Hello.
I'd like to ask, to @devbisme, the creator of such wonderful utilities and board designs: is there any means you could share here the codebase for the XSUSB firmware (i.e., the 'user' and 'boot' portions for the PIC18F4455 interfacing uC between the USB and LPT ports) ?
It would be so great to have it for me to analyze/ recompile in Ubuntu MPLAB-X, which I just did with the XuLA-2 firmware, included in this repo. But most importantly, I need the PIC source to be able to debug appropriately what might be going on with my personal "port" of XSTOOLS to the 64-bit Ubuntu realm and toolchain/s...
My purpose is to find means to analyze trouble I am having with my squeaky clean, GCC 5+, Ubuntu/Debian cleanly and smoothly "makeable", modern Autotools compatible tree of the XSTOOLS source I've painstakingly derived in recent weeks from the Windows XSTOOLS.
The code of these "clean port to modern 64bit Ubuntu" of XSTOOLS works allright except that: it seems it can't get to program the 'xessjtag.svf' successfully upon the XC9572XL CPLD via the PIC, e.g. via 'xstest' when an attached XSA-3S1000 has the CPLD witout the "<5>!" signature yet, or just lost it.
The capabilities, port initialization, reading of CPLD previous USERCODE after flashing the '.svf' on Windows environment (which works), and all that works for this Ubuntu "port". It's just the new '.svf' writting phase that is breaking, and I don't know why. In Windows this step works smoothly, on Linux it doesn't somehow...
To sum up:
- I've got a squeaky clean, Autotools-compliant, modern 64bit Ubuntu/ Debian source tree of XSTOOLS working for me (except the support for XSUSB)
- XSUSB utils e.g. like 'xstest' work all good if we try to flash 'xessjtag.svf' on CPLD of XSA-3S1000 via de XSUSB, in Windows system.
- I've tried to "port" / recode the 'xsusb' alongside my code in point 1) above, but doesn't work. If the CPLD was just flashed in Windows, then the Linux port for 'xstest' reads good the '<5>!' signature, but if it was not there to begin with, the Ubuntu/ Linux code is unable to write the '.svf' (same exact .svf) via LIBUSB.
- All other LIBUSB calls and stuff seem to work great (e.g. it reads the signature when it's there, things like that), so it's not the issue some people refer to when comparing e.g. LIUSB 1.x vs. LIBUSB 0.x, difference which is known to be a PITA for some Linux environments.
Thank you for all insight you all can give me about this.
Regards from Spain