Re: [Bluez-users] l2cap connection timeout

[Marcel Holtmann <marcel@holtmann.org>]

Hi Uichin,

> I just realized from the spec that BT has C-Plane (for device control
> service, i.e., HCI functions) and U-Plane (transport control and data
> services, etc.). I thought I could use HCI functions not only for
> device discovery, but also for connection and data transfer instead of
> using L2CAP. But according to this, although we can create a
> connection through hci_create_connection(), this connection can't be
> used for data transfer, thus useless. (is it true?)

you can use low-level ACL channels to transport data if you control both
sides of the connection. In general it is useless and hard to do it
right. The only user of ACL low-level connection should be L2CAP.

> When we call L2CAP connection, it does in fact just create a L2CAP
> connection request message and send it via hci_send_acl() to the lower
> layer (i.e., LMP??). BlueZ kernel simply manages this L2CAP connection
> information as far as I understand correctly.

The L2CAP is exposed as a SOCK_SEQSTREAM to the user space and really
easy to use. It takes care of everything inside HCI.

> When I try to connect to a remote device after inquiry step, how can I
> choose a RIGHT timeout value to the remote device? I tried various
> values from 1s to 5s: sometimes, it works well with 2s, but sometimes
> it doesn't. Spec. talks about SR mode, R0 (<1.28s), R1(<1.28s),
> R2(<2.56s) for page scan and thus accordingly, page for asynchronous
> link should be N_page>1, >128, >256 respectively. Will it be pretty
> much the same as the inquiry procedure, or should it be faster than
> that since we've already got the clock offset from inquiry?

I will be a little bit faster if you connect directly after an inquiry,
but you don't have to worry about this details. The kernel keeps track
of these things and if useful, it will use them, otherwise it won't.
There is nothing much you can influence and actually it never gave any
big advantage in real life.

I am also guessing that some chips are correcting wrong values in the
background for you. Only guess.

Regards

Marcel



-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Bluez-users mailing list
Bluez-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-users