[Qemu-devel] [PATCH v5 0/6] Memory hotplug for PowerPC sPAPR guests

[Qemu-devel] [PATCH v5 0/6] Memory hotplug for PowerPC sPAPR guests

[Bharata B Rao]

Hi,

This is v5 of memory hotplug support patchset for PowerPC
sPAPR guests.

This patchset applies on spapr-next branch of David Gibson's tree with
the other prerequisite patchset applied. Pre-requistes patchset was
posted at:
https://lists.gnu.org/archive/html/qemu-devel/2015-06/msg05157.html

Changes in v5
-------------
- Get rid of hotplug_mem_size alignment completely and start hotplug base
  right after RAM. (01/06)
- Disable LMB DR and revert to older style of memory DT representation
  if RAM size, maxmem size or individual node memory sizes aren't aligned
  to SPAPR_MEMORY_BLOCK_SIZE (256M). (02/06)
- Use dr_lmb_enabled from sPAPRMachineClass itself instead of copying
  it to sPAPRMachineState. (02/06)
- Ensure memory@0 is created even when -numa option isn't specified. (03/06)
- Calling pc_dimm_memory_unplug() when pc_dimm_memory_plug() fails isn't
  required. (05/06)
- Added a new patch to prevent memory hotplug to memory-less nodes (06/06)

v4: https://lists.gnu.org/archive/html/qemu-devel/2015-06/msg05242.html
v3: https://lists.nongnu.org/archive/html/qemu-devel/2015-04/msg02910.html

*** BLURB HERE ***

Bharata B Rao (6):
  spapr: Initialize hotplug memory address space
  spapr: Add LMB DR connectors
  spapr: Support ibm,dynamic-reconfiguration-memory
  spapr: Make hash table size a factor of maxram_size
  spapr: Memory hotplug support
  spapr: Don't allow memory hotplug to memory less nodes

 default-configs/ppc64-softmmu.mak |   1 +
 docs/specs/ppc-spapr-hotplug.txt  |  48 ++++
 hw/ppc/spapr.c                    | 460 ++++++++++++++++++++++++++++++++++----
 hw/ppc/spapr_events.c             |   8 +-
 hw/ppc/spapr_hcall.c              |  51 ++++-
 include/hw/ppc/spapr.h            |  28 ++-
 6 files changed, 542 insertions(+), 54 deletions(-)

-- 
2.1.0

[Qemu-devel] [PATCH v5 1/6] spapr: Initialize hotplug memory address space

[Bharata B Rao]

Initialize a hotplug memory region under which all the hotplugged
memory is accommodated. Also enable memory hotplug by setting
CONFIG_MEM_HOTPLUG.

Modelled on i386 memory hotplug.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
---
 default-configs/ppc64-softmmu.mak |  1 +
 hw/ppc/spapr.c                    | 18 ++++++++++++++++++
 include/hw/ppc/spapr.h            | 12 ++++++++++++
 3 files changed, 31 insertions(+)

diff --git a/default-configs/ppc64-softmmu.mak b/default-configs/ppc64-softmmu.mak
index ab62cc7..e77cb1a 100644
--- a/default-configs/ppc64-softmmu.mak
+++ b/default-configs/ppc64-softmmu.mak
@@ -52,3 +52,4 @@ CONFIG_XICS_KVM=$(and $(CONFIG_PSERIES),$(CONFIG_KVM))
 # For PReP
 CONFIG_MC146818RTC=y
 CONFIG_ISA_TESTDEV=y
+CONFIG_MEM_HOTPLUG=y
diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 5ca817c..241ecad 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -1549,6 +1549,24 @@ static void ppc_spapr_init(MachineState *machine)
         memory_region_add_subregion(sysmem, 0, rma_region);
     }
 
+    /* initialize hotplug memory address space */
+    if (machine->ram_size < machine->maxram_size) {
+        ram_addr_t hotplug_mem_size = machine->maxram_size - machine->ram_size;
+
+        if (machine->ram_slots > SPAPR_MAX_RAM_SLOTS) {
+            error_report("unsupported amount of memory slots: %"PRIu64,
+                          machine->ram_slots);
+            exit(EXIT_FAILURE);
+        }
+
+        spapr->hotplug_memory.base = ROUND_UP(machine->ram_size,
+                                              SPAPR_HOTPLUG_MEM_ALIGN);
+        memory_region_init(&spapr->hotplug_memory.mr, OBJECT(spapr),
+                           "hotplug-memory", hotplug_mem_size);
+        memory_region_add_subregion(sysmem, spapr->hotplug_memory.base,
+                                    &spapr->hotplug_memory.mr);
+    }
+
     filename = qemu_find_file(QEMU_FILE_TYPE_BIOS, "spapr-rtas.bin");
     if (!filename) {
         error_report("Could not find LPAR rtas '%s'", "spapr-rtas.bin");
diff --git a/include/hw/ppc/spapr.h b/include/hw/ppc/spapr.h
index 91a61ab..8a1929b 100644
--- a/include/hw/ppc/spapr.h
+++ b/include/hw/ppc/spapr.h
@@ -5,6 +5,7 @@
 #include "hw/boards.h"
 #include "hw/ppc/xics.h"
 #include "hw/ppc/spapr_drc.h"
+#include "hw/mem/pc-dimm.h"
 
 struct VIOsPAPRBus;
 struct sPAPRPHBState;
@@ -76,6 +77,7 @@ struct sPAPRMachineState {
 
     /*< public >*/
     char *kvm_type;
+    MemoryHotplugState hotplug_memory;
 };
 
 #define H_SUCCESS         0
@@ -609,4 +611,14 @@ int spapr_rtc_import_offset(DeviceState *dev, int64_t legacy_offset);
 
 #define SPAPR_MEMORY_BLOCK_SIZE (1 << 28) /* 256MB */
 
+/*
+ * This defines the maximum number of DIMM slots we can have for sPAPR
+ * guest. This is not defined by sPAPR but we are defining it to 32 slots
+ * based on default number of slots provided by PowerPC kernel.
+ */
+#define SPAPR_MAX_RAM_SLOTS     32
+
+/* 1GB alignment for hotplug memory region */
+#define SPAPR_HOTPLUG_MEM_ALIGN (1ULL << 30)
+
 #endif /* !defined (__HW_SPAPR_H__) */
-- 
2.1.0

[Qemu-devel] [PATCH v5 2/6] spapr: Add LMB DR connectors

[Bharata B Rao]

Enable memory hotplug for pseries 2.4 and add LMB DR connectors.
With memory hotplug, enforce RAM size, NUMA node memory size and maxmem
to be a multiple of SPAPR_MEMORY_BLOCK_SIZE (256M) since that's the
granularity in which LMBs are represented and hot-added.

LMB DR connectors will be used by the memory hotplug code.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
               [spapr_drc_reset implementation]
---
 hw/ppc/spapr.c         | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++
 include/hw/ppc/spapr.h |  1 +
 2 files changed, 80 insertions(+)

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 241ecad..d7e0e44 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -59,6 +59,7 @@
 #include "hw/nmi.h"
 
 #include "hw/compat.h"
+#include "qemu-common.h"
 
 #include <libfdt.h>
 
@@ -1436,10 +1437,76 @@ static void spapr_cpu_init(sPAPRMachineState *spapr, PowerPCCPU *cpu)
     qemu_register_reset(spapr_cpu_reset, cpu);
 }
 
