L2CAP Flush_Timeout Error

L2CAP Flush_Timeout Error

[Ajay]

Hi,
     Im currently working in an A2DP stuff, where i noticed an error. 
Here i wanted to monitor the flushed packets due to timeout.

  I used the  l2test code for my testing purpose

               On client side   "l2test  -R  -S <bdadddr> "
               and i set the flush timeout as 100 (opts.flush_to = 100)

        But  later  when i print the getsockopt value of flush_to, it 
was showing some default value (65532)  harcoded at kernel side.

     so  how can i flush some packets in the sender side,  without 
enabling any retransmission  and  able to monitor the sequence number in 
the receiver side.


   --
  Thanks & regards

AJAY KV
8892753703

Re: L2CAP Flush_Timeout Error

[Ajay]

On Monday 08 October 2012 02:41 PM, Andrei Emeltchenko wrote:
> On Tue, Oct 02, 2012 at 11:48:56PM +0530, Ajay wrote:
>> Hi,
>>      Im currently working in an A2DP stuff, where i noticed an error.
>> Here i wanted to monitor the flushed packets due to timeout.
>>
>>   I used the  l2test code for my testing purpose
>>
>>                On client side   "l2test  -R  -S <bdadddr> "
>>                and i set the flush timeout as 100 (opts.flush_to = 100)
>>
>>         But  later  when i print the getsockopt value of flush_to, it
>> was showing some default value (65532)  harcoded at kernel side.
>
> Try the patch in this thread.
>
> Best regards
> Andrei Emeltchenko
>

    Thanks for your patch , its working fine. Now even  if i set the 
flush_timeout = 5 , i couldnt  able to see any packets missing at the 
receiver side . Here in the sender  side  im giving proper sequence 
number, and these same sequence is getting at the receiver side . i 
doubted even though its flushing , controller is doing some 
retransmission job...please help . My kernel is 2.6.39 , and im using 
BASIC MODE of transmission. so shall i change it to STEAMING MODE ????

-- 
  Thanks & regards

AJAY KV
8892753703

Re: L2CAP Flush_Timeout Error

[Ajay]

> On Tue, Oct 02, 2012 at 11:48:56PM +0530, Ajay wrote:
>> Hi,
>>      Im currently working in an A2DP stuff, where i noticed an error.
>> Here i wanted to monitor the flushed packets due to timeout.
>>
>>   I used the  l2test code for my testing purpose
>>
>>                On client side   "l2test  -R  -S <bdadddr> "
>>                and i set the flush timeout as 100 (opts.flush_to = 100)
>>
>>         But  later  when i print the getsockopt value of flush_to, it
>> was showing some default value (65532)  harcoded at kernel side.
>
> Try the patch in this thread.
>
> Best regards
> Andrei Emeltchenko
>

    Thanks for your patch , its working fine. Now even  if i set the
flush_timeout = 5 , i couldnt  able to see any packets missing at the
receiver side . Here in the sender  side  im giving proper sequence
number, and these same sequence is getting at the receiver side . i
doubted even though its flushing , controller is doing some
retransmission job...please help . My kernel is 2.6.39 , and im using
BASIC MODE of transmission. so shall i change it to STEAMING MODE ????

-- 
  Thanks & regards

AJAY KV
8892753703

[PATCH] Bluetooth: Allow to set flush timeout

[Andrei Emeltchenko]

From: Andrei Emeltchenko <andrei.emeltchenko@intel.com>


Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
---
 net/bluetooth/l2cap_sock.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/net/bluetooth/l2cap_sock.c b/net/bluetooth/l2cap_sock.c
index 0c05d7d..70578d8 100644
--- a/net/bluetooth/l2cap_sock.c
+++ b/net/bluetooth/l2cap_sock.c
@@ -536,6 +536,7 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, char __us
 		chan->fcs  = opts.fcs;
 		chan->max_tx = opts.max_tx;
 		chan->tx_win = opts.txwin_size;
+		chan->flush_to = opts.flush_to;
 		break;
 
 	case L2CAP_LM:
-- 
1.7.9.5

Re: L2CAP Flush_Timeout Error

[Andrei Emeltchenko]

On Tue, Oct 02, 2012 at 11:48:56PM +0530, Ajay wrote:
> Hi,
>     Im currently working in an A2DP stuff, where i noticed an error.
> Here i wanted to monitor the flushed packets due to timeout.
> 
>  I used the  l2test code for my testing purpose
> 
>               On client side   "l2test  -R  -S <bdadddr> "
>               and i set the flush timeout as 100 (opts.flush_to = 100)
> 
>        But  later  when i print the getsockopt value of flush_to, it
> was showing some default value (65532)  harcoded at kernel side.

Try the patch in this thread.

Best regards 
Andrei Emeltchenko