Return-Path: Content-Type: text/plain; charset=iso-8859-15; format=flowed; delsp=yes To: "Peter Hurley" Cc: "linux-bluetooth@vger.kernel.org" Subject: Re: A couple of questions related to a rfcomm server Reply-To: santoscapela@gmail.com References: <1311524784.2657.21.camel@THOR> Date: Sun, 24 Jul 2011 23:08:00 +0100 MIME-Version: 1.0 From: =?iso-8859-15?Q?S=E9rgio_Capela?= Message-ID: In-Reply-To: <1311524784.2657.21.camel@THOR> Sender: linux-bluetooth-owner@vger.kernel.org List-ID: Hi Peter, On Sun, 24 Jul 2011 17:26:24 +0100, Peter Hurley wrote: > Hi S?rgio, > > On Wed, 2011-07-13 at 06:51 -0400, S?rgio Capela wrote: >> Hello, >> > ... >> Is it possible to accept connections at the same time a inquiry is >> running >> or is the inquiry a blocking operation at the stack level? With two >> adapters my server accepts remote connections and receives and sends >> data >> but when an inquiry is running it stops accepting connections. The >> inquiry >> is done in a separate thread. > > That's handled in the controller at the baseband level. I don't know of > any controllers that will accept connections while in the Inquiry > substate - I'm not even sure if it's possible. > > Why are you initiating discovery from the server-side? > Well, this is a test where I'm trying to see if content delivery is viable through bluetooth and in this case the content depends on context. Imagine a number of rooms which should serve differentiated content accordingly. The server side discovery would try to automatically infer which content should be delivered to a given possible client. For this purpose I perform inquiries with rssi and thus the servers would act like tracking devices which would transmit inquiry data to a central database. These are only trials for now and I know that inquiries with rssi values are not that good for this type of decision so in the future this operation can be dropped in favor of other solution even if it end up as a choice at the user side. > If you have to have server-side discovery, you could only perform > inquiry with one controller at a time. That way, the other controller's > available for new rfcomm connections (assuming your client load-balances > like that...). > Yep, that is what I'm doing. I bind one adapter to the rfcomm server and use the other one for discovery and content delivery. The main process thread is the accept server and both the inquiry and the content delivery are done in threads launched for the purpose. I also made more tests recently and what is happening is that as soon as the inquiry starts the server stops accepting connections and starts processing inquiry result events, however, after a couple of minutes I can see the server both accepting and processing new connections and new inquiry result events. The end result is that launching an inquiry leads me to loose possible connections during a few minutes which is excessive for this case where inquiries would be launched regularly. >> Another question is related to a rfcomm client getting into D state and >> thus not killable. Due to my setup (two adapters in the same computer) I >> can use the computer as both the server and the client. For test >> purposes >> only I made a simple rfcomm client and I put it in a shell cycle thus >> stressing the system regarding bluetooth connections. In this situation >> the rfcomm client application becomes unresponsive after some time due >> to >> a non exiting disk sleep. I then end up with multiple instances of the >> same application and I can't kill them. By launching the application >> after >> this I can see that the code blocks at a socket connection command >> entering in a disk sleep state from which it doesn't exit. When this >> happens I have to unplug the bluetooth adapter to get it working again >> since it starts to give timeouts with hciconfig commands like "hciconfig >> reset". >> >> Any ideas about this? The problem itself is not serious since it >> shouldn't >> arise in normal conditions but it does require a reboot because of the >> state D processes and makes the adapter unusable unless it is unplugged. > > Although I'm not certain that they would fix the problem you're having, > I just posted a patch series that resolves some issues with lost wakeups > in rfcomm and l2cap. They need testing. You'd need patches 1~4 of 7. > > Please feel free to reply direct regarding whether you can test these > and what that outcome was. I suppose those patches would be available in the git sources right? At the moment I can't test it but in a week or two I will try to compile the sources and unless I encounter some problem I will reply in this topic. Regards, S?rgio > > Regards, > Peter -- S?rgio Capela