Edward Bockhoefer wrote:
Thank
you for the reply David,
I have started the montavista 10/100 enet port from the xilinx edk
generated 2.4 driver to the 2.6 driver. Unfortunately, I can't post the
driver port publicly because it's not GLP'd, as far as I know. We only
needed a 10/100 mac, so we didn't bother with a TEMAC. I will work on
this more today. If you want to send me your TEMAC code, please do so.
I can always use the reference.
The original message for the GB TEMAC 2.6 patch is below. It is
from Andrei Konovalov.
I think it is just a newer version of the 2.4 Driver. Regardless
it just worked for me. I had to make a very minor change - I added an
"u32 spd" to the private data structure, initialized it to 10 (I am on
a 10Mhz Net), and found every reference to 1000Mhz replaced it with
lp->spd; There was one place I needed a switch statement to load the
appropriate value. I think there were only 3 really trivial changes. I
would post my changes but I am not setup to post them as a patch right
now. If you really want them I will just send my adaptor.c and you can
use everything else from the patch referenced below.
I have also attached my Local Link TEMAC code. That is basically a
single file driver. Again I am not able to provide it as a patch at the
moment. It (much like the PLB TEMAC driver) requires a platform data
entry for the LL TEMAC in your Board support code. Though it could be
fairly trivially changed to work from defines. I beleive there are only
2 critical values - the DCR base address, the TEMAC memory address.
There is an IRQ value, but I can only get the LL TEMAC to generate PHY
interupts, so the driver is polled. It does work albeit with lots of
dropped packets, but that is probably debugging code and optimizing the
poll rate.
The driver is solely my own work based loosely on several other
drivers, and is GPL'd. According to my client LL TEMAC performance has
tested abysmally in comparision to the PLB TEMAC, however the kicker
was the inability to generate rx interupts. We are working on other OS
ports, and some do not have strictly polled options, so for now the LL
TEMAC is on the back burner.
Of course I have learned an incredible amount about Linux Network
Drivers as well as now having a reasonable understanding of the
differences between the assorted Xilinx TEMAC's
-------- Original Message --------
Here:
http://source.mvista.com/~ank/paulus-powerpc/20060309/
is an StGIT patchset against
516450179454de9e689e0a53ed8f34b896e8651c
branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
virtex-ppc_sys_devices-fix.patch - is the same fix posted by Grant Likely few days ago (only formatting is different),
ppc32_xilinx_edk_temac.patch - the low level code from EDK-7.1i used by the this driver
ppc32_xilinx_edk_temac_mods.patch - few modifications to EDK-7.1i code to get rid of compiling the xparameters.h
#defines into the driver. These mods should not be needed after EDK-8.1 SP2 is released.
ppc32_xilinx_temac.patch - the linux driver itself
ppc32_xilinx_ml403_temac.patch - modifications to the ML403 platform code to add TEMAC.
temac.paulus-powerpc.20060309.tgz - all the patches put together into a tarball (just in case).
temac.config - .config used.
This is not the final version yet (no MII support; SGDMA is operational, but seems to need some additional work).
And I would probably wait for EDK-8.1 SP2 (up to one month or so) where both the TEMAC IP and the EDK code
should be updated.
But it would be nice to know if there is anything to rework in the current code to get the driver
accepted into linux/kernel/git/paulus/powerpc.git or linux/kernel/git/torvalds/linux-2.6.git tree.
And probably someone could start playing with the current driver now. At least I was able to mount
the root over NFS and to run several netperf tests.
Thanks,
Andrei
_______________________________________________
Linuxppc-embedded mailing list
Linuxppc-embedded@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
--
Dave Lynch DLA Systems
Software Development: Embedded Linux
717.627.3770 dhlii@dlasys.net http://www.dlasys.net
fax: 1.253.369.9244 Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.
"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein