Merge tag 'patchew/20200219160953.13771-1-imammedo@redhat.com' of htt…

…ps://github.com/patchew-project/qemu into HEAD

This series removes ad hoc RAM allocation API (memory_region_allocate_system_memory)
and consolidates it around hostmem backend. It allows to

* resolve conflicts between global -mem-prealloc and hostmem's "policy" option,
  fixing premature allocation before binding policy is applied

* simplify complicated memory allocation routines which had to deal with 2 ways
  to allocate RAM.

* reuse hostmem backends of a choice for main RAM without adding extra CLI
  options to duplicate hostmem features.  A recent case was -mem-shared, to
  enable vhost-user on targets that don't support hostmem backends [1] (ex: s390)

* move RAM allocation from individual boards into generic machine code and
  provide them with prepared MemoryRegion.

* clean up deprecated NUMA features which were tied to the old API (see patches)
  - "numa: remove deprecated -mem-path fallback to anonymous RAM"
  - (POSTPONED, waiting on libvirt side) "forbid '-numa node,mem' for 5.0 and newer machine types"
  - (POSTPONED) "numa: remove deprecated implicit RAM distribution between nodes"

Introduce a new machine.memory-backend property and wrapper code that aliases
global -mem-path and -mem-alloc into automatically created hostmem backend
properties (provided memory-backend was not set explicitly given by user).
A bulk of trivial patches then follow to incrementally convert individual
boards to using machine.memory-backend provided MemoryRegion.

Board conversion typically involves:

* providing MachineClass::default_ram_size and MachineClass::default_ram_id
  so generic code could create default backend if user didn't explicitly provide
  memory-backend or -m options

* dropping memory_region_allocate_system_memory() call

* using convenience MachineState::ram MemoryRegion, which points to MemoryRegion
   allocated by ram-memdev

On top of that for some boards:

* missing ram_size checks are added (typically it were boards with fixed ram size)

* ram_size fixups are replaced by checks and hard errors, forcing user to
  provide correct "-m" values instead of ignoring it and continuing running.

After all boards are converted, the old API is removed and memory allocation
routines are cleaned up.

backends/hostmem-file.c

@@ -18,13 +18,6 @@
 #include "sysemu/sysemu.h"
 #include "qom/object_interfaces.h"
 
-/* hostmem-file.c */
-/**
- * @TYPE_MEMORY_BACKEND_FILE:
- * name of backend that uses mmap on a file descriptor
- */
-#define TYPE_MEMORY_BACKEND_FILE "memory-backend-file"
-
 #define MEMORY_BACKEND_FILE(obj) \
     OBJECT_CHECK(HostMemoryBackendFile, (obj), TYPE_MEMORY_BACKEND_FILE)
 
@@ -58,7 +51,6 @@ file_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
         return;
     }
 
-    backend->force_prealloc = mem_prealloc;
     name = host_memory_backend_get_name(backend);
     memory_region_init_ram_from_file(&backend->mr, OBJECT(backend),
                                      name,

backends/hostmem-memfd.c

@@ -45,7 +45,6 @@ memfd_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
         return;
     }
 
-    backend->force_prealloc = mem_prealloc;
     fd = qemu_memfd_create(TYPE_MEMORY_BACKEND_MEMFD, backend->size,
                            m->hugetlb, m->hugetlbsize, m->seal ?
                            F_SEAL_GROW | F_SEAL_SHRINK | F_SEAL_SEAL : 0,

backends/hostmem-ram.c

@@ -16,8 +16,6 @@
 #include "qemu/module.h"
 #include "qom/object_interfaces.h"
 
-#define TYPE_MEMORY_BACKEND_RAM "memory-backend-ram"
-
 static void
 ram_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
 {

backends/hostmem.c

@@ -215,23 +215,14 @@ static bool host_memory_backend_get_prealloc(Object *obj, Error **errp)
 {
     HostMemoryBackend *backend = MEMORY_BACKEND(obj);
 
