AVR Freaks

If there's been a bug report, it's likely to have been fixed, but
I'm not sure. PDI on the AVRISP mkII is one of the few things I
didn't test, but I've tested PDI on the STK600, the AVR Dragon,
and the JTAG ICE mkII.

Sorry, I had to re-release AVRDUDE, so there's now version 5.10 out.

The reason was that 5.9 had a serious bug in handling intel hex files
with more than 128 KiB of flash contents. This has been fixed.

Also fixed are Cygwin build for recent Cygwin versions where the compiler
doesn't implement the -mno-cygwin option anymore. This situation is
detected and handled properly, although it's suggested to resort to the
mingw32-gcc compiler (from the Cygwin collection!) instead, in order to
avoid getting a depdency from cygwin1.dll.

Finally, the JTAG ICE now attempts to pull the external reset pin if the
initial JTAG activation failed with an "Invalid JTAG ID". This allows
easy programming of target MCUs that are asleep, or that have disabled
the JTAG interface through JTD at run-time.

(I edited the thread title to make everyone aware of the version number
bump.)

I've just tested it on an STK600, using the ATmega2560 piggyback
board the STK ships with. I cannot get it to communicate at
8 MHz ISP clock at all, even if I change the CPU clock to 20 MHz.
Presumably, the ISP path between the ISP connector and the
controller itself causes enough reflections and signal distortions
to be more of a limiting factor than the actual programmer speed.

With 4 MHz ISP clock, I can write 64 KiB of flash within 4.36 s.
It doesn't make a difference whether I use the STK600 itself
then or an external AVRISPmkII (which is not a big surprise as
both are identical in their communication and ISP firmware blocks,
for all practical purposes).