[En-Nut-Discussion] TCP FFlush slow on Ethernut 3

Ole Reinhardt ole.reinhardt at embedded-it.de
Sun Feb 15 21:54:53 CET 2009

Hi Martin,

> That explains the difference between ROM and RAM. I don´t really know what
> all is needed for tcp but i suppose the time (even in RAM) doesn´t  depend
> on caculations the cpu does. May be something waits there for an event or
> so. Why does it took first times 8ms and then 2 ms for the same
> transmission. 

Indeed that's quite interesting. One Idea that's still rumoring in my
head might be "NutGetMillis()". I don't know if it's used directly or
indirectly in the TCP stuff. I noticed some problems concerning the time
handling functions on the Ethernut3 board some time ago. The problem
was, that everytime a time handling function was called, the RTC clock
chip was queried by I2C if the CPU frequence was not configured hard by
the configurator but was calculated dynamicaly.

I don't remember if Harald had changed something in the time framework
since then. I solved the problem by defining the CPU frequency as
constant value in the configurator.

Just an idea...



|                                                             |
| Embedded-IT          Hard- und Softwarelösungen             |
|                                                             |
| Ole Reinhardt        Tel. / Fax:        +49 (0)271  7420433 |
| Luisenstraße 29      Mobil:             +49 (0)177  7420433 |
| 57076 Siegen         eMail:    ole.reinhardt at embedded-it.de |
| Germany              Web:         http://www.embedded-it.de |
|                      UstID / VAT:       DE198944716         |

More information about the En-Nut-Discussion mailing list