qemu-devel@nongnu.org; Fri, 31 Jul 2015 05:33:10 -0400 Received: from mailout3.w1.samsung.com ([210.118.77.13]:57957) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZL6gc-0005gW-Ab for qemu-devel@nongnu.org; Fri, 31 Jul 2015 05:33:06 -0400 Received: from eucpsbgm2.samsung.com (unknown [203.254.199.245]) by mailout3.w1.samsung.com (Oracle Communications Messaging Server 7.0.5.31.0 64bit (built May 5 2014)) with ESMTP id <0NSC00EIPIJ1RG60@mailout3.w1.samsung.com> for qemu-devel@nongnu.org; Fri, 31 Jul 2015 10:33:01 +0100 (BST) From: Pavel Fedin References: <1dbab1e73e54f0fca4d6d05d50986d2dd396f867.1437731107.git.p.fedin@samsung.com> In-reply-to: Date: Fri, 31 Jul 2015 12:32:59 +0300 Message-id: <026901d0cb73$e4708000$ad518000$@samsung.com> MIME-version: 1.0 Content-type: text/plain; charset=UTF-8 Content-transfer-encoding: quoted-printable Content-language: ru Subject: Re: [Qemu-devel] [PATCH v7 5/6] Initial implementation of vGICv3 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: 'Peter Maydell' Cc: 'Shlomo Pongratz' , 'Shlomo Pongratz' , 'QEMU Developers' , 'Christoffer Dall' , 'Eric Auger' Hello! > On 24 July 2015 at 10:55, Pavel Fedin wrote: > > Get/put routines are missing, live migration is not possible. >=20 > This commit message could do with being made a bit less terse. Ok. Subject was self-explanatory, so i didn't have much to add. > > +static void kvm_arm_gicv3_reset(DeviceState *dev) > > +{ > > + GICv3State *s =3D ARM_GICV3_COMMON(dev); > > + KVMARMGICv3Class *kgc =3D KVM_ARM_GICV3_GET_CLASS(s); > > + > > + DPRINTF("Reset\n"); > > + > > + kgc->parent_reset(dev); > > + kvm_arm_gicv3_put(s); > > +} >=20 > If we don't currently do anything in reset then does the GIC just > go wrong on a VM reset? No it doesn't, reset works. > > + /* Try to create the device via the device control API */ > > + s->dev_fd =3D -1; > > + ret =3D kvm_create_device(kvm_state, KVM_DEV_TYPE_ARM_VGIC_V3, = false); > > + if (ret >=3D 0) { > > + s->dev_fd =3D ret; > > + } else if (ret !=3D -ENODEV && ret !=3D -ENOTSUP) { >=20 > Why aren't ENODEV and ENOTSUP fatal errors like other errnos? Stupid leftover from vGICv2 code, copypasted. I have already removed it = in my repo, just decided to delay the respin until i know the fate of = this: = https://lists.gnu.org/archive/html/qemu-devel/2015-07/msg05778.html >=20 > > + error_setg_errno(errp, -ret, "error creating in-kernel = VGIC"); > > + return; > > + } > > + > > + if (kvm_gic_supports_attr(s->dev_fd, = KVM_DEV_ARM_VGIC_GRP_NR_IRQS, 0)) { >=20 > Is there any kernel which supports GICv3 but does not support > this attribute? I would hope not, in which case we can skip the > conditional check for support. >=20 > > + uint32_t numirqs =3D s->num_irq; > > + DPRINTF("KVM_DEV_ARM_VGIC_GRP_NR_IRQS =3D %u\n", numirqs); > > + kvm_gic_access(s->dev_fd, KVM_DEV_ARM_VGIC_GRP_NR_IRQS, > > + 0, 0, &numirqs, 1); > > + } > > + > > + /* Tell the kernel to complete VGIC initialization now */ > > + if (kvm_gic_supports_attr(s->dev_fd, KVM_DEV_ARM_VGIC_GRP_CTRL, > > + KVM_DEV_ARM_VGIC_CTRL_INIT)) { >=20 > Ditto. I intentionally put some tracing to these conditions. On my system = KVM_DEV_ARM_VGIC_GRP_NR_IRQS is supported and KVM_DEV_ARM_VGIC_CTRL_INIT = is not. So will it always be this way? Kind regards, Pavel Fedin Expert Engineer Samsung Electronics Research center Russia