-    return backend->prealloc || backend->force_prealloc;
+    return backend->prealloc;
 }
 
 static void host_memory_backend_set_prealloc(Object *obj, bool value,
                                              Error **errp)
 {
     Error *local_err = NULL;
     HostMemoryBackend *backend = MEMORY_BACKEND(obj);
-    MachineState *ms = MACHINE(qdev_get_machine());
-
-    if (backend->force_prealloc) {
-        if (value) {
-            error_setg(errp,
-                       "remove -mem-prealloc to use the prealloc property");
-            return;
-        }
-    }
 
     if (!host_memory_backend_mr_inited(backend)) {
         backend->prealloc = value;
@@ -243,7 +234,7 @@ static void host_memory_backend_set_prealloc(Object *obj, bool value,
         void *ptr = memory_region_get_ram_ptr(&backend->mr);
         uint64_t sz = memory_region_size(&backend->mr);
 
-        os_mem_prealloc(fd, ptr, sz, ms->smp.cpus, &local_err);
+        os_mem_prealloc(fd, ptr, sz, backend->prealloc_threads, &local_err);
         if (local_err) {
             error_propagate(errp, local_err);
             return;
@@ -252,14 +243,43 @@ static void host_memory_backend_set_prealloc(Object *obj, bool value,
     }
 }
 
+static void host_memory_backend_get_prealloc_threads(Object *obj, Visitor *v,
+    const char *name, void *opaque, Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+    visit_type_uint32(v, name, &backend->prealloc_threads, errp);
+}
+
+static void host_memory_backend_set_prealloc_threads(Object *obj, Visitor *v,
+    const char *name, void *opaque, Error **errp)
+{
+    HostMemoryBackend *backend = MEMORY_BACKEND(obj);
+    Error *local_err = NULL;
+    uint32_t value;
+
+    visit_type_uint32(v, name, &value, &local_err);
+    if (local_err) {
+        goto out;
+    }
+    if (value <= 0) {
+        error_setg(&local_err,
+                   "property '%s' of %s doesn't take value '%d'",
+                   name, object_get_typename(obj), value);
+        goto out;
+    }
+    backend->prealloc_threads = value;
+out:
+    error_propagate(errp, local_err);
+}
+
 static void host_memory_backend_init(Object *obj)
 {
     HostMemoryBackend *backend = MEMORY_BACKEND(obj);
     MachineState *machine = MACHINE(qdev_get_machine());
 
+    /* TODO: convert access to globals to compat properties */
     backend->merge = machine_mem_merge(machine);
     backend->dump = machine_dump_guest_core(machine);
-    backend->prealloc = mem_prealloc;
 }
 
 static void host_memory_backend_post_init(Object *obj)
@@ -313,7 +333,6 @@ host_memory_backend_memory_complete(UserCreatable *uc, Error **errp)
 {
     HostMemoryBackend *backend = MEMORY_BACKEND(uc);
     HostMemoryBackendClass *bc = MEMORY_BACKEND_GET_CLASS(uc);
-    MachineState *ms = MACHINE(qdev_get_machine());
     Error *local_err = NULL;
     void *ptr;
     uint64_t sz;
@@ -378,7 +397,7 @@ host_memory_backend_memory_complete(UserCreatable *uc, Error **errp)
          */
         if (backend->prealloc) {
             os_mem_prealloc(memory_region_get_fd(&backend->mr), ptr, sz,
-                            ms->smp.cpus, &local_err);
+                            backend->prealloc_threads, &local_err);
             if (local_err) {
                 goto out;
             }
@@ -456,6 +475,12 @@ host_memory_backend_class_init(ObjectClass *oc, void *data)
         host_memory_backend_set_prealloc, &error_abort);
     object_class_property_set_description(oc, "prealloc",
         "Preallocate memory", &error_abort);
+    object_class_property_add(oc, "prealloc-threads", "int",
+        host_memory_backend_get_prealloc_threads,
+        host_memory_backend_set_prealloc_threads,
+        NULL, NULL, &error_abort);
+    object_class_property_set_description(oc, "prealloc-threads",
+        "Number of CPU threads to use for prealloc", &error_abort);
     object_class_property_add(oc, "size", "int",
         host_memory_backend_get_size,
         host_memory_backend_set_size,

