Created attachment 123086 [details] backtrace Since gtask spicy hangs when it should be disconnected by another client Steps to reproduce: 1) connect with spicy client to a guest which has webdav 2) connect with another spicy client to the same guest Actual results: the first client hangs, the second client is connected Expected results: the first client is disconnected by the second client Additional info: There is already a FIXME in the code telling what is going on - "disconnecting in read_cancelled() causes a deadlock" [0] Fabiano's gtask-helper [1] avoids the issue, but imo we should rewrite the code to fit better to gtask [0] https://cgit.freedesktop.org/spice/spice-gtk/tree/src/vmcstream.c#n236 [1] https://lists.freedesktop.org/archives/spice-devel/2016-March/027561.html
> There is already a FIXME in the code telling what is going on - > "disconnecting in read_cancelled() causes a deadlock" [0] (...) > [0] https://cgit.freedesktop.org/spice/spice-gtk/tree/src/vmcstream.c#n236 The reference for the bug in the comment is: https://bugzilla.gnome.org/show_bug.cgi?id=705395
> The reference for the bug in the comment is: > https://bugzilla.gnome.org/show_bug.cgi?id=705395 This is an bug indeed and sadly needs good handling on cancel_cb in order to avoid the deadlock. I had the same issue around spice-pulse.c on the volume-sync stuff [2] https://cgit.freedesktop.org/spice/spice-gtk/tree/src/spice-pulse.c#n939 But that was before the gtask stuff. Let's see.
Fixed by: https://cgit.freedesktop.org/spice/spice-gtk/commit/?id=9e58e5024741da99dfda78496717e3863e8ef6bc
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.