diff options
author | Anthony Liguori <aliguori@us.ibm.com> | 2012-07-30 09:58:41 -0500 |
---|---|---|
committer | Anthony Liguori <aliguori@us.ibm.com> | 2012-07-30 09:58:48 -0500 |
commit | e6a76719987e5fcd63da552f7cf32d837b0a5cea (patch) | |
tree | 7e23d3f0f664c8f11613140c996ee4ca9cf6d117 | |
parent | 7162ab21fe8e82f924002951cd8e87f69358f8b5 (diff) | |
parent | 6c779f22a93cc6e4565b940ef616e3efc5b50ba5 (diff) | |
download | hqemu-e6a76719987e5fcd63da552f7cf32d837b0a5cea.zip hqemu-e6a76719987e5fcd63da552f7cf32d837b0a5cea.tar.gz |
Merge commit 'quintela/migration-next-v5' into staging
* commit '6c779f22a93cc6e4565b940ef616e3efc5b50ba5':
Change ram_save_block to return -1 if there are no more changes
ram: save_live_setup() we don't need to synchronize the dirty bitmap.
ram: iterate phase
ram: save_live_complete() only do one loop
ram: save_live_setup() don't need to sent pages
savevm: split save_live into stage2 and stage3
savevm: split save_live_setup from save_live_state
savevm: introduce is_active method
savevm: Refactor cancel operation in its own operation
savevm: remove SaveLiveStateHandler
savevm: remove SaveSetParamsHandler
savevm: Live migration handlers register the struct directly
savevm: Use a struct to pass all handlers
-rw-r--r-- | arch_init.c | 137 | ||||
-rw-r--r-- | block-migration.c | 153 | ||||
-rw-r--r-- | migration.h | 4 | ||||
-rw-r--r-- | savevm.c | 77 | ||||
-rw-r--r-- | vl.c | 3 | ||||
-rw-r--r-- | vmstate.h | 18 |
6 files changed, 245 insertions, 147 deletions
diff --git a/arch_init.c b/arch_init.c index ebdad3a..26f30ef 100644 --- a/arch_init.c +++ b/arch_init.c @@ -186,11 +186,19 @@ static void save_block_hdr(QEMUFile *f, RAMBlock *block, ram_addr_t offset, static RAMBlock *last_block; static ram_addr_t last_offset; +/* + * ram_save_block: Writes a page of memory to the stream f + * + * Returns: 0: if the page hasn't changed + * -1: if there are no more dirty pages + * n: the amount of bytes written in other case + */ + static int ram_save_block(QEMUFile *f) { RAMBlock *block = last_block; ram_addr_t offset = last_offset; - int bytes_sent = 0; + int bytes_sent = -1; MemoryRegion *mr; if (!block) @@ -298,51 +306,56 @@ static void migration_end(void) memory_global_dirty_log_stop(); } +static void ram_migration_cancel(void *opaque) +{ + migration_end(); +} + #define MAX_WAIT 50 /* ms, half buffered_file limit */ -int ram_save_live(QEMUFile *f, int stage, void *opaque) +static int ram_save_setup(QEMUFile *f, void *opaque) { ram_addr_t addr; - uint64_t bytes_transferred_last; - double bwidth = 0; - int ret; - int i; - - if (stage < 0) { - migration_end(); - return 0; - } + RAMBlock *block; - memory_global_sync_dirty_bitmap(get_system_memory()); + bytes_transferred = 0; + last_block = NULL; + last_offset = 0; + sort_ram_list(); - if (stage == 1) { - RAMBlock *block; - bytes_transferred = 0; - last_block = NULL; - last_offset = 0; - sort_ram_list(); - - /* Make sure all dirty bits are set */ - QLIST_FOREACH(block, &ram_list.blocks, next) { - for (addr = 0; addr < block->length; addr += TARGET_PAGE_SIZE) { - if (!memory_region_get_dirty(block->mr, addr, TARGET_PAGE_SIZE, - DIRTY_MEMORY_MIGRATION)) { - memory_region_set_dirty(block->mr, addr, TARGET_PAGE_SIZE); - } + /* Make sure all dirty bits are set */ + QLIST_FOREACH(block, &ram_list.blocks, next) { + for (addr = 0; addr < block->length; addr += TARGET_PAGE_SIZE) { + if (!memory_region_get_dirty(block->mr, addr, TARGET_PAGE_SIZE, + DIRTY_MEMORY_MIGRATION)) { + memory_region_set_dirty(block->mr, addr, TARGET_PAGE_SIZE); } } + } - memory_global_dirty_log_start(); + memory_global_dirty_log_start(); - qemu_put_be64(f, ram_bytes_total() | RAM_SAVE_FLAG_MEM_SIZE); + qemu_put_be64(f, ram_bytes_total() | RAM_SAVE_FLAG_MEM_SIZE); - QLIST_FOREACH(block, &ram_list.blocks, next) { - qemu_put_byte(f, strlen(block->idstr)); - qemu_put_buffer(f, (uint8_t *)block->idstr, strlen(block->idstr)); - qemu_put_be64(f, block->length); - } + QLIST_FOREACH(block, &ram_list.blocks, next) { + qemu_put_byte(f, strlen(block->idstr)); + qemu_put_buffer(f, (uint8_t *)block->idstr, strlen(block->idstr)); + qemu_put_be64(f, block->length); } + qemu_put_be64(f, RAM_SAVE_FLAG_EOS); + + return 0; +} + +static int ram_save_iterate(QEMUFile *f, void *opaque) +{ + uint64_t bytes_transferred_last; + double bwidth = 0; + int ret; + int i; + uint64_t expected_time; + bytes_transferred_last = bytes_transferred; bwidth = qemu_get_clock_ns(rt_clock); @@ -351,10 +364,11 @@ int ram_save_live(QEMUFile *f, int stage, void *opaque) int bytes_sent; bytes_sent = ram_save_block(f); - bytes_transferred += bytes_sent; - if (bytes_sent == 0) { /* no more blocks */ + /* no more blocks to sent */ + if (bytes_sent < 0) { break; } + bytes_transferred += bytes_sent; /* we want to check in the 1st loop, just in case it was the 1st time and we had to sync the dirty bitmap. qemu_get_clock_ns() is a bit expensive, so we only check each some @@ -384,28 +398,43 @@ int ram_save_live(QEMUFile *f, int stage, void *opaque) bwidth = 0.000001; } + qemu_put_be64(f, RAM_SAVE_FLAG_EOS); + + expected_time = ram_save_remaining() * TARGET_PAGE_SIZE / bwidth; + + DPRINTF("ram_save_live: expected(" PRIu64 ") <= max(" PRIu64 ")?\n", + expected_time, migrate_max_downtime()); + + if (expected_time <= migrate_max_downtime()) { + memory_global_sync_dirty_bitmap(get_system_memory()); + expected_time = ram_save_remaining() * TARGET_PAGE_SIZE / bwidth; + + return expected_time <= migrate_max_downtime(); + } + return 0; +} + +static int ram_save_complete(QEMUFile *f, void *opaque) +{ + memory_global_sync_dirty_bitmap(get_system_memory()); + /* try transferring iterative blocks of memory */ - if (stage == 3) { + + /* flush all remaining blocks regardless of rate limiting */ + while (true) { int bytes_sent; - /* flush all remaining blocks regardless of rate limiting */ - while ((bytes_sent = ram_save_block(f)) != 0) { - bytes_transferred += bytes_sent; + bytes_sent = ram_save_block(f); + /* no more blocks to sent */ + if (bytes_sent < 0) { + break; } - memory_global_dirty_log_stop(); + bytes_transferred += bytes_sent; } + memory_global_dirty_log_stop(); qemu_put_be64(f, RAM_SAVE_FLAG_EOS); - if (stage == 2) { - uint64_t expected_time; - expected_time = ram_save_remaining() * TARGET_PAGE_SIZE / bwidth; - - DPRINTF("ram_save_live: expected(" PRIu64 ") <= max(" PRIu64 ")?