[uClinux-dev] espd 4510b/samsung s3c4510b related network problems
greghol at sbcglobal.net
Thu Aug 4 16:42:47 EDT 2005
Why do you want to run the board in big endian mode? Stap the 4510B in
little endian mode and be done with it. Is there something that I'm am
missing in understanding what you really need to run big endian?
The EPSD and SND100 little endian builds runs on all of the 4510B little
endian boards I've ran into with the little endian tools and libs. I had one
big endian board and I strapped to little endian and it work also.
----- Original Message -----
From: "Håkan Kvist" <hagar at df.lth.se>
To: <uclinux-dev at uclinux.org>
Sent: Thursday, July 21, 2005 4:38 AM
Subject: [uClinux-dev] espd 4510b/samsung s3c4510b related network problems
I have modified the ESPD 4510b target (little endian) so that it works
on a big endian custom board.
There were only some structs that had to be define in opposite order
in big endian mode...
The SoC's ethernet works in the beginning...
incomming ethernet data is placed in a ringbuffer by the hardware, and
interrupts are generated.
The buffer can hold 16 packet at a time.
When the number of packets in the buffer are < 16 everything works fine.
When the buffer is full, everything still looks ok, but after the
interrupt routine has processed the 16 elements in the buffer, no more
ethernet interrupts are generated.
Outgoing data still works fine though (arp requests are still beeing
Does anyone know if the original ESPD 4510b sources running on the
original ESPD 4510b board behaves in the same way?
I'm using the lates uclinux-test, kernel replaced with the 2.6.9-hc0
kernel and gcc-3.4.1 compiled with softfloat,big endian options
(built with crosstool)
uClinux-dev mailing list
uClinux-dev at uclinux.org
This message was resent by uclinux-dev at uclinux.org
More information about the uClinux-dev