NUMA series, and hostmem improvements

index 7bbe153..7f87adb 100644
@@ -1313,20 +1313,6 @@ static void tcg_exec_all(void)
     exit_request = 0;
 }
 
-void set_numa_modes(void)
-{
-    CPUState *cpu;
-    int i;
-
-    CPU_FOREACH(cpu) {
-        for (i = 0; i < nb_numa_nodes; i++) {
-            if (test_bit(cpu->cpu_index, node_cpumask[i])) {
-                cpu->numa_node = i;
-            }
-        }
-    }
-}
-
 void list_cpus(FILE *f, fprintf_function cpu_fprintf, const char *optarg)
 {
     /* XXX: implement xxx_cpu_list for targets that still miss it */

index fb649a4..758a928 100644
@@ -414,8 +414,6 @@ CPUArchState *cpu_copy(CPUArchState *env);
 
 /* memory API */
 
-extern ram_addr_t ram_size;
-
 /* RAM is pre-allocated and passed into qemu_ram_alloc_from_ptr */
 #define RAM_PREALLOC_MASK   (1 << 0)
 

index a21b65a..e8c7970 100644
@@ -45,6 +45,8 @@ typedef uintptr_t ram_addr_t;
 #  define RAM_ADDR_FMT "%" PRIxPTR
 #endif
 
+extern ram_addr_t ram_size;
+
 /* memory API */
 
 typedef void CPUWriteMemoryFunc(void *opaque, hwaddr addr, uint32_t value);

index 6502488..4f79081 100644
@@ -23,7 +23,6 @@ extern int smp_threads;
 #define smp_threads 1
 #endif
 
-void set_numa_modes(void);
 void list_cpus(FILE *f, fprintf_function cpu_fprintf, const char *optarg);
 
 #endif

index ba5c7f8..565c8f6 100644
@@ -144,6 +144,9 @@ extern QEMUClockType rtc_clock;
 extern int nb_numa_nodes;
 extern uint64_t node_mem[MAX_NODES];
 extern unsigned long *node_cpumask[MAX_NODES];
+void numa_add(const char *optarg);
+void set_numa_nodes(void);
+void set_numa_modes(void);
 
 #define MAX_OPTION_ROMS 16
 typedef struct QEMUOptionRom {

186 lines not on this page · raw message

151 lines not on this page · raw message

index 395c14f..1d2f761 100644
@@ -27,6 +27,8 @@
 #include "exec/cpu-common.h"
 #include "qemu/bitmap.h"
 #include "qom/cpu.h"
+#include "qemu/error-report.h"
+#include "include/exec/cpu-common.h" /* for RAM_ADDR_FMT */
 
 static void numa_node_parse_cpus(int nodenr, const char *cpus)
 {

@@ -127,6 +129,7 @@ void numa_add(const char *optarg)
 void set_numa_nodes(void)
 {
     if (nb_numa_nodes > 0) {
+        uint64_t numa_total;
         int i;
 
         if (nb_numa_nodes > MAX_NODES) {

@@ -154,6 +157,17 @@ void set_numa_nodes(void)
             node_mem[i] = ram_size - usedmem;
         }
 
+        numa_total = 0;
+        for (i = 0; i < nb_numa_nodes; i++) {
+            numa_total += node_mem[i];
+        }
+        if (numa_total != ram_size) {
+            error_report("qemu: total memory size for NUMA nodes (%" PRIu64 ")"
+                         " should equal to ram_size (" RAM_ADDR_FMT ")\n",
+                         numa_total, ram_size);
+            exit(1);
+        }
+
         for (i = 0; i < nb_numa_nodes; i++) {
             if (!bitmap_empty(node_cpumask[i], MAX_CPUMASK_BITS)) {
                 break;

index 14f0d91..249e504 100644
@@ -675,14 +675,14 @@ static FWCfgState *bochs_bios_init(void)
         unsigned int apic_id = x86_cpu_apic_id_from_index(i);
         assert(apic_id < apic_id_limit);
         for (j = 0; j < nb_numa_nodes; j++) {
-            if (test_bit(i, node_cpumask[j])) {
+            if (test_bit(i, numa_info[j].node_cpu)) {
                 numa_fw_cfg[apic_id + 1] = cpu_to_le64(j);
                 break;
             }
         }
     }
     for (i = 0; i < nb_numa_nodes; i++) {
-        numa_fw_cfg[apic_id_limit + 1 + i] = cpu_to_le64(node_mem[i]);
+        numa_fw_cfg[apic_id_limit + 1 + i] = cpu_to_le64(numa_info[i].node_mem);
     }
     fw_cfg_add_bytes(fw_cfg, FW_CFG_NUMA, numa_fw_cfg,
                      (1 + apic_id_limit + nb_numa_nodes) *

@@ -1093,8 +1093,12 @@ PcGuestInfo *pc_guest_info_init(ram_addr_t below_4g_mem_size,
     guest_info->apic_id_limit = pc_apic_id_limit(max_cpus);
     guest_info->apic_xrupt_override = kvm_allows_irq0_override();
     guest_info->numa_nodes = nb_numa_nodes;
-    guest_info->node_mem = g_memdup(node_mem, guest_info->numa_nodes *
+    guest_info->node_mem = g_malloc0(guest_info->numa_nodes *
                                     sizeof *guest_info->node_mem);
+    for (i = 0; i < nb_numa_nodes; i++) {
+        guest_info->node_mem[i] = numa_info[i].node_mem;
+    }
+
     guest_info->node_cpu = g_malloc0(guest_info->apic_id_limit *
                                      sizeof *guest_info->node_cpu);
 

@@ -1102,7 +1106,7 @@ PcGuestInfo *pc_guest_info_init(ram_addr_t below_4g_mem_size,
         unsigned int apic_id = x86_cpu_apic_id_from_index(i);
         assert(apic_id < guest_info->apic_id_limit);
         for (j = 0; j < nb_numa_nodes; j++) {
-            if (test_bit(i, node_cpumask[j])) {
+            if (test_bit(i, numa_info[j].node_cpu)) {
                 guest_info->node_cpu[apic_id] = j;
                 break;
             }

index a11e121..89f71a8 100644
@@ -538,8 +538,8 @@ static int spapr_populate_memory(sPAPREnvironment *spapr, void *fdt)
     int i, off;
 
     /* memory node(s) */
-    if (nb_numa_nodes > 1 && node_mem[0] < ram_size) {
-        node0_size = node_mem[0];
+    if (nb_numa_nodes > 1 && numa_info[0].node_mem < ram_size) {
+        node0_size = numa_info[0].node_mem;
     } else {
         node0_size = ram_size;
     }

@@ -577,7 +577,7 @@ static int spapr_populate_memory(sPAPREnvironment *spapr, void *fdt)
         if (mem_start >= ram_size) {
             node_size = 0;
         } else {
-            node_size = node_mem[i];
+            node_size = numa_info[i].node_mem;
             if (node_size > ram_size - mem_start) {
                 node_size = ram_size - mem_start;
             }

@@ -722,7 +722,8 @@ static void spapr_reset_htab(sPAPREnvironment *spapr)
 
     /* Update the RMA size if necessary */
     if (spapr->vrma_adjust) {
-        hwaddr node0_size = (nb_numa_nodes > 1) ? node_mem[0] : ram_size;
+        hwaddr node0_size = (nb_numa_nodes > 1) ?
+            numa_info[0].node_mem : ram_size;
         spapr->rma_size = kvmppc_rma_size(node0_size, spapr->htab_shift);
     }
 }

@@ -1155,7 +1156,7 @@ static void ppc_spapr_init(QEMUMachineInitArgs *args)
     MemoryRegion *sysmem = get_system_memory();
     MemoryRegion *ram = g_new(MemoryRegion, 1);
     hwaddr rma_alloc_size;
-    hwaddr node0_size = (nb_numa_nodes > 1) ? node_mem[0] : ram_size;
+    hwaddr node0_size = (nb_numa_nodes > 1) ? numa_info[0].node_mem : ram_size;
     uint32_t initrd_base = 0;
     long kernel_size = 0, initrd_size = 0;
     long load_limit, rtas_limit, fw_size;

index 565c8f6..3a9308b 100644
@@ -9,6 +9,7 @@
 #include "qapi-types.h"
 #include "qemu/notify.h"
 #include "qemu/main-loop.h"
+#include "qemu/bitmap.h"
 
 /* vl.c */
 

@@ -142,8 +143,11 @@ extern QEMUClockType rtc_clock;
 #define MAX_CPUMASK_BITS 255
 
 extern int nb_numa_nodes;
-extern uint64_t node_mem[MAX_NODES];
-extern unsigned long *node_cpumask[MAX_NODES];
+typedef struct node_info {
+    uint64_t node_mem;
+    DECLARE_BITMAP(node_cpu, MAX_CPUMASK_BITS);
+} NodeInfo;
+extern NodeInfo numa_info[MAX_NODES];
 void numa_add(const char *optarg);
 void set_numa_nodes(void);
 void set_numa_modes(void);

index 1266ba0..db17a31 100644
@@ -2005,7 +2005,7 @@ static void do_info_numa(Monitor *mon, const QDict *qdict)
         }
         monitor_printf(mon, "\n");
         monitor_printf(mon, "node %d size: %" PRId64 " MB\n", i,
-            node_mem[i] >> 20);
+            numa_info[i].node_mem >> 20);
     }
 }
 

index 1d2f761..48d8bcc 100644
@@ -66,7 +66,7 @@ static void numa_node_parse_cpus(int nodenr, const char *cpus)
         goto error;
     }
 
-    bitmap_set(node_cpumask[nodenr], value, endvalue-value+1);
+    bitmap_set(numa_info[nodenr].node_cpu, value, endvalue-value+1);
     return;
 
 error:

@@ -106,7 +106,7 @@ void numa_add(const char *optarg)
         }
 
         if (get_param_value(option, 128, "mem", optarg) == 0) {
-            node_mem[nodenr] = 0;
+            numa_info[nodenr].node_mem = 0;
         } else {
             int64_t sval;
             sval = strtosz(option, &endptr);

@@ -114,7 +114,7 @@ void numa_add(const char *optarg)
                 fprintf(stderr, "qemu: invalid numa mem size: %s\n", optarg);
                 exit(1);
             }
-            node_mem[nodenr] = sval;
+            numa_info[nodenr].node_mem = sval;
         }
         if (get_param_value(option, 128, "cpus", optarg) != 0) {
             numa_node_parse_cpus(nodenr, option);

