Hi,
Only replying to the crash bit for now ;)
On Wed, Jan 09, 2019 at 06:29:51PM -0500, james harvey wrote:
> > Please let me know if it looks like this crash is related to
> > the clipboard problems, or if I should create a bug report on
> > it. My best guess is it's related to the clipboard problems.
> > I'm going back to the v1 patch, but will do so on top of
> > current git (dbdf692) to try to see if I get any more
> > crashing.
>
> I'm more confused on this now. After sending the email, I
> reproduced a few more identical crashes using v2 of the patch.
> I went back to v1 on top of current git (dbdf692), and had
> gotten 18 hours using that without any crashing. I was
> literally typing this part of this email and was saying it must
> be v2 causing the crash. Then, I had an identical crash. So,
> it's not v2.
Yep
> No idea how it lasted 18 hours then now crashed.
> I was away for 10 hours, but 8 of the 18 the host (running the
> client) was being used, and occasionally the VM's.
It is a broken pipe. I can't see any warnings on the log you
provided with the force shutdown before.
Besides what you highlighted before:
| GSpice-DEBUG: 22:12:43.460: channel-main.c:1362 main-1:0: Ignoring clipboard grab
| GSpice-DEBUG: 22:12:43.460: spice-channel.c:801 main-1:0: Send error Error sending data: Broken pipe
| GSpice-DEBUG: 22:12:43.460: spice-channel.c:837 main-1:0: Closing the channel: spice_channel_flush 32
| GSpice-DEBUG: 22:12:43.460: spice-channel.c:2680 main-1:0: Coroutine exit main-1:0
| GSpice-DEBUG: 22:12:43.460: spice-channel.c:2871 main-1:0: reset
Several other channels have failures
| GSpice-DEBUG: 22:12:43.460: spice-channel.c:1066 usbredir-9:1: Closing the connection: spice_channel_read() - ret=0
| GSpice-DEBUG: 22:12:43.460: spice-channel.c:1066 record-6:0: Closing the connection: spice_channel_read() - ret=0
| GSpice-DEBUG: 22:12:43.461: spice-channel.c:1066 usbredir-9:0: Closing the connection: spice_channel_read() - ret=0
| GSpice-DEBUG: 22:12:43.461: spice-channel.c:1066 playback-5:0: Closing the connection: spice_channel_read() - ret=0
...
This leads to a reset on all channels, which eventually emits
SpiceSession::disconnected signal and your Remote-viewer is
calling spice_session_disconnect()
| GSpice-DEBUG: 22:12:43.471: spice-session.c:1994 session: disconnecting 0
And the session is disconnected a bit later
| GSpice-DEBUG: 22:12:43.477: spice-session.c:1994 session: disconnecting 0
| GSpice-DEBUG: 22:12:43.477: spice-channel.c:2888 main-1:0: channel disconnect 0
| GSpice-DEBUG: 22:12:43.477: spice-channel.c:159 main-1:0: spice_channel_dispose 0x565032b1a400
| GSpice-DEBUG: 22:12:43.477: spice-channel.c:2888 main-1:0: channel disconnect 12
| GSpice-DEBUG: 22:12:43.477: spice-channel.c:177 main-1:0: spice_channel_finalize 0x565032b1a400
| GSpice-DEBUG: 22:12:43.477: spice-session.c:1802 no migration in progress
| GSpice-DEBUG: 22:12:43.477: spice-session.c:324 session dispose
(...)
On broken pipe we might try to reconnect but might be hard if
Remote-viewer calls spice_session_disconnect(). Perhaps some more
checks needs to be done before emitting the
SpiceSession::disconnected ...
Btw, that's pretty new in remote-viewer code
https://pagure.io/virt-viewer/c/65ef66e42a6db2a9826fffef0db49920a02d358f
Thanks for helping/testing!
Victor
Attachment:
signature.asc
Description: PGP signature
_______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel