On Mon, Mar 1, 2010 at 5:17 PM, Øyvind Harboe <oyvind.har...@zylin.com> wrote: >> But the official PIC32 GCC compiler comes with proprietary >> C libraries from Microchip (together with MIPS). > > I tried to build some test code and immediately ran into this, > which quelled my enthusiasm for doing any further OpenOCD > work.
Ah oh. Hopefully the work for OpenOCD PIC32 will continue. I see Spen is doing some work on it. Someone just need to ports newlib or things like that to GCC to get the missing link. Microchip's C32 does not have code size limitation though, but rather optimization limit (related to the library). So it is still usable for many people. >> Personally I think Microchip is wrong to choose the MIPS core. > > It was cheaper they thought & faster? I do not know the reasons. There are rumors that one potential reason is that MIPS allows Microchip to add proprietary ICSP protocol and trace functionality to the MIPS M4K core whereas ARM does not. Microchip wants to support existing tools like ICD 2, ICD 3 and Real ICE which do not speak JTAG. > They may be regretting this decision already and the > marketing department may be tearing their hairs out trying to a) > fix the problem b) present a happy face outwards c) trying > to figure out how to spin stuff now so they don't look like complete ... > when they finally do come to their sense and license ARM. :-) I think PIC32 should still have appeals to those who upgrade from PIC18 and PIC24. So Microchip can still do reasonably well from PIC32, but not the success it gets in the 8-bit market. OpenOCD for PIC32 should be able to help Microchip to get into more users' hands. -- Xiaofan http://mcuee.blogspot.com _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