@@ -140,7 +140,7 @@ void set_numa_nodes(void)
          * and distribute the available memory equally across all nodes
          */
         for (i = 0; i < nb_numa_nodes; i++) {
-            if (node_mem[i] != 0) {
+            if (numa_info[i].node_mem != 0) {
                 break;
             }
         }

@@ -151,15 +151,16 @@ void set_numa_nodes(void)
              * the final node gets the rest.
              */
             for (i = 0; i < nb_numa_nodes - 1; i++) {
-                node_mem[i] = (ram_size / nb_numa_nodes) & ~((1 << 23UL) - 1);
-                usedmem += node_mem[i];
+                numa_info[i].node_mem = (ram_size / nb_numa_nodes) &
+                                        ~((1 << 23UL) - 1);
+                usedmem += numa_info[i].node_mem;
             }
-            node_mem[i] = ram_size - usedmem;
+            numa_info[i].node_mem = ram_size - usedmem;
         }
 
         numa_total = 0;
         for (i = 0; i < nb_numa_nodes; i++) {
-            numa_total += node_mem[i];
+            numa_total += numa_info[i].node_mem;
         }
         if (numa_total != ram_size) {
             error_report("qemu: total memory size for NUMA nodes (%" PRIu64 ")"

@@ -169,7 +170,7 @@ void set_numa_nodes(void)
         }
 
         for (i = 0; i < nb_numa_nodes; i++) {
-            if (!bitmap_empty(node_cpumask[i], MAX_CPUMASK_BITS)) {
+            if (!bitmap_empty(numa_info[i].node_cpu, MAX_CPUMASK_BITS)) {
                 break;
             }
         }

@@ -179,7 +180,7 @@ void set_numa_nodes(void)
          */
         if (i == nb_numa_nodes) {
             for (i = 0; i < max_cpus; i++) {
-                set_bit(i, node_cpumask[i % nb_numa_nodes]);
+                set_bit(i, numa_info[i % nb_numa_nodes].node_cpu);
             }
         }
     }

@@ -192,7 +193,7 @@ void set_numa_modes(void)
 
     CPU_FOREACH(cpu) {
         for (i = 0; i < nb_numa_nodes; i++) {
-            if (test_bit(cpu->cpu_index, node_cpumask[i])) {
+            if (test_bit(cpu->cpu_index, numa_info[i].node_cpu)) {
                 cpu->numa_node = i;
             }
         }

index 303884c..77e77b5 100644
@@ -195,8 +195,7 @@ static QTAILQ_HEAD(, FWBootEntry) fw_boot_order =
     QTAILQ_HEAD_INITIALIZER(fw_boot_order);
 
 int nb_numa_nodes;
-uint64_t node_mem[MAX_NODES];
-unsigned long *node_cpumask[MAX_NODES];
+NodeInfo numa_info[MAX_NODES];
 
 uint8_t qemu_uuid[16];
 bool qemu_uuid_set;

@@ -2929,8 +2928,8 @@ int main(int argc, char **argv, char **envp)
     translation = BIOS_ATA_TRANSLATION_AUTO;
 
     for (i = 0; i < MAX_NODES; i++) {
-        node_mem[i] = 0;
-        node_cpumask[i] = bitmap_new(MAX_CPUMASK_BITS);
+        numa_info[i].node_mem = 0;
+        bitmap_zero(numa_info[i].node_cpu, MAX_CPUMASK_BITS);
     }
 
     nb_numa_nodes = 0;

index 3a9308b..4102be3 100644
@@ -148,9 +148,10 @@ typedef struct node_info {
     DECLARE_BITMAP(node_cpu, MAX_CPUMASK_BITS);
 } NodeInfo;
 extern NodeInfo numa_info[MAX_NODES];
-void numa_add(const char *optarg);
 void set_numa_nodes(void);
 void set_numa_modes(void);
+extern QemuOptsList qemu_numa_opts;
+int numa_init_func(QemuOpts *opts, void *opaque);
 
 #define MAX_OPTION_ROMS 16
 typedef struct QEMUOptionRom {

171 lines not on this page · raw message

index 77e77b5..78144c4 100644
@@ -2907,6 +2907,7 @@ int main(int argc, char **argv, char **envp)
     qemu_add_opts(&qemu_realtime_opts);
     qemu_add_opts(&qemu_msg_opts);
     qemu_add_opts(&qemu_name_opts);
+    qemu_add_opts(&qemu_numa_opts);
 
     runstate_init();
 

@@ -3103,7 +3104,10 @@ int main(int argc, char **argv, char **envp)
                 }
                 break;
             case QEMU_OPTION_numa:
-                numa_add(optarg);
+                opts = qemu_opts_parse(qemu_find_opts("numa"), optarg, 1);
+                if (!opts) {
+                    exit(1);
+                }
                 break;
             case QEMU_OPTION_display:
                 display_type = select_display(optarg);

@@ -4225,6 +4229,11 @@ int main(int argc, char **argv, char **envp)
 
     register_savevm_live(NULL, "ram", 0, 4, &savevm_ram_handlers, NULL);
 
+    if (qemu_opts_foreach(qemu_find_opts("numa"), numa_init_func,
+                          NULL, 1) != 0) {
+        exit(1);
+    }
+
     set_numa_nodes();
 
     if (qemu_opts_foreach(qemu_find_opts("mon"), mon_init_func, NULL, 1) != 0) {

index 4102be3..423d49e 100644
@@ -133,7 +133,7 @@ extern size_t boot_splash_filedata_size;
 extern uint8_t qemu_extra_params_fw[2];
 extern QEMUClockType rtc_clock;
 
-#define MAX_NODES 64
+#define MAX_NODES 128
 
 /* The following shall be true for all CPUs:
  *   cpu->cpu_index < max_cpus <= MAX_CPUMASK_BITS

index 78144c4..167856e 100644
@@ -115,8 +115,7 @@ int main(int argc, char **argv)
 #include "qemu/osdep.h"
 
 #include "ui/qemu-spice.h"
-#include "qapi/string-input-visitor.h"
-#include "qom/object_interfaces.h"
+#include "qapi/opts-visitor.h"
 
 #define DEFAULT_RAM_SIZE 128
 

@@ -2768,69 +2767,53 @@ static void free_and_trace(gpointer mem)
     free(mem);
 }
 
-static int object_set_property(const char *name, const char *value, void *opaque)
-{
-    Object *obj = OBJECT(opaque);
-    StringInputVisitor *siv;
-    Error *local_err = NULL;
-
-    if (strcmp(name, "qom-type") == 0 || strcmp(name, "id") == 0) {
-        return 0;
-    }
-
-    siv = string_input_visitor_new(value);
-    object_property_set(obj, string_input_get_visitor(siv), name, &local_err);
-    string_input_visitor_cleanup(siv);
-
-    if (local_err) {
-        qerror_report_err(local_err);
-        error_free(local_err);
-        return -1;
-    }
-
-    return 0;
-}
-
 static int object_create(QemuOpts *opts, void *opaque)
 {
-    const char *type = qemu_opt_get(opts, "qom-type");
-    const char *id = qemu_opts_id(opts);
-    Error *local_err = NULL;
-    Object *obj;
-
-    g_assert(type != NULL);
-
-    if (id == NULL) {
-        qerror_report(QERR_MISSING_PARAMETER, "id");
-        return -1;
+    Error *err = NULL;
+    char *type = NULL;
+    char *id = NULL;
+    void *dummy = NULL;
+    OptsVisitor *ov;
+    QDict *pdict;
+
+    ov = opts_visitor_new(opts);
+    pdict = qemu_opts_to_qdict(opts, NULL);
+
+    visit_start_struct(opts_get_visitor(ov), &dummy, NULL, NULL, 0, &err);
+    if (err) {
+        goto out;
     }
 
-    obj = object_new(type);
-    if (qemu_opt_foreach(opts, object_set_property, obj, 1) < 0) {
-        object_unref(obj);
-        return -1;
+    qdict_del(pdict, "qom-type");
+    visit_type_str(opts_get_visitor(ov), &type, "qom-type", &err);
+    if (err) {
+        goto out;
     }
 
-    if (!object_dynamic_cast(obj, TYPE_USER_CREATABLE)) {
-        error_setg(&local_err, "object '%s' isn't supported by -object",
-                   id);
+    qdict_del(pdict, "id");
+    visit_type_str(opts_get_visitor(ov), &id, "id", &err);
+    if (err) {
         goto out;
     }
 
-    user_creatable_complete(obj, &local_err);
-    if (local_err) {
+    object_add(type, id, pdict, opts_get_visitor(ov), &err);
+    if (err) {
         goto out;
     }
-
-    object_property_add_child(container_get(object_get_root(), "/objects"),
-                              id, obj, &local_err);
+    visit_end_struct(opts_get_visitor(ov), &err);
+    if (err) {
+        qmp_object_del(id, NULL);
+    }
 
 out:
-    object_unref(obj);
-    if (local_err) {
-        qerror_report_err(local_err);
-        error_free(local_err);
-        return -1;
+    opts_visitor_cleanup(ov);
+
+    QDECREF(pdict);
+    g_free(id);
+    g_free(type);
+    g_free(dummy);
+    if (err) {
+        qerror_report_err(err);
     }
     return 0;
 }

index 74107be..ee99d3a 100644
@@ -571,13 +571,18 @@ void object_add(const char *type, const char *id, const QDict *qdict,
         }
     }
 
-    user_creatable_complete(obj, &local_err);
+    object_property_add_child(container_get(object_get_root(), "/objects"),
+                              id, obj, &local_err);
     if (local_err) {
         goto out;
     }
 
-    object_property_add_child(container_get(object_get_root(), "/objects"),
-                              id, obj, &local_err);
+    user_creatable_complete(obj, &local_err);
+    if (local_err) {
+        object_property_del(container_get(object_get_root(), "/objects"),
+                            id, &error_abort);
+        goto out;
+    }
 out:
     if (local_err) {
         error_propagate(errp, local_err);

index ee99d3a..636e6f2 100644
@@ -546,7 +546,8 @@ void object_add(const char *type, const char *id, const QDict *qdict,
 
     klass = object_class_by_name(type);
     if (!klass) {
-        error_setg(errp, "invalid class name");
+        error_set(errp, QERR_INVALID_PARAMETER_VALUE,
+                  "qom-type", "a valid class name");
         return;
     }
 

index 249e504..d5fb884 100644
@@ -1161,10 +1161,8 @@ void pc_acpi_init(const char *default_dsdt)
     }
 }
 
