I was referring to the other part of the hardware setup: the dongle, the wiring to the chip and the flash chip itself. Testing different setups would hopefully give us more insight into what's the real problem. There must be something odd going on... because in general it should work AFAIK.
Yes, sure, we have to test on a different external hardware too. Probably I'll do this later or someone else will do.
Not in general, no, it is certainly possible to support it correctly. Just not with this quick hack.
Unfortunately I hadn't said this before but this patch is not for release. It's just the first try to support the discussion about performance. It could be better to post it under that thread.
The loop in serprog is a bug not a feature. I have just not removed it
Ok :-)