It appears that my mistake was bringing up the interface manually at the console before trying to enable bluetooth via the UI. It works if I don't do that.
Paired it with my cell phone, and tried transferring a file, but the connection failed. I don't know much about bluetooth, but I wonder if Android is set up for file transfer(a profile issue?), or maybe there is a permission problem somewhere. I've had issues with those before. Going to see if I can figure out where to look, pointers would be appreciated. Thanks for your help in getting it going. Mike On Thu, Oct 21, 2010 at 6:16 PM, Pavan Savoy <pavan.sa...@gmail.com> wrote: > I thought you removed the whole hciattach starting from the bluedroid, > because you have USB device? > hciattach is for UART devices... > > or you can have your init.c say hciattach -l for the hciattach service. > The important step I suppose here is ioctl HCIDEVUP.. so leave the > ioctl alone and try commenting hciattach, I guess that should work.. > > On Thu, Oct 21, 2010 at 3:45 PM, MikeS <mstic...@gmail.com> wrote: > > > > > > On Thu, Oct 21, 2010 at 2:20 PM, Pavan Savoy <pavan.sa...@gmail.com> > wrote: > >> > >> I suppose the basic BOARD_HAVE_BLUETOOTH flag is already set? > >> and the JNI layer actually make calls to bluedroid !!! > >> > > Yep, BOARD_HAVE_BLUETOOTH is set. > > > >> > >> Can you put in couple of debugs in the > >> /frameworks/base/core/jni/android_server_BluetoothService.cpp" ? just > >> to make sure you are in-fact building the right things? > >> > > Did this, and now it is calling bluedroid (not sure why it didn't before > > ... ) > > Error I get is now > > E/bluedroid( 895): bt_enable: Timeout waiting for HCI device to > come > > up > > > > This means the ioctl call to the socket isn't succeeding. > > BUT, just above this is the following comment: > > // Try for 10 seconds, this can only succeed once hciattach has > sent > > the > > // firmware and then turned on hci device via HCIUARTSETPROTO > ioctl > > > > So when I commented out hciattach, it looks like I also disabled some > > initializing of the device. > > Will go try and hunt down HCIUARTSETPROTO in hciattach and what it is > doing > > to "turn it on". > > I'm assuming (hoping) that the socket is going directly to hci0, and I > just > > have to set something > > via the ioctl to turn it on. > > > > Mike > > > > > > -- > --Pavan Savoy > -- unsubscribe: android-porting+unsubscr...@googlegroups.com website: http://groups.google.com/group/android-porting