+/*
+ * Reset routine for LMB DR devices.
+ *
+ * Unlike PCI DR devices, LMB DR devices explicitly register this reset
+ * routine. Reset for PCI DR devices will be handled by PHB reset routine
+ * when it walks all its children devices. LMB devices reset occurs
+ * as part of spapr_ppc_reset().
+ */
+static void spapr_drc_reset(void *opaque)
+{
+    sPAPRDRConnector *drc = opaque;
+    DeviceState *d = DEVICE(drc);
+
+    if (d) {
+        device_reset(d);
+    }
+}
+
+static void spapr_create_lmb_dr_connectors(sPAPRMachineState *spapr)
+{
+    MachineState *machine = MACHINE(spapr);
+    uint64_t lmb_size = SPAPR_MEMORY_BLOCK_SIZE;
+    uint32_t nr_rma_lmbs = spapr->rma_size/lmb_size;
+    uint32_t nr_lmbs = machine->maxram_size/lmb_size - nr_rma_lmbs;
+    uint32_t nr_assigned_lmbs = machine->ram_size/lmb_size - nr_rma_lmbs;
+    int i;
+
+    for (i = 0; i < nr_lmbs; i++) {
+        sPAPRDRConnector *drc;
+        uint64_t addr;
+
+        if (i < nr_assigned_lmbs) {
+            addr = (i + nr_rma_lmbs) * lmb_size;
+        } else {
+            addr = (i - nr_assigned_lmbs) * lmb_size +
+                    spapr->hotplug_memory.base;
+        }
+        drc = spapr_dr_connector_new(OBJECT(spapr), SPAPR_DR_CONNECTOR_TYPE_LMB,
+                                     addr/lmb_size);
+        qemu_register_reset(spapr_drc_reset, drc);
+    }
+}
+
+/*
+ * If RAM size, maxmem size and individual node mem sizes aren't aligned
+ * to SPAPR_MEMORY_BLOCK_SIZE(256MB), then disable LMB DR feature.
+ */
+static void spapr_validate_node_memory(MachineState *machine)
+{
+    int i;
+    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(machine);
+
+    if (machine->maxram_size % SPAPR_MEMORY_BLOCK_SIZE ||
+        machine->ram_size % SPAPR_MEMORY_BLOCK_SIZE) {
+        smc->dr_lmb_enabled = false;
+    }
+
+    for (i = 0; i < nb_numa_nodes; i++) {
+        if (numa_info[i].node_mem &&
+            numa_info[i].node_mem % SPAPR_MEMORY_BLOCK_SIZE) {
+            smc->dr_lmb_enabled = false;
+        }
+    }
+}
+
 /* pSeries LPAR / sPAPR hardware init */
 static void ppc_spapr_init(MachineState *machine)
 {
     sPAPRMachineState *spapr = SPAPR_MACHINE(machine);
+    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(machine);
     const char *kernel_filename = machine->kernel_filename;
     const char *kernel_cmdline = machine->kernel_cmdline;
     const char *initrd_filename = machine->initrd_filename;
@@ -1518,6 +1585,10 @@ static void ppc_spapr_init(MachineState *machine)
                                                smp_threads),
                                   XICS_IRQS);
 
+    if (smc->dr_lmb_enabled) {
+        spapr_validate_node_memory(machine);
+    }
+
     /* init CPUs */
     if (machine->cpu_model == NULL) {
         machine->cpu_model = kvm_enabled() ? "host" : "POWER7";
@@ -1567,6 +1638,10 @@ static void ppc_spapr_init(MachineState *machine)
                                     &spapr->hotplug_memory.mr);
     }
 
+    if (smc->dr_lmb_enabled) {
+        spapr_create_lmb_dr_connectors(spapr);
+    }
+
     filename = qemu_find_file(QEMU_FILE_TYPE_BIOS, "spapr-rtas.bin");
     if (!filename) {
         error_report("Could not find LPAR rtas '%s'", "spapr-rtas.bin");
@@ -1840,6 +1915,7 @@ static void spapr_nmi(NMIState *n, int cpu_index, Error **errp)
 static void spapr_machine_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    sPAPRMachineClass *smc = SPAPR_MACHINE_CLASS(oc);
     FWPathProviderClass *fwc = FW_PATH_PROVIDER_CLASS(oc);
     NMIClass *nc = NMI_CLASS(oc);
 
@@ -1853,6 +1929,7 @@ static void spapr_machine_class_init(ObjectClass *oc, void *data)
     mc->kvm_type = spapr_kvm_type;
     mc->has_dynamic_sysbus = true;
 
+    smc->dr_lmb_enabled = false;
     fwc->get_dev_path = spapr_get_fw_dev_path;
     nc->nmi_monitor_handler = spapr_nmi;
 }
@@ -1988,11 +2065,13 @@ static const TypeInfo spapr_machine_2_3_info = {
 static void spapr_machine_2_4_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    sPAPRMachineClass *smc = SPAPR_MACHINE_CLASS(oc);
 
     mc->name = "pseries-2.4";
     mc->desc = "pSeries Logical Partition (PAPR compliant) v2.4";
     mc->alias = "pseries";
     mc->is_default = 1;
+    smc->dr_lmb_enabled = true;
 }
 
 static const TypeInfo spapr_machine_2_4_info = {
diff --git a/include/hw/ppc/spapr.h b/include/hw/ppc/spapr.h
index 8a1929b..a142e91 100644
--- a/include/hw/ppc/spapr.h
+++ b/include/hw/ppc/spapr.h
@@ -35,6 +35,7 @@ struct sPAPRMachineClass {
     MachineClass parent_class;
 
     /*< public >*/
+    bool dr_lmb_enabled; /* enable dynamic-reconfig/hotplug of LMBs */
 };
 
 /**
-- 
2.1.0

[Qemu-devel] [PATCH v5 3/6] spapr: Support ibm, dynamic-reconfiguration-memory

[Bharata B Rao]

Parse ibm,architecture.vec table obtained from the guest and enable
memory node configuration via ibm,dynamic-reconfiguration-memory if guest
supports it. This is in preparation to support memory hotplug for
sPAPR guests.

This changes the way memory node configuration is done. Currently all
memory nodes are built upfront. But after this patch, only memory@0 node
for RMA is built upfront. Guest kernel boots with just that and rest of
the memory nodes (via memory@XXX or ibm,dynamic-reconfiguration-memory)
are built when guest does ibm,client-architecture-support call.

Note: This patch needs a SLOF enhancement which is already part of
SLOF binary in QEMU.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
---
 docs/specs/ppc-spapr-hotplug.txt |  48 +++++++++
 hw/ppc/spapr.c                   | 210 +++++++++++++++++++++++++++++++--------
 hw/ppc/spapr_hcall.c             |  51 ++++++++--
 include/hw/ppc/spapr.h           |  15 ++-
 4 files changed, 274 insertions(+), 50 deletions(-)

diff --git a/docs/specs/ppc-spapr-hotplug.txt b/docs/specs/ppc-spapr-hotplug.txt
index 46e0719..9d574b5 100644
--- a/docs/specs/ppc-spapr-hotplug.txt
+++ b/docs/specs/ppc-spapr-hotplug.txt
@@ -302,4 +302,52 @@ consisting of <phys>, <size> and <maxcpus>.
 pseries guests use this property to note the maximum allowed CPUs for the
 guest.
 
