nvdimm acpi: build ACPI NFIT table

NFIT is defined in ACPI 6.0: 5.2.25 NVDIMM Firmware Interface Table (NFIT)

Currently, we only support PMEM mode. Each device has 3 structures:
- SPA structure, defines the PMEM region info

- MEM DEV structure, it has the @handle which is used to associate specified
  ACPI NVDIMM  device we will introduce in later patch.
  Also we can happily ignored the memory device's interleave, the real
  nvdimm hardware access is hidden behind host

- DCR structure, it defines vendor ID used to associate specified vendor
  nvdimm driver. Since we only implement PMEM mode this time, Command
  window and Data window are not needed

The NVDIMM functionality is controlled by the parameter, 'nvdimm', which
is introduced for the machine, there is a example to enable it:
-machine pc,nvdimm -m 8G,maxmem=100G,slots=100  -object \
memory-backend-file,id=mem1,share,mem-path=/tmp/nvdimm1,size=10G -device \
nvdimm,memdev=mem1,id=nv1

It is disabled on default

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

9 files changed, 425 insertions, 1 deletions

diff --git a/default-configs/i386-softmmu.mak b/default-configs/i386-softmmu.mak
index b0355993e1..b177e52104 100644
--- a/default-configs/i386-softmmu.mak
+++ b/default-configs/i386-softmmu.mak
@@ -52,6 +52,7 @@ CONFIG_IOAPIC=y
 CONFIG_PVPANIC=y
 CONFIG_MEM_HOTPLUG=y
 CONFIG_NVDIMM=y
+CONFIG_ACPI_NVDIMM=y
 CONFIG_XIO3130=y
 CONFIG_IOH3420=y
 CONFIG_I82801B11=y
diff --git a/default-configs/x86_64-softmmu.mak b/default-configs/x86_64-softmmu.mak
index 27669bec58..6e3b312c5f 100644
--- a/default-configs/x86_64-softmmu.mak
+++ b/default-configs/x86_64-softmmu.mak
@@ -52,6 +52,7 @@ CONFIG_IOAPIC=y
 CONFIG_PVPANIC=y
 CONFIG_MEM_HOTPLUG=y
 CONFIG_NVDIMM=y
+CONFIG_ACPI_NVDIMM=y
 CONFIG_XIO3130=y
 CONFIG_IOH3420=y
 CONFIG_I82801B11=y
diff --git a/hw/acpi/Makefile.objs b/hw/acpi/Makefile.objs
index 7d3230c2a5..095597f791 100644
--- a/hw/acpi/Makefile.objs
+++ b/hw/acpi/Makefile.objs
@@ -2,6 +2,7 @@ common-obj-$(CONFIG_ACPI_X86) += core.o piix4.o pcihp.o
 common-obj-$(CONFIG_ACPI_X86_ICH) += ich9.o tco.o
 common-obj-$(CONFIG_ACPI_CPU_HOTPLUG) += cpu_hotplug.o
 common-obj-$(CONFIG_ACPI_MEMORY_HOTPLUG) += memory_hotplug.o
+common-obj-$(CONFIG_ACPI_NVDIMM) += nvdimm.o
 common-obj-$(CONFIG_ACPI) += acpi_interface.o
 common-obj-$(CONFIG_ACPI) += bios-linker-loader.o
 common-obj-$(CONFIG_ACPI) += aml-build.o
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
new file mode 100644
index 0000000000..98c004d1a5
--- /dev/null
+++ b/hw/acpi/nvdimm.c
@@ -0,0 +1,382 @@
+/*
+ * NVDIMM ACPI Implementation
+ *
+ * Copyright(C) 2015 Intel Corporation.
+ *
+ * Author:
+ *  Xiao Guangrong <guangrong.xiao@linux.intel.com>
+ *
+ * NFIT is defined in ACPI 6.0: 5.2.25 NVDIMM Firmware Interface Table (NFIT)
+ * and the DSM specification can be found at:
+ *       http://pmem.io/documents/NVDIMM_DSM_Interface_Example.pdf
+ *
+ * Currently, it only supports PMEM Virtualization.
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, see <http://www.gnu.org/licenses/>
