joel.sherrill at oarcorp.com
Thu Dec 6 19:04:52 UTC 2007
Allan Hessenflow wrote:
> Hoan Hoang wrote:
>> I'm planning to make a BSP for the EZKIT-BF561 but haven't got time to
>> get there yet also I don't have a GDB compatible Jtag. I have the
>> HPUSB-ICE but it not compatible with GDB. I'm looking at USBProg, don't
>> know if that could work. The price tag for the Icebear is too high and
>> the Igloo is too slow. Any recomandation ? or anyone know how to make
>> the HPUSB-ICE to work with GDB.
> I have no idea on the HPUSB-ICE. The only JTAG pods I own myself are
> Xilinx, one parallel and one USB. From the discussions at
> blackfin.uclinux.org it sounds like there's some hope for the Xilinx
> parallel port one (the stuff in the google cache I found are only
> reports of problems, with an implication that it is supported. At the
> moment blackfin.uclinux.org seems to be dead so I can't look there
> further). However I doubt that would be any faster than the Igloo.
>> I think it is a good idea to have a common Blackfin architecture to make
>> the port easier for other Blackfin processors as well. Can we use the
>> .h defined in VDSP or may be base on that?
> I don't have VDSP and so haven't seen what they did, but I suspect
> there may be license issues with using that.
> Joel Sherrill wrote:
>> Without a lot other than experience on other architectures
>> to go by, I would lean to the second option -- a base
>> architecture .h and a model specific one. It could even
>> be structured like this:
>> + define feature flags
>> include bfin shared header file
>> add more definitions
>> Then BSPs include a specific bfXXX.h and the shared code
>> for maintenance is invisible to them.
> This brings up one more point. Nothing in cpukit cares about any of
> the peripherals that are currently defined in bf533.h. I assume the
> file is in there because the peripherals are on the same chip as the
> cpu, but I think it would make more sense to move those definitions
> somewhere else, like under libbsp/bfin/shared.
Sounds like a logical alternative to me. It may also
be logical to place it in libcpu/bfin.
More information about the users