exec.c

@@ -1668,59 +1668,18 @@ static int find_max_backend_pagesize(Object *obj, void *opaque)
 long qemu_minrampagesize(void)
 {
     long hpsize = LONG_MAX;
-    long mainrampagesize;
-    Object *memdev_root;
-    MachineState *ms = MACHINE(qdev_get_machine());
-
-    mainrampagesize = qemu_mempath_getpagesize(mem_path);
-
-    /* it's possible we have memory-backend objects with
-     * hugepage-backed RAM. these may get mapped into system
-     * address space via -numa parameters or memory hotplug
-     * hooks. we want to take these into account, but we
-     * also want to make sure these supported hugepage
-     * sizes are applicable across the entire range of memory
-     * we may boot from, so we take the min across all
-     * backends, and assume normal pages in cases where a
-     * backend isn't backed by hugepages.
-     */
-    memdev_root = object_resolve_path("/objects", NULL);
-    if (memdev_root) {
-        object_child_foreach(memdev_root, find_min_backend_pagesize, &hpsize);
-    }
-    if (hpsize == LONG_MAX) {
-        /* No additional memory regions found ==> Report main RAM page size */
-        return mainrampagesize;
-    }
-
-    /* If NUMA is disabled or the NUMA nodes are not backed with a
-     * memory-backend, then there is at least one node using "normal" RAM,
-     * so if its page size is smaller we have got to report that size instead.
-     */
-    if (hpsize > mainrampagesize &&
-        (ms->numa_state == NULL ||
-         ms->numa_state->num_nodes == 0 ||
-         ms->numa_state->nodes[0].node_memdev == NULL)) {
-        static bool warned;
-        if (!warned) {
-            error_report("Huge page support disabled (n/a for main memory).");
-            warned = true;
-        }
-        return mainrampagesize;
-    }
+    Object *memdev_root = object_resolve_path("/objects", NULL);
 
+    object_child_foreach(memdev_root, find_min_backend_pagesize, &hpsize);
     return hpsize;
 }
 
 long qemu_maxrampagesize(void)
 {
-    long pagesize = qemu_mempath_getpagesize(mem_path);
+    long pagesize = 0;
     Object *memdev_root = object_resolve_path("/objects", NULL);
 
-    if (memdev_root) {
-        object_child_foreach(memdev_root, find_max_backend_pagesize,
-                             &pagesize);
-    }
+    object_child_foreach(memdev_root, find_max_backend_pagesize, &pagesize);
     return pagesize;
 }
 #else
@@ -1843,8 +1802,6 @@ static void *file_ram_alloc(RAMBlock *block,
                             bool truncate,
                             Error **errp)
 {
-    Error *err = NULL;
-    MachineState *ms = MACHINE(qdev_get_machine());
     void *area;
 
     block->page_size = qemu_fd_getpagesize(fd);
@@ -1900,15 +1857,6 @@ static void *file_ram_alloc(RAMBlock *block,
         return NULL;
     }
 
