From nobody Sun Feb 8 23:05:23 2026 Delivered-To: importer@patchew.org Received-SPF: pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) client-ip=208.118.235.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Authentication-Results: mx.zohomail.com; spf=pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org Return-Path: Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) by mx.zohomail.com with SMTPS id 1505853267984925.0588682703374; Tue, 19 Sep 2017 13:34:27 -0700 (PDT) Received: from localhost ([::1]:45278 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1duPDv-0001OL-6B for importer@patchew.org; Tue, 19 Sep 2017 16:34:27 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:42330) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1duOzZ-0005KX-3V for qemu-devel@nongnu.org; Tue, 19 Sep 2017 16:19:38 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1duOzT-0004g0-Ue for qemu-devel@nongnu.org; Tue, 19 Sep 2017 16:19:37 -0400 Received: from mx1.redhat.com ([209.132.183.28]:34014) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1duOzT-0004f9-LE for qemu-devel@nongnu.org; Tue, 19 Sep 2017 16:19:31 -0400 Received: from smtp.corp.redhat.com (int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id C756F7D0CB; Tue, 19 Sep 2017 20:19:30 +0000 (UTC) Received: from localhost (ovpn-116-23.gru2.redhat.com [10.97.116.23]) by smtp.corp.redhat.com (Postfix) with ESMTP id BAEC9600C2; Tue, 19 Sep 2017 20:19:24 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com C756F7D0CB Authentication-Results: ext-mx03.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com Authentication-Results: ext-mx03.extmail.prod.ext.phx2.redhat.com; spf=fail smtp.mailfrom=ehabkost@redhat.com From: Eduardo Habkost To: Peter Maydell , qemu-devel@nongnu.org Date: Tue, 19 Sep 2017 17:18:47 -0300 Message-Id: <20170919201850.14772-10-ehabkost@redhat.com> In-Reply-To: <20170919201850.14772-1-ehabkost@redhat.com> References: <20170919201850.14772-1-ehabkost@redhat.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.11 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.27]); Tue, 19 Sep 2017 20:19:30 +0000 (UTC) X-detected-operating-system: by eggs.gnu.org: GNU/Linux 2.2.x-3.x [generic] [fuzzy] X-Received-From: 209.132.183.28 Subject: [Qemu-devel] [PULL 09/12] numa: cpu: calculate/set default node-ids after all -numa CLI options are parsed X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Marcel Apfelbaum , Igor Mammedov Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" X-ZohoMail: RSF_0 Z_629925259 SPT_0 Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" From: Igor Mammedov Calculating default node-ids for CPUs in possible_cpu_arch_ids() is rather fragile since defaults calculation uses nb_numa_nodes but callback might be potentially called early before all -numa CLI options are parsed, which would lead to cpus assigned only upto nb_numa_nodes at the time possible_cpu_arch_ids() is called. Issue was introduced by (7c88e65 numa: mirror cpu to node mapping in MachineState::possible_cpus) and for example CLI: -smp 4 -numa node,cpus=3D0 -numa node would set props.node-id in possible_cpus array for every non explicitly mapped CPU to the first node. Issue is not visible to guest nor to mgmt interface due to 1) implictly mapped cpus are forced to the first node in case of partial mapping 2) in case of default mapping possible_cpu_arch_ids() is