+ */
+
+#include "hw/acpi/acpi.h"
+#include "hw/acpi/aml-build.h"
+#include "hw/mem/nvdimm.h"
+
+static int nvdimm_plugged_device_list(Object *obj, void *opaque)
+{
+    GSList **list = opaque;
+
+    if (object_dynamic_cast(obj, TYPE_NVDIMM)) {
+        DeviceState *dev = DEVICE(obj);
+
+        if (dev->realized) { /* only realized NVDIMMs matter */
+            *list = g_slist_append(*list, DEVICE(obj));
+        }
+    }
+
+    object_child_foreach(obj, nvdimm_plugged_device_list, opaque);
+    return 0;
+}
+
+/*
+ * inquire plugged NVDIMM devices and link them into the list which is
+ * returned to the caller.
+ *
+ * Note: it is the caller's responsibility to free the list to avoid
+ * memory leak.
+ */
+static GSList *nvdimm_get_plugged_device_list(void)
+{
+    GSList *list = NULL;
+
+    object_child_foreach(qdev_get_machine(), nvdimm_plugged_device_list,
+                         &list);
+    return list;
+}
+
+#define NVDIMM_UUID_LE(a, b, c, d0, d1, d2, d3, d4, d5, d6, d7)             \
+   { (a) & 0xff, ((a) >> 8) & 0xff, ((a) >> 16) & 0xff, ((a) >> 24) & 0xff, \
+     (b) & 0xff, ((b) >> 8) & 0xff, (c) & 0xff, ((c) >> 8) & 0xff,          \
+     (d0), (d1), (d2), (d3), (d4), (d5), (d6), (d7) }
+
+/*
+ * define Byte Addressable Persistent Memory (PM) Region according to
+ * ACPI 6.0: 5.2.25.1 System Physical Address Range Structure.
+ */
+static const uint8_t nvdimm_nfit_spa_uuid[] =
+      NVDIMM_UUID_LE(0x66f0d379, 0xb4f3, 0x4074, 0xac, 0x43, 0x0d, 0x33,
+                     0x18, 0xb7, 0x8c, 0xdb);
+
+/*
+ * NVDIMM Firmware Interface Table
+ * @signature: "NFIT"
+ *
+ * It provides information that allows OSPM to enumerate NVDIMM present in
+ * the platform and associate system physical address ranges created by the
+ * NVDIMMs.
+ *
+ * It is defined in ACPI 6.0: 5.2.25 NVDIMM Firmware Interface Table (NFIT)
+ */
+struct NvdimmNfitHeader {
+    ACPI_TABLE_HEADER_DEF
+    uint32_t reserved;
+} QEMU_PACKED;
+typedef struct NvdimmNfitHeader NvdimmNfitHeader;
+
+/*
+ * define NFIT structures according to ACPI 6.0: 5.2.25 NVDIMM Firmware
+ * Interface Table (NFIT).
+ */
+
+/*
+ * System Physical Address Range Structure
+ *
+ * It describes the system physical address ranges occupied by NVDIMMs and
+ * the types of the regions.
+ */
+struct NvdimmNfitSpa {
+    uint16_t type;
+    uint16_t length;
+    uint16_t spa_index;
+    uint16_t flags;
+    uint32_t reserved;
+    uint32_t proximity_domain;
+    uint8_t type_guid[16];
+    uint64_t spa_base;
+    uint64_t spa_length;
+    uint64_t mem_attr;
+} QEMU_PACKED;
+typedef struct NvdimmNfitSpa NvdimmNfitSpa;
+
+/*
+ * Memory Device to System Physical Address Range Mapping Structure
+ *
+ * It enables identifying each NVDIMM region and the corresponding SPA
+ * describing the memory interleave
+ */
+struct NvdimmNfitMemDev {
+    uint16_t type;
+    uint16_t length;
+    uint32_t nfit_handle;
+    uint16_t phys_id;
+    uint16_t region_id;
+    uint16_t spa_index;
+    uint16_t dcr_index;
+    uint64_t region_len;
+    uint64_t region_offset;
+    uint64_t region_dpa;
+    uint16_t interleave_index;
+    uint16_t interleave_ways;
+    uint16_t flags;
+    uint16_t reserved;
+} QEMU_PACKED;
+typedef struct NvdimmNfitMemDev NvdimmNfitMemDev;
+
+/*
+ * NVDIMM Control Region Structure
+ *
+ * It describes the NVDIMM and if applicable, Block Control Window.
+ */
+struct NvdimmNfitControlRegion {
+    uint16_t type;
+    uint16_t length;
+    uint16_t dcr_index;
+    uint16_t vendor_id;
+    uint16_t device_id;
+    uint16_t revision_id;
+    uint16_t sub_vendor_id;
+    uint16_t sub_device_id;
+    uint16_t sub_revision_id;
+    uint8_t reserved[6];
+    uint32_t serial_number;
+    uint16_t fic;
+    uint16_t num_bcw;
+    uint64_t bcw_size;
+    uint64_t cmd_offset;
+    uint64_t cmd_size;
+    uint64_t status_offset;
+    uint64_t status_size;
+    uint16_t flags;
+    uint8_t reserved2[6];
+} QEMU_PACKED;
+typedef struct NvdimmNfitControlRegion NvdimmNfitControlRegion;
+
+/*
+ * Module serial number is a unique number for each device. We use the
+ * slot id of NVDIMM device to generate this number so that each device
+ * associates with a different number.
+ *
+ * 0x123456 is a magic number we arbitrarily chose.
+ */
+static uint32_t nvdimm_slot_to_sn(int slot)
+{
+    return 0x123456 + slot;
+}
+
+/*
+ * handle is used to uniquely associate nfit_memdev structure with NVDIMM
+ * ACPI device - nfit_memdev.nfit_handle matches with the value returned
+ * by ACPI device _ADR method.
+ *
+ * We generate the handle with the slot id of NVDIMM device and reserve
+ * 0 for NVDIMM root device.
+ */
+static uint32_t nvdimm_slot_to_handle(int slot)
+{
+    return slot + 1;
+}
+
+/*
+ * index uniquely identifies the structure, 0 is reserved which indicates
+ * that the structure is not valid or the associated structure is not
+ * present.
+ *
+ * Each NVDIMM device needs two indexes, one for nfit_spa and another for
+ * nfit_dc which are generated by the slot id of NVDIMM device.
+ */
+static uint16_t nvdimm_slot_to_spa_index(int slot)
+{
+    return (slot + 1) << 1;
+}
+
+/* See the comments of nvdimm_slot_to_spa_index(). */
+static uint32_t nvdimm_slot_to_dcr_index(int slot)
+{
+    return nvdimm_slot_to_spa_index(slot) + 1;
+}
+
+/* ACPI 6.0: 5.2.25.1 System Physical Address Range Structure */
+static void
+nvdimm_build_structure_spa(GArray *structures, DeviceState *dev)
+{
+    NvdimmNfitSpa *nfit_spa;
+    uint64_t addr = object_property_get_int(OBJECT(dev), PC_DIMM_ADDR_PROP,
+                                            NULL);
+    uint64_t size = object_property_get_int(OBJECT(dev), PC_DIMM_SIZE_PROP,
+                                            NULL);
+    uint32_t node = object_property_get_int(OBJECT(dev), PC_DIMM_NODE_PROP,
+                                            NULL);
+    int slot = object_property_get_int(OBJECT(dev), PC_DIMM_SLOT_PROP,
+                                            NULL);
+
+    nfit_spa = acpi_data_push(structures, sizeof(*nfit_spa));
+
+    nfit_spa->type = cpu_to_le16(0 /* System Physical Address Range
+                                      Structure */);
+    nfit_spa->length = cpu_to_le16(sizeof(*nfit_spa));
+    nfit_spa->spa_index = cpu_to_le16(nvdimm_slot_to_spa_index(slot));
+
+    /*
+     * Control region is strict as all the device info, such as SN, index,
+     * is associated with slot id.
+     */
+    nfit_spa->flags = cpu_to_le16(1 /* Control region is strictly for
+                                       management during hot add/online
+                                       operation */ |
+                                  2 /* Data in Proximity Domain field is
+                                       valid*/);
+
+    /* NUMA node. */
+    nfit_spa->proximity_domain = cpu_to_le32(node);
+    /* the region reported as PMEM. */
+    memcpy(nfit_spa->type_guid, nvdimm_nfit_spa_uuid,
+           sizeof(nvdimm_nfit_spa_uuid));
+
+    nfit_spa->spa_base = cpu_to_le64(addr);
+    nfit_spa->spa_length = cpu_to_le64(size);
+
+    /* It is the PMEM and can be cached as writeback. */
+    nfit_spa->mem_attr = cpu_to_le64(0x8ULL /* EFI_MEMORY_WB */ |
+                                     0x8000ULL /* EFI_MEMORY_NV */);
+}
+
+/*
+ * ACPI 6.0: 5.2.25.2 Memory Device to System Physical Address Range Mapping
+ * Structure
+ */
+static void
+nvdimm_build_structure_memdev(GArray *structures, DeviceState *dev)
+{
+    NvdimmNfitMemDev *nfit_memdev;
+    uint64_t addr = object_property_get_int(OBJECT(dev), PC_DIMM_ADDR_PROP,
+                                            NULL);
+    uint64_t size = object_property_get_int(OBJECT(dev), PC_DIMM_SIZE_PROP,
+                                            NULL);
+    int slot = object_property_get_int(OBJECT(dev), PC_DIMM_SLOT_PROP,
+                                            NULL);
+    uint32_t handle = nvdimm_slot_to_handle(slot);
+
+    nfit_memdev = acpi_data_push(structures, sizeof(*nfit_memdev));
+
+    nfit_memdev->type = cpu_to_le16(1 /* Memory Device to System Address
+                                         Range Map Structure*/);
+    nfit_memdev->length = cpu_to_le16(sizeof(*nfit_memdev));
+    nfit_memdev->nfit_handle = cpu_to_le32(handle);
+
+    /*
+     * associate memory device with System Physical Address Range
+     * Structure.
+     */
+    nfit_memdev->spa_index = cpu_to_le16(nvdimm_slot_to_spa_index(slot));
+    /* associate memory device with Control Region Structure. */
+    nfit_memdev->dcr_index = cpu_to_le16(nvdimm_slot_to_dcr_index(slot));
+
+    /* The memory region on the device. */
+    nfit_memdev->region_len = cpu_to_le64(size);
+    nfit_memdev->region_dpa = cpu_to_le64(addr);
+
+    /* Only one interleave for PMEM. */
+    nfit_memdev->interleave_ways = cpu_to_le16(1);
+}
+
+/*
+ * ACPI 6.0: 5.2.25.5 NVDIMM Control Region Structure.
+ */
+static void nvdimm_build_structure_dcr(GArray *structures, DeviceState *dev)
+{
+    NvdimmNfitControlRegion *nfit_dcr;
+    int slot = object_property_get_int(OBJECT(dev), PC_DIMM_SLOT_PROP,
+                                       NULL);
+    uint32_t sn = nvdimm_slot_to_sn(slot);
+
