Re: [Qemu-devel] [RFC 26/29] migration: synchronize dirty bitmap for resume

[Dr. David Alan Gilbert]

* Peter Xu (address@hidden) wrote:
> This patch implements the first part of core RAM resume logic for
> postcopy. ram_resume_prepare() is provided for the work.
> 
> When the migration is interrupted by network failure, the dirty bitmap
> on the source side will be meaningless, because even the dirty bit is
> cleared, it is still possible that the sent page was lost along the way
> to destination. Here instead of continue the migration with the old
> dirty bitmap on source, we ask the destination side to send back its
> received bitmap, then invert it to be our initial dirty bitmap.
> 
> The source side send thread will issue the MIG_CMD_RECV_BITMAP requests,
> once per ramblock, to ask for the received bitmap. On destination side,
> MIG_RP_MSG_RECV_BITMAP will be issued, along with the requested bitmap.
> Data will be received on the return-path thread of source, and the main
> migration thread will be notified when all the ramblock bitmaps are
> synchronized.
> 
> One issue to be solved here is how to synchronize the source send thread
> and return-path thread. Semaphore cannot really work here since we
> cannot guarantee the order of wait/post (it's possible that the reply is
> very fast, even before send thread starts to wait). So conditional
> variable is used to make sure the ordering is always correct.
> 
> Signed-off-by: Peter Xu <address@hidden>
> ---
>  migration/migration.c  |  4 +++
>  migration/migration.h  |  4 +++
>  migration/ram.c        | 68 
> ++++++++++++++++++++++++++++++++++++++++++++++++++
>  migration/trace-events |  1 +
>  4 files changed, 77 insertions(+)
> 
> diff --git a/migration/migration.c b/migration/migration.c
> index 6cb0ad3..93fbc96 100644
> --- a/migration/migration.c
> +++ b/migration/migration.c
> @@ -1093,6 +1093,8 @@ static void migrate_fd_cleanup(void *opaque)
>  
>      qemu_sem_destroy(&s->postcopy_pause_sem);
>      qemu_sem_destroy(&s->postcopy_pause_rp_sem);
> +    qemu_mutex_destroy(&s->resume_lock);
> +    qemu_cond_destroy(&s->resume_cond);
>  }
>  
>  void migrate_fd_error(MigrationState *s, const Error *error)
> @@ -1238,6 +1240,8 @@ MigrationState *migrate_init(void)
>      s->error = NULL;
>      qemu_sem_init(&s->postcopy_pause_sem, 0);
>      qemu_sem_init(&s->postcopy_pause_rp_sem, 0);
> +    qemu_mutex_init(&s->resume_lock);
> +    qemu_cond_init(&s->resume_cond);
>  
>      migrate_set_state(&s->state, MIGRATION_STATUS_NONE, 
> MIGRATION_STATUS_SETUP);
>  
> diff --git a/migration/migration.h b/migration/migration.h
> index 2a3f905..c270f4c 100644
> --- a/migration/migration.h
> +++ b/migration/migration.h
> @@ -159,6 +159,10 @@ struct MigrationState
>      /* Needed by postcopy-pause state */
>      QemuSemaphore postcopy_pause_sem;
>      QemuSemaphore postcopy_pause_rp_sem;
> +
> +    /* Used to sync-up between main send thread and rp-thread */
> +    QemuMutex resume_lock;
> +    QemuCond resume_cond;
>  };
>  
>  void migrate_set_state(int *state, int old_state, int new_state);
> diff --git a/migration/ram.c b/migration/ram.c
> index d543483..c695b13 100644
> --- a/migration/ram.c
> +++ b/migration/ram.c
> @@ -46,6 +46,7 @@
>  #include "exec/ram_addr.h"
>  #include "qemu/rcu_queue.h"
>  #include "migration/colo.h"
> +#include "savevm.h"
>  
>  /***********************************************************/
>  /* ram save/restore */
> @@ -256,6 +257,8 @@ struct RAMState {
>      RAMBlock *last_req_rb;
>      /* Queue of outstanding page requests from the destination */
>      QemuMutex src_page_req_mutex;
> +    /* Ramblock counts to sync dirty bitmap. Only used for recovery */
> +    int ramblock_to_sync;
>      QSIMPLEQ_HEAD(src_page_requests, RAMSrcPageRequest) src_page_requests;
>  };
>  typedef struct RAMState RAMState;
> @@ -2731,6 +2734,57 @@ static int ram_load(QEMUFile *f, void *opaque, int 
> version_id)
>      return ret;
>  }
>  
> +/* Sync all the dirty bitmap with destination VM.  */
> +static int ram_dirty_bitmap_sync_all(MigrationState *s, RAMState *rs)
> +{
> +    RAMBlock *block;
> +    QEMUFile *file = s->to_dst_file;
> +    int ramblock_count = 0;
> +
> +    trace_ram_dirty_bitmap_sync("start");

