On Wed, Jun 11, 2014 at 08:58:08PM +0200, Paolo Bonzini wrote: > Il 11/06/2014 19:25, Hani Benhabiles ha scritto: > >Due to an incomplete initialization, adding a usb-bt-dongle device through > >HMP > >or QMP will cause a segmentation fault. > > > >Signed-off-by: Hani Benhabiles <h...@linux.com> > >--- > > > >Not sure about the exact policy of qemu-stable. CC'ing it as this bug > >results in > >a segfault. > > > > hw/usb/dev-bluetooth.c | 6 ++++++ > > 1 file changed, 6 insertions(+) > > > >diff --git a/hw/usb/dev-bluetooth.c b/hw/usb/dev-bluetooth.c > >index a9661d2..6d02343 100644 > >--- a/hw/usb/dev-bluetooth.c > >+++ b/hw/usb/dev-bluetooth.c > >@@ -506,6 +506,12 @@ static int usb_bt_initfn(USBDevice *dev) > > > > usb_desc_create_serial(dev); > > usb_desc_init(dev); > >+ s->dev.opaque = s; > >+ s->hci = bt_new_hci(qemu_find_bt_vlan(0)); > >+ s->hci->opaque = s; > >+ s->hci->evt_recv = usb_bt_out_hci_packet_event; > >+ s->hci->acl_recv = usb_bt_out_hci_packet_acl; > >+ usb_bt_handle_reset(&s->dev); > > > All lines but the s->hci assignment should be removed from usb_bt_init too. > > As to s->hci, I suggest inlining usb_create_simple into usb_bt_init, and > initializing s->hci there before doing the qdev_init() call. > > Then here you can wrap the assignment under "if (!s->hci)".
I am afraid I don't quite understand what you want to achieve with this and why. Could you please explain how is usb_bt_init() relevant to this case ? Thanks > > Thanks for TLC of this little-used piece of code. > > Paolo > > > s->intr = usb_ep_get(dev, USB_TOKEN_IN, USB_EVT_EP); > > > > return 0; > > >
