Blame SOURCES/kvm-migration-multifd-fix-destroyed-mutex-access-in-term.patch

22c213
From 2c14a6831954a59256cc8d1980da0ad705a3a3fa Mon Sep 17 00:00:00 2001
22c213
From: Juan Quintela <quintela@redhat.com>
22c213
Date: Tue, 3 Mar 2020 14:51:37 +0000
22c213
Subject: [PATCH 05/18] migration/multifd: fix destroyed mutex access in
22c213
 terminating multifd threads
22c213
22c213
RH-Author: Juan Quintela <quintela@redhat.com>
22c213
Message-id: <20200303145143.149290-5-quintela@redhat.com>
22c213
Patchwork-id: 94119
22c213
O-Subject: [RHEL-AV-8.2.0 qemu-kvm PATCH v2 04/10] migration/multifd: fix destroyed mutex access in terminating multifd threads
22c213
Bugzilla: 1738451
22c213
RH-Acked-by: Laurent Vivier <lvivier@redhat.com>
22c213
RH-Acked-by: Peter Xu <peterx@redhat.com>
22c213
RH-Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
22c213
22c213
From: Jiahui Cen <cenjiahui@huawei.com>
22c213
22c213
One multifd will lock all the other multifds' IOChannel mutex to inform them
22c213
to quit by setting p->quit or shutting down p->c. In this senario, if some
22c213
multifds had already been terminated and multifd_load_cleanup/multifd_save_cleanup
22c213
had destroyed their mutex, it could cause destroyed mutex access when trying
22c213
lock their mutex.
22c213
22c213
Here is the coredump stack:
22c213
    #0  0x00007f81a2794437 in raise () from /usr/lib64/libc.so.6
22c213
    #1  0x00007f81a2795b28 in abort () from /usr/lib64/libc.so.6
22c213
    #2  0x00007f81a278d1b6 in __assert_fail_base () from /usr/lib64/libc.so.6
22c213
    #3  0x00007f81a278d262 in __assert_fail () from /usr/lib64/libc.so.6
22c213
    #4  0x000055eb1bfadbd3 in qemu_mutex_lock_impl (mutex=0x55eb1e2d1988, file=<optimized out>, line=<optimized out>) at util/qemu-thread-posix.c:64
22c213
    #5  0x000055eb1bb4564a in multifd_send_terminate_threads (err=<optimized out>) at migration/ram.c:1015
22c213
    #6  0x000055eb1bb4bb7f in multifd_send_thread (opaque=0x55eb1e2d19f8) at migration/ram.c:1171
22c213
    #7  0x000055eb1bfad628 in qemu_thread_start (args=0x55eb1e170450) at util/qemu-thread-posix.c:502
22c213
    #8  0x00007f81a2b36df5 in start_thread () from /usr/lib64/libpthread.so.0
22c213
    #9  0x00007f81a286048d in clone () from /usr/lib64/libc.so.6
22c213
22c213
To fix it up, let's destroy the mutex after all the other multifd threads had
22c213
been terminated.
22c213
22c213
Signed-off-by: Jiahui Cen <cenjiahui@huawei.com>
22c213
Signed-off-by: Ying Fang <fangying1@huawei.com>
22c213
Reviewed-by: Juan Quintela <quintela@redhat.com>
22c213
Signed-off-by: Juan Quintela <quintela@redhat.com>
22c213
(cherry picked from commit 9560a48ecc0c20d87bc458a6db77fba651605819)
22c213
Signed-off-by: Danilo C. L. de Paula <ddepaula@redhat.com>
22c213
---
22c213
 migration/ram.c | 8 ++++++++
22c213
 1 file changed, 8 insertions(+)
22c213
22c213
diff --git a/migration/ram.c b/migration/ram.c
22c213
index 860f781..6c55c5d 100644
22c213
--- a/migration/ram.c
22c213
+++ b/migration/ram.c
22c213
@@ -1052,6 +1052,10 @@ void multifd_save_cleanup(void)
22c213
         if (p->running) {
22c213
             qemu_thread_join(&p->thread);
22c213
         }
22c213
+    }
22c213
+    for (i = 0; i < migrate_multifd_channels(); i++) {
22c213
+        MultiFDSendParams *p = &multifd_send_state->params[i];
22c213
+
22c213
         socket_send_channel_destroy(p->c);
22c213
         p->c = NULL;
22c213
         qemu_mutex_destroy(&p->mutex);
22c213
@@ -1335,6 +1339,10 @@ int multifd_load_cleanup(Error **errp)
22c213
             qemu_sem_post(&p->sem_sync);
22c213
             qemu_thread_join(&p->thread);
22c213
         }
22c213
+    }
22c213
+    for (i = 0; i < migrate_multifd_channels(); i++) {
22c213
+        MultiFDRecvParams *p = &multifd_recv_state->params[i];
22c213
+
22c213
         object_unref(OBJECT(p->c));
22c213
         p->c = NULL;
22c213
         qemu_mutex_destroy(&p->mutex);
22c213
-- 
22c213
1.8.3.1
22c213