+== ibm,dynamic-reconfiguration-memory ==
+
+ibm,dynamic-reconfiguration-memory is a device tree node that represents
+dynamically reconfigurable logical memory blocks (LMB). This node
+is generated only when the guest advertises the support for it via
+ibm,client-architecture-support call. Memory that is not dynamically
+reconfigurable is represented by /memory nodes. The properties of this
+node that are of interest to the sPAPR memory hotplug implementation
+in QEMU are described here.
+
+ibm,lmb-size
+
+This 64bit integer defines the size of each dynamically reconfigurable LMB.
+
+ibm,associativity-lookup-arrays
+
+This property defines a lookup array in which the NUMA associativity
+information for each LMB can be found. It is a property encoded array
+that begins with an integer M, the number of associativity lists followed
+by an integer N, the number of entries per associativity list and terminated
+by M associativity lists each of length N integers.
+
+This property provides the same information as given by ibm,associativity
+property in a /memory node. Each assigned LMB has an index value between
+0 and M-1 which is used as an index into this table to select which
+associativity list to use for the LMB. This index value for each LMB
+is defined in ibm,dynamic-memory property.
+
+ibm,dynamic-memory
+
+This property describes the dynamically reconfigurable memory. It is a
+property endoded array that has an integer N, the number of LMBs followed
+by N LMB list entires.
+
+Each LMB list entry consists of the following elements:
+
+- Logical address of the start of the LMB encoded as a 64bit integer. This
+  corresponds to reg property in /memory node.
+- DRC index of the LMB that corresponds to ibm,my-drc-index property
+  in a /memory node.
+- Four bytes reserved for expansion.
+- Associativity list index for the LMB that is used an index into
+  ibm,associativity-lookup-arrays property described earlier. This
+  is used to retrieve the right associativity list to be used for this
+  LMB.
+- A 32bit flags word. The bit at bit position 0x00000008 defines whether
+  the LMB is assigned to the the partition as of boot time.
+
 [1] http://thread.gmane.org/gmane.linux.ports.ppc.embedded/75350/focus=106867
diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index d7e0e44..aee8156 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -495,44 +495,7 @@ static void *spapr_create_fdt_skel(hwaddr initrd_base,
     return fdt;
 }
 
-int spapr_h_cas_compose_response(sPAPRMachineState *spapr,
-                                 target_ulong addr, target_ulong size)
-{
-    void *fdt, *fdt_skel;
-    sPAPRDeviceTreeUpdateHeader hdr = { .version_id = 1 };
-
-    size -= sizeof(hdr);
-
-    /* Create sceleton */
-    fdt_skel = g_malloc0(size);
-    _FDT((fdt_create(fdt_skel, size)));
-    _FDT((fdt_begin_node(fdt_skel, "")));
-    _FDT((fdt_end_node(fdt_skel)));
-    _FDT((fdt_finish(fdt_skel)));
-    fdt = g_malloc0(size);
-    _FDT((fdt_open_into(fdt_skel, fdt, size)));
-    g_free(fdt_skel);
-
-    /* Fix skeleton up */
-    _FDT((spapr_fixup_cpu_dt(fdt, spapr)));
-
-    /* Pack resulting tree */
-    _FDT((fdt_pack(fdt)));
-
-    if (fdt_totalsize(fdt) + sizeof(hdr) > size) {
-        trace_spapr_cas_failed(size);
-        return -1;
-    }
-
-    cpu_physical_memory_write(addr, &hdr, sizeof(hdr));
-    cpu_physical_memory_write(addr + sizeof(hdr), fdt, fdt_totalsize(fdt));
-    trace_spapr_cas_continue(fdt_totalsize(fdt) + sizeof(hdr));
-    g_free(fdt);
-
-    return 0;
-}
-
-static void spapr_populate_memory_node(void *fdt, int nodeid, hwaddr start,
+static int spapr_populate_memory_node(void *fdt, int nodeid, hwaddr start,
                                        hwaddr size)
 {
     uint32_t associativity[] = {
@@ -555,6 +518,7 @@ static void spapr_populate_memory_node(void *fdt, int nodeid, hwaddr start,
                       sizeof(mem_reg_property))));
     _FDT((fdt_setprop(fdt, off, "ibm,associativity", associativity,
                       sizeof(associativity))));
+    return off;
 }
 
 static int spapr_populate_memory(sPAPRMachineState *spapr, void *fdt)
@@ -586,7 +550,6 @@ static int spapr_populate_memory(sPAPRMachineState *spapr, void *fdt)
         }
         if (!mem_start) {
             /* ppc_spapr_init() checks for rma_size <= node0_size already */
-            spapr_populate_memory_node(fdt, i, 0, spapr->rma_size);
             mem_start += spapr->rma_size;
             node_size -= spapr->rma_size;
         }
@@ -738,6 +701,154 @@ static void spapr_populate_cpus_dt_node(void *fdt, sPAPRMachineState *spapr)
 
 }
 
+/*
+ * Adds ibm,dynamic-reconfiguration-memory node.
+ * Refer to docs/specs/ppc-spapr-hotplug.txt for the documentation
+ * of this device tree node.
+ */
+static int spapr_populate_drconf_memory(sPAPRMachineState *spapr, void *fdt)
+{
+    MachineState *machine = MACHINE(spapr);
+    int ret, i, offset;
+    uint64_t lmb_size = SPAPR_MEMORY_BLOCK_SIZE;
+    uint32_t prop_lmb_size[] = {0, cpu_to_be32(lmb_size)};
+    uint32_t nr_rma_lmbs = spapr->rma_size/lmb_size;
+    uint32_t nr_lmbs = machine->maxram_size/lmb_size - nr_rma_lmbs;
+    uint32_t nr_assigned_lmbs = machine->ram_size/lmb_size - nr_rma_lmbs;
+    uint32_t *int_buf, *cur_index, buf_len;
+
+    /* Allocate enough buffer size to fit in ibm,dynamic-memory */
+    buf_len = nr_lmbs * SPAPR_DR_LMB_LIST_ENTRY_SIZE * sizeof(uint32_t) +
+                sizeof(uint32_t);
+    cur_index = int_buf = g_malloc0(buf_len);
+
+    offset = fdt_add_subnode(fdt, 0, "ibm,dynamic-reconfiguration-memory");
+
+    ret = fdt_setprop(fdt, offset, "ibm,lmb-size", prop_lmb_size,
+                    sizeof(prop_lmb_size));
+    if (ret < 0) {
+        goto out;
+    }
+
+    ret = fdt_setprop_cell(fdt, offset, "ibm,memory-flags-mask", 0xff);
+    if (ret < 0) {
+        goto out;
+    }
+
+    ret = fdt_setprop_cell(fdt, offset, "ibm,memory-preservation-time", 0x0);
+    if (ret < 0) {
+        goto out;
+    }
+
+    /* ibm,dynamic-memory */
+    int_buf[0] = cpu_to_be32(nr_lmbs);
+    cur_index++;
+    for (i = 0; i < nr_lmbs; i++) {
+        sPAPRDRConnector *drc;
+        sPAPRDRConnectorClass *drck;
+        uint64_t addr;
+        uint32_t *dynamic_memory = cur_index;
+
+        if (i < nr_assigned_lmbs) {
+            addr = (i + nr_rma_lmbs) * lmb_size;
+        } else {
+            addr = (i - nr_assigned_lmbs) * lmb_size +
+                    spapr->hotplug_memory.base;
+        }
+        drc = spapr_dr_connector_by_id(SPAPR_DR_CONNECTOR_TYPE_LMB,
+                                       addr/lmb_size);
+        g_assert(drc);
+        drck = SPAPR_DR_CONNECTOR_GET_CLASS(drc);
+
+        dynamic_memory[0] = cpu_to_be32(addr >> 32);
+        dynamic_memory[1] = cpu_to_be32(addr & 0xffffffff);
+        dynamic_memory[2] = cpu_to_be32(drck->get_index(drc));
+        dynamic_memory[3] = cpu_to_be32(0); /* reserved */
+        dynamic_memory[4] = cpu_to_be32(numa_get_node(addr, NULL));
+        if (addr < machine->ram_size ||
+                    memory_region_present(get_system_memory(), addr)) {
+            dynamic_memory[5] = cpu_to_be32(SPAPR_LMB_FLAGS_ASSIGNED);
+        } else {
+            dynamic_memory[5] = cpu_to_be32(0);
+        }
+
+        cur_index += SPAPR_DR_LMB_LIST_ENTRY_SIZE;
+    }
+    ret = fdt_setprop(fdt, offset, "ibm,dynamic-memory", int_buf, buf_len);
+    if (ret < 0) {
+        goto out;
+    }
+
+    /* ibm,associativity-lookup-arrays */
+    cur_index = int_buf;
+    int_buf[0] = cpu_to_be32(nb_numa_nodes);
+    int_buf[1] = cpu_to_be32(4); /* Number of entries per associativity list */
+    cur_index += 2;
+    for (i = 0; i < nb_numa_nodes; i++) {
+        uint32_t associativity[] = {
+            cpu_to_be32(0x0),
+            cpu_to_be32(0x0),
+            cpu_to_be32(0x0),
+            cpu_to_be32(i)
+        };
+        memcpy(cur_index, associativity, sizeof(associativity));
+        cur_index += 4;
+    }
+    ret = fdt_setprop(fdt, offset, "ibm,associativity-lookup-arrays", int_buf,
+            (cur_index - int_buf) * sizeof(uint32_t));
+out:
+    g_free(int_buf);
+    return ret;
+}
+
+int spapr_h_cas_compose_response(sPAPRMachineState *spapr,
+                                 target_ulong addr, target_ulong size,
+                                 bool cpu_update, bool memory_update)
+{
+    void *fdt, *fdt_skel;
+    sPAPRDeviceTreeUpdateHeader hdr = { .version_id = 1 };
+    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(qdev_get_machine());
+
+    size -= sizeof(hdr);
+
+    /* Create sceleton */
+    fdt_skel = g_malloc0(size);
+    _FDT((fdt_create(fdt_skel, size)));
+    _FDT((fdt_begin_node(fdt_skel, "")));
+    _FDT((fdt_end_node(fdt_skel)));
+    _FDT((fdt_finish(fdt_skel)));
+    fdt = g_malloc0(size);
+    _FDT((fdt_open_into(fdt_skel, fdt, size)));
+    g_free(fdt_skel);
+
+    /* Fixup cpu nodes */
+    if (cpu_update) {
+        _FDT((spapr_fixup_cpu_dt(fdt, spapr)));
+    }
+
+    /* Generate memory nodes or ibm,dynamic-reconfiguration-memory node */
+    if (memory_update && smc->dr_lmb_enabled) {
+        _FDT((spapr_populate_drconf_memory(spapr, fdt)));
+    } else {
+        _FDT((spapr_populate_memory(spapr, fdt)));
+    }
+
+    /* Pack resulting tree */
+    _FDT((fdt_pack(fdt)));
+
+    if (fdt_totalsize(fdt) + sizeof(hdr) > size) {
+        trace_spapr_cas_failed(size);
+        return -1;
+    }
+
+    cpu_physical_memory_write(addr, &hdr, sizeof(hdr));
+    cpu_physical_memory_write(addr + sizeof(hdr), fdt, fdt_totalsize(fdt));
+    trace_spapr_cas_continue(fdt_totalsize(fdt) + sizeof(hdr));
+    g_free(fdt);
+
+    return 0;
+}
+
 static void spapr_finalize_fdt(sPAPRMachineState *spapr,
                                hwaddr fdt_addr,
                                hwaddr rtas_addr,
@@ -756,10 +867,23 @@ static void spapr_finalize_fdt(sPAPRMachineState *spapr,
     /* open out the base tree into a temp buffer for the final tweaks */
     _FDT((fdt_open_into(spapr->fdt_skel, fdt, FDT_MAX_SIZE)));
 