-FWCfgState *pc_memory_init(MemoryRegion *system_memory,
-                           const char *kernel_filename,
-                           const char *kernel_cmdline,
-                           const char *initrd_filename,
+FWCfgState *pc_memory_init(QEMUMachineInitArgs *args,
+                           MemoryRegion *system_memory,
                            ram_addr_t below_4g_mem_size,
                            ram_addr_t above_4g_mem_size,
                            MemoryRegion *rom_memory,

@@ -1176,7 +1174,7 @@ FWCfgState *pc_memory_init(MemoryRegion *system_memory,
     MemoryRegion *ram_below_4g, *ram_above_4g;
     FWCfgState *fw_cfg;
 
-    linux_boot = (kernel_filename != NULL);
+    linux_boot = (args->kernel_filename != NULL);
 
     /* Allocate RAM.  We allocate it as a single memory region and use
      * aliases to address portions of it, mostly for backwards compatibility

@@ -1217,7 +1215,8 @@ FWCfgState *pc_memory_init(MemoryRegion *system_memory,
     rom_set_fw(fw_cfg);
 
     if (linux_boot) {
-        load_linux(fw_cfg, kernel_filename, initrd_filename, kernel_cmdline, below_4g_mem_size);
+        load_linux(fw_cfg, args->kernel_filename, args->initrd_filename,
+                   args->kernel_cmdline, below_4g_mem_size);
     }
 
     for (i = 0; i < nb_option_roms; i++) {

index 7930a26..3bd8821 100644
@@ -151,11 +151,9 @@ static void pc_init1(QEMUMachineInitArgs *args,
 
     /* allocate ram and load rom/bios */
     if (!xen_enabled()) {
-        fw_cfg = pc_memory_init(system_memory,
-                       args->kernel_filename, args->kernel_cmdline,
-                       args->initrd_filename,
-                       below_4g_mem_size, above_4g_mem_size,
-                       rom_memory, &ram_memory, guest_info);
+        fw_cfg = pc_memory_init(args, system_memory,
+                                below_4g_mem_size, above_4g_mem_size,
+                                rom_memory, &ram_memory, guest_info);
     }
 
     gsi_state = g_malloc0(sizeof(*gsi_state));

index c844dc2..01794fc 100644
@@ -138,9 +138,7 @@ static void pc_q35_init(QEMUMachineInitArgs *args)
 
     /* allocate ram and load rom/bios */
     if (!xen_enabled()) {
-        pc_memory_init(get_system_memory(),
-                       args->kernel_filename, args->kernel_cmdline,
-                       args->initrd_filename,
+        pc_memory_init(args, get_system_memory(),
                        below_4g_mem_size, above_4g_mem_size,
                        rom_memory, &ram_memory, guest_info);
     }

index 9010246..8fc0527 100644
@@ -3,6 +3,7 @@
 
 #include "qemu-common.h"
 #include "exec/memory.h"
+#include "hw/boards.h"
 #include "hw/isa/isa.h"
 #include "hw/block/fdc.h"
 #include "net/net.h"

@@ -134,10 +135,8 @@ PcGuestInfo *pc_guest_info_init(ram_addr_t below_4g_mem_size,
 void pc_pci_as_mapping_init(Object *owner, MemoryRegion *system_memory,
                             MemoryRegion *pci_address_space);
 
-FWCfgState *pc_memory_init(MemoryRegion *system_memory,
-                           const char *kernel_filename,
-                           const char *kernel_cmdline,
-                           const char *initrd_filename,
+FWCfgState *pc_memory_init(QEMUMachineInitArgs *args,
+                           MemoryRegion *system_memory,
                            ram_addr_t below_4g_mem_size,
                            ram_addr_t above_4g_mem_size,
                            MemoryRegion *rom_memory,

index d5fb884..571dc72 100644
@@ -1181,9 +1181,7 @@ FWCfgState *pc_memory_init(QEMUMachineInitArgs *args,
      * with older qemus that used qemu_ram_alloc().
      */
     ram = g_malloc(sizeof(*ram));
-    memory_region_init_ram(ram, NULL, "pc.ram",
-                           below_4g_mem_size + above_4g_mem_size);
-    vmstate_register_ram_global(ram);
+    memory_region_allocate_system_memory(ram, NULL, "pc.ram", args->ram_size);
     *ram_memory = ram;
     ram_below_4g = g_malloc(sizeof(*ram_below_4g));
     memory_region_init_alias(ram_below_4g, NULL, "ram-below-4g", ram,

index dd2c70d..1ea52c0 100644
@@ -50,9 +50,13 @@ struct QEMUMachine {
     const char *hw_version;
 };
 
-#define TYPE_MACHINE_SUFFIX "-machine"
+void memory_region_allocate_system_memory(MemoryRegion *mr, Object *owner,
+                                          const char *name,
+                                          uint64_t ram_size);
+
 int qemu_register_machine(QEMUMachine *m);
 
+#define TYPE_MACHINE_SUFFIX "-machine"
 #define TYPE_MACHINE "machine"
 #undef MACHINE  /* BSD defines it and QEMU does not use it */
 #define MACHINE(obj) \

index 423d49e..caf88dd 100644
@@ -10,6 +10,7 @@
 #include "qemu/notify.h"
 #include "qemu/main-loop.h"
 #include "qemu/bitmap.h"
+#include "qom/object.h"
 
 /* vl.c */
 

index 439df87..bcd7b04 100644
@@ -33,6 +33,7 @@
 #include "qapi/opts-visitor.h"
 #include "qapi/dealloc-visitor.h"
 #include "qapi/qmp/qerror.h"
+#include "hw/boards.h"
 
 QemuOptsList qemu_numa_opts = {
     .name = "numa",

@@ -194,3 +195,11 @@ void set_numa_modes(void)
         }
     }
 }
+
+void memory_region_allocate_system_memory(MemoryRegion *mr, Object *owner,
+                                          const char *name,
+                                          uint64_t ram_size)
+{
+    memory_region_init_ram(mr, owner, name, ram_size);
+    vmstate_register_ram_global(mr);
+}

54 lines not on this page · raw message

109 lines not on this page · raw message

58 lines not on this page · raw message

index caf88dd..1e141e3 100644
@@ -147,6 +147,7 @@ extern int nb_numa_nodes;
 typedef struct node_info {
     uint64_t node_mem;
     DECLARE_BITMAP(node_cpu, MAX_CPUMASK_BITS);
+    struct HostMemoryBackend *node_memdev;
 } NodeInfo;
 extern NodeInfo numa_info[MAX_NODES];
 void set_numa_nodes(void);

94 lines not on this page · raw message

index 91513c6..8ae8b95 100644
@@ -1247,56 +1247,30 @@ static int memory_try_enable_merging(void *addr, size_t len)
     return qemu_madvise(addr, len, QEMU_MADV_MERGEABLE);
 }
 
-ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
-                                   MemoryRegion *mr)
+static ram_addr_t ram_block_add(RAMBlock *new_block)
 {
-    RAMBlock *block, *new_block;
+    RAMBlock *block;
     ram_addr_t old_ram_size, new_ram_size;
 
     old_ram_size = last_ram_offset() >> TARGET_PAGE_BITS;
 
-    size = TARGET_PAGE_ALIGN(size);
-    new_block = g_malloc0(sizeof(*new_block));
-    new_block->fd = -1;
-
     /* This assumes the iothread lock is taken here too.  */
     qemu_mutex_lock_ramlist();
-    new_block->mr = mr;
-    new_block->offset = find_ram_offset(size);
-    if (host) {
-        new_block->host = host;
-        new_block->flags |= RAM_PREALLOC_MASK;
-    } else if (xen_enabled()) {
-        if (mem_path) {
-            fprintf(stderr, "-mem-path not supported with Xen\n");
-            exit(1);
-        }
-        xen_ram_alloc(new_block->offset, size, mr);
-    } else {
-        if (mem_path) {
-            if (phys_mem_alloc != qemu_anon_ram_alloc) {
-                /*
-                 * file_ram_alloc() needs to allocate just like
-                 * phys_mem_alloc, but we haven't bothered to provide
-                 * a hook there.
-                 */
-                fprintf(stderr,
-                        "-mem-path not supported with this accelerator\n");
-                exit(1);
-            }
-            new_block->host = file_ram_alloc(new_block, size, mem_path);
-        }
-        if (!new_block->host) {
-            new_block->host = phys_mem_alloc(size);
+    new_block->offset = find_ram_offset(new_block->length);
+
+    if (!new_block->host) {
+        if (xen_enabled()) {
+            xen_ram_alloc(new_block->offset, new_block->length, new_block->mr);
+        } else {
+            new_block->host = phys_mem_alloc(new_block->length);
             if (!new_block->host) {
                 fprintf(stderr, "Cannot set up guest memory '%s': %s\n",
                         new_block->mr->name, strerror(errno));
                 exit(1);
             }
-            memory_try_enable_merging(new_block->host, size);
+            memory_try_enable_merging(new_block->host, new_block->length);
         }
     }
-    new_block->length = size;
 
     /* Keep the list sorted from biggest to smallest block.  */
     QTAILQ_FOREACH(block, &ram_list.blocks, next) {

@@ -1324,18 +1298,65 @@ ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
                                    old_ram_size, new_ram_size);
        }
     }
-    cpu_physical_memory_set_dirty_range(new_block->offset, size);
+    cpu_physical_memory_set_dirty_range(new_block->offset, new_block->length);
 
-    qemu_ram_setup_dump(new_block->host, size);
-    qemu_madvise(new_block->host, size, QEMU_MADV_HUGEPAGE);
-    qemu_madvise(new_block->host, size, QEMU_MADV_DONTFORK);
+    qemu_ram_setup_dump(new_block->host, new_block->length);
+    qemu_madvise(new_block->host, new_block->length, QEMU_MADV_HUGEPAGE);
+    qemu_madvise(new_block->host, new_block->length, QEMU_MADV_DONTFORK);
 