called after all -numa options are parsed (resulting in correct mapping). However it's fragile to rely on late execution of possible_cpu_arch_ids(), therefore add machine specific callback that returns node-id for CPU and use it to calculate/ set defaults at machine_numa_finish_init() time when all -numa options are parsed. Reported-by: Eduardo Habkost Signed-off-by: Igor Mammedov Message-Id: <1496314408-163972-1-git-send-email-imammedo@redhat.com> Reviewed-by: Eduardo Habkost Signed-off-by: Eduardo Habkost --- include/hw/boards.h | 4 ++++ hw/arm/virt.c | 14 ++++++-------- hw/core/machine.c | 1 + hw/i386/pc.c | 20 +++++++++++--------- hw/ppc/spapr.c | 15 ++++++--------- 5 files changed, 28 insertions(+), 26 deletions(-) diff --git a/include/hw/boards.h b/include/hw/boards.h index 6b67adaef6..156e0a5701 100644 --- a/include/hw/boards.h +++ b/include/hw/boards.h @@ -123,6 +123,9 @@ typedef struct { * Returns an array of @CPUArchId architecture-dependent CPU IDs * which includes CPU IDs for present and possible to hotplug CPUs. * Caller is responsible for freeing returned list. + * @get_default_cpu_node_id: + * returns default board specific node_id value for CPU slot specified = by + * index @idx in @ms->possible_cpus[] * @has_hotpluggable_cpus: * If true, board supports CPUs creation with -device/device_add. * @default_cpu_type: @@ -196,6 +199,7 @@ struct MachineClass { CpuInstanceProperties (*cpu_index_to_instance_props)(MachineState *mac= hine, unsigned cpu_inde= x); const CPUArchIdList *(*possible_cpu_arch_ids)(MachineState *machine); + int64_t (*get_default_cpu_node_id)(const MachineState *ms, int idx); }; =20 /** diff --git a/hw/arm/virt.c b/hw/arm/virt.c index 65d68bc50d..9e18b410d7 100644 --- a/hw/arm/virt.c +++ b/hw/arm/virt.c @@ -1554,6 +1554,11 @@ virt_cpu_index_to_props(MachineState *ms, unsigned c= pu_index) return possible_cpus->cpus[cpu_index].props; } =20 +static int64_t virt_get_default_cpu_node_id(const MachineState *ms, int id= x) +{ + return idx % nb_numa_nodes; +} + static const CPUArchIdList *virt_possible_cpu_arch_ids(MachineState *ms) { int n; @@ -1572,14 +1577,6 @@ static const CPUArchIdList *virt_possible_cpu_arch_i= ds(MachineState *ms) virt_cpu_mp_affinity(vms, n); ms->possible_cpus->cpus[n].props.has_thread_id =3D true; ms->possible_cpus->cpus[n].props.thread_id =3D n; - - /* default distribution of CPUs over NUMA nodes */ - if (nb_numa_nodes) { - /* preset values but do not enable them i.e. 'has_node_id =3D = false', - * numa init code will enable them later if manual mapping was= n't - * present on CLI */ - ms->possible_cpus->cpus[n].props.node_id =3D n % nb_numa_nodes; - } } return ms->possible_cpus; } @@ -1603,6 +1600,7 @@ static void virt_machine_class_init(ObjectClass *oc, = void *data) mc->possible_cpu_arch_ids =3D virt_possible_cpu_arch_ids; mc->cpu_index_to_instance_props =3D virt_cpu_index_to_props; mc->default_cpu_type =3D ARM_CPU_TYPE_NAME("cortex-a15"); + mc->get_default_cpu_node_id =3D virt_get_default_cpu_node_id; } =20 static const TypeInfo virt_machine_info =3D { diff --git a/hw/core/machine.c b/hw/core/machine.c index 41b53a17ad..80647edc2a 100644 --- a/hw/core/machine.c +++ b/hw/core/machine.c @@ -724,6 +724,7 @@ static void machine_numa_finish_init(MachineState *mach= ine) /* fetch default mapping from board and enable it */ CpuInstanceProperties props =3D cpu_slot->props; =20 + props.node_id =3D mc->get_default_cpu_node_id(machine, i); if (!default_mapping) { /* record slots with not set mapping, * TODO: make it hard error in future */ diff --git a/hw/i386/pc.c b/hw/i386/pc.c index 2247ac0a01..610c65aeab 100644 --- a/hw/i386/pc.c +++ b/hw/i386/pc.c @@ -2234,6 +2234,16 @@ pc_cpu_index_to_props(MachineState *ms, unsigned cpu= _index) return possible_cpus->cpus[cpu_index].props; } =20 +static int64_t pc_get_default_cpu_node_id(const MachineState *ms, int idx) +{ + X86CPUTopoInfo topo; + + assert(idx < ms->possible_cpus->len); + x86_topo_ids_from_apicid(ms->possible_cpus->cpus[idx].arch_id, + smp_cores, smp_threads, &topo); + return topo.pkg_id % nb_numa_nodes; +} + static const CPUArchIdList *pc_possible_cpu_arch_ids(MachineState *ms) { int i; @@ -2263,15 +2273,6 @@ static const CPUArchIdList *pc_possible_cpu_arch_ids= (MachineState *ms) ms->possible_cpus->cpus[i].props.core_id =3D topo.core_id; ms->possible_cpus->cpus[i].props.has_thread_id =3D true; ms->possible_cpus->cpus[i].props.thread_id =3D topo.smt_id; - - /* default distribution of CPUs over NUMA nodes */ - if (nb_numa_nodes) { - /* preset values but do not enable them i.e. 'has_node_id =3D = false', - * numa init code will enable them later if manual mapping was= n't - * present on CLI */ - ms->possible_cpus->cpus[i].props.node_id =3D - topo.pkg_id % nb_numa_nodes; - } } return ms->possible_cpus; } @@ -2316,6 +2317,7 @@ static void pc_machine_class_init(ObjectClass *oc, vo= id *data) pcmc->linuxboot_dma_enabled =3D true; mc->get_hotplug_handler =3D pc_get_hotpug_handler; mc->cpu_index_to_instance_props =3D pc_cpu_index_to_props; + mc->get_default_cpu_node_id =3D pc_get_default_cpu_node_id; mc->possible_cpu_arch_ids =3D pc_possible_cpu_arch_ids; mc->has_hotpluggable_cpus =3D true; mc->default_boot_order =3D "cad"; diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c index f680f28a15..17ea77618c 100644 --- a/hw/ppc/spapr.c +++ b/hw/ppc/spapr.c @@ -3404,6 +3404,11 @@ spapr_cpu_index_to_props(MachineState *machine, unsi= gned cpu_index) return core_slot->props; } =20 +static int64_t spapr_get_default_cpu_node_id(const MachineState *ms, int i= dx) +{ + return idx / smp_cores % nb_numa_nodes; +} + static const CPUArchIdList *spapr_possible_cpu_arch_ids(MachineState *mach= ine) { int i; @@ -3428,15 +3433,6 @@ static const CPUArchIdList *spapr_possible_cpu_arch_= ids(MachineState *machine) machine->possible_cpus->cpus[i].arch_id =3D core_id; machine->possible_cpus->cpus[i].props.has_core_id =3D true; machine->possible_cpus->cpus[i].props.core_id =3D core_id; - - /* default distribution of CPUs over NUMA nodes */ - if (nb_numa_nodes) { - /* preset values but do not enable them i.e. 'has_node_id =3D = false', - * numa init code will enable them later if manual mapping was= n't - * present on CLI */ - machine->possible_cpus->cpus[i].props.node_id =3D - core_id / smp_threads / smp_cores % nb_numa_nodes; - } } return machine->possible_cpus; } @@ -3587,6 +3583,7 @@ static void spapr_machine_class_init(ObjectClass *oc,= void *data) hc->pre_plug =3D spapr_machine_device_pre_plug; hc->plug =3D spapr_machine_device_plug; mc->cpu_index_to_instance_props =3D spapr_cpu_index_to_props; + mc->get_default_cpu_node_id =3D spapr_get_default_cpu_node_id; mc->possible_cpu_arch_ids =3D spapr_possible_cpu_arch_ids; hc->unplug_request =3D spapr_machine_device_unplug_request; =20 --=20 2.13.5