-    ret = spapr_populate_memory(spapr, fdt);
-    if (ret < 0) {
-        fprintf(stderr, "couldn't setup memory nodes in fdt\n");
-        exit(1);
+    /*
+     * Add memory@0 node to represent RMA. Rest of the memory is either
+     * represented by memory nodes or ibm,dynamic-reconfiguration-memory
+     * node later during ibm,client-architecture-support call.
+     *
+     * If NUMA is configured, ensure that memory@0 ends up in the
+     * first memory-less node.
+     */
+    if (nb_numa_nodes) {
+        for (i = 0; i < nb_numa_nodes; ++i) {
+            if (numa_info[i].node_mem) {
+                spapr_populate_memory_node(fdt, i, 0, spapr->rma_size);
+                break;
+            }
+        }
+    } else {
+        spapr_populate_memory_node(fdt, 0, 0, spapr->rma_size);
     }
 
     ret = spapr_populate_vdevice(spapr->vio_bus, fdt);
diff --git a/hw/ppc/spapr_hcall.c b/hw/ppc/spapr_hcall.c
index 652ddf6..2caac82 100644
--- a/hw/ppc/spapr_hcall.c
+++ b/hw/ppc/spapr_hcall.c
@@ -808,6 +808,32 @@ static target_ulong h_set_mode(PowerPCCPU *cpu, sPAPRMachineState *spapr,
     return ret;
 }
 
+/*
+ * Return the offset to the requested option vector @vector in the
+ * option vector table @table.
+ */
+static target_ulong cas_get_option_vector(int vector, target_ulong table)
+{
+    int i;
+    char nr_vectors, nr_entries;
+
+    if (!table) {
+        return 0;
+    }
+
+    nr_vectors = (ldl_phys(&address_space_memory, table) >> 24) + 1;
+    if (!vector || vector > nr_vectors) {
+        return 0;
+    }
+    table++; /* skip nr option vectors */
+
+    for (i = 0; i < vector - 1; i++) {
+        nr_entries = ldl_phys(&address_space_memory, table) >> 24;
+        table += nr_entries + 2;
+    }
+    return table;
+}
+
 typedef struct {
     PowerPCCPU *cpu;
     uint32_t cpu_version;
@@ -828,19 +854,22 @@ static void do_set_compat(void *arg)
     ((cpuver) == CPU_POWERPC_LOGICAL_2_06_PLUS) ? 2061 : \
     ((cpuver) == CPU_POWERPC_LOGICAL_2_07) ? 2070 : 0)
 
+#define OV5_DRCONF_MEMORY 0x20
+
 static target_ulong h_client_architecture_support(PowerPCCPU *cpu_,
                                                   sPAPRMachineState *spapr,
                                                   target_ulong opcode,
                                                   target_ulong *args)
 {
-    target_ulong list = args[0];
+    target_ulong list = args[0], ov_table;
     PowerPCCPUClass *pcc_ = POWERPC_CPU_GET_CLASS(cpu_);
     CPUState *cs;
-    bool cpu_match = false;
+    bool cpu_match = false, cpu_update = true, memory_update = false;
     unsigned old_cpu_version = cpu_->cpu_version;
     unsigned compat_lvl = 0, cpu_version = 0;
     unsigned max_lvl = get_compat_level(cpu_->max_compat);
     int counter;
+    char ov5_byte2;
 
     /* Parse PVR list */
     for (counter = 0; counter < 512; ++counter) {
@@ -890,8 +919,6 @@ static target_ulong h_client_architecture_support(PowerPCCPU *cpu_,
         }
     }
 
-    /* For the future use: here @list points to the first capability */
-
     /* Parsing finished */
     trace_spapr_cas_pvr(cpu_->cpu_version, cpu_match,
                         cpu_version, pcc_->pcr_mask);
@@ -915,14 +942,26 @@ static target_ulong h_client_architecture_support(PowerPCCPU *cpu_,
     }
 
     if (!cpu_version) {
-        return H_SUCCESS;
+        cpu_update = false;
     }
 
+    /* For the future use: here @ov_table points to the first option vector */
+    ov_table = list;
+
+    list = cas_get_option_vector(5, ov_table);
     if (!list) {
         return H_SUCCESS;
     }
 
-    if (spapr_h_cas_compose_response(spapr, args[1], args[2])) {
+    /* @list now points to OV 5 */
+    list += 2;
+    ov5_byte2 = rtas_ld(list, 0) >> 24;
+    if (ov5_byte2 & OV5_DRCONF_MEMORY) {
+        memory_update = true;
+    }
+
+    if (spapr_h_cas_compose_response(spapr, args[1], args[2],
+                                     cpu_update, memory_update)) {
         qemu_system_reset_request();
     }
 
diff --git a/include/hw/ppc/spapr.h b/include/hw/ppc/spapr.h
index a142e91..154b853 100644
--- a/include/hw/ppc/spapr.h
+++ b/include/hw/ppc/spapr.h
@@ -580,7 +580,8 @@ struct sPAPREventLogEntry {
 void spapr_events_init(sPAPRMachineState *sm);
 void spapr_events_fdt_skel(void *fdt, uint32_t epow_irq);
 int spapr_h_cas_compose_response(sPAPRMachineState *sm,
-                                 target_ulong addr, target_ulong size);
+                                 target_ulong addr, target_ulong size,
+                                 bool cpu_update, bool memory_update);
 sPAPRTCETable *spapr_tce_new_table(DeviceState *owner, uint32_t liobn,
                                    uint64_t bus_offset,
                                    uint32_t page_shift,
@@ -622,4 +623,16 @@ int spapr_rtc_import_offset(DeviceState *dev, int64_t legacy_offset);
 /* 1GB alignment for hotplug memory region */
 #define SPAPR_HOTPLUG_MEM_ALIGN (1ULL << 30)
 
+/*
+ * Number of 32 bit words in each LMB list entry in ibm,dynamic-memory
+ * property under ibm,dynamic-reconfiguration-memory node.
+ */
+#define SPAPR_DR_LMB_LIST_ENTRY_SIZE 6
+
+/*
+ * This flag value defines the LMB as assigned in ibm,dynamic-memory
+ * property under ibm,dynamic-reconfiguration-memory node.
+ */
+#define SPAPR_LMB_FLAGS_ASSIGNED 0x00000008
+
 #endif /* !defined (__HW_SPAPR_H__) */
-- 
2.1.0

[Qemu-devel] [PATCH v5 4/6] spapr: Make hash table size a factor of maxram_size

[Bharata B Rao]

The hash table size is dependent on ram_size, but since with hotplug
the memory can grow till maxram_size. Hence make hash table size dependent
on maxram_size.

This allows to hotplug huge amounts of memory to the guest.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/ppc/spapr.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index aee8156..0ab78fa 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -1697,7 +1697,7 @@ static void ppc_spapr_init(MachineState *machine)
      * more than needed for the Linux guests we support. */
     spapr->htab_shift = 18; /* Minimum architected size */
     while (spapr->htab_shift <= 46) {
-        if ((1ULL << (spapr->htab_shift + 7)) >= machine->ram_size) {
+        if ((1ULL << (spapr->htab_shift + 7)) >= machine->maxram_size) {
             break;
         }
         spapr->htab_shift++;
-- 
2.1.0

[Qemu-devel] [PATCH v5 5/6] spapr: Memory hotplug support

[Bharata B Rao]

Make use of pc-dimm infrastructure to support memory hotplug
for PowerPC.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
---
 hw/ppc/spapr.c        | 129 ++++++++++++++++++++++++++++++++++++++++++++++++++
 hw/ppc/spapr_events.c |   8 ++--
 2 files changed, 134 insertions(+), 3 deletions(-)

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 0ab78fa..f4792b8 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -33,6 +33,7 @@
 #include "sysemu/block-backend.h"
 #include "sysemu/cpus.h"
 #include "sysemu/kvm.h"
+#include "sysemu/device_tree.h"
 #include "kvm_ppc.h"
 #include "mmu-hash64.h"
 #include "qom/cpu.h"
@@ -855,6 +856,7 @@ static void spapr_finalize_fdt(sPAPRMachineState *spapr,
                                hwaddr rtas_size)
 {
     MachineState *machine = MACHINE(qdev_get_machine());
+    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(machine);
     const char *boot_device = machine->boot_order;
     int ret, i;
     size_t cb = 0;
@@ -938,6 +940,10 @@ static void spapr_finalize_fdt(sPAPRMachineState *spapr,
         spapr_populate_chosen_stdout(fdt, spapr->vio_bus);
     }
 
+    if (smc->dr_lmb_enabled) {
+        _FDT(spapr_drc_populate_dt(fdt, 0, NULL, SPAPR_DR_CONNECTOR_TYPE_LMB));
+    }
+
     _FDT((fdt_pack(fdt)));
 
     if (fdt_totalsize(fdt) > FDT_MAX_SIZE) {
@@ -2036,12 +2042,131 @@ static void spapr_nmi(NMIState *n, int cpu_index, Error **errp)
     }
 }
 
+static void spapr_add_lmbs(DeviceState *dev, uint64_t addr, uint64_t size,
+                           uint32_t node, Error **errp)
+{
+    sPAPRDRConnector *drc;
+    sPAPRDRConnectorClass *drck;
+    uint32_t nr_lmbs = size/SPAPR_MEMORY_BLOCK_SIZE;
+    int i, fdt_offset, fdt_size;
+    void *fdt;
+    Error *local_err = NULL;
+
+    if (size % SPAPR_MEMORY_BLOCK_SIZE) {
+        error_setg(errp, "Hotplugged memory size must be a multiple of "
+                      "%lld MB", SPAPR_MEMORY_BLOCK_SIZE/M_BYTE);
+        return;
+    }
+
+    /*
+     * Check for DRC connectors and send hotplug notification to the
+     * guest only in case of hotplugged memory. This allows cold plugged
+     * memory to be specified at boot time.
+     */
+    if (!dev->hotplugged) {
+        return;
+    }
+
+    for (i = 0; i < nr_lmbs; i++) {
+        drc = spapr_dr_connector_by_id(SPAPR_DR_CONNECTOR_TYPE_LMB,
+                addr/SPAPR_MEMORY_BLOCK_SIZE);
+        g_assert(drc);
+
+        fdt = create_device_tree(&fdt_size);
+        fdt_offset = spapr_populate_memory_node(fdt, node, addr,
+                                                SPAPR_MEMORY_BLOCK_SIZE);
+
+        drck = SPAPR_DR_CONNECTOR_GET_CLASS(drc);
+        drck->attach(drc, dev, fdt, fdt_offset, !dev->hotplugged, &local_err);
+        if (local_err) {
+            g_free(fdt);
+            error_propagate(errp, local_err);
+            return;
+        }
+
+        spapr_hotplug_req_add_event(drc);
+        addr += SPAPR_MEMORY_BLOCK_SIZE;
+    }
+}
+
+static void spapr_memory_plug(HotplugHandler *hotplug_dev, DeviceState *dev,
+                              uint32_t node, Error **errp)
+{
+    Error *local_err = NULL;
+    sPAPRMachineState *ms = SPAPR_MACHINE(hotplug_dev);
+    PCDIMMDevice *dimm = PC_DIMM(dev);
+    PCDIMMDeviceClass *ddc = PC_DIMM_GET_CLASS(dimm);
+    MemoryRegion *mr = ddc->get_memory_region(dimm);
+    uint64_t align = memory_region_get_alignment(mr);
+    uint64_t addr;
+
+    pc_dimm_memory_plug(dev, &ms->hotplug_memory, mr, align, &local_err);
+    if (local_err) {
+        goto out;
+    }
+
+    addr = object_property_get_int(OBJECT(dimm), PC_DIMM_ADDR_PROP, &local_err);
+    if (local_err) {
+        goto out_unplug;
+    }
+
+    spapr_add_lmbs(dev, addr, memory_region_size(mr), node, &local_err);
+    if (local_err) {
+        goto out_unplug;
+    }
+    return;
+
+out_unplug:
+    pc_dimm_memory_unplug(dev, &ms->hotplug_memory, mr);
+out:
+    error_propagate(errp, local_err);
+}
+
+static void spapr_machine_device_plug(HotplugHandler *hotplug_dev,
+                                      DeviceState *dev, Error **errp)
+{
+    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(qdev_get_machine());
+
+    if (object_dynamic_cast(OBJECT(dev), TYPE_PC_DIMM)) {
+        uint32_t node;
+
+        if (!smc->dr_lmb_enabled) {
+            error_setg(errp, "Memory hotplug not supported for this machine");
+            return;
+        }
+        node = object_property_get_int(OBJECT(dev), PC_DIMM_NODE_PROP, errp);
+        if (*errp) {
+            return;
+        }
+
+        spapr_memory_plug(hotplug_dev, dev, node, errp);
+    }
+}
+
+static void spapr_machine_device_unplug(HotplugHandler *hotplug_dev,
+                                      DeviceState *dev, Error **errp)
+{
+    if (object_dynamic_cast(OBJECT(dev), TYPE_PC_DIMM)) {
+        error_setg(errp, "Memory hot unplug not supported by sPAPR");
+    }
+}
+
+static HotplugHandler *spapr_get_hotpug_handler(MachineState *machine,
+                                             DeviceState *dev)
+{
+    if (object_dynamic_cast(OBJECT(dev), TYPE_PC_DIMM)) {
+        return HOTPLUG_HANDLER(machine);
+    }
+    return NULL;
+}
+
 static void spapr_machine_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
     sPAPRMachineClass *smc = SPAPR_MACHINE_CLASS(oc);
     FWPathProviderClass *fwc = FW_PATH_PROVIDER_CLASS(oc);
     NMIClass *nc = NMI_CLASS(oc);
+    HotplugHandlerClass *hc = HOTPLUG_HANDLER_CLASS(oc);
 