-    if (kvm_enabled())
-        kvm_setup_guest_memory(new_block->host, size);
+    if (kvm_enabled()) {
+        kvm_setup_guest_memory(new_block->host, new_block->length);
+    }
 
     return new_block->offset;
 }
 
+ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
+                                    const char *mem_path)
+{
+    RAMBlock *new_block;
+
+    if (xen_enabled()) {
+        fprintf(stderr, "-mem-path not supported with Xen\n");
+        exit(1);
+    }
+
+    if (phys_mem_alloc != qemu_anon_ram_alloc) {
+        /*
+         * file_ram_alloc() needs to allocate just like
+         * phys_mem_alloc, but we haven't bothered to provide
+         * a hook there.
+         */
+        fprintf(stderr,
+                "-mem-path not supported with this accelerator\n");
+        exit(1);
+    }
+
+    size = TARGET_PAGE_ALIGN(size);
+    new_block = g_malloc0(sizeof(*new_block));
+    new_block->mr = mr;
+    new_block->length = size;
+    new_block->host = file_ram_alloc(new_block, size, mem_path);
+    return ram_block_add(new_block);
+}
+
+ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
+                                   MemoryRegion *mr)
+{
+    RAMBlock *new_block;
+
+    size = TARGET_PAGE_ALIGN(size);
+    new_block = g_malloc0(sizeof(*new_block));
+    new_block->mr = mr;
+    new_block->length = size;
+    new_block->fd = -1;
+    new_block->host = host;
+    if (host) {
+        new_block->flags |= RAM_PREALLOC_MASK;
+    }
+    return ram_block_add(new_block);
+}
+
 ram_addr_t qemu_ram_alloc(ram_addr_t size, MemoryRegion *mr)
 {
     return qemu_ram_alloc_from_ptr(size, NULL, mr);

index 758a928..dfdd071 100644
@@ -442,9 +442,6 @@ typedef struct RAMList {
 } RAMList;
 extern RAMList ram_list;
 
-extern const char *mem_path;
-extern int mem_prealloc;
-
 /* Flags stored in the low bits of the TLB virtual address.  These are
    defined so that fast path ram access is all zeros.  */
 /* Zero if TLB entry is valid.  */

index 2edfa96..dedb258 100644
@@ -22,6 +22,8 @@
 #ifndef CONFIG_USER_ONLY
 #include "hw/xen/xen.h"
 
+ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
+                                    const char *mem_path);
 ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
                                    MemoryRegion *mr);
 ram_addr_t qemu_ram_alloc(ram_addr_t size, MemoryRegion *mr);

index 1e141e3..277230d 100644
@@ -133,6 +133,8 @@ extern uint8_t *boot_splash_filedata;
 extern size_t boot_splash_filedata_size;
 extern uint8_t qemu_extra_params_fw[2];
 extern QEMUClockType rtc_clock;
+extern const char *mem_path;
+extern int mem_prealloc;
 
 #define MAX_NODES 128
 

index 3f1df23..daaeb7e 100644
@@ -23,6 +23,7 @@
 
 #include "exec/memory-internal.h"
 #include "exec/ram_addr.h"
+#include "sysemu/sysemu.h"
 
 //#define DEBUG_UNASSIGNED
 

@@ -1016,7 +1017,11 @@ void memory_region_init_ram(MemoryRegion *mr,
     mr->ram = true;
     mr->terminates = true;
     mr->destructor = memory_region_destructor_ram;
-    mr->ram_addr = qemu_ram_alloc(size, mr);
+    if (mem_path) {
+        mr->ram_addr = qemu_ram_alloc_from_file(size, mr, mem_path);
+    } else {
+        mr->ram_addr = qemu_ram_alloc(size, mr);
+    }
 }
 
 void memory_region_init_ram_ptr(MemoryRegion *mr,

index 8ae8b95..379472d 100644
@@ -1130,14 +1130,6 @@ error:
     }
     return NULL;
 }
-#else
-static void *file_ram_alloc(RAMBlock *block,
-                            ram_addr_t memory,
-                            const char *path)
-{
-    fprintf(stderr, "-mem-path not supported on this host\n");
-    exit(1);
-}
 #endif
 
 static ram_addr_t find_ram_offset(ram_addr_t size)

@@ -1311,6 +1303,7 @@ static ram_addr_t ram_block_add(RAMBlock *new_block)
     return new_block->offset;
 }
 
+#ifdef __linux__
 ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
                                     const char *mem_path)
 {

@@ -1339,6 +1332,7 @@ ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
     new_block->host = file_ram_alloc(new_block, size, mem_path);
     return ram_block_add(new_block);
 }
+#endif
 
 ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
                                    MemoryRegion *mr)

index c084db2..75d4635 100644
@@ -311,6 +311,24 @@ void memory_region_init_ram(MemoryRegion *mr,
                             const char *name,
                             uint64_t size);
 
+#ifdef __linux__
+/**
+ * memory_region_init_ram_from_file:  Initialize RAM memory region with a
+ *                                    mmap-ed backend.
+ *
+ * @mr: the #MemoryRegion to be initialized.
+ * @owner: the object that tracks the region's reference count
+ * @name: the name of the region.
+ * @size: size of the region.
+ * @path: the path in which to allocate the RAM.
+ */
+void memory_region_init_ram_from_file(MemoryRegion *mr,
+                                      struct Object *owner,
+                                      const char *name,
+                                      uint64_t size,
+                                      const char *path);
+#endif
+
 /**
  * memory_region_init_ram_ptr:  Initialize RAM memory region from a
  *                              user-provided pointer.  Accesses into the

index daaeb7e..677b8f3 100644
@@ -1017,13 +1017,24 @@ void memory_region_init_ram(MemoryRegion *mr,
     mr->ram = true;
     mr->terminates = true;
     mr->destructor = memory_region_destructor_ram;
-    if (mem_path) {
-        mr->ram_addr = qemu_ram_alloc_from_file(size, mr, mem_path);
-    } else {
-        mr->ram_addr = qemu_ram_alloc(size, mr);
-    }
+    mr->ram_addr = qemu_ram_alloc(size, mr);
 }
 
+#ifdef __linux__
+void memory_region_init_ram_from_file(MemoryRegion *mr,
+                                      struct Object *owner,
+                                      const char *name,
+                                      uint64_t size,
+                                      const char *path)
+{
+    memory_region_init(mr, owner, name, size);
+    mr->ram = true;
+    mr->terminates = true;
+    mr->destructor = memory_region_destructor_ram;
+    mr->ram_addr = qemu_ram_alloc_from_file(size, mr, path);
+}
+#endif
+
 void memory_region_init_ram_ptr(MemoryRegion *mr,
                                 Object *owner,
                                 const char *name,

index b9850d7..e136612 100644
@@ -229,7 +229,16 @@ static void allocate_system_memory_nonnuma(MemoryRegion *mr, Object *owner,
                                            const char *name,
                                            uint64_t ram_size)
 {
-    memory_region_init_ram(mr, owner, name, ram_size);
+    if (mem_path) {
+#ifdef __linux__
+        memory_region_init_ram_from_file(mr, owner, name, ram_size, mem_path);
+#else
+        fprintf(stderr, "-mem-path not supported on this host\n");
+        exit(1);
+#endif
+    } else {
+        memory_region_init_ram(mr, owner, name, ram_size);
+    }
     vmstate_register_ram_global(mr);
 }
 

index 379472d..1e51b2f 100644
@@ -1021,7 +1021,8 @@ static void sigbus_handler(int signal)
 
 static void *file_ram_alloc(RAMBlock *block,
                             ram_addr_t memory,
-                            const char *path)
+                            const char *path,
+                            Error **errp)
 {
     char *filename;
     char *sanitized_name;

@@ -1040,7 +1041,8 @@ static void *file_ram_alloc(RAMBlock *block,
     }
 
     if (kvm_enabled() && !kvm_has_sync_mmu()) {
-        fprintf(stderr, "host lacks kvm mmu notifiers, -mem-path unsupported\n");
+        error_setg(errp,
+                   "host lacks kvm mmu notifiers, -mem-path unsupported\n");
         goto error;
     }
 

@@ -1057,7 +1059,8 @@ static void *file_ram_alloc(RAMBlock *block,
 
     fd = mkstemp(filename);
     if (fd < 0) {
-        perror("unable to create backing store for hugepages");
+        error_setg_errno(errp, errno,
+                         "unable to create backing store for hugepages");
         g_free(filename);
         goto error;
     }

@@ -1072,12 +1075,14 @@ static void *file_ram_alloc(RAMBlock *block,
      * If anything goes wrong with it under other filesystems,
      * mmap will fail.
      */
-    if (ftruncate(fd, memory))
+    if (ftruncate(fd, memory)) {
         perror("ftruncate");
+    }
 
     area = mmap(0, memory, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
     if (area == MAP_FAILED) {
-        perror("file_ram_alloc: can't mmap RAM pages");
+        error_setg_errno(errp, errno,
+                         "unable to map backing store for hugepages");
         close(fd);
         goto error;
     }

@@ -1305,13 +1310,14 @@ static ram_addr_t ram_block_add(RAMBlock *new_block)
 
 #ifdef __linux__
 ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
-                                    const char *mem_path)
+                                    const char *mem_path,
+                                    Error **errp)
 {
     RAMBlock *new_block;
 
     if (xen_enabled()) {
-        fprintf(stderr, "-mem-path not supported with Xen\n");
-        exit(1);
+        error_setg(errp, "-mem-path not supported with Xen\n");
+        return -1;
     }
 
     if (phys_mem_alloc != qemu_anon_ram_alloc) {

@@ -1320,16 +1326,22 @@ ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
          * phys_mem_alloc, but we haven't bothered to provide
          * a hook there.
          */
-        fprintf(stderr,
-                "-mem-path not supported with this accelerator\n");
-        exit(1);
+        error_setg(errp,
+                   "-mem-path not supported with this accelerator\n");
+        return -1;
     }
 
     size = TARGET_PAGE_ALIGN(size);
     new_block = g_malloc0(sizeof(*new_block));
     new_block->mr = mr;
     new_block->length = size;
-    new_block->host = file_ram_alloc(new_block, size, mem_path);
+    new_block->host = file_ram_alloc(new_block, size,
+                                     mem_path, errp);
+    if (!new_block->host) {
+        g_free(new_block);
+        return -1;
+    }
+
     return ram_block_add(new_block);
 }
 #endif

index 75d4635..4e606dd 100644
@@ -31,6 +31,7 @@
 #include "qemu/queue.h"
 #include "qemu/int128.h"
 #include "qemu/notify.h"
+#include "qapi/error.h"
 
