From: Greg Kurz <groug@kaod.org>
To: Alexey Kardashevskiy <aik@ozlabs.ru>
Cc: David Gibson <david@gibson.dropbear.id.au>,
nikunj@linux.vnet.ibm.com, mdroth@linux.vnet.ibm.com,
lvivier@redhat.com, thuth@redhat.com, qemu-ppc@nongnu.org,
qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [Qemu-ppc] [RFC 03/17] pseries: Always use core objects for CPU construction
Date: Fri, 4 Nov 2016 10:51:40 +0100 [thread overview]
Message-ID: <20161104105140.4703b2b2@bahia> (raw)
In-Reply-To: <e2c703e3-667d-3745-fffb-9259831def12@ozlabs.ru>
On Thu, 3 Nov 2016 19:11:48 +1100
Alexey Kardashevskiy <aik@ozlabs.ru> wrote:
> On 30/10/16 22:11, David Gibson wrote:
> > Currently the pseries machine has two paths for constructing CPUs. On
> > newer machine type versions, which support cpu hotplug, it constructs
> > cpu core objects, which in turn construct CPU threads. For older machine
> > versions it individually constructs the CPU threads.
> >
> > This division is going to make some future changes to the cpu construction
> > harder, so this patch unifies them. Now cpu core objects are always
> > created. This requires some updates to allow core objects to be created
> > without a full complement of threads (since older versions allowed a
> > number of cpus not a multiple of the threads-per-core). Likewise it needs
> > some changes to the cpu core hot/cold plug path so as not to choke on the
> > old machine types without hotplug support.
> >
> > For good measure, we move the cpu construction to its own subfunction,
> > spapr_init_cpus().
> >
> > Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> > ---
> > hw/ppc/spapr.c | 125 +++++++++++++++++++++++++++---------------------
> > hw/ppc/spapr_cpu_core.c | 30 +++++++-----
> > include/hw/ppc/spapr.h | 1 -
> > 3 files changed, 89 insertions(+), 67 deletions(-)
> >
> > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> > index c8e2921..ad68a9d 100644
> > --- a/hw/ppc/spapr.c
> > +++ b/hw/ppc/spapr.c
> > @@ -1688,11 +1688,80 @@ static void spapr_validate_node_memory(MachineState *machine, Error **errp)
> > }
> > }
> >
> > +static void spapr_init_cpus(sPAPRMachineState *spapr)
> > +{
> > + MachineState *machine = MACHINE(spapr);
> > + MachineClass *mc = MACHINE_GET_CLASS(machine);
> > + char *type = spapr_get_cpu_core_type(machine->cpu_model);
> > + int smt = kvmppc_smt_threads();
> > + int spapr_max_cores, spapr_cores;
> > + int i;
> > +
> > + if (!type) {
> > + error_report("Unable to find sPAPR CPU Core definition");
> > + exit(1);
> > + }
> > +
> > + if (mc->query_hotpluggable_cpus) {
> > + if (smp_cpus % smp_threads) {
> > + error_report("smp_cpus (%u) must be multiple of threads (%u)",
> > + smp_cpus, smp_threads);
> > + exit(1);
> > + }
> > + if (max_cpus % smp_threads) {
> > + error_report("max_cpus (%u) must be multiple of threads (%u)",
> > + max_cpus, smp_threads);
> > + exit(1);
> > + }
> > +
> > + spapr_max_cores = max_cpus / smp_threads;
> > + spapr_cores = smp_cpus / smp_threads;
> > + } else {
> > + if (max_cpus != smp_cpus) {
> > + error_report("This machine version does not support CPU hotplug");
> > + exit(1);
> > + }
> > +
> > + spapr_max_cores = QEMU_ALIGN_UP(smp_cpus, smp_threads) / smp_threads;
> > + spapr_cores = spapr_max_cores;
> > + }
> > +
> > + spapr->cores = g_new0(Object *, spapr_max_cores);
> > + for (i = 0; i < spapr_max_cores; i++) {
> > + int core_id = i * smp_threads;
> > +
> > + if (mc->query_hotpluggable_cpus) {
> > + sPAPRDRConnector *drc =
> > + spapr_dr_connector_new(OBJECT(spapr),
> > + SPAPR_DR_CONNECTOR_TYPE_CPU,
> > + (core_id / smp_threads) * smt);
> > +
> > + qemu_register_reset(spapr_drc_reset, drc);
> > + }
> > +
> > + if (i < spapr_cores) {
> > + Object *core = object_new(type);
> > + int nr_threads = smp_threads;
> > +
> > + /* Handle the partially filled core for older machine types */
> > + if ((i + 1) * smp_threads >= smp_cpus) {
> > + nr_threads = smp_cpus - i * smp_threads;
> > + }
>
>
> What is this exactly for? Older machines report "qemu-system-ppc64: threads
> must be 8" when I do "-smp 12,threads=8 -machine pseries-2.2".