     mc->init = ppc_spapr_init;
     mc->reset = ppc_spapr_reset;
@@ -2052,6 +2177,9 @@ static void spapr_machine_class_init(ObjectClass *oc, void *data)
     mc->default_ram_size = 512 * M_BYTE;
     mc->kvm_type = spapr_kvm_type;
     mc->has_dynamic_sysbus = true;
+    mc->get_hotplug_handler = spapr_get_hotpug_handler;
+    hc->plug = spapr_machine_device_plug;
+    hc->unplug = spapr_machine_device_unplug;
 
     smc->dr_lmb_enabled = false;
     fwc->get_dev_path = spapr_get_fw_dev_path;
@@ -2069,6 +2197,7 @@ static const TypeInfo spapr_machine_info = {
     .interfaces = (InterfaceInfo[]) {
         { TYPE_FW_PATH_PROVIDER },
         { TYPE_NMI },
+        { TYPE_HOTPLUG_HANDLER },
         { }
     },
 };
diff --git a/hw/ppc/spapr_events.c b/hw/ppc/spapr_events.c
index f626eb7..98bf7ae 100644
--- a/hw/ppc/spapr_events.c
+++ b/hw/ppc/spapr_events.c
@@ -427,14 +427,16 @@ static void spapr_hotplug_req_event(sPAPRDRConnector *drc, uint8_t hp_action)
     hp->hdr.section_length = cpu_to_be16(sizeof(*hp));
     hp->hdr.section_version = 1; /* includes extended modifier */
     hp->hotplug_action = hp_action;
-
+    hp->drc.index = cpu_to_be32(drck->get_index(drc));
+    hp->hotplug_identifier = RTAS_LOG_V6_HP_ID_DRC_INDEX;
 
     switch (drc_type) {
     case SPAPR_DR_CONNECTOR_TYPE_PCI:
-        hp->drc.index = cpu_to_be32(drck->get_index(drc));
-        hp->hotplug_identifier = RTAS_LOG_V6_HP_ID_DRC_INDEX;
         hp->hotplug_type = RTAS_LOG_V6_HP_TYPE_PCI;
         break;
+    case SPAPR_DR_CONNECTOR_TYPE_LMB:
+        hp->hotplug_type = RTAS_LOG_V6_HP_TYPE_MEMORY;
+        break;
     default:
         /* we shouldn't be signaling hotplug events for resources
          * that don't support them
-- 
2.1.0

[Qemu-devel] [RFC PATCH v5 6/6] spapr: Don't allow memory hotplug to memory less nodes

[Bharata B Rao]

Currently PowerPC kernel doesn't allow hot-adding memory to memory-less
node, but instead will silently add the memory to the first node that has
some memory. This causes two unexpected behaviours for the user.

Memory gets hotplugged to a different node than what the user specified.
Since pc-dimm subsystem in QEMU still thinks that memory belongs to
memory-less node, a reboot will set things accordingly and the previously
hotplugged memory now ends in the right node. This appears as if some
memory moved from one node to another.

So until kernel starts supporting memory hotplug to memory-less
nodes, just prevent such attempts upfront in QEMU.

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

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index f4792b8..003ef14 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -2139,6 +2139,28 @@ static void spapr_machine_device_plug(HotplugHandler *hotplug_dev,
             return;
         }
 
+        /*
+         * Currently PowerPC kernel doesn't allow hot-adding memory to
+         * memory-less node, but instead will silently add the memory
+         * to the first node that has some memory. This causes two
+         * unexpected behaviours for the user.
+         *
+         * - Memory gets hotplugged to a different node than what the user
+         *   specified.
+         * - Since pc-dimm subsystem in QEMU still thinks that memory belongs
+         *   to memory-less node, a reboot will set things accordingly
+         *   and the previously hotplugged memory now ends in the right node.
+         *   This appears as if some memory moved from one node to another.
+         *
+         * So until kernel starts supporting memory hotplug to memory-less
+         * nodes, just prevent such attempts upfront in QEMU.
+         */
+        if (nb_numa_nodes && !numa_info[node].node_mem) {
+            error_setg(errp, "Can't hotplug memory to memory-less node %d",
+                       node);
+            return;
+        }
+
         spapr_memory_plug(hotplug_dev, dev, node, errp);
     }
 }
-- 
2.1.0

Re: [Qemu-devel] [PATCH v5 1/6] spapr: Initialize hotplug memory address space

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 581 bytes --]

On Thu, Jun 25, 2015 at 11:44:10AM +0530, Bharata B Rao wrote:
> Initialize a hotplug memory region under which all the hotplugged
> memory is accommodated. Also enable memory hotplug by setting
> CONFIG_MEM_HOTPLUG.
> 
> Modelled on i386 memory hotplug.
> 
> Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [PATCH v5 3/6] spapr: Support ibm, dynamic-reconfiguration-memory

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1086 bytes --]

On Thu, Jun 25, 2015 at 11:44:12AM +0530, Bharata B Rao wrote:
> Parse ibm,architecture.vec table obtained from the guest and enable
> memory node configuration via ibm,dynamic-reconfiguration-memory if guest
> supports it. This is in preparation to support memory hotplug for
> sPAPR guests.
> 
> This changes the way memory node configuration is done. Currently all
> memory nodes are built upfront. But after this patch, only memory@0 node
> for RMA is built upfront. Guest kernel boots with just that and rest of
> the memory nodes (via memory@XXX or ibm,dynamic-reconfiguration-memory)
> are built when guest does ibm,client-architecture-support call.
> 
> Note: This patch needs a SLOF enhancement which is already part of
> SLOF binary in QEMU.
> 
> Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [PATCH v5 5/6] spapr: Memory hotplug support

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 464 bytes --]

On Thu, Jun 25, 2015 at 11:44:14AM +0530, Bharata B Rao wrote:
> Make use of pc-dimm infrastructure to support memory hotplug
> for PowerPC.
> 
> Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [RFC PATCH v5 6/6] spapr: Don't allow memory hotplug to memory less nodes

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1303 bytes --]

On Thu, Jun 25, 2015 at 11:44:15AM +0530, Bharata B Rao wrote:
> Currently PowerPC kernel doesn't allow hot-adding memory to memory-less
> node, but instead will silently add the memory to the first node that has
> some memory. This causes two unexpected behaviours for the user.
> 
> Memory gets hotplugged to a different node than what the user specified.
> Since pc-dimm subsystem in QEMU still thinks that memory belongs to
> memory-less node, a reboot will set things accordingly and the previously
> hotplugged memory now ends in the right node. This appears as if some
> memory moved from one node to another.
> 
> So until kernel starts supporting memory hotplug to memory-less
> nodes, just prevent such attempts upfront in QEMU.
> 
> Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>

I presume you're talking about host kernel limitations here?

If that does get changed in future we'll need some sort of ugly compat
test in qemu to see if the new kernel can handle it, but there's no
way to avoid that, so.