-    if (mem_prealloc) {
-        os_mem_prealloc(fd, area, memory, ms->smp.cpus, &err);
-        if (err) {
-            error_propagate(errp, err);
-            qemu_ram_munmap(fd, area, memory);
-            return NULL;
-        }
-    }
-
     block->fd = fd;
     return area;
 }
@@ -2356,9 +2304,9 @@ RAMBlock *qemu_ram_alloc_from_fd(ram_addr_t size, MemoryRegion *mr,
     size = HOST_PAGE_ALIGN(size);
     file_size = get_file_size(fd);
     if (file_size > 0 && file_size < size) {
-        error_setg(errp, "backing store %s size 0x%" PRIx64
+        error_setg(errp, "backing store size 0x%" PRIx64
                    " does not match 'size' option 0x" RAM_ADDR_FMT,
-                   mem_path, file_size, size);
+                   file_size, size);
         return NULL;
     }
 

hw/alpha/alpha_sys.h

@@ -11,7 +11,7 @@
 #include "hw/intc/i8259.h"
 
 
-PCIBus *typhoon_init(ram_addr_t, ISABus **, qemu_irq *, AlphaCPU *[4],
+PCIBus *typhoon_init(MemoryRegion *, ISABus **, qemu_irq *, AlphaCPU *[4],
                      pci_map_irq_fn);
 
 /* alpha_pci.c.  */

hw/alpha/dp264.c

@@ -75,7 +75,7 @@ static void clipper_init(MachineState *machine)
     cpus[0]->env.trap_arg2 = smp_cpus;
 
     /* Init the chipset.  */
-    pci_bus = typhoon_init(ram_size, &isa_bus, &rtc_irq, cpus,
+    pci_bus = typhoon_init(machine->ram, &isa_bus, &rtc_irq, cpus,
                            clipper_pci_map_irq);
 
     /* Since we have an SRM-compatible PALcode, use the SRM epoch.  */
@@ -183,6 +183,7 @@ static void clipper_machine_init(MachineClass *mc)
     mc->max_cpus = 4;
     mc->is_default = 1;
     mc->default_cpu_type = ALPHA_CPU_TYPE_NAME("ev67");
+    mc->default_ram_id = "ram";
 }
 
 DEFINE_MACHINE("clipper", clipper_machine_init)

hw/alpha/typhoon.c

@@ -58,7 +58,6 @@ typedef struct TyphoonState {
     TyphoonCchip cchip;
     TyphoonPchip pchip;
     MemoryRegion dchip_region;
-    MemoryRegion ram_region;
 } TyphoonState;
 
 /* Called when one of DRIR or DIM changes.  */
@@ -817,8 +816,7 @@ static void typhoon_alarm_timer(void *opaque)
     cpu_interrupt(CPU(s->cchip.cpu[cpu]), CPU_INTERRUPT_TIMER);
 }
 
-PCIBus *typhoon_init(ram_addr_t ram_size, ISABus **isa_bus,
-                     qemu_irq *p_rtc_irq,
+PCIBus *typhoon_init(MemoryRegion *ram, ISABus **isa_bus, qemu_irq *p_rtc_irq,
                      AlphaCPU *cpus[4], pci_map_irq_fn sys_map_irq)
 {
     MemoryRegion *addr_space = get_system_memory();
@@ -851,9 +849,7 @@ PCIBus *typhoon_init(ram_addr_t ram_size, ISABus **isa_bus,
 
     /* Main memory region, 0x00.0000.0000.  Real hardware supports 32GB,
        but the address space hole reserved at this point is 8TB.  */
-    memory_region_allocate_system_memory(&s->ram_region, OBJECT(s), "ram",
-                                         ram_size);
-    memory_region_add_subregion(addr_space, 0, &s->ram_region);
+    memory_region_add_subregion(addr_space, 0, ram);
 
     /* TIGbus, 0x801.0000.0000, 1GB.  */
     /* ??? The TIGbus is used for delivering interrupts, and access to

hw/arm/aspeed.c

@@ -35,7 +35,6 @@ static struct arm_boot_info aspeed_board_binfo = {
 struct AspeedBoardState {
     AspeedSoCState soc;
     MemoryRegion ram_container;
-    MemoryRegion ram;
     MemoryRegion max_ram;
 };
 