Most (but not all) of our trace_ uses have separate trace_ entries for
each step; e.g.   trace_ram_dirty_bitmap_sync_start9)

> +    /*
> +     * We need to take the resume lock to make sure that the send
> +     * thread (current thread) and the rp-thread will do their work in
> +     * order.
> +     */
> +    qemu_mutex_lock(&s->resume_lock);
> +
> +    /* Request for receive-bitmap for each block */
> +    RAMBLOCK_FOREACH(block) {
> +        ramblock_count++;
> +        qemu_savevm_send_recv_bitmap(file, block->idstr);
> +    }
> +
> +    /* Init the ramblock count to total */
> +    atomic_set(&rs->ramblock_to_sync, ramblock_count);
> +
> +    trace_ram_dirty_bitmap_sync("wait-bitmap");
> +
> +    /* Wait until all the ramblocks' dirty bitmap synced */
> +    while (rs->ramblock_to_sync) {
> +        qemu_cond_wait(&s->resume_cond, &s->resume_lock);
> +    }

Does the locking here get simpler if you:
  a) count the number of RAMBlocks 'n'
  b) Initialise a sempahore to -(n-1)
  c) Call qemu_savevm_send_recv_bitmap for each bitmap
  d) sem_wait on the semaphore - which is waiting for the semaphore to
     be >0

as you receive each bitmap do a sem_post; on the last one
it should go from 0->1 and the sem_wait should wake up?

Dave

> +    trace_ram_dirty_bitmap_sync("completed");
> +
> +    qemu_mutex_unlock(&s->resume_lock);
> +
> +    return 0;
> +}
> +
> +static void ram_dirty_bitmap_reload_notify(MigrationState *s)
> +{
> +    qemu_mutex_lock(&s->resume_lock);
> +    atomic_dec(&ram_state->ramblock_to_sync);
> +    if (ram_state->ramblock_to_sync == 0) {
> +        /* Make sure the other thread gets the latest */
> +        trace_ram_dirty_bitmap_sync("notify-send");
> +        qemu_cond_signal(&s->resume_cond);
> +    }
> +    qemu_mutex_unlock(&s->resume_lock);
> +}
> +
>  /*
>   * Read the received bitmap, revert it as the initial dirty bitmap.
>   * This is only used when the postcopy migration is paused but wants
> @@ -2776,9 +2830,22 @@ int ram_dirty_bitmap_reload(MigrationState *s, 
> RAMBlock *block)
>  
>      trace_ram_dirty_bitmap_reload(block->idstr);
>  
> +    /*
> +     * We succeeded to sync bitmap for current ramblock. If this is
> +     * the last one to sync, we need to notify the main send thread.
> +     */
> +    ram_dirty_bitmap_reload_notify(s);
> +
>      return 0;
>  }
>  
> +static int ram_resume_prepare(MigrationState *s, void *opaque)
> +{
> +    RAMState *rs = *(RAMState **)opaque;
> +
> +    return ram_dirty_bitmap_sync_all(s, rs);
> +}
> +
>  static SaveVMHandlers savevm_ram_handlers = {
>      .save_setup = ram_save_setup,
>      .save_live_iterate = ram_save_iterate,
> @@ -2789,6 +2856,7 @@ static SaveVMHandlers savevm_ram_handlers = {
>      .save_cleanup = ram_save_cleanup,
>      .load_setup = ram_load_setup,
>      .load_cleanup = ram_load_cleanup,
> +    .resume_prepare = ram_resume_prepare,
>  };
>  
>  void ram_mig_init(void)
> diff --git a/migration/trace-events b/migration/trace-events
> index 0fb2d1e..15ff1bf 100644
> --- a/migration/trace-events
> +++ b/migration/trace-events
> @@ -80,6 +80,7 @@ ram_postcopy_send_discard_bitmap(void) ""
>  ram_save_page(const char *rbname, uint64_t offset, void *host) "%s: offset: 
> %" PRIx64 " host: %p"
>  ram_save_queue_pages(const char *rbname, size_t start, size_t len) "%s: 
> start: %zx len: %zx"
>  ram_dirty_bitmap_reload(char *str) "%s"
> +ram_dirty_bitmap_sync(const char *str) "%s"
>  
>  # migration/migration.c
>  await_return_path_close_on_source_close(void) ""
> -- 
> 2.7.4
> 
--
Dr. David Alan Gilbert / address@hidden / Manchester, UK