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [PATCH v5 2/6] spapr: Add LMB DR connectors

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 4187 bytes --]

On Thu, Jun 25, 2015 at 11:44:11AM +0530, Bharata B Rao wrote:
> Enable memory hotplug for pseries 2.4 and add LMB DR connectors.
> With memory hotplug, enforce RAM size, NUMA node memory size and maxmem
> to be a multiple of SPAPR_MEMORY_BLOCK_SIZE (256M) since that's the
> granularity in which LMBs are represented and hot-added.
> 
> LMB DR connectors will be used by the memory hotplug code.
> 
> Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
> Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
>                [spapr_drc_reset implementation]
> ---
>  hw/ppc/spapr.c         | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  include/hw/ppc/spapr.h |  1 +
>  2 files changed, 80 insertions(+)
> 
> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> index 241ecad..d7e0e44 100644
> --- a/hw/ppc/spapr.c
> +++ b/hw/ppc/spapr.c
> @@ -59,6 +59,7 @@
>  #include "hw/nmi.h"
>  
>  #include "hw/compat.h"
> +#include "qemu-common.h"
>  
>  #include <libfdt.h>
>  
> @@ -1436,10 +1437,76 @@ static void spapr_cpu_init(sPAPRMachineState *spapr, PowerPCCPU *cpu)
>      qemu_register_reset(spapr_cpu_reset, cpu);
>  }
>  
> +/*
> + * Reset routine for LMB DR devices.
> + *
> + * Unlike PCI DR devices, LMB DR devices explicitly register this reset
> + * routine. Reset for PCI DR devices will be handled by PHB reset routine
> + * when it walks all its children devices. LMB devices reset occurs
> + * as part of spapr_ppc_reset().
> + */
> +static void spapr_drc_reset(void *opaque)
> +{
> +    sPAPRDRConnector *drc = opaque;
> +    DeviceState *d = DEVICE(drc);
> +
> +    if (d) {
> +        device_reset(d);
> +    }
> +}
> +
> +static void spapr_create_lmb_dr_connectors(sPAPRMachineState *spapr)
> +{
> +    MachineState *machine = MACHINE(spapr);
> +    uint64_t lmb_size = SPAPR_MEMORY_BLOCK_SIZE;
> +    uint32_t nr_rma_lmbs = spapr->rma_size/lmb_size;
> +    uint32_t nr_lmbs = machine->maxram_size/lmb_size - nr_rma_lmbs;
> +    uint32_t nr_assigned_lmbs = machine->ram_size/lmb_size - nr_rma_lmbs;
> +    int i;
> +
> +    for (i = 0; i < nr_lmbs; i++) {
> +        sPAPRDRConnector *drc;
> +        uint64_t addr;
> +
> +        if (i < nr_assigned_lmbs) {
> +            addr = (i + nr_rma_lmbs) * lmb_size;
> +        } else {
> +            addr = (i - nr_assigned_lmbs) * lmb_size +
> +                    spapr->hotplug_memory.base;
> +        }
> +        drc = spapr_dr_connector_new(OBJECT(spapr), SPAPR_DR_CONNECTOR_TYPE_LMB,
> +                                     addr/lmb_size);
> +        qemu_register_reset(spapr_drc_reset, drc);
> +    }
> +}
> +
> +/*
> + * If RAM size, maxmem size and individual node mem sizes aren't aligned
> + * to SPAPR_MEMORY_BLOCK_SIZE(256MB), then disable LMB DR feature.
> + */
> +static void spapr_validate_node_memory(MachineState *machine)
> +{
> +    int i;
> +    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(machine);
> +
> +    if (machine->maxram_size % SPAPR_MEMORY_BLOCK_SIZE ||
> +        machine->ram_size % SPAPR_MEMORY_BLOCK_SIZE) {
> +        smc->dr_lmb_enabled = false;
> +    }
> +
> +    for (i = 0; i < nb_numa_nodes; i++) {
> +        if (numa_info[i].node_mem &&
> +            numa_info[i].node_mem % SPAPR_MEMORY_BLOCK_SIZE) {
> +            smc->dr_lmb_enabled = false;

Hrm, little nit here.

First, I'm not sure that changing a value in the MachineClass after
initialization is a QOM-ly correct thing to do at all.

I'm also concerned that this will silently disable a feature the user
expected, based on what may not be an obvious problem in the specified
options.

This is only called from ppc_spapr_init() - not on every reset, so I
think it would be better to fail out completely (with a useful error
message), rather than carry on silently disabling memory hotplug.
(Obviously don't fail if the user has explicitly disabled mem hotplug
with their machine type choice).

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [RFC PATCH v5 6/6] spapr: Don't allow memory hotplug to memory less nodes

[Bharata B Rao]

On Fri, Jun 26, 2015 at 03:33:29PM +1000, David Gibson wrote:
> On Thu, Jun 25, 2015 at 11:44:15AM +0530, Bharata B Rao wrote:
> > Currently PowerPC kernel doesn't allow hot-adding memory to memory-less
> > node, but instead will silently add the memory to the first node that has
> > some memory. This causes two unexpected behaviours for the user.
> > 
> > Memory gets hotplugged to a different node than what the user specified.
> > Since pc-dimm subsystem in QEMU still thinks that memory belongs to
> > memory-less node, a reboot will set things accordingly and the previously
> > hotplugged memory now ends in the right node. This appears as if some
> > memory moved from one node to another.
> > 
> > So until kernel starts supporting memory hotplug to memory-less
> > nodes, just prevent such attempts upfront in QEMU.
> > 
> > Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
> 
> I presume you're talking about host kernel limitations here?

No, this is a guest kernel limitation. Pls refer to the related discussions
about this here:

https://lkml.org/lkml/2015/6/24/1

> 
> If that does get changed in future we'll need some sort of ugly compat
> test in qemu to see if the new kernel can handle it, but there's no
> way to avoid that, so.

Yeah, I see that.

Regards,
Bharata.

Re: [Qemu-devel] [PATCH v5 0/6] Memory hotplug for PowerPC sPAPR guests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 736 bytes --]

On Thu, Jun 25, 2015 at 11:44:09AM +0530, Bharata B Rao wrote:
> Hi,
> 
> This is v5 of memory hotplug support patchset for PowerPC
> sPAPR guests.
> 
> This patchset applies on spapr-next branch of David Gibson's tree with
> the other prerequisite patchset applied. Pre-requistes patchset was
> posted at:
> https://lists.gnu.org/archive/html/qemu-devel/2015-06/msg05157.html

I have one slight concern noted on patch 2/5.  Otherwise, I'm ready to
pull these into spapr-next as soon as the generic prereq patches are
merged.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [PATCH v5 2/6] spapr: Add LMB DR connectors

[Bharata B Rao]

On Fri, Jun 26, 2015 at 03:38:41PM +1000, David Gibson wrote:
> > + */
> > +static void spapr_validate_node_memory(MachineState *machine)
> > +{
> > +    int i;
> > +    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(machine);
> > +
> > +    if (machine->maxram_size % SPAPR_MEMORY_BLOCK_SIZE ||
> > +        machine->ram_size % SPAPR_MEMORY_BLOCK_SIZE) {
> > +        smc->dr_lmb_enabled = false;
> > +    }
> > +
> > +    for (i = 0; i < nb_numa_nodes; i++) {
> > +        if (numa_info[i].node_mem &&
> > +            numa_info[i].node_mem % SPAPR_MEMORY_BLOCK_SIZE) {
> > +            smc->dr_lmb_enabled = false;
> 
> Hrm, little nit here.
> 
> First, I'm not sure that changing a value in the MachineClass after
> initialization is a QOM-ly correct thing to do at all.
> 
> I'm also concerned that this will silently disable a feature the user
> expected, based on what may not be an obvious problem in the specified
> options.
> 
> This is only called from ppc_spapr_init() - not on every reset, so I
> think it would be better to fail out completely (with a useful error
> message), rather than carry on silently disabling memory hotplug.
> (Obviously don't fail if the user has explicitly disabled mem hotplug
> with their machine type choice).

