Re: [Qemu-devel] [PATCH] virtio-blk: assert on starting/stopping

[Cornelia Huck]

On Sun, 3 Apr 2016 23:13:28 +0200
Paolo Bonzini <address@hidden> wrote:

> On 03/04/2016 21:59, Christian Borntraeger wrote:
> > Thread 1 (Thread 0x3ffad25bb90 (LWP 41685)):
> > ---Type <return> to continue, or q <return> to quit---
> > #0  0x000003ffab5be2c0 in raise () at /lib64/libc.so.6
> > #1  0x000003ffab5bfc26 in abort () at /lib64/libc.so.6
> > #2  0x000003ffab5b5bce in __assert_fail_base () at /lib64/libc.so.6
> > #3  0x000003ffab5b5c5c in  () at /lib64/libc.so.6
> > #4  0x00000000800b79e4 in virtio_blk_data_plane_start (s=0x80b195a0) at 
> > /home/cborntra/REPOS/qemu/hw/block/dataplane/virtio-blk.c:210
> > #5  0x00000000800b57ba in virtio_blk_handle_output (vdev=0x80e4b0f8, 
> > vq=0x80eaa180) at /home/cborntra/REPOS/qemu/hw/block/virtio-blk.c:607
> > #6  0x00000000800f0d74 in virtio_queue_notify_vq (vq=0x80eaa180) at 
> > /home/cborntra/REPOS/qemu/hw/virtio/virtio.c:1108
> > #7  0x00000000800f376c in virtio_queue_host_notifier_read (n=0x80eaa1e0) at 
> > /home/cborntra/REPOS/qemu/hw/virtio/virtio.c:1820
> > #8  0x00000000800f381c in virtio_queue_set_host_notifier_fd_handler 
> > (vq=0x80eaa180, assign=false, set_handler=false) at 
> > /home/cborntra/REPOS/qemu/hw/virtio/virtio.c:1836
> > #9  0x000000008010b808 in virtio_ccw_set_guest2host_notifier 
> > (dev=0x80e49fb0, n=0, assign=false, set_handler=false) at 
> > /home/cborntra/REPOS/qemu/hw/s390x/virtio-ccw.c:98
> > #10 0x000000008010baaa in virtio_ccw_stop_ioeventfd (dev=0x80e49fb0) at 
> > /home/cborntra/REPOS/qemu/hw/s390x/virtio-ccw.c:155
> > #11 0x000000008010f162 in virtio_ccw_set_host_notifier (d=0x80e49fb0, n=0, 
> > assign=true) at /home/cborntra/REPOS/qemu/hw/s390x/virtio-ccw.c:1212
> > #12 0x00000000800b7ab0 in virtio_blk_data_plane_start (s=0x80b195a0) at 
> > /home/cborntra/REPOS/qemu/hw/block/dataplane/virtio-blk.c:225
> > #13 0x00000000800b57ba in virtio_blk_handle_output (vdev=0x80e4b0f8, 
> > vq=0x80eaa180) at /home/cborntra/REPOS/qemu/hw/block/virtio-blk.c:607
> > #14 0x00000000800f0d74 in virtio_queue_notify_vq (vq=0x80eaa180) at 
> > /home/cborntra/REPOS/qemu/hw/virtio/virtio.c:1108
> > #15 0x00000000800f376c in virtio_queue_host_notifier_read (n=0x80eaa1e0) at 
> > /home/cborntra/REPOS/qemu/hw/virtio/virtio.c:1820
> > #16 0x00000000802f1a0c in aio_dispatch (ctx=0x80abae30) at 
> > /home/cborntra/REPOS/qemu/aio-posix.c:327
> > #17 0x00000000802df4d4 in aio_ctx_dispatch (source=0x80abae30, 
> > callback=0x0, user_data=0x0) at /home/cborntra/REPOS/qemu/async.c:233
> > #18 0x000003ffabfd1c0a in g_main_context_dispatch () at 
> > /lib64/libglib-2.0.so.0
> > #19 0x00000000802ee70e in glib_pollfds_poll () at 
> > /home/cborntra/REPOS/qemu/main-loop.c:213
> > #20 0x00000000802ee84a in os_host_main_loop_wait (timeout=1472000000) at 
> > /home/cborntra/REPOS/qemu/main-loop.c:258
> > #21 0x00000000802ee956 in main_loop_wait (nonblocking=0) at 
> > /home/cborntra/REPOS/qemu/main-loop.c:506
> > #22 0x000000008017dc0c in main_loop () at 
> > /home/cborntra/REPOS/qemu/vl.c:1934
> > #23 0x00000000801860e0 in main (argc=72, argv=0x3ffd60fe838, 
> > envp=0x3ffd60fea80) at /home/cborntra/REPOS/qemu/vl.c:4652
> 
> This will be fixed by Cornelia's rework, and is an example of why I
> think patch 1/9 is a good idea (IOW, assign=false is harmful).

So what do we want to do for 2.6? The aio handler rework (without the
cleanup) is needed. Do we want to include the minimal version of my
"keep handler assigned" patch (the one without the api rework) as well,
as it fixes a latent bug?