 #define MAX_PHYS_ADDR_SPACE_BITS 62
 #define MAX_PHYS_ADDR            (((hwaddr)1 << MAX_PHYS_ADDR_SPACE_BITS) - 1)

@@ -321,12 +322,14 @@ void memory_region_init_ram(MemoryRegion *mr,
  * @name: the name of the region.
  * @size: size of the region.
  * @path: the path in which to allocate the RAM.
+ * @errp: pointer to Error*, to store an error if it happens.
  */
 void memory_region_init_ram_from_file(MemoryRegion *mr,
                                       struct Object *owner,
                                       const char *name,
                                       uint64_t size,
-                                      const char *path);
+                                      const char *path,
+                                      Error **errp);
 #endif
 
 /**

index dedb258..f9518a6 100644
@@ -23,7 +23,7 @@
 #include "hw/xen/xen.h"
 
 ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
-                                    const char *mem_path);
+                                    const char *mem_path, Error **errp);
 ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
                                    MemoryRegion *mr);
 ram_addr_t qemu_ram_alloc(ram_addr_t size, MemoryRegion *mr);

index 677b8f3..6b0c116 100644
@@ -1025,13 +1025,14 @@ void memory_region_init_ram_from_file(MemoryRegion *mr,
                                       struct Object *owner,
                                       const char *name,
                                       uint64_t size,
-                                      const char *path)
+                                      const char *path,
+                                      Error **errp)
 {
     memory_region_init(mr, owner, name, size);
     mr->ram = true;
     mr->terminates = true;
     mr->destructor = memory_region_destructor_ram;
-    mr->ram_addr = qemu_ram_alloc_from_file(size, mr, path);
+    mr->ram_addr = qemu_ram_alloc_from_file(size, mr, path, errp);
 }
 #endif
 

index e136612..966cacc 100644
@@ -231,7 +231,18 @@ static void allocate_system_memory_nonnuma(MemoryRegion *mr, Object *owner,
 {
     if (mem_path) {
 #ifdef __linux__
-        memory_region_init_ram_from_file(mr, owner, name, ram_size, mem_path);
+        Error *err = NULL;
+        memory_region_init_ram_from_file(mr, owner, name, ram_size,
+                                         mem_path, &err);
+
+        /* Legacy behavior: if allocation failed, fall back to
+         * regular RAM allocation.
+         */
+        if (!memory_region_size(mr)) {
+            qerror_report_err(err);
+            error_free(err);
+            memory_region_init_ram(mr, owner, name, ram_size);
+        }
 #else
         fprintf(stderr, "-mem-path not supported on this host\n");
         exit(1);

index 1e51b2f..710e025 100644
@@ -1012,13 +1012,6 @@ static long gethugepagesize(const char *path)
     return fs.f_bsize;
 }
 
-static sigjmp_buf sigjump;
-
-static void sigbus_handler(int signal)
-{
-    siglongjmp(sigjump, 1);
-}
-
 static void *file_ram_alloc(RAMBlock *block,
                             ram_addr_t memory,
                             const char *path,

@@ -1088,42 +1081,7 @@ static void *file_ram_alloc(RAMBlock *block,
     }
 
     if (mem_prealloc) {
-        int ret, i;
-        struct sigaction act, oldact;
-        sigset_t set, oldset;
-
-        memset(&act, 0, sizeof(act));
-        act.sa_handler = &sigbus_handler;
-        act.sa_flags = 0;
-
-        ret = sigaction(SIGBUS, &act, &oldact);
-        if (ret) {
-            perror("file_ram_alloc: failed to install signal handler");
-            exit(1);
-        }
-
-        /* unblock SIGBUS */
-        sigemptyset(&set);
-        sigaddset(&set, SIGBUS);
-        pthread_sigmask(SIG_UNBLOCK, &set, &oldset);
-
-        if (sigsetjmp(sigjump, 1)) {
-            fprintf(stderr, "file_ram_alloc: failed to preallocate pages\n");
-            exit(1);
-        }
-
-        /* MAP_POPULATE silently ignores failures */
-        for (i = 0; i < (memory/hpagesize); i++) {
-            memset(area + (hpagesize*i), 0, 1);
-        }
-
-        ret = sigaction(SIGBUS, &oldact, NULL);
-        if (ret) {
-            perror("file_ram_alloc: failed to reinstall signal handler");
-            exit(1);
-        }
-
-        pthread_sigmask(SIG_SETMASK, &oldset, NULL);
+        os_mem_prealloc(fd, area, memory);
     }
 
     block->fd = fd;

index ffb2966..9c1a119 100644
@@ -251,4 +251,6 @@ void qemu_init_auxval(char **envp);
 
 void qemu_set_tty_echo(int fd, bool echo);
 
+void os_mem_prealloc(int fd, char *area, size_t sz);
+
 #endif

91 lines not on this page · raw message

index 710e025..691d216 100644
@@ -70,6 +70,9 @@ AddressSpace address_space_memory;
 MemoryRegion io_mem_rom, io_mem_notdirty;
 static MemoryRegion io_mem_unassigned;
 
+/* RAM is pre-allocated and passed into qemu_ram_alloc_from_ptr */
+#define RAM_PREALLOC   (1 << 0)
+
 #endif
 
 struct CPUTailQ cpus = QTAILQ_HEAD_INITIALIZER(cpus);

@@ -1316,7 +1319,7 @@ ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
     new_block->fd = -1;
     new_block->host = host;
     if (host) {
-        new_block->flags |= RAM_PREALLOC_MASK;
+        new_block->flags |= RAM_PREALLOC;
     }
     return ram_block_add(new_block);
 }

@@ -1355,7 +1358,7 @@ void qemu_ram_free(ram_addr_t addr)
             QTAILQ_REMOVE(&ram_list.blocks, block, next);
             ram_list.mru_block = NULL;
             ram_list.version++;
-            if (block->flags & RAM_PREALLOC_MASK) {
+            if (block->flags & RAM_PREALLOC) {
                 ;
             } else if (xen_enabled()) {
                 xen_invalidate_map_cache_entry(block->host);

@@ -1387,7 +1390,7 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length)
         offset = addr - block->offset;
         if (offset < block->length) {
             vaddr = block->host + offset;
-            if (block->flags & RAM_PREALLOC_MASK) {
+            if (block->flags & RAM_PREALLOC) {
                 ;
             } else if (xen_enabled()) {
                 abort();

index dfdd071..6960544 100644
@@ -414,9 +414,6 @@ CPUArchState *cpu_copy(CPUArchState *env);
 
 /* memory API */
 
-/* RAM is pre-allocated and passed into qemu_ram_alloc_from_ptr */
-#define RAM_PREALLOC_MASK   (1 << 0)
-
 typedef struct RAMBlock {
     struct MemoryRegion *mr;
     uint8_t *host;

108 lines not on this page · raw message

index e22bcef..75a165a 100644
@@ -30,10 +30,9 @@ struct HostMemoryBackendFile {
 };
 
 static void
-file_backend_memory_init(UserCreatable *uc, Error **errp)
+file_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
 {
-    HostMemoryBackend *backend = MEMORY_BACKEND(uc);
-    HostMemoryBackendFile *fb = MEMORY_BACKEND_FILE(uc);
+    HostMemoryBackendFile *fb = MEMORY_BACKEND_FILE(backend);
 
     if (!backend->size) {
         error_setg(errp, "can't create backend with size 0");

@@ -58,9 +57,9 @@ file_backend_memory_init(UserCreatable *uc, Error **errp)
 static void
 file_backend_class_init(ObjectClass *oc, void *data)
 {
-    UserCreatableClass *ucc = USER_CREATABLE_CLASS(oc);
+    HostMemoryBackendClass *bc = MEMORY_BACKEND_CLASS(oc);
 
-    ucc->complete = file_backend_memory_init;
+    bc->alloc = file_backend_memory_alloc;
 }
 
 static char *get_mem_path(Object *o, Error **errp)

index cbf7e5a..7cbc051 100644
@@ -16,9 +16,8 @@
 
 
 static void
-ram_backend_memory_init(UserCreatable *uc, Error **errp)
+ram_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
 {
-    HostMemoryBackend *backend = MEMORY_BACKEND(uc);
     char *path;
 
     if (!backend->size) {

@@ -35,9 +34,9 @@ ram_backend_memory_init(UserCreatable *uc, Error **errp)
 static void
 ram_backend_class_init(ObjectClass *oc, void *data)
 {
-    UserCreatableClass *ucc = USER_CREATABLE_CLASS(oc);
+    HostMemoryBackendClass *bc = MEMORY_BACKEND_CLASS(oc);
 
-    ucc->complete = ram_backend_memory_init;
+    bc->alloc = ram_backend_memory_alloc;
 }
 
 static const TypeInfo ram_backend_info = {

index 6f26605..5b2117d 100644
@@ -68,8 +68,16 @@ static void host_memory_backend_finalize(Object *obj)
 static void
 host_memory_backend_memory_init(UserCreatable *uc, Error **errp)
 {
-    error_setg(errp, "memory_init is not implemented for type [%s]",
-               object_get_typename(OBJECT(uc)));
+    HostMemoryBackend *backend = MEMORY_BACKEND(uc);
+    HostMemoryBackendClass *bc = MEMORY_BACKEND_GET_CLASS(uc);
+
+    if (!bc->alloc) {
+        error_setg(errp, "memory_alloc is not implemented for type [%s]",
+                   object_get_typename(OBJECT(uc)));
+        return;
+    }
+
+    bc->alloc(backend, errp);
 }
 
 MemoryRegion *

index d396fd8..42c98bd 100644
@@ -32,6 +32,8 @@ typedef struct HostMemoryBackendClass HostMemoryBackendClass;
  */
 struct HostMemoryBackendClass {
     ObjectClass parent_class;
+
+    void (*alloc)(HostMemoryBackend *backend, Error **errp);
 };
 
 /**

index 5b2117d..3e5ebfc 100644
@@ -49,8 +49,74 @@ host_memory_backend_set_size(Object *obj, Visitor *v, void *opaque,
     backend->size = value;
 }
 
