Blame SOURCES/kvm-virtio-blk-Cancel-the-pending-BH-when-the-dataplane-.patch

016a62
From 93a7832d5dfd83f170119e7130f3968fe37fa8e6 Mon Sep 17 00:00:00 2001
016a62
From: =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= <philmd@redhat.com>
016a62
Date: Fri, 13 Sep 2019 14:16:25 +0100
016a62
Subject: [PATCH 08/22] virtio-blk: Cancel the pending BH when the dataplane is
016a62
 reset
016a62
MIME-Version: 1.0
016a62
Content-Type: text/plain; charset=UTF-8
016a62
Content-Transfer-Encoding: 8bit
016a62
016a62
RH-Author: Philippe Mathieu-Daudé <philmd@redhat.com>
016a62
Message-id: <20190913141625.12521-2-philmd@redhat.com>
016a62
Patchwork-id: 90453
016a62
O-Subject: [RHEL-7.7.z qemu-kvm-rhev + RHEL-8.1.0 qemu-kvm + RHEL-AV-8.1.0 qemu-kvm PATCH v2 1/1] virtio-blk: Cancel the pending BH when the dataplane is reset
016a62
Bugzilla: 1708459
016a62
RH-Acked-by: John Snow <jsnow@redhat.com>
016a62
RH-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
016a62
RH-Acked-by: Danilo de Paula <ddepaula@redhat.com>
016a62
016a62
When 'system_reset' is called, the main loop clear the memory
016a62
region cache before the BH has a chance to execute. Later when
016a62
the deferred function is called, some assumptions that were
016a62
made when scheduling them are no longer true when they actually
016a62
execute.
016a62
016a62
This is what happens using a virtio-blk device (fresh RHEL7.8 install):
016a62
016a62
 $ (sleep 12.3; echo system_reset; sleep 12.3; echo system_reset; sleep 1; echo q) \
016a62
   | qemu-system-x86_64 -m 4G -smp 8 -boot menu=on \
016a62
     -device virtio-blk-pci,id=image1,drive=drive_image1 \
016a62
     -drive file=/var/lib/libvirt/images/rhel78.qcow2,if=none,id=drive_image1,format=qcow2,cache=none \
016a62
     -device virtio-net-pci,netdev=net0,id=nic0,mac=52:54:00:c4:e7:84 \
016a62
     -netdev tap,id=net0,script=/bin/true,downscript=/bin/true,vhost=on \
016a62
     -monitor stdio -serial null -nographic
016a62
  (qemu) system_reset
016a62
  (qemu) system_reset
016a62
  (qemu) qemu-system-x86_64: hw/virtio/virtio.c:225: vring_get_region_caches: Assertion `caches != NULL' failed.
016a62
  Aborted
016a62
016a62
  (gdb) bt
016a62
  Thread 1 (Thread 0x7f109c17b680 (LWP 10939)):
016a62
  #0  0x00005604083296d1 in vring_get_region_caches (vq=0x56040a24bdd0) at hw/virtio/virtio.c:227
016a62
  #1  0x000056040832972b in vring_avail_flags (vq=0x56040a24bdd0) at hw/virtio/virtio.c:235
016a62
  #2  0x000056040832d13d in virtio_should_notify (vdev=0x56040a240630, vq=0x56040a24bdd0) at hw/virtio/virtio.c:1648
016a62
  #3  0x000056040832d1f8 in virtio_notify_irqfd (vdev=0x56040a240630, vq=0x56040a24bdd0) at hw/virtio/virtio.c:1662
016a62
  #4  0x00005604082d213d in notify_guest_bh (opaque=0x56040a243ec0) at hw/block/dataplane/virtio-blk.c:75
016a62
  #5  0x000056040883dc35 in aio_bh_call (bh=0x56040a243f10) at util/async.c:90
016a62
  #6  0x000056040883dccd in aio_bh_poll (ctx=0x560409161980) at util/async.c:118
016a62
  #7  0x0000560408842af7 in aio_dispatch (ctx=0x560409161980) at util/aio-posix.c:460
016a62
  #8  0x000056040883e068 in aio_ctx_dispatch (source=0x560409161980, callback=0x0, user_data=0x0) at util/async.c:261
016a62
  #9  0x00007f10a8fca06d in g_main_context_dispatch () at /lib64/libglib-2.0.so.0
016a62
  #10 0x0000560408841445 in glib_pollfds_poll () at util/main-loop.c:215
016a62
  #11 0x00005604088414bf in os_host_main_loop_wait (timeout=0) at util/main-loop.c:238
016a62
  #12 0x00005604088415c4 in main_loop_wait (nonblocking=0) at util/main-loop.c:514
016a62
  #13 0x0000560408416b1e in main_loop () at vl.c:1923
016a62
  #14 0x000056040841e0e8 in main (argc=20, argv=0x7ffc2c3f9c58, envp=0x7ffc2c3f9d00) at vl.c:4578
016a62
016a62
Fix this by cancelling the BH when the virtio dataplane is stopped.
016a62
016a62
[This is version of the patch was modified as discussed with Philippe on
016a62
the mailing list thread.
016a62
--Stefan]
016a62
016a62
Reported-by: Yihuang Yu <yihyu@redhat.com>
016a62
Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
016a62
Fixes: https://bugs.launchpad.net/qemu/+bug/1839428
016a62
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
016a62
Message-Id: <20190816171503.24761-1-philmd@redhat.com>
016a62
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
016a62
(cherry picked from commit ebb6ff25cd888a52a64a9adc3692541c6d1d9a42)
016a62
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
016a62
Signed-off-by: Danilo C. L. de Paula <ddepaula@redhat.com>
016a62
---
016a62
 hw/block/dataplane/virtio-blk.c | 3 +++
016a62
 1 file changed, 3 insertions(+)
016a62
016a62
diff --git a/hw/block/dataplane/virtio-blk.c b/hw/block/dataplane/virtio-blk.c
016a62
index 101f32c..23e4022 100644
016a62
--- a/hw/block/dataplane/virtio-blk.c
016a62
+++ b/hw/block/dataplane/virtio-blk.c
016a62
@@ -292,6 +292,9 @@ void virtio_blk_data_plane_stop(VirtIODevice *vdev)
016a62
         virtio_bus_cleanup_host_notifier(VIRTIO_BUS(qbus), i);
016a62
     }
016a62
 
016a62
+    qemu_bh_cancel(s->bh);
016a62
+    notify_guest_bh(s); /* final chance to notify guest */
016a62
+
016a62
     /* Clean up guest notifier (irq) */
016a62
     k->set_guest_notifiers(qbus->parent, nvqs, false);
016a62
 
016a62
-- 
016a62
1.8.3.1
016a62