Since we validate these memory sizes only for newer pseries machines
that support LMB DR, it should be fine to just terminate instead
of silently disabling memory hotplug. Here is the replacement patch
which does this. If you prefer a complete re-spin due to this, I can
do that as well.

Regards,
Bharata.
------------------------------

spapr: Add LMB DR connectors

From: Bharata B Rao <bharata@linux.vnet.ibm.com>

Enable memory hotplug for pseries 2.4 and add LMB DR connectors.
With memory hotplug, enforce RAM size, NUMA node memory size and maxmem
to be a multiple of SPAPR_MEMORY_BLOCK_SIZE (256M) since that's the
granularity in which LMBs are represented and hot-added.

LMB DR connectors will be used by the memory hotplug code.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
               [spapr_drc_reset implementation]
---
 hw/ppc/spapr.c         |   88 ++++++++++++++++++++++++++++++++++++++++++++++++
 include/hw/ppc/spapr.h |    1 +
 2 files changed, 89 insertions(+)

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 241ecad..bee868c 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -59,6 +59,7 @@
 #include "hw/nmi.h"
 
 #include "hw/compat.h"
+#include "qemu-common.h"
 
 #include <libfdt.h>
 
@@ -1436,10 +1437,85 @@ static void spapr_cpu_init(sPAPRMachineState *spapr, PowerPCCPU *cpu)
     qemu_register_reset(spapr_cpu_reset, cpu);
 }
 
+/*
+ * Reset routine for LMB DR devices.
+ *
+ * Unlike PCI DR devices, LMB DR devices explicitly register this reset
+ * routine. Reset for PCI DR devices will be handled by PHB reset routine
+ * when it walks all its children devices. LMB devices reset occurs
+ * as part of spapr_ppc_reset().
+ */
+static void spapr_drc_reset(void *opaque)
+{
+    sPAPRDRConnector *drc = opaque;
+    DeviceState *d = DEVICE(drc);
+
+    if (d) {
+        device_reset(d);
+    }
+}
+
+static void spapr_create_lmb_dr_connectors(sPAPRMachineState *spapr)
+{
+    MachineState *machine = MACHINE(spapr);
+    uint64_t lmb_size = SPAPR_MEMORY_BLOCK_SIZE;
+    uint32_t nr_rma_lmbs = spapr->rma_size/lmb_size;
+    uint32_t nr_lmbs = machine->maxram_size/lmb_size - nr_rma_lmbs;
+    uint32_t nr_assigned_lmbs = machine->ram_size/lmb_size - nr_rma_lmbs;
+    int i;
+
+    for (i = 0; i < nr_lmbs; i++) {
+        sPAPRDRConnector *drc;
+        uint64_t addr;
+
+        if (i < nr_assigned_lmbs) {
+            addr = (i + nr_rma_lmbs) * lmb_size;
+        } else {
+            addr = (i - nr_assigned_lmbs) * lmb_size +
+                    spapr->hotplug_memory.base;
+        }
+        drc = spapr_dr_connector_new(OBJECT(spapr), SPAPR_DR_CONNECTOR_TYPE_LMB,
+                                     addr/lmb_size);
+        qemu_register_reset(spapr_drc_reset, drc);
+    }
+}
+
+/*
+ * If RAM size, maxmem size and individual node mem sizes aren't aligned
+ * to SPAPR_MEMORY_BLOCK_SIZE(256MB), then refuse to start the guest
+ * since we can't support such unaligned sizes with DRCONF_MEMORY.
+ */
+static void spapr_validate_node_memory(MachineState *machine)
+{
+    int i;
+
+    if (machine->maxram_size % SPAPR_MEMORY_BLOCK_SIZE ||
+        machine->ram_size % SPAPR_MEMORY_BLOCK_SIZE) {
+        error_report("Can't support memory configuration where RAM size "
+                     "0x" RAM_ADDR_FMT " or maxmem size "
+                     "0x" RAM_ADDR_FMT " isn't aligned to %lld MB",
+                     machine->ram_size, machine->maxram_size,
+                     SPAPR_MEMORY_BLOCK_SIZE/M_BYTE);
+        exit(EXIT_FAILURE);
+    }
+
+    for (i = 0; i < nb_numa_nodes; i++) {
+        if (numa_info[i].node_mem &&
+            numa_info[i].node_mem % SPAPR_MEMORY_BLOCK_SIZE) {
+            error_report("Can't support memory configuration where memory "
+                         "size %lx of node %d isn't aligned to %lld MB",
+                         numa_info[i].node_mem, i,
+                         SPAPR_MEMORY_BLOCK_SIZE/M_BYTE);
+            exit(EXIT_FAILURE);
+        }
+    }
+}
+
 /* pSeries LPAR / sPAPR hardware init */
 static void ppc_spapr_init(MachineState *machine)
 {
     sPAPRMachineState *spapr = SPAPR_MACHINE(machine);
+    sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(machine);
     const char *kernel_filename = machine->kernel_filename;
     const char *kernel_cmdline = machine->kernel_cmdline;
     const char *initrd_filename = machine->initrd_filename;
@@ -1518,6 +1594,10 @@ static void ppc_spapr_init(MachineState *machine)
                                                smp_threads),
                                   XICS_IRQS);
 
+    if (smc->dr_lmb_enabled) {
+        spapr_validate_node_memory(machine);
+    }
+
     /* init CPUs */
     if (machine->cpu_model == NULL) {
         machine->cpu_model = kvm_enabled() ? "host" : "POWER7";
@@ -1567,6 +1647,10 @@ static void ppc_spapr_init(MachineState *machine)
                                     &spapr->hotplug_memory.mr);
     }
 
+    if (smc->dr_lmb_enabled) {
+        spapr_create_lmb_dr_connectors(spapr);
+    }
+
     filename = qemu_find_file(QEMU_FILE_TYPE_BIOS, "spapr-rtas.bin");
     if (!filename) {
         error_report("Could not find LPAR rtas '%s'", "spapr-rtas.bin");
@@ -1840,6 +1924,7 @@ static void spapr_nmi(NMIState *n, int cpu_index, Error **errp)
 static void spapr_machine_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    sPAPRMachineClass *smc = SPAPR_MACHINE_CLASS(oc);
     FWPathProviderClass *fwc = FW_PATH_PROVIDER_CLASS(oc);
     NMIClass *nc = NMI_CLASS(oc);
 
@@ -1853,6 +1938,7 @@ static void spapr_machine_class_init(ObjectClass *oc, void *data)
     mc->kvm_type = spapr_kvm_type;
     mc->has_dynamic_sysbus = true;
 
+    smc->dr_lmb_enabled = false;
     fwc->get_dev_path = spapr_get_fw_dev_path;
     nc->nmi_monitor_handler = spapr_nmi;
 }
@@ -1988,11 +2074,13 @@ static const TypeInfo spapr_machine_2_3_info = {
 static void spapr_machine_2_4_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    sPAPRMachineClass *smc = SPAPR_MACHINE_CLASS(oc);
 
     mc->name = "pseries-2.4";
     mc->desc = "pSeries Logical Partition (PAPR compliant) v2.4";
     mc->alias = "pseries";
     mc->is_default = 1;
+    smc->dr_lmb_enabled = true;
 }
 
 static const TypeInfo spapr_machine_2_4_info = {
diff --git a/include/hw/ppc/spapr.h b/include/hw/ppc/spapr.h
index 8a1929b..a142e91 100644
--- a/include/hw/ppc/spapr.h
+++ b/include/hw/ppc/spapr.h
@@ -35,6 +35,7 @@ struct sPAPRMachineClass {
     MachineClass parent_class;
 
     /*< public >*/
+    bool dr_lmb_enabled; /* enable dynamic-reconfig/hotplug of LMBs */
 };
 
 /**