I encounter a wierd problem when testing my cell-phone via usb-redirect. Test environment: Client OS: win7-32bit Client: remote-viewer VM OS: win7-32bit Tool-chain: zadig-2.1 (use libusbk) usbredir-0.7 spice-gtk-0.27 (use gtk-2.0) virt-viewer-0.6 (use gtk-2.0) Test device: IPhone 5s Google Nexus 3 Problem: It seems that almost all USB device can be redirected successfully included cellphone. However, when my IPhone appeared in the VM OS, I can double click the phone icon, see its storage card. Then, I clicked 'Trust' button on my phone. Then, the whole window of remote-viewer is stuck, and cannot be closed normally( one way to close it is kill the process from task manager). The same problem can be reproduced on Android phone, take my Nexus 3 as an example. The phone screen is locked with pattern lock before redirect, and unlock it after redirect. Then, remote-viewer is stuck. Accidentally, I found a way to fix the bug. I add a printf in the beginning of usbredir_handle_msg() function, as below: static void usbredir_handle_msg(SpiceChannel *c, SpiceMsgIn *in) { SpiceUsbredirChannel *channel = SPICE_USBREDIR_CHANNEL(c); SpiceUsbredirChannelPrivate *priv = channel->priv; device_error_data data; int r, size; uint8_t *buf; g_return_if_fail(priv->host != NULL); /* No recursion allowed! */ g_return_if_fail(priv->read_buf == NULL); buf = spice_msg_in_raw(in, &size); priv->read_buf = buf; priv->read_buf_size = size; printf("msg\n"); r = usbredirhost_read_guest_data(priv->host); I don't know why it works. And I try to use sleep(0) to instead, it doesn't work. Then, use sleep(1), it works, but decreased the performance of handling usb message.
Is it an issue similar to http://lists.freedesktop.org/archives/spice-devel/2014-December/018408.html ? According to http://lists.freedesktop.org/archives/spice-devel/2014-December/018430.html this is a libusb bug which is being fixed upstream.
(In reply to Christophe Fergeau from comment #1) > Is it an issue similar to > http://lists.freedesktop.org/archives/spice-devel/2014-December/018408.html > ? According to > http://lists.freedesktop.org/archives/spice-devel/2014-December/018430.html > this is a libusb bug which is being fixed upstream. Yes, they are same. Thanks a lot!
Ok, closing this bug then, thanks for the followup!
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.