qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Bharata B Rao <bharata@linux.vnet.ibm.com>
To: qemu-devel@nongnu.org
Cc: mdroth@linux.vnet.ibm.com, Bharata B Rao <bharata@linux.vnet.ibm.com>
Subject: [Qemu-devel] [RFC PATCH v0 15/15] ppc: Allow hotplugging of CPU cores only
Date: Thu,  4 Sep 2014 11:36:25 +0530	[thread overview]
Message-ID: <1409810785-12391-16-git-send-email-bharata@linux.vnet.ibm.com> (raw)
In-Reply-To: <1409810785-12391-1-git-send-email-bharata@linux.vnet.ibm.com>

pseries kernels support hot-plugging CPUs in core granularity. i,e., when
a core is hotplugged, guest kernel tries to bring up all the threads in
that core. This doesn't match 1-to-1 with QEMU hotplug semantics where
cpu-add monitor command hotplugs 1 CPU at a time.

While it is still an open question as to what kind of CPU hotplug semantics
we should support here, this patch restricts cpu-add monitor command
to allow only CPU cores to be added. All the threads of the hot-plugged
core at onlined at once.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
---
 hw/ppc/spapr.c | 27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 9a3d1ca..96fb11b 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -1717,7 +1717,7 @@ static int spapr_populate_cpu_dt(CPUState *cs, void *fdt, int offset)
     return 0;
 }
 
-static void spapr_cpu_hotplug_add(CPUState *cs)
+static int spapr_cpu_hotplug_add(CPUState *cs)
 {
     int i, j;
     sPAPRDrcEntry *drc_entry;
@@ -1740,6 +1740,11 @@ static void spapr_cpu_hotplug_add(CPUState *cs)
     PowerPCCPUClass *pcc = POWERPC_CPU_GET_CLASS(cs);
     char *nodename;
     int index = ppc_get_vcpu_dt_id(cpu);
+    int smt = kvmppc_smt_threads();
+
+    if ((index % smt) != 0) {
+        return -1;
+    }
 
     drc_entry = spapr_cpu_to_drc_entry(cpu->cpu_dt_id + SPAPR_DRC_CPU_ID_BASE);
     g_assert(drc_entry);
@@ -1797,14 +1802,17 @@ static void spapr_cpu_hotplug_add(CPUState *cs)
     ccs->fdt = fdt;
     ccs->offset_start = offset;
     ccs->state = CC_STATE_PENDING;
+
+    return 0;
 }
 
 static void spapr_cpu_added_req(Notifier *n, void *opaque)
 {
     CPUState *cs = CPU(opaque);
 
-    spapr_cpu_hotplug_add(cs);
-    spapr_cpu_hotplug_add_event(cs);
+    if (!spapr_cpu_hotplug_add(cs)) {
+        spapr_cpu_hotplug_add_event(cs);
+    }
     return;
 }
 
@@ -2193,6 +2201,7 @@ static void ppc_hot_add_cpu(const int64_t id, Error **errp)
 {
     CPUState *cs;
     PowerPCCPU *cpu;
+    int i;
 
     if (id < 0) {
         error_setg(errp, "Invalid CPU id: %" PRIi64, id);
@@ -2213,8 +2222,16 @@ static void ppc_hot_add_cpu(const int64_t id, Error **errp)
         return;
     }
 
-    cpu = ppc_new_cpu(current_cpu_model);
-    spapr_cpu_reset(cpu);
+    if ((id % smp_threads) != 0) {
+        error_setg(errp, "Invalid CPU id: %" PRIi64
+                   ", only cores can be added", id);
+        return;
+    }
+
+    for (i = 0; i < smp_threads; i++) {
+        cpu = ppc_new_cpu(current_cpu_model);
+        spapr_cpu_reset(cpu);
+    }
 }
 
 static void spapr_machine_class_init(ObjectClass *oc, void *data)
-- 
1.7.11.7

      parent reply	other threads:[~2014-09-04  6:07 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-09-04  6:06 [Qemu-devel] [RFC PATCH v0 00/15] CPU hotplug support of PowerPC sPAPR guests Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 01/15] ppc: Store dr entity state bits at the right bit offset Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 02/15] ppc: Rename SPAPR_DRC_TABLE_SIZE to SPAPR_DRC_PHB_TABLE_SIZE Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 03/15] ppc: Rename sPAPRDrcEntry.phb_buid to sPAPRDrcEntry.id Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 04/15] ppc: Make creation of DRC entries in FDT endian safe Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 05/15] ppc: Accommodate CPU DRC entries in DRC table Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 06/15] ppc: stop after getting first unused DR slot " Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 07/15] ppc: Initialize DRC table before initializing CPUs Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 08/15] ppc: Add CPU dynamic reconfiguration (DR) support Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 09/15] ppc: Consider max_cpus during xics initialization Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 10/15] ppc: Factor out CPU initialization code to a new routine Bharata B Rao
2014-09-26 15:29   ` Igor Mammedov
2014-09-29  3:00     ` Bharata B Rao
2014-09-29  8:49       ` Igor Mammedov
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 11/15] ppc: Move RTAS indicator defines to a header file Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 12/15] ppc: Support ibm, lrdr-capacity device tree property Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 13/15] ppc: Make ibm, configure-connector endian-safe Bharata B Rao
2014-09-04  6:06 ` [Qemu-devel] [RFC PATCH v0 14/15] ppc: Add CPU hotplug support for sPAPR guests Bharata B Rao
2014-09-05 21:51   ` Tyrel Datwyler
2014-09-04  6:06 ` Bharata B Rao [this message]

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=1409810785-12391-16-git-send-email-bharata@linux.vnet.ibm.com \
    --to=bharata@linux.vnet.ibm.com \
    --cc=mdroth@linux.vnet.ibm.com \
    --cc=qemu-devel@nongnu.org \
    /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).