Hi Ytai, got one step further, so the program now runs. The checkout from the App-IOIO0400 tag gave me the same problem, but checking out from trunk after your 1273 commit, I got the bootloader working with the original app from the board. By examining the difference between the working and non-working hexfiles, I found that programming address 54E20 with the value 3C instead of unprogrammed, got everything working, haven't checked yet what it is, maybe USB clock calibration value.
But I have one big issue left, I am not able to debug using the Pickit 3. I can use it for small programs like the blinker, but if I try starting the bootloader, it fails during the chip programming (I know the trick, and have to use it, moving the highlight from the Pickit 3 serial number to the Pickit3 in the project properties !!!). If I use the 3 "Descrete Debugger Operation" steps instead of just "debug project", all three are successful, but there is no symbolic info/linenumbers loaded. Single stepping goes from address 0->2->0->2 endlessly. Clicking the "set PC at cursor" does not change the PC. I suspect it has something to do with the protection mechanism you have made for avoiding overwriting the bootloader via IOIODude/IOIOManager. So if you have any advice or hints to get debugging functioning, it will be greatly appreciated. I will probably end up having only the app, not the bootloader in the final program, but I am not that far yet. Best regards, Johny Den tirsdag den 7. januar 2014 02.08.56 UTC+1 skrev Ytai: > > Your build process is correct, this is what I meant. Personally, I > normally just build all the libraries in all the configurations (there's an > option in the "Run" menu for batch build) and not have to worry about it, > or otherwise use tools/make-all to do that from command line. A later > version of make-all, which I haven't yet pushed even auto-generated the > makefiles from the project files, but otherwise you just need to build once > from MPLAB to get them generated. > > Anyway, if you're getting different hex files than I did, the options are: > > - You're building from different sources. I propose you sync against > github and use the tagged versions. For example, I built V4.00 of the > firmware from the App-IOIO0400 tag, etc. By "sources" I also mean the > project configurations, etc. > - Your compiler is different in some way, for example, you're using a > version that has optimizations disabled or something. > > > > On Mon, Jan 6, 2014 at 10:30 AM, Johny Johansen > <[email protected]<javascript:> > > wrote: > >> Hi Ytai, >> >> thanks for the quick reply, and thanks for the great IOIO, have been >> flying with this a couple of years in my glider. >> >> Your suggestion is actually how I started, but it would not link. >> >> I have the sources in SVN, so just made a fresh checkout with the >> original sources, and compiled all libraries without problems. But when >> building device_bootloader, I get: >> c:\programmer\microchip\mplab c30\bin\bin\..\bin/pic30-coff-ld.exe: >> cannot find -lusb >> >> The linker command has the following: >> -l"usb",-L"../libusb/dist/PIC24FJ256GB206_CDC/production" >> >> but when examining the libusb project, its active configuration >> is PIC24FJ128DA106 >> So I don't see how I can build without changing this MPLAB project >> setting. >> >> Just to be sure, I downloaded the ioio-master.zip file once more, same >> result. >> >> Then tried setting the libusb active conf. to PIC24FJ256GB206_CDC, and >> rebuilding libusb and device_bootloader. Now it links OK. >> >> For the following projects, I also had to change the active >> configurations like this: >> AppLayerV1: IOIO0030 >> libconn, libbtstack and libadb: PIC24FJ256GB206 >> >> Recompiling these 3 libs and the AppLayerV1 now results in: >> c:\programmer\microchip\mplab c30\bin\bin\..\bin/pic30-coff-ld.exe: >> cannot find -lusb >> >> Having this in the linker command: >> -L"../libusb/dist/PIC24FJ256GB206_OTG/production" >> >> So it seems the bootloader and app requires two different libusb configs. >> Is this correct? >> >> So changed the libusb conf. now to PIC24FJ256GB206_OTG, and rebuild >> libusb and AppLayerV1, which now succeeds. >> >> The sad thing: When comparing the two hexfiles produced this way, with >> the ones I have build previously, they are identical. >> >> Is it the starting point that is wrong, downloading the ioio-master.zip >> from the https://github.com/ytai/ioio >> ? >> >> Sorry for the long mail. >> >> Regards, Johny >> >>> >>> -- >> You received this message because you are subscribed to the Google Groups >> "ioio-users" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected] <javascript:>. >> To post to this group, send email to [email protected]<javascript:> >> . >> Visit this group at http://groups.google.com/group/ioio-users. >> For more options, visit https://groups.google.com/groups/opt_out. >> > > -- You received this message because you are subscribed to the Google Groups "ioio-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/ioio-users. For more options, visit https://groups.google.com/groups/opt_out.