+    nfit_dcr = acpi_data_push(structures, sizeof(*nfit_dcr));
+
+    nfit_dcr->type = cpu_to_le16(4 /* NVDIMM Control Region Structure */);
+    nfit_dcr->length = cpu_to_le16(sizeof(*nfit_dcr));
+    nfit_dcr->dcr_index = cpu_to_le16(nvdimm_slot_to_dcr_index(slot));
+
+    /* vendor: Intel. */
+    nfit_dcr->vendor_id = cpu_to_le16(0x8086);
+    nfit_dcr->device_id = cpu_to_le16(1);
+
+    /* The _DSM method is following Intel's DSM specification. */
+    nfit_dcr->revision_id = cpu_to_le16(1 /* Current Revision supported
+                                             in ACPI 6.0 is 1. */);
+    nfit_dcr->serial_number = cpu_to_le32(sn);
+    nfit_dcr->fic = cpu_to_le16(0x201 /* Format Interface Code. See Chapter
+                                         2: NVDIMM Device Specific Method
+                                         (DSM) in DSM Spec Rev1.*/);
+}
+
+static GArray *nvdimm_build_device_structure(GSList *device_list)
+{
+    GArray *structures = g_array_new(false, true /* clear */, 1);
+
+    for (; device_list; device_list = device_list->next) {
+        DeviceState *dev = device_list->data;
+
+        /* build System Physical Address Range Structure. */
+        nvdimm_build_structure_spa(structures, dev);
+
+        /*
+         * build Memory Device to System Physical Address Range Mapping
+         * Structure.
+         */
+        nvdimm_build_structure_memdev(structures, dev);
+
+        /* build NVDIMM Control Region Structure. */
+        nvdimm_build_structure_dcr(structures, dev);
+    }
+
+    return structures;
+}
+
+static void nvdimm_build_nfit(GSList *device_list, GArray *table_offsets,
+                              GArray *table_data, GArray *linker)
+{
+    GArray *structures = nvdimm_build_device_structure(device_list);
+    void *header;
+
+    acpi_add_table(table_offsets, table_data);
+
+    /* NFIT header. */
+    header = acpi_data_push(table_data, sizeof(NvdimmNfitHeader));
+    /* NVDIMM device structures. */
+    g_array_append_vals(table_data, structures->data, structures->len);
+
+    build_header(linker, table_data, header, "NFIT",
+                 sizeof(NvdimmNfitHeader) + structures->len, 1, NULL);
+    g_array_free(structures, true);
+}
+
+void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
+                       GArray *linker)
+{
+    GSList *device_list;
+
+    /* no NVDIMM device is plugged. */
+    device_list = nvdimm_get_plugged_device_list();
+    if (!device_list) {
+        return;
+    }
+    nvdimm_build_nfit(device_list, table_offsets, table_data, linker);
+    g_slist_free(device_list);
+}
diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
index 27d2caf144..479b11ec4d 100644
--- a/hw/i386/acpi-build.c
+++ b/hw/i386/acpi-build.c
@@ -39,6 +39,7 @@
 #include "hw/loader.h"
 #include "hw/isa/isa.h"
 #include "hw/acpi/memory_hotplug.h"
+#include "hw/mem/nvdimm.h"
 #include "sysemu/tpm.h"
 #include "hw/acpi/tpm.h"
 #include "sysemu/tpm_backend.h"
@@ -1683,6 +1684,13 @@ static bool acpi_has_iommu(void)
     return intel_iommu && !ambiguous;
 }
 
