RE: [lwip-users] LWIP -> Microblaze

[Lachlan Grogan]

Thanks Christiaan,

I need to clarify that Im using Xilinx Micro Kernel and using the MAC in
Interrupt Mode so the acutaly contrib file Im after is
Xemacif_intr_xmk.c.  (I believe this is copied and renamed as xemacif.c
during the libgen process).

My device implements a web site very similar to that of a residential
gateway device, however the device does not rely on the network connection
for any serious amounts of data transfer.  

In your opinion, would I be better looking at using the RAW_API?

Many Thanks
Lachlan. 

-----Original Message-----
From: address@hidden
[mailto:address@hidden On Behalf Of
Christiaan Simons
Sent: Wednesday, 26 April 2006 8:17 PM
To: Mailing list for lwIP users
Subject: Re: [lwip-users] LWIP -> Microblaze


Lachlan wrote:

> Q:  Can anyone point me to an updated version of xemacif.c that would 
> correct these problems. ?

The authorative source is the contrib/ports/v2pro/netif/xemacif.c.
If that file is broken, we want to fix it there. I can have a quick peek,
but I can't guarantee it gets fixed. (I don't have the Xilinx development
environment)

> Q:  Has anyone experienced similar latency problems in the
read(...)function.?

I know the sequential / socket API performs slow and has some unclear bugs
and design flaws.
See also http://savannah.nongnu.org/bugs/?func=detailitem&item_id=15926

What I can tell fore sure this code is slow because there are too many
thread synchronisation waitstates.

When using the lower level raw-API, the achievable performance levels are
much better. Some users have implemented and alternative socket layer on top
of the raw-API.

Christiaan Simons

Hardware Designer
Axon Digital Design

http://www.axon.tv



_______________________________________________
lwip-users mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/lwip-users