@@ -197,15 +196,20 @@ static void aspeed_machine_init(MachineState *machine)
 
     memory_region_init(&bmc->ram_container, NULL, "aspeed-ram-container",
                        UINT32_MAX);
+    memory_region_add_subregion(&bmc->ram_container, 0, machine->ram);
 
     object_initialize_child(OBJECT(machine), "soc", &bmc->soc,
                             (sizeof(bmc->soc)), amc->soc_name, &error_abort,
                             NULL);
 
     sc = ASPEED_SOC_GET_CLASS(&bmc->soc);
 
+    /*
+     * This will error out if isize is not supported by memory controller.
+     */
     object_property_set_uint(OBJECT(&bmc->soc), ram_size, "ram-size",
-                             &error_abort);
+                             &error_fatal);
+
     object_property_set_int(OBJECT(&bmc->soc), amc->hw_strap1, "hw-strap1",
                             &error_abort);
     object_property_set_int(OBJECT(&bmc->soc), amc->hw_strap2, "hw-strap2",
@@ -228,15 +232,6 @@ static void aspeed_machine_init(MachineState *machine)
     object_property_set_bool(OBJECT(&bmc->soc), true, "realized",
                              &error_abort);
 
-    /*
-     * Allocate RAM after the memory controller has checked the size
-     * was valid. If not, a default value is used.
-     */
-    ram_size = object_property_get_uint(OBJECT(&bmc->soc), "ram-size",
-                                        &error_abort);
-
-    memory_region_allocate_system_memory(&bmc->ram, NULL, "ram", ram_size);
-    memory_region_add_subregion(&bmc->ram_container, 0, &bmc->ram);
     memory_region_add_subregion(get_system_memory(),
                                 sc->memmap[ASPEED_SDRAM],
                                 &bmc->ram_container);
@@ -439,6 +434,7 @@ static void aspeed_machine_class_init(ObjectClass *oc, void *data)
     mc->no_floppy = 1;
     mc->no_cdrom = 1;
     mc->no_parallel = 1;
+    mc->default_ram_id = "ram";
 
     aspeed_machine_class_props_init(oc);
 }

hw/arm/collie.c

@@ -10,6 +10,7 @@
  */
 #include "qemu/osdep.h"
 #include "qemu/units.h"
+#include "qemu/cutils.h"
 #include "hw/sysbus.h"
 #include "hw/boards.h"
 #include "strongarm.h"
@@ -27,13 +28,18 @@ static void collie_init(MachineState *machine)
 {
     StrongARMState *s;
     DriveInfo *dinfo;
-    MemoryRegion *sdram = g_new(MemoryRegion, 1);
+    MachineClass *mc = MACHINE_GET_CLASS(machine);
+
+    if (machine->ram_size != mc->default_ram_size) {
+        char *sz = size_to_str(mc->default_ram_size);
+        error_report("Invalid RAM size, should be %s", sz);
+        g_free(sz);
+        exit(EXIT_FAILURE);
+    }
 
     s = sa1110_init(machine->cpu_type);
 
-    memory_region_allocate_system_memory(sdram, NULL, "strongarm.sdram",
-                                         collie_binfo.ram_size);
-    memory_region_add_subregion(get_system_memory(), SA_SDCS0, sdram);
+    memory_region_add_subregion(get_system_memory(), SA_SDCS0, machine->ram);
 
     dinfo = drive_get(IF_PFLASH, 0, 0);
     pflash_cfi01_register(SA_CS0, "collie.fl1", 0x02000000,
@@ -57,6 +63,8 @@ static void collie_machine_init(MachineClass *mc)
     mc->init = collie_init;
     mc->ignore_memory_transaction_failures = true;
     mc->default_cpu_type = ARM_CPU_TYPE_NAME("sa1110");
+    mc->default_ram_size = 0x20000000;
+    mc->default_ram_id = "strongarm.sdram";
 }
 
 DEFINE_MACHINE("collie", collie_machine_init)