qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Juan Quintela <quintela@redhat.com>
To: qemu-devel@nongnu.org
Subject: [Qemu-devel] [PATCH 06/14] migration: stop all cpus correctly
Date: Fri, 21 Sep 2012 16:08:12 +0200	[thread overview]
Message-ID: <1348236500-2565-7-git-send-email-quintela@redhat.com> (raw)
In-Reply-To: <1348236500-2565-1-git-send-email-quintela@redhat.com>

You can only stop all cpus from the iothread or an vcpu.  As we want
to do it from the migration_thread, we need to do this dance with the
botton handlers.

This patch is a request for ideas.  I can move this function to cpus.c, but
wondered if there is an easy way of doing this?

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration.c | 29 +++++++++++++++++------------
 1 file changed, 17 insertions(+), 12 deletions(-)

diff --git a/migration.c b/migration.c
index 550c5c2..ee0519e 100644
--- a/migration.c
+++ b/migration.c
@@ -20,6 +20,7 @@
 #include "sysemu.h"
 #include "block.h"
 #include "qemu_socket.h"
+#include "qemu-thread.h"
 #include "block-migration.h"
 #include "qmp-commands.h"

@@ -320,11 +321,22 @@ ssize_t migrate_fd_put_buffer(MigrationState *s, const void *data,
 void migrate_fd_put_ready(MigrationState *s)
 {
     int ret;
+    static bool first_time = true;

     if (s->state != MIG_STATE_ACTIVE) {
         DPRINTF("put_ready returning because of non-active state\n");
         return;
     }
+    if (first_time) {
+        first_time = false;
+        DPRINTF("beginning savevm\n");
+        ret = qemu_savevm_state_begin(s->file, &s->params);
+        if (ret < 0) {
+            DPRINTF("failed, %d\n", ret);
+            migrate_fd_error(s);
+            return;
+        }
+    }

     DPRINTF("iterate\n");
     ret = qemu_savevm_state_iterate(s->file);
@@ -337,7 +349,11 @@ void migrate_fd_put_ready(MigrationState *s)
         DPRINTF("done iterating\n");
         start_time = qemu_get_clock_ms(rt_clock);
         qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
-        vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
+        if (old_vm_running) {
+            vm_stop(RUN_STATE_FINISH_MIGRATE);
+        } else {
+            vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
+        }

         if (qemu_savevm_state_complete(s->file) < 0) {
             migrate_fd_error(s);
@@ -426,19 +442,8 @@ bool migration_has_failed(MigrationState *s)

 void migrate_fd_connect(MigrationState *s)
 {
-    int ret;
-
     s->state = MIG_STATE_ACTIVE;
     qemu_fopen_ops_buffered(s);
-
-    DPRINTF("beginning savevm\n");
-    ret = qemu_savevm_state_begin(s->file, &s->params);
-    if (ret < 0) {
-        DPRINTF("failed, %d\n", ret);
-        migrate_fd_error(s);
-        return;
-    }
-    migrate_fd_put_ready(s);
 }

 static MigrationState *migrate_init(const MigrationParams *params)
-- 
1.7.11.4

  parent reply	other threads:[~2012-09-21 14:54 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-09-21 14:08 [Qemu-devel] [RFC 00/14] Migration thread Juan Quintela
2012-09-21 14:08 ` [Qemu-devel] [PATCH 01/14] split MRU ram list Juan Quintela
2012-09-21 14:38   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 02/14] add a version number to ram_list Juan Quintela
2012-09-21 14:42   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 03/14] protect the ramlist with a separate mutex Juan Quintela
2012-09-21 14:44   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 04/14] buffered_file: Move from using a timer to use a thread Juan Quintela
2012-09-21 15:29   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 05/14] migration: make qemu_fopen_ops_buffered() return void Juan Quintela
2012-09-21 14:57   ` Paolo Bonzini
2012-09-21 14:08 ` Juan Quintela [this message]
2012-09-21 15:20   ` [Qemu-devel] [PATCH 06/14] migration: stop all cpus correctly Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 07/14] migration: make writes blocking Juan Quintela
2012-09-21 15:31   ` Paolo Bonzini
2012-12-14 12:40     ` Juan Quintela
2012-09-21 14:08 ` [Qemu-devel] [PATCH 08/14] migration: remove unfreeze logic Juan Quintela
2012-09-21 14:58   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 09/14] migration: take finer locking Juan Quintela
2012-09-21 15:33   ` Paolo Bonzini
2012-12-14 12:44     ` Juan Quintela
2012-09-21 14:08 ` [Qemu-devel] [PATCH 10/14] buffered_file: Unfold the trick to restart generating migration data Juan Quintela
2012-09-21 15:00   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 11/14] buffered_file: don't flush on put buffer Juan Quintela
2012-09-21 15:34   ` Paolo Bonzini
2012-09-21 21:12     ` Eric Blake
2012-09-21 14:08 ` [Qemu-devel] [PATCH 12/14] buffered_file: unfold buffered_append in buffered_put_buffer Juan Quintela
2012-09-21 17:07   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 13/14] savevm: New save live migration method: pending Juan Quintela
2012-09-21 15:11   ` Paolo Bonzini
2012-09-21 14:08 ` [Qemu-devel] [PATCH 14/14] migration: print times for end phase Juan Quintela
2012-09-21 15:13   ` Paolo Bonzini
2012-09-21 15:27     ` Juan Quintela

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1348236500-2565-7-git-send-email-quintela@redhat.com \
    --to=quintela@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).