Hi, >-----Original Message----- >From: Prabhakaran M.C [mailto:mcprabhakaran@xxxxxxxxx] >Sent: Thursday, August 05, 2010 12:38 PM >To: Rymarkiewicz Waldemar >Cc: linux-bluetooth@xxxxxxxxxxxxxxx >Subject: Re: L2cap Security And Role Switch > >Hi Waldek, > >On Thu, Aug 5, 2010 at 3:37 PM, ><Waldemar.Rymarkiewicz@xxxxxxxxx> wrote: >> Hi, >> >>>-----Original Message----- >>>From: linux-bluetooth-owner@xxxxxxxxxxxxxxx >>>[mailto:linux-bluetooth-owner@xxxxxxxxxxxxxxx] On Behalf Of >>>Prabhakaran M.C >>>Sent: Wednesday, August 04, 2010 4:09 PM >>>To: linux-bluetooth@xxxxxxxxxxxxxxx >>>Subject: Reg: L2cap Security And Role Switch >>> >>>Hello All, >>> >>> Whenever L2cap security is HIGH and remote device does role switch, >>>Bluez accepts the Role switch and L2cap disconnects the channel >>>because of HIGH security. >>> >>> For PAN profile, I would like to keep the L2cap security to HIGH >>>since it involves internet browsing but the Widcomm stack >always does >>>a role switch in PAN connection and Bluez disconnects l2cap channel. >>> >>> Can someone please point in specification about the l2cap security >>>level and Role switch relation. I tried to find out this but I could >>>not get this behavior described in specification. Please >provide your >>>comments and inputs. Thanks in Advance. >>> >>>Thanks, >>>Prabhakaran. >>>-- >> >> Note that HIGH sec level requires encription on the link. >Role switch procedure turn off the encription before it >starts switching roles and turn on it again after all. In 2.1 >spec the controller handles switching off/on encription (pause/resume). >> >> Thanks, >> /Waldek > > From the logs, the link was authenticated and encrypted. >Then Widcomm stack disables the encryption, does a role >switch, enables the encryption. >After role switch bluez kernel disconnects l2cap channel due >to HIGH security. I just want to know where this disconnection >part is defined in spec. Or Bluez has to just reject the role >switch operation instead of disconnection? > >-- >Thanks, >Prabhakaran. As far as I know it's not defined in the spec. It's simply Bluez design. What I would like to see, the bluez should block outgoing data flow in l2cap for the period of role switch. It's done in rfcomm this way, I guess (?). I agree, in my view the current desing in this use case is not perfect. Thanks, /Waldek -- To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
