This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Samsung SNDS100 Eval. Board...


On Tue, Oct 15, 2002 at 01:03:51PM -0700, Jay Foster wrote:

> I'm beginning a new project with the S3C4510 part and have the
> SNDS100 eval board running RedBoot and the hello world eCos
> application.  My impression of the processor are that the UARTs
> are feature poor

Yup.  No modem control lines, no flow control, no FIFOs.  We
use them for console/debugging purposes and add external UARTs
for "real" application use.  Like you said, it depends on the
application.

> My SNDS100 eval board came with NO documentation.  I was able
> to download a schematic (for the old board) to get some idea of
> what's on it and how to configure it.

I've got some example code that is (I believe) the pSOS BSP
stuff for the older SNDS100.  I've also got what I think is
source for some sort of board test SW for the SNDS100, but it's
really badly written. 

> Michael - you stated that you found a few bugs with the KS32C5000.  

Actually I think it was I who said that.  The main problem we
found in the '5000 (now called the S3C4500) is that the DMA
controller associated with the Ethernet MAC will corrupt a
receive packet once in a while. This happens after the Ethernet
CRC has been validated.  We had to add a software CRC check of
receive packets to our Ethernet driver.  This problem seems to
be fixed in the 4510.

Another minor bit of weirdness in the KS32C5000 is that _some_
of the parts had undocumented FIFOs on the UARTs.  That feature
seems to have "gone away" as silently as it appeared.

I _think_ I've seen the 4510 corrupt transmit packets (again it
appears to be the DMA controller, so they get corrupted before
they get to the CRC logic in the MAC).  It's hard to get it to
happen in a predictable manner, and TCP (or UDP w/ checksums)
will take care of it anyway.

> If it's not too much bother could you reply with a short list
> or point me to the errata (if there is any) on the Samsung web
> site?

I've never been able to find squat on the Samsung web site.
It's a completely awful site.  Most of the time you can put in
valid Samsung part numbers and it finds 0 hits.

Another gripe: they (Samsung) never publish errata in a public
manner, and the FAE here in the 'States will always deny any
knowledge of any bugs in the parts.  If you're insistent enough
and threaten your distributor, you can usually get errata out
of them.

I've placed copies of the example source code and the errata I
have for the 4510 at ftp://ftp.visi.com/users/grante/Samsung.
Good luck trying to read them.  ;)

-- 
Grant Edwards
grante@visi.com

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]