\n", - expected_time, migrate_max_downtime()); - - return expected_time <= migrate_max_downtime(); - } return 0; } @@ -439,7 +468,7 @@ static inline void *host_from_stream_offset(QEMUFile *f, return NULL; } -int ram_load(QEMUFile *f, void *opaque, int version_id) +static int ram_load(QEMUFile *f, void *opaque, int version_id) { ram_addr_t addr; int flags, ret = 0; @@ -536,6 +565,14 @@ done: return ret; } +SaveVMHandlers savevm_ram_handlers = { + .save_live_setup = ram_save_setup, + .save_live_iterate = ram_save_iterate, + .save_live_complete = ram_save_complete, + .load_state = ram_load, + .cancel = ram_migration_cancel, +}; + #ifdef HAS_AUDIO struct soundhw { const char *name; diff --git a/block-migration.c b/block-migration.c index b95b4e1..7def8ab 100644 --- a/block-migration.c +++ b/block-migration.c @@ -536,30 +536,44 @@ static void blk_mig_cleanup(void) } } -static int block_save_live(QEMUFile *f, int stage, void *opaque) +static void block_migration_cancel(void *opaque) +{ + blk_mig_cleanup(); +} + +static int block_save_setup(QEMUFile *f, void *opaque) { int ret; - DPRINTF("Enter save live stage %d submitted %d transferred %d\n", - stage, block_mig_state.submitted, block_mig_state.transferred); + DPRINTF("Enter save live setup submitted %d transferred %d\n", + block_mig_state.submitted, block_mig_state.transferred); + + init_blk_migration(f); + + /* start track dirty blocks */ + set_dirty_tracking(1); - if (stage < 0) { + flush_blks(f); + + ret = qemu_file_get_error(f); + if (ret) { blk_mig_cleanup(); - return 0; + return ret; } - if (block_mig_state.blk_enable != 1) { - /* no need to migrate storage */ - qemu_put_be64(f, BLK_MIG_FLAG_EOS); - return 1; - } + blk_mig_reset_dirty_cursor(); - if (stage == 1) { - init_blk_migration(f); + qemu_put_be64(f, BLK_MIG_FLAG_EOS); - /* start track dirty blocks */ - set_dirty_tracking(1); - } + return 0; +} + +static int block_save_iterate(QEMUFile *f, void *opaque) +{ + int ret; + + DPRINTF("Enter save live iterate submitted %d transferred %d\n", + block_mig_state.submitted, block_mig_state.transferred); flush_blks(f); @@ -571,56 +585,76 @@ static int block_save_live(QEMUFile *f, int stage, void *opaque) blk_mig_reset_dirty_cursor(); - if (stage == 2) { - /* control the rate of transfer */ - while ((block_mig_state.submitted + - block_mig_state.read_done) * BLOCK_SIZE < - qemu_file_get_rate_limit(f)) { - if (block_mig_state.bulk_completed == 0) { - /* first finish the bulk phase */ - if (blk_mig_save_bulked_block(f) == 0) { - /* finished saving bulk on all devices */ - block_mig_state.bulk_completed = 1; - } - } else { - if (blk_mig_save_dirty_block(f, 1) == 0) { - /* no more dirty blocks */ - break; - } + /* control the rate of transfer */ + while ((block_mig_state.submitted + + block_mig_state.read_done) * BLOCK_SIZE < + qemu_file_get_rate_limit(f)) { + if (block_mig_state.bulk_completed == 0) { + /* first finish the bulk phase */ + if (blk_mig_save_bulked_block(f) == 0) { + /* finished saving bulk on all devices */ + block_mig_state.bulk_completed = 1; + } + } else { + if (blk_mig_save_dirty_block(f, 1) == 0) { + /* no more dirty blocks */ + break; } } + } - flush_blks(f); + flush_blks(f); - ret = qemu_file_get_error(f); - if (ret) { - blk_mig_cleanup(); - return ret; - } + ret = qemu_file_get_error(f); + if (ret) { + blk_mig_cleanup(); + return ret; } - if (stage == 3) { - /* we know for sure that save bulk is completed and - all async read completed */ - assert(block_mig_state.submitted == 0); + qemu_put_be64(f, BLK_MIG_FLAG_EOS); + + return is_stage2_completed(); +} + +static int block_save_complete(QEMUFile *f, void *opaque) +{ + int ret; + + DPRINTF("Enter save live complete submitted %d transferred %d\n", + block_mig_state.submitted, block_mig_state.transferred); + + flush_blks(f); - while (blk_mig_save_dirty_block(f, 0) != 0); + ret = qemu_file_get_error(f); + if (ret) { blk_mig_cleanup(); + return ret; + } - /* report completion */ - qemu_put_be64(f, (100 << BDRV_SECTOR_BITS) | BLK_MIG_FLAG_PROGRESS); + blk_mig_reset_dirty_cursor(); - ret = qemu_file_get_error(f); - if (ret) { - return ret; - } + /* we know for sure that save bulk is completed and + all async read completed */ + assert(block_mig_state.submitted == 0); - DPRINTF("Block migration completed\n"); + while (blk_mig_save_dirty_block(f, 0) != 0) { + /* Do nothing */ + } + blk_mig_cleanup(); + + /* report completion */ + qemu_put_be64(f, (100 << BDRV_SECTOR_BITS) | BLK_MIG_FLAG_PROGRESS); + + ret = qemu_file_get_error(f); + if (ret) { + return ret; } + DPRINTF("Block migration completed\n"); + qemu_put_be64(f, BLK_MIG_FLAG_EOS); - return ((stage == 2) && is_stage2_completed()); + return 0; } static int block_load(QEMUFile *f, void *opaque, int version_id) @@ -709,11 +743,26 @@ static void block_set_params(const MigrationParams *params, void *opaque) block_mig_state.blk_enable |= params->shared; } +static bool block_is_active(void *opaque) +{ + return block_mig_state.blk_enable == 1; +} + +SaveVMHandlers savevm_block_handlers = { + .set_params = block_set_params, + .save_live_setup = block_save_setup, + .save_live_iterate = block_save_iterate, + .save_live_complete = block_save_complete, + .load_state = block_load, + .cancel = block_migration_cancel, + .is_active = block_is_active, +}; + void blk_mig_init(void) { QSIMPLEQ_INIT(&block_mig_state.bmds_list); QSIMPLEQ_INIT(&block_mig_state.blk_list); - register_savevm_live(NULL, "block", 0, 1, block_set_params, - block_save_live, NULL, block_load, &block_mig_state); + register_savevm_live(NULL, "block", 0, 1, &savevm_block_handlers, + &block_mig_state); } diff --git a/migration.h b/migration.h index de13004..57572a6 100644 --- a/migration.h +++ b/migration.h @@ -18,6 +18,7 @@ #include "qemu-common.h" #include "notify.h" #include "error.h" +#include "vmstate.h" struct MigrationParams { bool blk; @@ -81,8 +82,7 @@ uint64_t ram_bytes_remaining(void); uint64_t ram_bytes_transferred(void); uint64_t ram_bytes_total(void); -int ram_save_live(QEMUFile *f, int stage, void *opaque); -int ram_load(QEMUFile *f, void *opaque, int version_id); +extern SaveVMHandlers savevm_ram_handlers; /** * @migrate_add_blocker - prevent migration from proceeding @@ -1171,10 +1171,7 @@ typedef struct SaveStateEntry { int alias_id; int version_id; int section_id; - SaveSetParamsHandler *set_params; - SaveLiveStateHandler *save_live_state; - SaveStateHandler *save_state; - LoadStateHandler *load_state; + SaveVMHandlers *ops; const VMStateDescription *vmsd; void *opaque; CompatEntry *compat; @@ -1226,10 +1223,7 @@ int register_savevm_live(DeviceState *dev, const char *idstr, int instance_id, int version_id, - SaveSetParamsHandler *set_params, - SaveLiveStateHandler *save_live_state, - SaveStateHandler *save_state, - LoadStateHandler *load_state, + SaveVMHandlers *ops, void *opaque) { SaveStateEntry *se; @@ -1237,15 +1231,12 @@ int register_savevm_live(DeviceState *dev, se = g_malloc0(sizeof(SaveStateEntry)); se->version_id = version_id; se->section_id = global_section_id++; - se->set_params = set_params; - se->save_live_state = save_live_state; - se->save_state = save_state; - se->load_state = load_state; + se->ops = ops; se->opaque = opaque; se->vmsd = NULL; se->no_migrate = 0; /* if this is a live_savem then set is_ram */ - if (save_live_state != NULL) { + if (ops->save_live_setup != NULL) { se->is_ram = 1; } @@ -1284,8 +1275,11 @@ int register_savevm(DeviceState *dev, LoadStateHandler *load_state, void *opaque) { + SaveVMHandlers *ops = g_malloc0(sizeof(SaveVMHandlers)); + ops->save_state = save_state; + ops->load_state = load_state; return register_savevm_live(dev, idstr, instance_id, version_id, - NULL, NULL, save_state, load_state, opaque); + ops, opaque); } void unregister_savevm(DeviceState *dev, const char *idstr, void *opaque) @@ -1309,6 +1303,7 @@ void unregister_savevm(DeviceState *dev, const char *idstr, void *opaque) if (se->compat) { g_free(se->compat); } + g_free(se->ops); g_free(se); } } @@ -1327,9 +1322,6 @@ int vmstate_register_with_alias_id(DeviceState *dev, int instance_id, se = g_malloc0(sizeof(SaveStateEntry)); se->version_id = vmsd->version_id; se->section_id = global_section_id++; - se->save_live_state = NULL; - se->save_state = NULL; - se->load_state = NULL; se->opaque = opaque; se->vmsd = vmsd; se->alias_id = alias_id; @@ -1524,7 +1516,7 @@ void vmstate_save_state(QEMUFile *f, const VMStateDescription *vmsd, static int vmstate_load(QEMUFile *f, SaveStateEntry *se, int version_id) { if (!se->vmsd) { /* Old style */ - return se->load_state(f, se->opaque, version_id); + return se->ops->load_state(f, se->opaque, version_id); } return vmstate_load_state(f, se->vmsd, se->opaque, version_id); } @@ -1532,7 +1524,7 @@ static int vmstate_load(QEMUFile *f, SaveStateEntry *se, int version_id) static void vmstate_save(QEMUFile *f, SaveStateEntry *se) { if (!se->vmsd) { /* Old style */ - se->save_state(f, se->opaque); + se->ops->save_state(f, se->opaque); return; } vmstate_save_state(f,se->vmsd, se->opaque); @@ -1569,10 +1561,10 @@ int qemu_savevm_state_begin(QEMUFile *f, int ret; QTAILQ_FOREACH(se, &savevm_handlers, entry) { - if(se->set_params == NULL) { + if (!se->ops || !se->ops->set_params) { continue; } - se->set_params(params, se->opaque); + se->ops->set_params(params, se->opaque); } qemu_put_be32(f, QEMU_VM_FILE_MAGIC); @@ -1581,9 +1573,14 @@ int qemu_savevm_state_begin(QEMUFile *f, QTAILQ_FOREACH(se, &savevm_handlers, entry) { int len; - if (se->save_live_state == NULL) + if (!se->ops || !se->ops->save_live_setup) { continue; - + } + if (se->ops && se->ops->is_active) { + if (!se->ops->is_active(se->opaque)) { + continue; + } + } /* Section type */ qemu_put_byte(f, QEMU_VM_SECTION_START); qemu_put_be32(f, se->section_id); @@ -1596,7 +1593,7 @@ int qemu_savevm_state_begin(QEMUFile *f, qemu_put_be32(f, se->instance_id); qemu_put_be32(f, se->version_id); - ret = se->save_live_state(f, QEMU_VM_SECTION_START, se->opaque); + ret = se->ops->save_live_setup(f, se->opaque); if (ret < 0) { qemu_savevm_state_cancel(f); return ret; @@ -1623,9 +1620,14 @@ int qemu_savevm_state_iterate(QEMUFile *f) int ret = 1; QTAILQ_FOREACH(se, &savevm_handlers, entry) { - if (se->save_live_state == NULL) + if (!se->ops || !se->ops->save_live_iterate) { continue; - + } + if (se->ops && se->ops->is_active) { + if (!se->ops->is_active(se->opaque)) { + continue; + } + } if (qemu_file_rate_limit(f)) { return 