qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Juan Quintela <quintela@redhat.com>
To: qemu-devel@nongnu.org
Cc: dgilbert@redhat.com, lvivier@redhat.com, peterx@redhat.com
Subject: [Qemu-devel] [PULL 19/41] migration: allow fault thread to pause
Date: Wed,  9 May 2018 13:23:44 +0200	[thread overview]
Message-ID: <20180509112406.6183-20-quintela@redhat.com> (raw)
In-Reply-To: <20180509112406.6183-1-quintela@redhat.com>

From: Peter Xu <peterx@redhat.com>

Allows the fault thread to stop handling page faults temporarily. When
network failure happened (and if we expect a recovery afterwards), we
should not allow the fault thread to continue sending things to source,
instead, it should halt for a while until the connection is rebuilt.

When the dest main thread noticed the failure, it kicks the fault thread
to switch to pause state.

Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180502104740.12123-7-peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/migration.c    |  1 +
 migration/migration.h    |  1 +
 migration/postcopy-ram.c | 54 +++++++++++++++++++++++++++++++++++++---
 migration/savevm.c       |  3 +++
 migration/trace-events   |  2 ++
 5 files changed, 57 insertions(+), 4 deletions(-)

diff --git a/migration/migration.c b/migration/migration.c
index f9c58d2511..3d3b0a5b4a 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -160,6 +160,7 @@ MigrationIncomingState *migration_incoming_get_current(void)
         qemu_mutex_init(&mis_current.rp_mutex);
         qemu_event_init(&mis_current.main_thread_load_event, false);
         qemu_sem_init(&mis_current.postcopy_pause_sem_dst, 0);
+        qemu_sem_init(&mis_current.postcopy_pause_sem_fault, 0);
 
         init_dirty_bitmap_incoming_migration();
 
diff --git a/migration/migration.h b/migration/migration.h
index 32fd50e9be..4ea5949104 100644
--- a/migration/migration.h
+++ b/migration/migration.h
@@ -76,6 +76,7 @@ struct MigrationIncomingState {
 
     /* notify PAUSED postcopy incoming migrations to try to continue */
     QemuSemaphore postcopy_pause_sem_dst;
+    QemuSemaphore postcopy_pause_sem_fault;
 };
 
 MigrationIncomingState *migration_incoming_get_current(void);
diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c
index 8ceeaa2a93..658b750a8e 100644
--- a/migration/postcopy-ram.c
+++ b/migration/postcopy-ram.c
@@ -830,6 +830,17 @@ static void mark_postcopy_blocktime_end(uintptr_t addr)
                                       affected_cpu);
 }
 
+static bool postcopy_pause_fault_thread(MigrationIncomingState *mis)
+{
+    trace_postcopy_pause_fault_thread();
+
+    qemu_sem_wait(&mis->postcopy_pause_sem_fault);
+
+    trace_postcopy_pause_fault_thread_continued();
+
+    return true;
+}
+
 /*
  * Handle faults detected by the USERFAULT markings
  */
@@ -880,6 +891,22 @@ static void *postcopy_ram_fault_thread(void *opaque)
             break;
         }
 
+        if (!mis->to_src_file) {
+            /*
+             * Possibly someone tells us that the return path is
+             * broken already using the event. We should hold until
+             * the channel is rebuilt.
+             */
+            if (postcopy_pause_fault_thread(mis)) {
+                mis->last_rb = NULL;
+                /* Continue to read the userfaultfd */
+            } else {
+                error_report("%s: paused but don't allow to continue",
+                             __func__);
+                break;
+            }
+        }
+
         if (pfd[1].revents) {
             uint64_t tmp64 = 0;
 
@@ -942,18 +969,37 @@ static void *postcopy_ram_fault_thread(void *opaque)
                     (uintptr_t)(msg.arg.pagefault.address),
                                 msg.arg.pagefault.feat.ptid, rb);
 
+retry:
             /*
              * Send the request to the source - we want to request one
              * of our host page sizes (which is >= TPS)
              */
             if (rb != mis->last_rb) {
                 mis->last_rb = rb;
-                migrate_send_rp_req_pages(mis, qemu_ram_get_idstr(rb),
-                                         rb_offset, qemu_ram_pagesize(rb));
+                ret = migrate_send_rp_req_pages(mis,
+                                                qemu_ram_get_idstr(rb),
+                                                rb_offset,
+                                                qemu_ram_pagesize(rb));
             } else {
                 /* Save some space */
-                migrate_send_rp_req_pages(mis, NULL,
-                                         rb_offset, qemu_ram_pagesize(rb));
+                ret = migrate_send_rp_req_pages(mis,
+                                                NULL,
+                                                rb_offset,
+                                                qemu_ram_pagesize(rb));
+            }
+
+            if (ret) {
+                /* May be network failure, try to wait for recovery */
+                if (ret == -EIO && postcopy_pause_fault_thread(mis)) {
+                    /* We got reconnected somehow, try to continue */
+                    mis->last_rb = NULL;
+                    goto retry;
+                } else {
+                    /* This is a unavoidable fault */
+                    error_report("%s: migrate_send_rp_req_pages() get %d",
+                                 __func__, ret);
+                    break;
+                }
             }
         }
 
diff --git a/migration/savevm.c b/migration/savevm.c
index 8ad99b1eaa..6ee69d8283 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2083,6 +2083,9 @@ static bool postcopy_pause_incoming(MigrationIncomingState *mis)
     mis->to_src_file = NULL;
     qemu_mutex_unlock(&mis->rp_mutex);
 
