Merge remote-tracking branch 'remotes/amit-migration/tags/migration-for-2.6-5' into staging

migration pull
 - fix a qcow2 assert
 - fix for older distros (CentOS 5)
 - documentation for vmstate flags
 - minor code rearrangement

# gpg: Signature made Fri 26 Feb 2016 15:15:15 GMT using RSA key ID 854083B6
# gpg: Good signature from "Amit Shah <amit@amitshah.net>"
# gpg:                 aka "Amit Shah <amit@kernel.org>"
# gpg:                 aka "Amit Shah <amitshah@gmx.net>"

* remotes/amit-migration/tags/migration-for-2.6-5:
  migration (postcopy): move bdrv_invalidate_cache_all of of coroutine context
  migration (ordinary): move bdrv_invalidate_cache_all of of coroutine context
  migration/vmstate: document VMStateFlags
  MAINTAINERS: Add docs/migration.txt to the "Migration" section
  migration/postcopy-ram: Guard use of sys/eventfd.h with CONFIG_EVENTFD
  migration: reorder code to make it symmetric

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

6 files changed, 165 insertions, 65 deletions

diff --git a/MAINTAINERS b/MAINTAINERS
index 606d9c08b5..5fafa815ad 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1241,6 +1241,7 @@ F: include/migration/
 F: migration/
 F: scripts/vmstate-static-checker.py
 F: tests/vmstate-static-checker-data/
+F: docs/migration.txt
 
 Seccomp
 M: Eduardo Otubo <eduardo.otubo@profitbricks.com>
diff --git a/include/migration/migration.h b/include/migration/migration.h
index 85b6026d10..ac2c12c2a5 100644
--- a/include/migration/migration.h
+++ b/include/migration/migration.h
@@ -104,6 +104,8 @@ struct MigrationIncomingState {
     QemuMutex rp_mutex;    /* We send replies from multiple threads */
     void     *postcopy_tmp_page;
 
+    QEMUBH *bh;
+
     int state;
     /* See savevm.c */
     LoadStateEntry_Head loadvm_handlers;
diff --git a/include/migration/vmstate.h b/include/migration/vmstate.h
index 7246f29afe..84ee355ceb 100644
--- a/include/migration/vmstate.h
+++ b/include/migration/vmstate.h
@@ -88,21 +88,101 @@ struct VMStateInfo {
 };
 
