missing ack in modified xilinx sockets example

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

missing ack in modified xilinx sockets example

kyle treleaven
Hi folks,

My problem manifests itself first as a missed ACK at the end of a tcp
connection handshake.  I'm writing the client, so it goes SYN, gets a
SYN ACK response, and only ACKs after a second SYN ACK from the server
approx. 3-4 seconds later.  I used Ethereal to determine that.  This
symptom is 100% reliable-- it happens every time.  I also receive an
abornormally high amount of retransmit traffic, which I'm led to
believe stems from the same fundamental problem.  My first guess is
that transmit data gets shoved into a buffer somewhere and then
summarily forgotten until some kind of other event occurs.

I'm using the Xilinx port of lwIP to the PPC405, and Xilkernel OS.
I've modified the AppNote example directly, if anybody here has
experience with that.  In fact, I should probably talk to Xilinx about
this, but I figured I'd try here.

I guess I should mention that the program returns "successful" from
lwip_connect after the first SYN ACK response from the server, but
BEFORE actually transmitting an ACK.  That's what leads me to my
buffer theory.

Anybody else seen this kind of thing?

Thanks,
Kyle


_______________________________________________
lwip-users mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/lwip-users
Reply | Threaded
Open this post in threaded view
|

Re: missing ack in modified xilinx sockets example

Sathya Thammanur
Hi Kyle,
I havent seen this issue with the example. It would be better to contact Xilinx with the symptoms that you mentioned.
 
Sathya


 
On 3/11/06, kyle treleaven <[hidden email]> wrote:
Hi folks,

My problem manifests itself first as a missed ACK at the end of a tcp
connection handshake.  I'm writing the client, so it goes SYN, gets a
SYN ACK response, and only ACKs after a second SYN ACK from the server
approx. 3-4 seconds later.  I used Ethereal to determine that.  This
symptom is 100% reliable-- it happens every time.  I also receive an
abornormally high amount of retransmit traffic, which I'm led to
believe stems from the same fundamental problem.  My first guess is
that transmit data gets shoved into a buffer somewhere and then
summarily forgotten until some kind of other event occurs.

I'm using the Xilinx port of lwIP to the PPC405, and Xilkernel OS.
I've modified the AppNote example directly, if anybody here has
experience with that.  In fact, I should probably talk to Xilinx about
this, but I figured I'd try here.

I guess I should mention that the program returns "successful" from
lwip_connect after the first SYN ACK response from the server, but
BEFORE actually transmitting an ACK.  That's what leads me to my
buffer theory.

Anybody else seen this kind of thing?

Thanks,
Kyle


_______________________________________________
lwip-users mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/lwip-users


_______________________________________________
lwip-users mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/lwip-users