+    /* Notify the fault thread for the invalidated file handle */
+    postcopy_fault_thread_notify(mis);
+
     error_report("Detected IO failure for postcopy. "
                  "Migration paused.");
 
diff --git a/migration/trace-events b/migration/trace-events
index cd971bf9fe..7f836499d1 100644
--- a/migration/trace-events
+++ b/migration/trace-events
@@ -101,6 +101,8 @@ open_return_path_on_source_continue(void) ""
 postcopy_start(void) ""
 postcopy_pause_return_path(void) ""
 postcopy_pause_return_path_continued(void) ""
+postcopy_pause_fault_thread(void) ""
+postcopy_pause_fault_thread_continued(void) ""
 postcopy_pause_continued(void) ""
 postcopy_pause_incoming(void) ""
 postcopy_pause_incoming_continued(void) ""
-- 
2.17.0

  parent reply	other threads:[~2018-05-09 11:24 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-09 11:23 [Qemu-devel] [PULL 00/41] Migration queue Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 01/41] migration: fix saving normal page even if it's been compressed Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 02/41] tests: Add migration precopy test Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 03/41] tests: Add migration xbzrle test Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 04/41] tests: Migration ppc now inlines its program Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 05/41] migration: Set error state in case of error Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 06/41] migration: Introduce multifd_recv_new_channel() Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 07/41] migration: terminate_* can be called for other threads Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 08/41] migration: Be sure all recv channels are created Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 09/41] migration: Export functions to create send channels Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 10/41] migration: Create multifd channels Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 11/41] migration: Delay start of migration main routines Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 12/41] migration: Transmit initial package through the multifd channels Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 13/41] migration: Define MultifdRecvParams sooner Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 14/41] migration: let incoming side use thread context Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 15/41] migration: new postcopy-pause state Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 16/41] migration: implement "postcopy-pause" src logic Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 17/41] migration: allow dst vm pause on postcopy Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 18/41] migration: allow src return path to pause Juan Quintela
2018-05-09 11:23 ` Juan Quintela [this message]
2018-05-09 11:23 ` [Qemu-devel] [PULL 20/41] qmp: hmp: add migrate "resume" option Juan Quintela
2018-05-09 12:57   ` Eric Blake
2018-05-09 11:23 ` [Qemu-devel] [PULL 21/41] migration: rebuild channel on source Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 22/41] migration: new state "postcopy-recover" Juan Quintela
2018-05-09 12:57   ` Eric Blake
2018-05-09 11:23 ` [Qemu-devel] [PULL 23/41] migration: wakeup dst ram-load-thread for recover Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 24/41] migration: new cmd MIG_CMD_RECV_BITMAP Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 25/41] migration: new message MIG_RP_MSG_RECV_BITMAP Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 26/41] migration: new cmd MIG_CMD_POSTCOPY_RESUME Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 27/41] migration: new message MIG_RP_MSG_RESUME_ACK Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 28/41] migration: introduce SaveVMHandlers.resume_prepare Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 29/41] migration: synchronize dirty bitmap for resume Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 30/41] migration: setup ramstate " Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 31/41] migration: final handshake for the resume Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 32/41] migration: init dst in migration_object_init too Juan Quintela
2018-05-09 11:23 ` [Qemu-devel] [PULL 33/41] qmp/migration: new command migrate-recover Juan Quintela
2018-05-09 12:59   ` Eric Blake
2018-05-09 11:23 ` [Qemu-devel] [PULL 34/41] hmp/migration: add migrate_recover command Juan Quintela
2018-05-09 11:24 ` [Qemu-devel] [PULL 35/41] migration: introduce lock for to_dst_file Juan Quintela
2018-05-09 11:24 ` [Qemu-devel] [PULL 36/41] migration/qmp: add command migrate-pause Juan Quintela
2018-05-09 12:59   ` Eric Blake
2018-05-09 11:24 ` [Qemu-devel] [PULL 37/41] migration/hmp: add migrate_pause command Juan Quintela
2018-05-09 11:24 ` [Qemu-devel] [PULL 38/41] migration: update docs Juan Quintela
2018-05-09 11:24 ` [Qemu-devel] [PULL 39/41] migration: update index field when delete or qsort RDMALocalBlock Juan Quintela
2018-05-09 11:24 ` [Qemu-devel] [PULL 40/41] migration: Textual fixups for blocktime Juan Quintela
2018-05-09 11:24 ` [Qemu-devel] [PULL 41/41] Migration+TLS: Fix crash due to double cleanup Juan Quintela
2018-05-11 13:41 ` [Qemu-devel] [PULL 00/41] Migration queue Peter Maydell
2018-05-11 14:20   ` Dr. David Alan Gilbert
2018-05-11 14:22     ` Peter Maydell
2018-05-18 10:19   ` Peter Maydell
2018-05-18 10:22     ` Peter Maydell

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=20180509112406.6183-20-quintela@redhat.com \
    --to=quintela@redhat.com \
    --cc=dgilbert@redhat.com \
    --cc=lvivier@redhat.com \
    --cc=peterx@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).