>
IIUC, this lowers nr_threads for the last core to end up with the requested
number of vCPUs... but spapr_core_pre_plug() doesn't like partially filled
cores.
if (cc->nr_threads != smp_threads) {
error_setg(&local_err, "threads must be %d", smp_threads);
goto out;
}
BTW, this error message looks weird when ones has passed "-smp threads=8"...
It should better reads:
"unsupported partially filled core (%d threads, should have %d)"
If this check is removed, then we hit:
qemu-system-ppc64: core id 8 out of range
because of:
int spapr_max_cores = max_cpus / smp_threads;
index = cc->core_id / smp_threads;
if (index < 0 || index >= spapr_max_cores) {
error_setg(&local_err, "core id %d out of range", cc->core_id);
goto out;
}
Since the cc->core_id / smp_threads pattern is only used on the plug/unplug
paths, maybe these checks in spapr_core_pre_plug() should only be done
when mc->query_hotpluggable_cpus != NULL ?
>
>
> > +
> > + object_property_set_int(core, nr_threads, "nr-threads",
> > + &error_fatal);
> > + object_property_set_int(core, core_id, CPU_CORE_PROP_CORE_ID,
> > + &error_fatal);
> > + object_property_set_bool(core, true, "realized", &error_fatal);
> > + }
> > + }
> > + g_free(type);
> > +}
> > +
> > /* pSeries LPAR / sPAPR hardware init */
> > static void ppc_spapr_init(MachineState *machine)
> > {
> > sPAPRMachineState *spapr = SPAPR_MACHINE(machine);
> > - MachineClass *mc = MACHINE_GET_CLASS(machine);
> > sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(machine);
> > const char *kernel_filename = machine->kernel_filename;
> > const char *initrd_filename = machine->initrd_filename;
> > @@ -1707,21 +1776,6 @@ static void ppc_spapr_init(MachineState *machine)
> > long load_limit, fw_size;
> > char *filename;
> > int smt = kvmppc_smt_threads();
> > - int spapr_cores = smp_cpus / smp_threads;
> > - int spapr_max_cores = max_cpus / smp_threads;
> > -
> > - if (mc->query_hotpluggable_cpus) {
> > - if (smp_cpus % smp_threads) {
> > - error_report("smp_cpus (%u) must be multiple of threads (%u)",
> > - smp_cpus, smp_threads);
> > - exit(1);
> > - }
> > - if (max_cpus % smp_threads) {
> > - error_report("max_cpus (%u) must be multiple of threads (%u)",
> > - max_cpus, smp_threads);
> > - exit(1);
> > - }
> > - }
> >
> > msi_nonbroken = true;
> >
> > @@ -1801,44 +1855,7 @@ static void ppc_spapr_init(MachineState *machine)
> >
> > ppc_cpu_parse_features(machine->cpu_model);
> >
> > - if (mc->query_hotpluggable_cpus) {
> > - char *type = spapr_get_cpu_core_type(machine->cpu_model);
> > -
> > - if (type == NULL) {
> > - error_report("Unable to find sPAPR CPU Core definition");
> > - exit(1);
> > - }
> > -
> > - spapr->cores = g_new0(Object *, spapr_max_cores);
> > - for (i = 0; i < spapr_max_cores; i++) {
> > - int core_id = i * smp_threads;
> > - sPAPRDRConnector *drc =
> > - spapr_dr_connector_new(OBJECT(spapr),
> > - SPAPR_DR_CONNECTOR_TYPE_CPU,
> > - (core_id / smp_threads) * smt);
> > -
> > - qemu_register_reset(spapr_drc_reset, drc);
> > -
> > - if (i < spapr_cores) {
> > - Object *core = object_new(type);
> > - object_property_set_int(core, smp_threads, "nr-threads",
> > - &error_fatal);
> > - object_property_set_int(core, core_id, CPU_CORE_PROP_CORE_ID,
> > - &error_fatal);
> > - object_property_set_bool(core, true, "realized", &error_fatal);
> > - }
> > - }
> > - g_free(type);
> > - } else {
> > - for (i = 0; i < smp_cpus; i++) {
> > - PowerPCCPU *cpu = cpu_ppc_init(machine->cpu_model);
> > - if (cpu == NULL) {
> > - error_report("Unable to find PowerPC CPU definition");
> > - exit(1);
> > - }
> > - spapr_cpu_init(spapr, cpu, &error_fatal);
> > - }
> > - }
> > + spapr_init_cpus(spapr);
> >
> > if (kvm_enabled()) {
> > /* Enable H_LOGICAL_CI_* so SLOF can talk to in-kernel devices */
> > diff --git a/hw/ppc/spapr_cpu_core.c b/hw/ppc/spapr_cpu_core.c
> > index e0c14f6..1357293 100644
> > --- a/hw/ppc/spapr_cpu_core.c
> > +++ b/hw/ppc/spapr_cpu_core.c
> > @@ -46,7 +46,8 @@ static void spapr_cpu_destroy(PowerPCCPU *cpu)
> > qemu_unregister_reset(spapr_cpu_reset, cpu);
> > }
> >
> > -void spapr_cpu_init(sPAPRMachineState *spapr, PowerPCCPU *cpu, Error **errp)
> > +static void spapr_cpu_init(sPAPRMachineState *spapr, PowerPCCPU *cpu,
> > + Error **errp)
> > {
> > CPUPPCState *env = &cpu->env;
> > CPUState *cs = CPU(cpu);
> > @@ -166,6 +167,7 @@ void spapr_core_plug(HotplugHandler *hotplug_dev, DeviceState *dev,
> > Error **errp)
> > {
> > sPAPRMachineState *spapr = SPAPR_MACHINE(OBJECT(hotplug_dev));
> > + MachineClass *mc = MACHINE_GET_CLASS(spapr);
> > sPAPRCPUCore *core = SPAPR_CPU_CORE(OBJECT(dev));
> > CPUCore *cc = CPU_CORE(dev);
> > CPUState *cs = CPU(core->threads);
> > @@ -180,7 +182,7 @@ void spapr_core_plug(HotplugHandler *hotplug_dev, DeviceState *dev,
> > drc = spapr_dr_connector_by_id(SPAPR_DR_CONNECTOR_TYPE_CPU, index * smt);
> > spapr->cores[index] = OBJECT(dev);
> >
> > - g_assert(drc);
> > + g_assert(drc || !mc->query_hotpluggable_cpus);
> >
> > /*
> > * Setup CPU DT entries only for hotplugged CPUs. For boot time or
> > @@ -190,13 +192,15 @@ void spapr_core_plug(HotplugHandler *hotplug_dev, DeviceState *dev,
> > fdt = spapr_populate_hotplug_cpu_dt(cs, &fdt_offset, spapr);
> > }
> >
> > - drck = SPAPR_DR_CONNECTOR_GET_CLASS(drc);
> > - drck->attach(drc, dev, fdt, fdt_offset, !dev->hotplugged, &local_err);
> > - if (local_err) {
> > - g_free(fdt);
> > - spapr->cores[index] = NULL;
> > - error_propagate(errp, local_err);
> > - return;
> > + if (drc) {
> > + drck = SPAPR_DR_CONNECTOR_GET_CLASS(drc);
> > + drck->attach(drc, dev, fdt, fdt_offset, !dev->hotplugged, &local_err);
> > + if (local_err) {
> > + g_free(fdt);
> > + spapr->cores[index] = NULL;
> > + error_propagate(errp, local_err);
> > + return;
> > + }
> > }
> >
> > if (dev->hotplugged) {
> > @@ -209,8 +213,10 @@ void spapr_core_plug(HotplugHandler *hotplug_dev, DeviceState *dev,
> > /*
> > * Set the right DRC states for cold plugged CPU.
> > */
> > - drck->set_allocation_state(drc, SPAPR_DR_ALLOCATION_STATE_USABLE);
> > - drck->set_isolation_state(drc, SPAPR_DR_ISOLATION_STATE_UNISOLATED);
> > + if (drc) {
> > + drck->set_allocation_state(drc, SPAPR_DR_ALLOCATION_STATE_USABLE);
> > + drck->set_isolation_state(drc, SPAPR_DR_ISOLATION_STATE_UNISOLATED);
> > + }
> > }
> > }
> >
> > @@ -227,7 +233,7 @@ void spapr_core_pre_plug(HotplugHandler *hotplug_dev, DeviceState *dev,
> > char *base_core_type = spapr_get_cpu_core_type(machine->cpu_model);
> > const char *type = object_get_typename(OBJECT(dev));
> >
> > - if (!mc->query_hotpluggable_cpus) {
> > + if (dev->hotplugged && !mc->query_hotpluggable_cpus) {
> > error_setg(&local_err, "CPU hotplug not supported for this machine");
> > goto out;
> > }
> > diff --git a/include/hw/ppc/spapr.h b/include/hw/ppc/spapr.h
> > index bd5bcf7..f8d444d 100644
> > --- a/include/hw/ppc/spapr.h
> > +++ b/include/hw/ppc/spapr.h
> > @@ -614,7 +614,6 @@ void spapr_hotplug_req_add_by_count_indexed(sPAPRDRConnectorType drc_type,
> > uint32_t count, uint32_t index);
> > void spapr_hotplug_req_remove_by_count_indexed(sPAPRDRConnectorType drc_type,
> > uint32_t count, uint32_t index);
> > -void spapr_cpu_init(sPAPRMachineState *spapr, PowerPCCPU *cpu, Error **errp);
> > void *spapr_populate_hotplug_cpu_dt(CPUState *cs, int *fdt_offset,
> > sPAPRMachineState *spapr);
> >
> >
>
>
next prev parent reply other threads:[~2016-11-04 9:52 UTC|newest]
Thread overview: 75+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-30 11:11 [Qemu-devel] [RFC 00/17] Clean up compatibility mode handling David Gibson
2016-10-30 11:11 ` [Qemu-devel] [RFC 01/17] ppc: Remove some stub POWER6 models David Gibson
2016-10-31 7:38 ` Thomas Huth
2016-10-31 8:37 ` David Gibson
2016-11-08 3:40 ` David Gibson
2016-10-30 11:11 ` [Qemu-devel] [RFC 02/17] powernv: CPU compatibility modes don't make sense for powernv David Gibson
2016-10-31 7:46 ` Thomas Huth
2016-10-31 8:38 ` David Gibson
2016-10-31 10:35 ` Greg Kurz
2016-10-30 11:11 ` [Qemu-devel] [RFC 03/17] pseries: Always use core objects for CPU construction David Gibson
2016-11-03 8:11 ` Alexey Kardashevskiy
2016-11-04 9:51 ` Greg Kurz [this message]
2016-11-08 5:34 ` [Qemu-devel] [Qemu-ppc] " David Gibson
2016-10-30 11:11 ` [Qemu-devel] [RFC 04/17] pseries: Make cpu_update during CAS unconditional David Gibson
2016-11-03 8:24 ` Alexey Kardashevskiy
2016-11-04 10:45 ` Thomas Huth
2016-11-08 3:44 ` David Gibson
2016-10-30 11:11 ` [Qemu-devel] [RFC 05/17] ppc: Clean up and QOMify hypercall emulation David Gibson
2016-11-03 8:50 ` Alexey Kardashevskiy
2016-10-30 11:11 ` [Qemu-devel] [RFC 06/17] ppc: Rename cpu_version to compat_pvr David Gibson
2016-11-04 2:26 ` Alexey Kardashevskiy
2016-11-08 3:48 ` David Gibson
2016-11-04 10:51 ` Thomas Huth
2016-10-30 11:11 ` [Qemu-devel] [RFC 07/17] ppc: Rewrite ppc_set_compat() David Gibson