+static bool host_memory_backend_get_merge(Object *obj, Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+
+    return backend->merge;
+}
+
+static void host_memory_backend_set_merge(Object *obj, bool value, Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+
+    if (!memory_region_size(&backend->mr)) {
+        backend->merge = value;
+        return;
+    }
+
+    if (value != backend->merge) {
+        void *ptr = memory_region_get_ram_ptr(&backend->mr);
+        uint64_t sz = memory_region_size(&backend->mr);
+
+        qemu_madvise(ptr, sz,
+                     value ? QEMU_MADV_MERGEABLE : QEMU_MADV_UNMERGEABLE);
+        backend->merge = value;
+    }
+}
+
+static bool host_memory_backend_get_dump(Object *obj, Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+
+    return backend->dump;
+}
+
+static void host_memory_backend_set_dump(Object *obj, bool value, Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+
+    if (!memory_region_size(&backend->mr)) {
+        backend->dump = value;
+        return;
+    }
+
+    if (value != backend->dump) {
+        void *ptr = memory_region_get_ram_ptr(&backend->mr);
+        uint64_t sz = memory_region_size(&backend->mr);
+
+        qemu_madvise(ptr, sz,
+                     value ? QEMU_MADV_DODUMP : QEMU_MADV_DONTDUMP);
+        backend->dump = value;
+    }
+}
+
+
 static void host_memory_backend_initfn(Object *obj)
 {
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+
+    backend->merge = qemu_opt_get_bool(qemu_get_machine_opts(),
+                                       "mem-merge", true);
+    backend->dump = qemu_opt_get_bool(qemu_get_machine_opts(),
+                                      "dump-guest-core", true);
+
+    object_property_add_bool(obj, "merge",
+                        host_memory_backend_get_merge,
+                        host_memory_backend_set_merge, NULL);
+    object_property_add_bool(obj, "dump",
+                        host_memory_backend_get_dump,
+                        host_memory_backend_set_dump, NULL);
     object_property_add(obj, "size", "int",
                         host_memory_backend_get_size,
                         host_memory_backend_set_size, NULL, NULL, NULL);

@@ -70,6 +136,9 @@ host_memory_backend_memory_init(UserCreatable *uc, Error **errp)
 {
     HostMemoryBackend *backend = MEMORY_BACKEND(uc);
     HostMemoryBackendClass *bc = MEMORY_BACKEND_GET_CLASS(uc);
+    Error *local_err = NULL;
+    void *ptr;
+    uint64_t sz;
 
     if (!bc->alloc) {
         error_setg(errp, "memory_alloc is not implemented for type [%s]",

@@ -77,7 +146,21 @@ host_memory_backend_memory_init(UserCreatable *uc, Error **errp)
         return;
     }
 
-    bc->alloc(backend, errp);
+    bc->alloc(backend, &local_err);
+    if (local_err) {
+        error_propagate(errp, local_err);
+        return;
+    }
+
+    ptr = memory_region_get_ram_ptr(&backend->mr);
+    sz = memory_region_size(&backend->mr);
+
+    if (backend->merge) {
+        qemu_madvise(ptr, sz, QEMU_MADV_MERGEABLE);
+    }
+    if (!backend->dump) {
+        qemu_madvise(ptr, sz, QEMU_MADV_DONTDUMP);
+    }
 }
 
 MemoryRegion *

index 9c1a119..820c5d0 100644
@@ -116,6 +116,16 @@ void qemu_anon_ram_free(void *ptr, size_t size);
 #else
 #define QEMU_MADV_MERGEABLE QEMU_MADV_INVALID
 #endif
+#ifdef MADV_UNMERGEABLE
+#define QEMU_MADV_UNMERGEABLE MADV_UNMERGEABLE
+#else
+#define QEMU_MADV_UNMERGEABLE QEMU_MADV_INVALID
+#endif
+#ifdef MADV_DODUMP
+#define QEMU_MADV_DODUMP MADV_DODUMP
+#else
+#define QEMU_MADV_DODUMP QEMU_MADV_INVALID
+#endif
 #ifdef MADV_DONTDUMP
 #define QEMU_MADV_DONTDUMP MADV_DONTDUMP
 #else

index 42c98bd..a6b96eb 100644
@@ -50,6 +50,7 @@ struct HostMemoryBackend {
 
     /* protected */
     uint64_t size;
+    bool merge, dump;
 
     MemoryRegion mr;
 };

index 75a165a..67328b9 100644
@@ -9,7 +9,9 @@
  * This work is licensed under the terms of the GNU GPL, version 2 or later.
  * See the COPYING file in the top-level directory.
  */
+#include "qemu-common.h"
 #include "sysemu/hostmem.h"
+#include "sysemu/sysemu.h"
 #include "qom/object_interfaces.h"
 
 /* hostmem-file.c */

@@ -46,6 +48,7 @@ file_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
     error_setg(errp, "-mem-path not supported on this host");
 #else
     if (!memory_region_size(&backend->mr)) {
+        backend->force_prealloc = mem_prealloc;
         memory_region_init_ram_from_file(&backend->mr, OBJECT(backend),
                                  object_get_canonical_path(OBJECT(backend)),
                                  backend->size,

66 lines not on this page · raw message

index 691d216..fab45b5 100644
@@ -1433,6 +1433,13 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length)
 }
 #endif /* !_WIN32 */
 
+int qemu_get_ram_fd(ram_addr_t addr)
+{
+    RAMBlock *block = qemu_get_ram_block(addr);
+
+    return block->fd;
+}
+
 /* Return a host pointer to ram allocated with qemu_ram_alloc.
    With the exception of the softmmu code in this file, this should
    only be used for local memory (e.g. video ram) that the device owns,

index 4e606dd..846834e 100644
@@ -534,6 +534,16 @@ bool memory_region_is_logging(MemoryRegion *mr);
 bool memory_region_is_rom(MemoryRegion *mr);
 
 /**
+ * memory_region_get_fd: Get a file descriptor backing a RAM memory region.
+ *
+ * Returns a file descriptor backing a file-based RAM memory region,
+ * or -1 if the region is not a file-based RAM memory region.
+ *
+ * @mr: the RAM or alias memory region being queried.
+ */
+int memory_region_get_fd(MemoryRegion *mr);
+
+/**
  * memory_region_get_ram_ptr: Get a pointer into a RAM memory region.
  *
  * Returns a host pointer to a RAM memory region (created with

index f9518a6..d352f60 100644
@@ -27,6 +27,7 @@ ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
 ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
                                    MemoryRegion *mr);
 ram_addr_t qemu_ram_alloc(ram_addr_t size, MemoryRegion *mr);
+int qemu_get_ram_fd(ram_addr_t addr);
 void *qemu_get_ram_ptr(ram_addr_t addr);
 void qemu_ram_free(ram_addr_t addr);
 void qemu_ram_free_from_ptr(ram_addr_t addr);

index a6b96eb..819b72d 100644
@@ -51,6 +51,7 @@ struct HostMemoryBackend {
     /* protected */
     uint64_t size;
     bool merge, dump;
+    bool prealloc, force_prealloc;
 
     MemoryRegion mr;
 };

index 6b0c116..118704d 100644
@@ -1258,6 +1258,17 @@ void memory_region_reset_dirty(MemoryRegion *mr, hwaddr addr,
     cpu_physical_memory_reset_dirty(mr->ram_addr + addr, size, client);
 }
 
+int memory_region_get_fd(MemoryRegion *mr)
+{
+    if (mr->alias) {
+        return memory_region_get_fd(mr->alias);
+    }
+
+    assert(mr->terminates);
+
+    return qemu_get_ram_fd(mr->ram_addr & TARGET_PAGE_MASK);
+}
+
 void *memory_region_get_ram_ptr(MemoryRegion *mr)
 {
     if (mr->alias) {

index 67328b9..7f30011 100644
@@ -28,6 +28,8 @@ typedef struct HostMemoryBackendFile HostMemoryBackendFile;
 
 struct HostMemoryBackendFile {
     HostMemoryBackend parent_obj;
+
+    bool share;
     char *mem_path;
 };
 

@@ -51,7 +53,7 @@ file_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
         backend->force_prealloc = mem_prealloc;
         memory_region_init_ram_from_file(&backend->mr, OBJECT(backend),
                                  object_get_canonical_path(OBJECT(backend)),
-                                 backend->size,
+                                 backend->size, fb->share,
                                  fb->mem_path, errp);
     }
 #endif

@@ -87,9 +89,31 @@ static void set_mem_path(Object *o, const char *str, Error **errp)
     fb->mem_path = g_strdup(str);
 }
 
+static bool file_memory_backend_get_share(Object *o, Error **errp)
+{
+    HostMemoryBackendFile *fb = MEMORY_BACKEND_FILE(o);
+
+    return fb->share;
+}
+
+static void file_memory_backend_set_share(Object *o, bool value, Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(o);
+    HostMemoryBackendFile *fb = MEMORY_BACKEND_FILE(o);
+
+    if (memory_region_size(&backend->mr)) {
+        error_setg(errp, "cannot change property value");
+        return;
+    }
+    fb->share = value;
+}
+
 static void
 file_backend_instance_init(Object *o)
 {
+    object_property_add_bool(o, "share",
+                        file_memory_backend_get_share,
+                        file_memory_backend_set_share, NULL);
     object_property_add_str(o, "mem-path", get_mem_path,
                             set_mem_path, NULL);
 }

index fab45b5..85fb85c 100644
@@ -73,6 +73,9 @@ static MemoryRegion io_mem_unassigned;
 /* RAM is pre-allocated and passed into qemu_ram_alloc_from_ptr */
 #define RAM_PREALLOC   (1 << 0)
 
+/* RAM is mmap-ed with MAP_SHARED */
+#define RAM_SHARED     (1 << 1)
+
 #endif
 
 struct CPUTailQ cpus = QTAILQ_HEAD_INITIALIZER(cpus);

@@ -1075,7 +1078,9 @@ static void *file_ram_alloc(RAMBlock *block,
         perror("ftruncate");
     }
 
-    area = mmap(0, memory, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
+    area = mmap(0, memory, PROT_READ | PROT_WRITE,
+                (block->flags & RAM_SHARED ? MAP_SHARED : MAP_PRIVATE),
+                fd, 0);
     if (area == MAP_FAILED) {
         error_setg_errno(errp, errno,
                          "unable to map backing store for hugepages");