 enum VMStateFlags {
+    /* Ignored */
     VMS_SINGLE           = 0x001,
+
+    /* The struct member at opaque + VMStateField.offset is a pointer
+     * to the actual field (e.g. struct a { uint8_t *b;
+     * }). Dereference the pointer before using it as basis for
+     * further pointer arithmetic (see e.g. VMS_ARRAY). Does not
+     * affect the meaning of VMStateField.num_offset or
+     * VMStateField.size_offset; see VMS_VARRAY* and VMS_VBUFFER for
+     * those. */
     VMS_POINTER          = 0x002,
+
+    /* The field is an array of fixed size. VMStateField.num contains
+     * the number of entries in the array. The size of each entry is
+     * given by VMStateField.size and / or opaque +
+     * VMStateField.size_offset; see VMS_VBUFFER and
+     * VMS_MULTIPLY. Each array entry will be processed individually
+     * (VMStateField.info.get()/put() if VMS_STRUCT is not set,
+     * recursion into VMStateField.vmsd if VMS_STRUCT is set). May not
+     * be combined with VMS_VARRAY*. */
     VMS_ARRAY            = 0x004,
+
+    /* The field is itself a struct, containing one or more
+     * fields. Recurse into VMStateField.vmsd. Most useful in
+     * combination with VMS_ARRAY / VMS_VARRAY*, recursing into each
+     * array entry. */
     VMS_STRUCT           = 0x008,
-    VMS_VARRAY_INT32     = 0x010,  /* Array with size in int32_t field*/
-    VMS_BUFFER           = 0x020,  /* static sized buffer */
+
+    /* The field is an array of variable size. The int32_t at opaque +
+     * VMStateField.num_offset contains the number of entries in the
+     * array. See the VMS_ARRAY description regarding array handling
+     * in general. May not be combined with VMS_ARRAY or any other
+     * VMS_VARRAY*. */
+    VMS_VARRAY_INT32     = 0x010,
+
+    /* Ignored */
+    VMS_BUFFER           = 0x020,
+
+    /* The field is a (fixed-size or variable-size) array of pointers
+     * (e.g. struct a { uint8_t *b[]; }). Dereference each array entry
+     * before using it. Note: Does not imply any one of VMS_ARRAY /
+     * VMS_VARRAY*; these need to be set explicitly. */
     VMS_ARRAY_OF_POINTER = 0x040,
-    VMS_VARRAY_UINT16    = 0x080,  /* Array with size in uint16_t field */
-    VMS_VBUFFER          = 0x100,  /* Buffer with size in int32_t field */
-    VMS_MULTIPLY         = 0x200,  /* multiply "size" field by field_size */
-    VMS_VARRAY_UINT8     = 0x400,  /* Array with size in uint8_t field*/
-    VMS_VARRAY_UINT32    = 0x800,  /* Array with size in uint32_t field*/
-    VMS_MUST_EXIST       = 0x1000, /* Field must exist in input */
-    VMS_ALLOC            = 0x2000, /* Alloc a buffer on the destination */
-    VMS_MULTIPLY_ELEMENTS = 0x4000, /* multiply varray size by field->num */
+
+    /* The field is an array of variable size. The uint16_t at opaque
+     * + VMStateField.num_offset (subject to VMS_MULTIPLY_ELEMENTS)
+     * contains the number of entries in the array. See the VMS_ARRAY
+     * description regarding array handling in general. May not be
+     * combined with VMS_ARRAY or any other VMS_VARRAY*. */
+    VMS_VARRAY_UINT16    = 0x080,
+
+    /* The size of the individual entries (a single array entry if
+     * VMS_ARRAY or any of VMS_VARRAY* are set, or the field itself if
+     * neither is set) is variable (i.e. not known at compile-time),
+     * but the same for all entries. Use the int32_t at opaque +
+     * VMStateField.size_offset (subject to VMS_MULTIPLY) to determine
+     * the size of each (and every) entry. */
+    VMS_VBUFFER          = 0x100,
+
+    /* Multiply the entry size given by the int32_t at opaque +
+     * VMStateField.size_offset (see VMS_VBUFFER description) with
+     * VMStateField.size to determine the number of bytes to be
+     * allocated. Only valid in combination with VMS_VBUFFER. */
+    VMS_MULTIPLY         = 0x200,
+
+    /* The field is an array of variable size. The uint8_t at opaque +
+     * VMStateField.num_offset (subject to VMS_MULTIPLY_ELEMENTS)
+     * contains the number of entries in the array. See the VMS_ARRAY
+     * description regarding array handling in general. May not be
+     * combined with VMS_ARRAY or any other VMS_VARRAY*. */
+    VMS_VARRAY_UINT8     = 0x400,
+
+    /* The field is an array of variable size. The uint32_t at opaque
+     * + VMStateField.num_offset (subject to VMS_MULTIPLY_ELEMENTS)
+     * contains the number of entries in the array. See the VMS_ARRAY
+     * description regarding array handling in general. May not be
+     * combined with VMS_ARRAY or any other VMS_VARRAY*. */
+    VMS_VARRAY_UINT32    = 0x800,
+
+    /* Fail loading the serialised VM state if this field is missing
+     * from the input. */
+    VMS_MUST_EXIST       = 0x1000,
+
+    /* When loading serialised VM state, allocate memory for the
+     * (entire) field. Only valid in combination with
+     * VMS_POINTER. Note: Not all combinations with other flags are
+     * currently supported, e.g. VMS_ALLOC|VMS_ARRAY_OF_POINTER won't
+     * cause the individual entries to be allocated. */
+    VMS_ALLOC            = 0x2000,
+
+    /* Multiply the number of entries given by the integer at opaque +
+     * VMStateField.num_offset (see VMS_VARRAY*) with VMStateField.num
+     * to determine the number of entries in the array. Only valid in
+     * combination with one of VMS_VARRAY*. */
+    VMS_MULTIPLY_ELEMENTS = 0x4000,
 };
 
 typedef struct {
diff --git a/migration/migration.c b/migration/migration.c
index fc5e50b0be..0129d9f420 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -323,10 +323,56 @@ void qemu_start_incoming_migration(const char *uri, Error **errp)
     }
 }
 
+static void process_incoming_migration_bh(void *opaque)
+{
+    Error *local_err = NULL;
+    MigrationIncomingState *mis = opaque;
+
+    /* Make sure all file formats flush their mutable metadata */
+    bdrv_invalidate_cache_all(&local_err);
+    if (local_err) {
+        migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
+                          MIGRATION_STATUS_FAILED);
+        error_report_err(local_err);
+        migrate_decompress_threads_join();
+        exit(EXIT_FAILURE);
+    }
+
+    /*
+     * This must happen after all error conditions are dealt with and
+     * we're sure the VM is going to be running on this host.
+     */
+    qemu_announce_self();
+
+    /* If global state section was not received or we are in running
+       state, we need to obey autostart. Any other state is set with
+       runstate_set. */
+
+    if (!global_state_received() ||
+        global_state_get_runstate() == RUN_STATE_RUNNING) {
+        if (autostart) {
+            vm_start();
+        } else {
+            runstate_set(RUN_STATE_PAUSED);
+        }
+    } else {
+        runstate_set(global_state_get_runstate());
+    }
+    migrate_decompress_threads_join();
+    /*
+     * This must happen after any state changes since as soon as an external
+     * observer sees this event they might start to prod at the VM assuming
+     * it's ready to use.
+     */
+    migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
+                      MIGRATION_STATUS_COMPLETED);
+    qemu_bh_delete(mis->bh);
+    migration_incoming_state_destroy();
+}
+
 static void process_incoming_migration_co(void *opaque)
 {
     QEMUFile *f = opaque;
-    Error *local_err = NULL;
     MigrationIncomingState *mis;
     PostcopyState ps;
     int ret;
@@ -369,45 +415,8 @@ static void process_incoming_migration_co(void *opaque)
         exit(EXIT_FAILURE);
     }
 
-    /* Make sure all file formats flush their mutable metadata */
-    bdrv_invalidate_cache_all(&local_err);
-    if (local_err) {
-        migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
-                          MIGRATION_STATUS_FAILED);
-        error_report_err(local_err);
-        migrate_decompress_threads_join();
-        exit(EXIT_FAILURE);
-    }
-
-    /*
-     * This must happen after all error conditions are dealt with and
-     * we're sure the VM is going to be running on this host.
-     */
-    qemu_announce_self();
-
-    /* If global state section was not received or we are in running
-       state, we need to obey autostart. Any other state is set with
-       runstate_set. */
-
-    if (!global_state_received() ||
-        global_state_get_runstate() == RUN_STATE_RUNNING) {
-        if (autostart) {
-            vm_start();
-        } else {
-            runstate_set(RUN_STATE_PAUSED);
-        }
-    } else {
-        runstate_set(global_state_get_runstate());
-    }
-    migrate_decompress_threads_join();
-    /*
-     * This must happen after any state changes since as soon as an external
-     * observer sees this event they might start to prod at the VM assuming
-     * it's ready to use.
-     */
-    migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
-                      MIGRATION_STATUS_COMPLETED);
-    migration_incoming_state_destroy();
+    mis->bh = qemu_bh_new(process_incoming_migration_bh, mis);
+    qemu_bh_schedule(mis->bh);
 }
 
 void process_incoming_migration(QEMUFile *f)
diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c
index 254c629d48..fbd0064fce 100644
--- a/migration/postcopy-ram.c
+++ b/migration/postcopy-ram.c
@@ -52,14 +52,14 @@ struct PostcopyDiscardState {
 #if defined(__linux__)
 