+static bool acpi_has_nvdimm(void)
+{
+    PCMachineState *pcms = PC_MACHINE(qdev_get_machine());
+
+    return pcms->nvdimm;
+}
+
 static
 void acpi_build(PcGuestInfo *guest_info, AcpiBuildTables *tables)
 {
@@ -1767,6 +1775,10 @@ void acpi_build(PcGuestInfo *guest_info, AcpiBuildTables *tables)
         build_dmar_q35(tables_blob, tables->linker);
     }
 
+    if (acpi_has_nvdimm()) {
+        nvdimm_build_acpi(table_offsets, tables_blob, tables->linker);
+    }
+
     /* Add tables supplied by user (if any) */
     for (u = acpi_table_first(); u; u = acpi_table_next(u)) {
         unsigned len = acpi_table_len(u);
diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index 48aa37a138..459260b4af 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -1865,6 +1865,20 @@ static void pc_machine_set_smm(Object *obj, Visitor *v, void *opaque,
     visit_type_OnOffAuto(v, &pcms->smm, name, errp);
 }
 
+static bool pc_machine_get_nvdimm(Object *obj, Error **errp)
+{
+    PCMachineState *pcms = PC_MACHINE(obj);
+
+    return pcms->nvdimm;
+}
+
+static void pc_machine_set_nvdimm(Object *obj, bool value, Error **errp)
+{
+    PCMachineState *pcms = PC_MACHINE(obj);
+
+    pcms->nvdimm = value;
+}
+
 static void pc_machine_initfn(Object *obj)
 {
     PCMachineState *pcms = PC_MACHINE(obj);
@@ -1899,6 +1913,11 @@ static void pc_machine_initfn(Object *obj)
     object_property_set_description(obj, PC_MACHINE_VMPORT,
                                     "Enable vmport (pc & q35)",
                                     &error_abort);
+
+    /* nvdimm is disabled on default. */
+    pcms->nvdimm = false;
+    object_property_add_bool(obj, PC_MACHINE_NVDIMM, pc_machine_get_nvdimm,
+                             pc_machine_set_nvdimm, &error_abort);
 }
 
 static void pc_machine_reset(void)
diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
index c6a777546a..b0d6283b25 100644
--- a/include/hw/i386/pc.h
+++ b/include/hw/i386/pc.h
@@ -55,6 +55,7 @@ struct PCMachineState {
     uint64_t max_ram_below_4g;
     OnOffAuto vmport;
     OnOffAuto smm;
+    bool nvdimm;
 
     /* RAM information (sizes, addresses, configuration): */
     ram_addr_t below_4g_mem_size, above_4g_mem_size;
@@ -65,6 +66,7 @@ struct PCMachineState {
 #define PC_MACHINE_MAX_RAM_BELOW_4G "max-ram-below-4g"
 #define PC_MACHINE_VMPORT           "vmport"
 #define PC_MACHINE_SMM              "smm"
+#define PC_MACHINE_NVDIMM           "nvdimm"
 
 /**
  * PCMachineClass:
diff --git a/include/hw/mem/nvdimm.h b/include/hw/mem/nvdimm.h
index dbfa8d6bd4..49183c126b 100644
--- a/include/hw/mem/nvdimm.h
+++ b/include/hw/mem/nvdimm.h
@@ -26,4 +26,7 @@
 #include "hw/mem/pc-dimm.h"
 
 #define TYPE_NVDIMM      "nvdimm"
+
+void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
+                       GArray *linker);
 #endif
diff --git a/qemu-options.hx b/qemu-options.hx
index 94ab8806fb..215d00ddd3 100644
--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -42,7 +42,8 @@ DEF("machine", HAS_ARG, QEMU_OPTION_machine, \
     "                igd-passthru=on|off controls IGD GFX passthrough support (default=off)\n"
     "                aes-key-wrap=on|off controls support for AES key wrapping (default=on)\n"
     "                dea-key-wrap=on|off controls support for DEA key wrapping (default=on)\n"
-    "                suppress-vmdesc=on|off disables self-describing migration (default=off)\n",
+    "                suppress-vmdesc=on|off disables self-describing migration (default=off)\n"
+    "                nvdimm=on|off controls NVDIMM support (default=off)\n",
     QEMU_ARCH_ALL)
 STEXI
 @item -machine [type=]@var{name}[,prop=@var{value}[,...]]
@@ -81,6 +82,8 @@ execution of AES cryptographic functions.  The default is on.
 Enables or disables DEA key wrapping support on s390-ccw hosts. This feature
 controls whether DEA wrapping keys will be created to allow
 execution of DEA cryptographic functions.  The default is on.
+@item nvdimm=on|off
+Enables or disables NVDIMM support. The default is off.
 @end table
 ETEXI