2016-11-04 2:57 ` Alexey Kardashevskiy
2016-11-08 3:49 ` David Gibson
2016-10-30 11:11 ` [Qemu-devel] [RFC 08/17] ppc: Rewrite ppc_get_compat_smt_threads() David Gibson
2016-11-04 3:37 ` Alexey Kardashevskiy
2016-11-08 5:13 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 09/17] ppc: Validate compatibility modes when setting David Gibson
2016-10-31 5:55 ` Alexey Kardashevskiy
2016-10-31 8:39 ` David Gibson
2016-11-04 3:45 ` Alexey Kardashevskiy
2016-11-08 5:14 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 10/17] pseries: Rewrite CAS PVR compatibility logic David Gibson
2016-10-31 5:00 ` Alexey Kardashevskiy
2016-10-31 5:44 ` David Gibson
2016-11-10 17:54 ` Michael Roth
2016-11-10 23:50 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 11/17] ppc: Add ppc_set_compat_all() David Gibson
2016-11-04 4:01 ` Alexey Kardashevskiy
2016-11-08 5:18 ` David Gibson
2016-11-09 1:27 ` Alexey Kardashevskiy
2016-11-09 3:52 ` David Gibson
2016-11-09 5:18 ` Alexey Kardashevskiy
2016-11-10 3:13 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 12/17] ppc: Migrate compatibility mode David Gibson
2016-11-04 5:58 ` Alexey Kardashevskiy
2016-11-08 5:19 ` David Gibson
2016-11-08 5:51 ` Alexey Kardashevskiy
2016-11-10 1:59 ` David Gibson
2016-11-10 23:55 ` Michael Roth
2016-11-14 1:15 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 13/17] pseries: Move CPU compatibility property to machine David Gibson
2016-11-04 7:43 ` Alexey Kardashevskiy
2016-11-08 5:26 ` David Gibson
2016-11-08 5:56 ` Alexey Kardashevskiy
2016-11-09 4:41 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 14/17] pseries: Reset CPU compatibility mode David Gibson
2016-11-04 7:50 ` Alexey Kardashevskiy
2016-10-30 11:12 ` [Qemu-devel] [RFC 15/17] ppc: Check that CPU model stays consistent across migration David Gibson
2016-11-04 7:54 ` Alexey Kardashevskiy
2016-11-08 5:29 ` David Gibson
2016-11-08 6:03 ` Alexey Kardashevskiy
2016-11-09 4:24 ` David Gibson
2016-11-09 6:06 ` Alexey Kardashevskiy
2016-11-09 6:40 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 16/17] ppc: Remove counter-productive "sanity checks" in migration David Gibson
2016-11-04 5:52 ` Alexey Kardashevskiy
2016-11-08 5:31 ` David Gibson
2016-11-11 18:13 ` [Qemu-devel] [Qemu-ppc] " Greg Kurz
2016-11-14 2:34 ` Alexey Kardashevskiy
2016-11-14 6:08 ` David Gibson
2016-10-30 11:12 ` [Qemu-devel] [RFC 17/17] pseries: Default to POWER8 compatibility mode David Gibson
2016-10-30 11:58 ` David Gibson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20161104105140.4703b2b2@bahia \
--to=groug@kaod.org \
--cc=aik@ozlabs.ru \
--cc=david@gibson.dropbear.id.au \
--cc=lvivier@redhat.com \
--cc=mdroth@linux.vnet.ibm.com \
--cc=nikunj@linux.vnet.ibm.com \
--cc=qemu-devel@nongnu.org \
--cc=qemu-ppc@nongnu.org \
--cc=thuth@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).