Hi all:
Since kernel.org/bugzilla is dead for a while, I sent the bug report
here in hope to get more insight.
This is a long-lasting bug and happens across several distros with
different combinations of bluez/kernel verisions but shares much the
same symptom that bluetooth is not working after the resume. As seen
in the bug report below, it happens on a variety of different laptops
too.
Redhat:
https://bugzilla.redhat.com/show_bug.cgi?id=727106
Debian:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=614929
Ubuntu:
https://launchpad.net/bugs/824144
https://launchpad.net/bugs/824633
https://launchpad.net/bugs/823467
https://launchpad.net/bugs/822705
Personally I did some investigation on the issue and found that
bluetoothd does not receive the HCI commands of read_bd_addr,
read_local_version, read_local_features, nor read_local_name. So in
bluez.plugins/hciops.c:device_devup_setup(), init_adapter() is not
called at resume.
It can be further confirmed by the attached patch of bluez. It is
ugly but can work around the bug as reported in some of the bug
trackers above.
I am looking for any advice to get it move on, for the debugging
and the fixing.
Thanks,
-kengyu
--- a/plugins/hciops.c
+++ b/plugins/hciops.c
@@ -2512,8 +2512,8 @@
hci_send_cmd(dev->sk, OGF_HOST_CTL, OCF_READ_STORED_LINK_KEY,
READ_STORED_LINK_KEY_CP_SIZE, &cp);
- if (!dev->pending)
- init_adapter(index);
+ DBG("dev->pending=%d", dev->pending);
+ init_adapter(index);
}
static void init_pending(int index)
