XCore Exchange

StartKit looks great, I look forward to getting it up and running here.

I checked the documentation, but I didn't see the JTAG loop brought out on any connectors.

Assuming XTimeComposer will not run on my computer, how can I connect my existing JTAG programmer to this board?

Thanks,
LyleHaze

The startKIT has been designed as a very low cost board and hence the XTAG has been integrated into the xCORE device. Due to the design of the system there is no simple way to access the external JTAG pins on the device to allow it to be used with an external programmer.

"No simple way"?

Challenge accepted.

segher wrote:"No simple way"?

Challenge accepted.

Good stuff, as a hint to start with the JTAG pins are looped back to ports on tile 1 as well as to the external programming header.

Yeah saw that header (since a few minutes I have a board --
looks great, love the packaging too (but yet another logo
for XMOS?!), it's not plugged in yet).

The plan is first to reverse the USB protocol to get the board
to run at all -- I cannot run the host tools...

LyleHaze wrote:StartKit looks great, I look forward to getting it up and running here.

I checked the documentation, but I didn't see the JTAG loop brought out on any connectors.

Assuming XTimeComposer will not run on my computer, how can I connect my existing JTAG programmer to this board?

Thanks,
LyleHaze

What OS are you running?

@AJB2K3

I am running AmigaOS4, which is neither Windows, Mac, nor Linux compatible.
My current computer is an X1000, which is full ATX sized PowerPC board.
The main processor is a PASemi PA6t, but it also has a XS1-L02 from XMOS with an interface directly to the fast localbus of the PA6T. An integrated JTAG port allows us to re-program the built in XMOS chip (and any additional ones we add) on the fly without needing any adapters or programmers.

With lots of help from Segher, I was able to get the built in JTAG up and running, and we have command line and GUI based tools to operate that easily.

Even though we have a dedicated XMOS chip on every board, we can not compile any code for it locally, because the tools from XMOS all require an X86 processor, running Windows or Linux. :(

Now the next gen board is being designed. This time it looks like we will be going with a more available freescale processor, but still embedding an XMOS chip in each board, and a dedicated slot for accessing its available I/O.

I find it odd that while A-Eon is building each new computer with an XMOS chip and slot built in to each board, the compiler tools from XMOS ignore this family of computers completely.

I traced the traces... Now of course we have to convince core #0 to not drive
those pins. Haven't tried yet.

I'm not very sharp today it seems...

After the power pins (G 5 33) all the other pins
(M I S K D O R G H) of course line up perfectly
with the XSYS pins 3..19, and then another G.
The G H are the uart pins :-)

So yeah then R is reset; it doesn't seem to actually
go to the reset input of the device though. Curious.

Still haven't booted the device; it turns out I don't
have any microusb cables :-P

The plot thickens.

Still no cable so doing all this "dry", just the way I like it.

It turns out there is a hardware manual for the startkit; I've
looked for it but couldn't find it on the XMOS website, but
there of course always is google. The manual contains the
schematic.

It seems the UART pins on XSYS are no more; TX (on the
target side) now is another GND (G on startkit), and RX
now is called OTP_HOLD (H on startkit), but not connected.

The debug core uses all 1-bit ports, and in a nice sane order,
except it skips 1I (XnD24, ball V13), and in fact that seems
to be the only pin missing from the schematic.