lwip-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[lwip-users] With logs ... Re: Delayed Ack causing problems? Where to ca


From: Richard Barry
Subject: [lwip-users] With logs ... Re: Delayed Ack causing problems? Where to call tcp_nagle_disable()?
Date: Mon, 12 Sep 2011 15:31:30 +0100
User-agent: Mozilla/5.0 (Windows NT 5.1; rv:6.0.2) Gecko/20110902 Thunderbird/6.0.2

Hi Simon, et al,

I have not got the Microblaze hardware out again, but I have tried the
same code on an LPC1769 now, and get the same result.  I had to change
lwipopts.h slightly to reduce the RAM footprint, but the C code (barring
the MAC driver) and all other lwIP configuration is as per the FreeRTOS
simulator code you already looked at.

Attached are two logs, one using the raw http server with no
modification, and one with the insertion of the tcp_nagle_disable(pcb)
call, as per my previous post.  If its not obvious, 192.168.0.200 is the
target that is running lwIP, and 192.168.0.100 is the host running WinXP.

The point of interest is the difference in the timing between each
frame, as you will see.


Regards,
Richard.

+ http://www.FreeRTOS.org
Designed for Microcontrollers.
More than 7000 downloads per month.




On 09/09/2011 09:54, FreeRTOS Info wrote:
> Hi Simon,
> 
> 
> On 09/09/2011 09:39, Simon Goldschmidt wrote:
>> FreeRTOS Info <address@hidden> wrote:
>>> The simulator is exactly that - it is running the FreeRTOS code, which
>>> is normally real time, in a non real time windows environment.  It is
>>> simulated in the sense that the only way to get the algorithms to run as
>>> expected [..] is to have simulated time move much more slowly that
>>> actual time.
>>
>> That's OK with me, I only wanted to point out that the httpd does *not* need 
>> nagle to be disabled, it's fast with it enabled and doesn't gain anything 
>> from disabling it.
> 
> Yes - the windows simulator does not really tell me anything about the
> timing of the TCP/IP applications, so I didn't recognise there was a
> problem when using the simulator.  I only recognised there was a problem
> when I moved the code onto a proper platform, which happened to be a
> Microblaze.
> 
> http://www.freertos.org/Free-RTOS-for-Xilinx-MicroBlaze-on-Spartan-6-FPGA.html
> 
> Note the microblaze example does not yet include the command console, so
> is only using lwIP with its raw interface.
> 
> I can get the hardware out again and retry it with/without nagle, just
> in case there was something else in the mix that was causing the
> perceived problem.  The source code running on the Microblaze (barring
> the command console) is the same as that running in the simulator -
> other than the MAC driver, naturally.
> 
> ---
> 
> Regarding the default mbox sizes, there was a definite reason why I had
> to change them.  It was so definite I cannot remember off hand exactly
> why the change was needed, but removing the new definitions would show
> it up again right away.
> 
> Regards,
> Richard.
> 
> + http://www.FreeRTOS.org
> Designed for Microcontrollers.
> More than 7000 downloads per month.

Attachment: get_logo_disabled_nagle.pcap
Description: Binary data

Attachment: get_logo_with_nagle.pcap
Description: Binary data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]