@@ -1271,7 +1276,7 @@ static ram_addr_t ram_block_add(RAMBlock *new_block)
 
 #ifdef __linux__
 ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
-                                    const char *mem_path,
+                                    bool share, const char *mem_path,
                                     Error **errp)
 {
     RAMBlock *new_block;

@@ -1296,6 +1301,7 @@ ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
     new_block = g_malloc0(sizeof(*new_block));
     new_block->mr = mr;
     new_block->length = size;
+    new_block->flags = share ? RAM_SHARED : 0;
     new_block->host = file_ram_alloc(new_block, size,
                                      mem_path, errp);
     if (!new_block->host) {

@@ -1398,12 +1404,8 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length)
                 flags = MAP_FIXED;
                 munmap(vaddr, length);
                 if (block->fd >= 0) {
-#ifdef MAP_POPULATE
-                    flags |= mem_prealloc ? MAP_POPULATE | MAP_SHARED :
-                        MAP_PRIVATE;
-#else
-                    flags |= MAP_PRIVATE;
-#endif
+                    flags |= (block->flags & RAM_SHARED ?
+                              MAP_SHARED : MAP_PRIVATE);
                     area = mmap(vaddr, length, PROT_READ | PROT_WRITE,
                                 flags, block->fd, offset);
                 } else {

index 846834e..21443bc 100644
@@ -321,6 +321,7 @@ void memory_region_init_ram(MemoryRegion *mr,
  * @owner: the object that tracks the region's reference count
  * @name: the name of the region.
  * @size: size of the region.
+ * @share: %true if memory must be mmaped with the MAP_SHARED flag
  * @path: the path in which to allocate the RAM.
  * @errp: pointer to Error*, to store an error if it happens.
  */

@@ -328,6 +329,7 @@ void memory_region_init_ram_from_file(MemoryRegion *mr,
                                       struct Object *owner,
                                       const char *name,
                                       uint64_t size,
+                                      bool share,
                                       const char *path,
                                       Error **errp);
 #endif

index d352f60..1d4ac74 100644
@@ -23,7 +23,8 @@
 #include "hw/xen/xen.h"
 
 ram_addr_t qemu_ram_alloc_from_file(ram_addr_t size, MemoryRegion *mr,
-                                    const char *mem_path, Error **errp);
+                                    bool share, const char *mem_path,
+                                    Error **errp);
 ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void *host,
                                    MemoryRegion *mr);
 ram_addr_t qemu_ram_alloc(ram_addr_t size, MemoryRegion *mr);

index 118704d..202869a 100644
@@ -1025,6 +1025,7 @@ void memory_region_init_ram_from_file(MemoryRegion *mr,
                                       struct Object *owner,
                                       const char *name,
                                       uint64_t size,
+                                      bool share,
                                       const char *path,
                                       Error **errp)
 {

@@ -1032,7 +1033,7 @@ void memory_region_init_ram_from_file(MemoryRegion *mr,
     mr->ram = true;
     mr->terminates = true;
     mr->destructor = memory_region_destructor_ram;
-    mr->ram_addr = qemu_ram_alloc_from_file(size, mr, path, errp);
+    mr->ram_addr = qemu_ram_alloc_from_file(size, mr, share, path, errp);
 }
 #endif
 

index 966cacc..fa117f0 100644
@@ -232,7 +232,7 @@ static void allocate_system_memory_nonnuma(MemoryRegion *mr, Object *owner,
     if (mem_path) {
 #ifdef __linux__
         Error *err = NULL;
-        memory_region_init_ram_from_file(mr, owner, name, ram_size,
+        memory_region_init_ram_from_file(mr, owner, name, ram_size, false,
                                          mem_path, &err);
 
         /* Legacy behavior: if allocation failed, fall back to

index 738bb31..d3f8476 100644
@@ -10,11 +10,20 @@
  * See the COPYING file in the top-level directory.
  */
 #include "sysemu/hostmem.h"
-#include "sysemu/sysemu.h"
 #include "qapi/visitor.h"
+#include "qapi-types.h"
+#include "qapi-visit.h"
 #include "qapi/qmp/qerror.h"
 #include "qom/object_interfaces.h"
 
+#ifdef CONFIG_NUMA
+#include <numaif.h>
+QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_DEFAULT != MPOL_DEFAULT);
+QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_PREFERRED != MPOL_PREFERRED);
+QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_BIND != MPOL_BIND);
+QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_INTERLEAVE != MPOL_INTERLEAVE);
+#endif
+
 static void
 host_memory_backend_get_size(Object *obj, Visitor *v, void *opaque,
                             const char *name, Error **errp)

@@ -49,6 +58,84 @@ host_memory_backend_set_size(Object *obj, Visitor *v, void *opaque,
     backend->size = value;
 }
 
+static void
+get_host_nodes(Object *obj, Visitor *v, void *opaque, const char *name,
+               Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+    uint16List *host_nodes = NULL;
+    uint16List **node = &host_nodes;
+    unsigned long value;
+
+    value = find_first_bit(backend->host_nodes, MAX_NODES);
+    if (value == MAX_NODES) {
+        return;
+    }
+
+    *node = g_malloc0(sizeof(**node));
+    (*node)->value = value;
+    node = &(*node)->next;
+
+    do {
+        value = find_next_bit(backend->host_nodes, MAX_NODES, value + 1);
+        if (value == MAX_NODES) {
+            break;
+        }
+
+        *node = g_malloc0(sizeof(**node));
+        (*node)->value = value;
+        node = &(*node)->next;
+    } while (true);
+
+    visit_type_uint16List(v, &host_nodes, name, errp);
+}
+
+static void
+set_host_nodes(Object *obj, Visitor *v, void *opaque, const char *name,
+               Error **errp)
+{
+#ifdef CONFIG_NUMA
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+    uint16List *l = NULL;
+
+    visit_type_uint16List(v, &l, name, errp);
+
+    while (l) {
+        bitmap_set(backend->host_nodes, l->value, 1);
+        l = l->next;
+    }
+#else
+    error_setg(errp, "NUMA node binding are not supported by this QEMU");
+#endif
+}
+
+static void
+get_policy(Object *obj, Visitor *v, void *opaque, const char *name,
+           Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+    int policy = backend->policy;
+
+    visit_type_enum(v, &policy, HostMemPolicy_lookup, NULL, name, errp);
+}
+
+static void
+set_policy(Object *obj, Visitor *v, void *opaque, const char *name,
+           Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+    int policy;
+
+    visit_type_enum(v, &policy, HostMemPolicy_lookup, NULL, name, errp);
+    backend->policy = policy;
+
+#ifndef CONFIG_NUMA
+    if (policy != HOST_MEM_POLICY_DEFAULT) {
+        error_setg(errp, "NUMA policies are not supported by this QEMU");
+    }
+#endif
+}
+
 static bool host_memory_backend_get_merge(Object *obj, Error **errp)
 {
     HostMemoryBackend *backend = MEMORY_BACKEND(obj);

@@ -159,6 +246,12 @@ static void host_memory_backend_initfn(Object *obj)
     object_property_add(obj, "size", "int",
                         host_memory_backend_get_size,
                         host_memory_backend_set_size, NULL, NULL, NULL);
+    object_property_add(obj, "host-nodes", "int",
+                        get_host_nodes,
+                        set_host_nodes, NULL, NULL, NULL);
+    object_property_add(obj, "policy", "str",
+                        get_policy,
+                        set_policy, NULL, NULL, NULL);
 }
 
 static void host_memory_backend_finalize(Object *obj)

@@ -203,6 +296,20 @@ host_memory_backend_memory_init(UserCreatable *uc, Error **errp)
     if (backend->prealloc) {
         os_mem_prealloc(memory_region_get_fd(&backend->mr), ptr, sz);
     }
+
+#ifdef CONFIG_NUMA
+    unsigned long maxnode = find_last_bit(backend->host_nodes, MAX_NODES);
+
+    /* This is a workaround for a long standing bug in Linux'
+     * mbind implementation, which cuts off the last specified
+     * node.
+     */
+    if (mbind(ptr, sz, backend->policy, backend->host_nodes, maxnode + 2, 0)) {
+        error_setg_errno(errp, errno,
+                         "cannot bind memory to host NUMA nodes");
+        return;
+    }
+#endif
 }
 
 MemoryRegion *

index 819b72d..4e96298 100644
@@ -12,8 +12,10 @@
 #ifndef QEMU_HOSTMEM_H
 #define QEMU_HOSTMEM_H
 
+#include "sysemu/sysemu.h" /* for MAX_NODES */
 #include "qom/object.h"
 #include "exec/memory.h"
+#include "qemu/bitmap.h"
 
 #define TYPE_MEMORY_BACKEND "memory"
 #define MEMORY_BACKEND(obj) \

@@ -52,6 +54,8 @@ struct HostMemoryBackend {
     uint64_t size;
     bool merge, dump;
     bool prealloc, force_prealloc;
+    DECLARE_BITMAP(host_nodes, MAX_NODES);
+    HostMemPolicy policy;
 
     MemoryRegion mr;
 };

index aae9720..8879f8a 100644
@@ -3,6 +3,7 @@
 
 #include <inttypes.h>
 #include <qemu/typedefs.h>
+#include "qemu/queue.h"
 
 /*
  * Operations on 64 bit address ranges.

