Return-Path: To: linux-hotplug@vger.kernel.org, linux-bluetooth@vger.kernel.org Subject: At wits end with a Broadcom Corp BCM92046DG-CL1ROM From: gene heskett Date: Mon, 3 Jan 2011 15:52:14 -0500 MIME-Version: 1.0 Content-Type: Text/Plain; charset="utf-8" Message-Id: <201101031552.15059.gheskett@wdtv.com> Sender: linux-hotplug-owner@vger.kernel.org List-ID: Greetings all; Ping? Repeat post with a few additions. This device, a bluetooth button/dongle, or the a7 eb101 it pairs with to create a virtual rs232 cable, have to be the most obstinate inanimate devices ever. AIUI the BT specs, 2 devices once paired, are to remain so, even after powerdowns, until such time as a new pairing is performed. Right/wrong? It takes about a days messing around to finally get it to work, usually accomplished by unplugging it long enough for linux to detect and do the cleanup. Then when plugging it back in, if one is quick enough to catch it before bluetoothd times out and goes away and gets blueman-manager running within this timeout, which appears to be in the 5 second range after its rediscovery, then, 1 time only per reboot of this machine, I can get it to work, and it works error free, with good signal strength showing in blueman-manager. For possibly 12 hours or so. At some point, blueman- manager goes away silently, and the link is dropped and cannot be re- established by the same procedure until I have rebooted both boxes again. Pertinent (maybe) info: current kernel: 2.6.37-rc8 current distro: PClos, 32 bit, 100% uptodate current hdwe: AMD quad core phenom, 4Gb dram, 4 Tb of drives current gui: KDE-4.5.4 I have edited /etc/bluetooth/rfcomm.conf to: # # RFCOMM configuration file. # # this was ALL commented out rfcomm0 { # # Automatically bind the device at startup bind yes; # # # Bluetooth address of the device device 00:0C:84:00:86:F8; # # # RFCOMM channel for the connection channel 1; # # # Description of the connection # comment "Example Bluetooth device"; comment "connection to coco3's eb-301 device"; } Added question: Pin seems to be a problem as I have to answer several requests for it, and it seems to work only if I press the enter key in the requester box, which seemingly has no effect, and then click the ok button, which puts 5 bytes in the buffer. And FWIW, I have noted that the 4 digit pin is stored as 5 bytes long no matter how I generate it, including deleting the file and building a new one. So my added question is, why is the return key not being accepted as the "go do it" key? And as of this instant /dev/rfcomm0 has not been removed. And the link is dead. And is again working after a reboot of both machines. No clue how long though. It seems to me there should be a foolproof method that will accomplish this from reboot to reboot, but I haven't found the method yet despite warming up googles servers searching for instructional material that does NOT appear to exist or is so dated that the 2nd or third example in any of those tuts is no longer valid. What can I do, or better yet, URL's where can I find the info that makes this Just Work(TM) for a more modern 2010-2011 distribution using todays kernels? Thanks all. -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author)