0; } @@ -1634,7 +1636,7 @@ int qemu_savevm_state_iterate(QEMUFile *f) qemu_put_byte(f, QEMU_VM_SECTION_PART); qemu_put_be32(f, se->section_id); - ret = se->save_live_state(f, QEMU_VM_SECTION_PART, se->opaque); + ret = se->ops->save_live_iterate(f, se->opaque); trace_savevm_section_end(se->section_id); if (ret <= 0) { @@ -1663,15 +1665,20 @@ int qemu_savevm_state_complete(QEMUFile *f) cpu_synchronize_all_states(); QTAILQ_FOREACH(se, &savevm_handlers, entry) { - if (se->save_live_state == NULL) + if (!se->ops || !se->ops->save_live_complete) { continue; - + } + if (se->ops && se->ops->is_active) { + if (!se->ops->is_active(se->opaque)) { + continue; + } + } trace_savevm_section_start(); /* Section type */ qemu_put_byte(f, QEMU_VM_SECTION_END); qemu_put_be32(f, se->section_id); - ret = se->save_live_state(f, QEMU_VM_SECTION_END, se->opaque); + ret = se->ops->save_live_complete(f, se->opaque); trace_savevm_section_end(se->section_id); if (ret < 0) { return ret; @@ -1681,9 +1688,9 @@ int qemu_savevm_state_complete(QEMUFile *f) QTAILQ_FOREACH(se, &savevm_handlers, entry) { int len; - if (se->save_state == NULL && se->vmsd == NULL) + if ((!se->ops || !se->ops->save_state) && !se->vmsd) { continue; - + } trace_savevm_section_start(); /* Section type */ qemu_put_byte(f, QEMU_VM_SECTION_FULL); @@ -1711,8 +1718,8 @@ void qemu_savevm_state_cancel(QEMUFile *f) SaveStateEntry *se; QTAILQ_FOREACH(se, &savevm_handlers, entry) { - if (se->save_live_state) { - se->save_live_state(f, -1, se->opaque); + if (se->ops && se->ops->cancel) { + se->ops->cancel(se->opaque); } } } @@ -1765,7 +1772,7 @@ static int qemu_save_device_state(QEMUFile *f) if (se->is_ram) { continue; } - if (se->save_state == NULL && se->vmsd == NULL) { + if ((!se->ops || !se->ops->save_state) && !se->vmsd) { continue; } @@ -3437,8 +3437,7 @@ int main(int argc, char **argv, char **envp) default_drive(default_sdcard, snapshot, machine->use_scsi, IF_SD, 0, SD_OPTS); - register_savevm_live(NULL, "ram", 0, 4, NULL, ram_save_live, NULL, - ram_load, NULL); + register_savevm_live(NULL, "ram", 0, 4, &savevm_ram_handlers, NULL); if (nb_numa_nodes > 0) { int i; @@ -26,11 +26,20 @@ #ifndef QEMU_VMSTATE_H #define QEMU_VMSTATE_H 1 -typedef void SaveSetParamsHandler(const MigrationParams *params, void * opaque); typedef void SaveStateHandler(QEMUFile *f, void *opaque); -typedef int SaveLiveStateHandler(QEMUFile *f, int stage, void *opaque); typedef int LoadStateHandler(QEMUFile *f, void *opaque, int version_id); +typedef struct SaveVMHandlers { + void (*set_params)(const MigrationParams *params, void * opaque); + SaveStateHandler *save_state; + int (*save_live_setup)(QEMUFile *f, void *opaque); + int (*save_live_iterate)(QEMUFile *f, void *opaque); + int (*save_live_complete)(QEMUFile *f, void *opaque); + void (*cancel)(void *opaque); + LoadStateHandler *load_state; + bool (*is_active)(void *opaque); +} SaveVMHandlers; + int register_savevm(DeviceState *dev, const char *idstr, int instance_id, @@ -43,10 +52,7 @@ int register_savevm_live(DeviceState *dev, const char *idstr, int instance_id, int version_id, - SaveSetParamsHandler *set_params, - SaveLiveStateHandler *save_live_state, - SaveStateHandler *save_state, - LoadStateHandler *load_state, + SaveVMHandlers *ops, void *opaque); void unregister_savevm(DeviceState *dev, const char *idstr, void *opaque); |