 #include <poll.h>
-#include <sys/eventfd.h>
 #include <sys/mman.h>
 #include <sys/ioctl.h>
 #include <sys/syscall.h>
 #include <asm/types.h> /* for __u64 */
 #endif
 
-#if defined(__linux__) && defined(__NR_userfaultfd)
+#if defined(__linux__) && defined(__NR_userfaultfd) && defined(CONFIG_EVENTFD)
+#include <sys/eventfd.h>
 #include <linux/userfaultfd.h>
 
 static bool ufd_version_check(int ufd)
diff --git a/migration/savevm.c b/migration/savevm.c
index 94f2894243..b45915612f 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -1088,12 +1088,11 @@ void qemu_savevm_state_complete_precopy(QEMUFile *f, bool iterable_only)
         json_prop_int(vmdesc, "instance_id", se->instance_id);
 
         save_section_header(f, se, QEMU_VM_SECTION_FULL);
-
         vmstate_save(f, se, vmdesc);
-
-        json_end_object(vmdesc);
         trace_savevm_section_end(se->idstr, se->section_id, 0);
         save_section_footer(f, se);
+
+        json_end_object(vmdesc);
     }
 
     if (!in_postcopy) {
@@ -1496,17 +1495,10 @@ static int loadvm_postcopy_handle_listen(MigrationIncomingState *mis)
     return 0;
 }
 
-/* After all discards we can start running and asking for pages */
-static int loadvm_postcopy_handle_run(MigrationIncomingState *mis)
+static void loadvm_postcopy_handle_run_bh(void *opaque)
 {
-    PostcopyState ps = postcopy_state_set(POSTCOPY_INCOMING_RUNNING);
     Error *local_err = NULL;
-
-    trace_loadvm_postcopy_handle_run();
-    if (ps != POSTCOPY_INCOMING_LISTENING) {
-        error_report("CMD_POSTCOPY_RUN in wrong postcopy state (%d)", ps);
-        return -1;
-    }
+    MigrationIncomingState *mis = opaque;
 
     /* TODO we should move all of this lot into postcopy_ram.c or a shared code
      * in migration.c
@@ -1519,7 +1511,6 @@ static int loadvm_postcopy_handle_run(MigrationIncomingState *mis)
     bdrv_invalidate_cache_all(&local_err);
     if (local_err) {
         error_report_err(local_err);
-        return -1;
     }
 
     trace_loadvm_postcopy_handle_run_cpu_sync();
@@ -1535,6 +1526,23 @@ static int loadvm_postcopy_handle_run(MigrationIncomingState *mis)
         runstate_set(RUN_STATE_PAUSED);
     }
 
+    qemu_bh_delete(mis->bh);
+}
+
+/* After all discards we can start running and asking for pages */
+static int loadvm_postcopy_handle_run(MigrationIncomingState *mis)
+{
+    PostcopyState ps = postcopy_state_set(POSTCOPY_INCOMING_RUNNING);
+
+    trace_loadvm_postcopy_handle_run();
+    if (ps != POSTCOPY_INCOMING_LISTENING) {
+        error_report("CMD_POSTCOPY_RUN in wrong postcopy state (%d)", ps);
+        return -1;
+    }
+
+    mis->bh = qemu_bh_new(loadvm_postcopy_handle_run_bh, NULL);
+    qemu_bh_schedule(mis->bh);
+
     /* We need to finish reading the stream from the package
      * and also stop reading anything more from the stream that loaded the
      * package (since it's now being read by the listener thread).