@@ -60,4 +61,127 @@ static inline int ranges_overlap(uint64_t first1, uint64_t len1,
     return !(last2 < first1 || last1 < first2);
 }
 
+typedef struct SignedRangeList SignedRangeList;
+
+typedef struct SignedRange {
+    int64_t start;
+    int64_t length;
+
+    QTAILQ_ENTRY(SignedRange) entry;
+} SignedRange;
+
+QTAILQ_HEAD(SignedRangeList, SignedRange);
+
+static inline int64_t s_range_end(int64_t start, int64_t length)
+{
+    return start + length - 1;
+}
+
+/* negative length or overflow */
+static inline bool s_range_overflow(int64_t start, int64_t length)
+{
+    return s_range_end(start, length) < start;
+}
+
+static inline SignedRange *s_range_new(int64_t start, int64_t length)
+{
+    SignedRange *range = NULL;
+
+    if (s_range_overflow(start, length)) {
+        return NULL;
+    }
+
+    range = g_malloc0(sizeof(*range));
+    range->start = start;
+    range->length = length;
+
+    return range;
+}
+
+static inline void s_range_free(SignedRange *range)
+{
+    g_free(range);
+}
+
+static inline bool s_range_overlap(int64_t start1, int64_t length1,
+                                   int64_t start2, int64_t length2)
+{
+    return !((start1 + length1) < start2 || (start2 + length2) < start1);
+}
+
+static inline int s_range_join(SignedRange *range,
+                               int64_t start, int64_t length)
+{
+    if (s_range_overflow(start, length)) {
+        return -1;
+    }
+
+    if (s_range_overlap(range->start, range->length, start, length)) {
+        int64_t end = s_range_end(range->start, range->length);
+        if (end < s_range_end(start, length)) {
+            end = s_range_end(start, length);
+        }
+        if (range->start > start) {
+            range->start = start;
+        }
+        range->length = end - range->start + 1;
+        return 0;
+    }
+
+    return -1;
+}
+
+static inline int s_range_compare(int64_t start1, int64_t length1,
+                                  int64_t start2, int64_t length2)
+{
+    if (start1 == start2 && length1 == length2) {
+        return 0;
+    } else if (s_range_end(start1, length1) <
+               s_range_end(start2, length2)) {
+        return -1;
+    } else {
+        return 1;
+    }
+}
+
+/* Add range to list. Keep them sorted, and merge ranges whenever possible */
+static inline bool range_list_add(SignedRangeList *list,
+                                  int64_t start, int64_t length)
+{
+    SignedRange *r, *next, *new_range = NULL, *cur = NULL;
+
+    if (s_range_overflow(start, length)) {
+        return false;
+    }
+
+    QTAILQ_FOREACH_SAFE(r, list, entry, next) {
+        if (s_range_overlap(r->start, r->length, start, length)) {
+            s_range_join(r, start, length);
+            break;
+        } else if (s_range_compare(start, length, r->start, r->length) < 0) {
+            cur = r;
+            break;
+        }
+    }
+
+    if (!r) {
+        new_range = s_range_new(start, length);
+        QTAILQ_INSERT_TAIL(list, new_range, entry);
+    } else if (cur) {
+        new_range = s_range_new(start, length);
+        QTAILQ_INSERT_BEFORE(cur, new_range, entry);
+    } else {
+        SignedRange *next = QTAILQ_NEXT(r, entry);
+        while (next && s_range_overlap(r->start, r->length,
+                                       next->start, next->length)) {
+            s_range_join(r, next->start, next->length);
+            QTAILQ_REMOVE(list, next, entry);
+            s_range_free(next);
+            next = QTAILQ_NEXT(r, entry);
+        }
+    }
+
+    return true;
+}
+
 #endif

index 793548a..0f6add7 100644
@@ -15,31 +15,182 @@
 #include "qapi/visitor-impl.h"
 #include "qapi/qmp/qerror.h"
 #include "qemu/option.h"
+#include "qemu/queue.h"
+#include "qemu/range.h"
+
 
 struct StringInputVisitor
 {
     Visitor visitor;
+
+    bool head;
+
+    SignedRangeList *ranges;
+    SignedRange *cur_range;
+    int64_t cur;
+
     const char *string;
 };
 
+static void parse_str(StringInputVisitor *siv, Error **errp)
+{
+    char *str = (char *) siv->string;
+    long long start, end;
+    SignedRange *r, *next;
+    char *endptr;
+
+    if (siv->ranges) {
+        return;
+    }
+
+    siv->ranges = g_malloc0(sizeof(*siv->ranges));
+    QTAILQ_INIT(siv->ranges);
+    errno = 0;
+    do {
+        start = strtoll(str, &endptr, 0);
+        if (errno == 0 && endptr > str && INT64_MIN <= start &&
+            start <= INT64_MAX) {
+            if (*endptr == '\0') {
+                if (!range_list_add(siv->ranges, start, 1)) {
+                    goto error;
+                }
+                str = NULL;
+            } else if (*endptr == '-') {
+                str = endptr + 1;
+                end = strtoll(str, &endptr, 0);
+                if (errno == 0 && endptr > str &&
+                    INT64_MIN <= end && end <= INT64_MAX && start <= end &&
+                    (start > INT64_MAX - 65536 ||
+                     end < start + 65536)) {
+                    if (*endptr == '\0') {
+                        if (!range_list_add(siv->ranges, start,
+                                            end - start + 1)) {
+                            goto error;
+                        }
+                        str = NULL;
+                    } else if (*endptr == ',') {
+                        str = endptr + 1;
+                        if (!range_list_add(siv->ranges, start,
+                                            end - start + 1)) {
+                            goto error;
+                        }
+                    } else {
+                        goto error;
+                    }
+                } else {
+                    goto error;
+                }
+            } else if (*endptr == ',') {
+                str = endptr + 1;
+                if (!range_list_add(siv->ranges, start, 1)) {
+                    goto error;
+                }
+            } else {
+                goto error;
+            }
+        } else {
+            goto error;
+        }
+    } while (str);
+
+    return;
+error:
+    if (siv->ranges) {
+        QTAILQ_FOREACH_SAFE(r, siv->ranges, entry, next) {
+            QTAILQ_REMOVE(siv->ranges, r, entry);
+            g_free(r);
+        }
+        g_free(siv->ranges);
+        siv->ranges = NULL;
+    }
+}
+
+static void
+start_list(Visitor *v, const char *name, Error **errp)
+{
+    StringInputVisitor *siv = DO_UPCAST(StringInputVisitor, visitor, v);
+
+    parse_str(siv, errp);
+
+    if (siv->ranges) {
+        siv->cur_range = QTAILQ_FIRST(siv->ranges);
+        if (siv->cur_range) {
+            siv->cur = siv->cur_range->start;
+        }
+    }
+}
+
+static GenericList *
+next_list(Visitor *v, GenericList **list, Error **errp)
+{
+    StringInputVisitor *siv = DO_UPCAST(StringInputVisitor, visitor, v);
+    GenericList **link;
+
+    if (!siv->ranges || !siv->cur_range) {
+        return NULL;
+    }
+
+    if (siv->cur < siv->cur_range->start ||
+        siv->cur >= (siv->cur_range->start + siv->cur_range->length)) {
+        siv->cur_range = QTAILQ_NEXT(siv->cur_range, entry);
+        if (siv->cur_range) {
+            siv->cur = siv->cur_range->start;
+        } else {
+            return NULL;
+        }
+    }
+
+    if (siv->head) {
+        link = list;
+        siv->head = false;
+    } else {
+        link = &(*list)->next;
+    }
+
+    *link = g_malloc0(sizeof **link);
+    return *link;
+}
+
+static void
+end_list(Visitor *v, Error **errp)
+{
+    StringInputVisitor *siv = DO_UPCAST(StringInputVisitor, visitor, v);
+    siv->head = true;
+}
+
 static void parse_type_int(Visitor *v, int64_t *obj, const char *name,
                            Error **errp)
 {
     StringInputVisitor *siv = DO_UPCAST(StringInputVisitor, visitor, v);
-    char *endp = (char *) siv->string;
-    long long val;
 
-    errno = 0;
-    if (siv->string) {
-        val = strtoll(siv->string, &endp, 0);
-    }
-    if (!siv->string || errno || endp == siv->string || *endp) {
+    if (!siv->string) {
         error_set(errp, QERR_INVALID_PARAMETER_TYPE, name ? name : "null",
                   "integer");
         return;
     }
 
-    *obj = val;
+    parse_str(siv, errp);
+
+    if (!siv->ranges) {
+        goto error;
+    }
+
+    if (!siv->cur_range) {
+        siv->cur_range = QTAILQ_FIRST(siv->ranges);
+        if (siv->cur_range) {
+            siv->cur = siv->cur_range->start;
+        } else {
+            goto error;
+        }
+    }
+
+    *obj = siv->cur;
+    siv->cur++;
+    return;
+
+error:
+    error_set(errp, QERR_INVALID_PARAMETER_VALUE, name,
+              "an int64 value or range");
 }
 
 static void parse_type_size(Visitor *v, uint64_t *obj, const char *name,

@@ -140,6 +291,16 @@ Visitor *string_input_get_visitor(StringInputVisitor *v)
 
 void string_input_visitor_cleanup(StringInputVisitor *v)
 {
+    SignedRange *r, *next;
+
+    if (v->ranges) {
+        QTAILQ_FOREACH_SAFE(r, v->ranges, entry, next) {
+            QTAILQ_REMOVE(v->ranges, r, entry);
+            g_free(r);
+        }
+        g_free(v->ranges);
+    }
+
     g_free(v);
 }
 

@@ -155,8 +316,12 @@ StringInputVisitor *string_input_visitor_new(const char *str)
     v->visitor.type_bool = parse_type_bool;
     v->visitor.type_str = parse_type_str;
     v->visitor.type_number = parse_type_number;
+    v->visitor.start_list = start_list;
+    v->visitor.next_list = next_list;
+    v->visitor.end_list = end_list;
     v->visitor.start_optional = parse_start_optional;
 
     v->string = str;
+    v->head = true;
     return v;
 }

index d406263..98210b4 100644
@@ -64,6 +64,35 @@ static void test_visitor_in_int(TestInputVisitorData *data,
     g_assert_cmpint(res, ==, value);
 }
 
+static void test_visitor_in_intList(TestInputVisitorData *data,
+                                    const void *unused)
+{
+    int64_t value[] = {-2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 20};
+    int16List *res = NULL, *tmp;
+    Error *errp = NULL;
+    Visitor *v;
+    int i = 0;
+
+    v = visitor_input_test_init(data, "1,2,-2-1,2-4,20,5-9,1-8");
+
+    visit_type_int16List(v, &res, NULL, &errp);
+    g_assert(!error_is_set(&errp));
+    tmp = res;
+    while (i < sizeof(value) / sizeof(value[0])) {
+        g_assert(tmp);
+        g_assert_cmpint(tmp->value, ==, value[i++]);
+        tmp = tmp->next;
+    }
+    g_assert(!tmp);
+
+    tmp = res;
+    while (tmp) {
+        res = res->next;
+        g_free(tmp);
+        tmp = res;
+    }
+}
+
 static void test_visitor_in_bool(TestInputVisitorData *data,
                                  const void *unused)
 {