On Wed, Nov 25, 2009 at 9:33 PM, David Brownell <[email protected]> wrote: > On Sunday 22 November 2009, Johnny Halfmoon wrote: >> > Consider a board: >> > >> > USB --> Cortex-M3 --> level shifting --> JTAG >> > >> ... >> > Come up with such a board with initial "it works" >> > firmware ... and OpenOCD could speed things up >> > over time. >> >> Now there's an interesting project. Do I hear a call for porting >> OpenOCD to an STM32 (got USB) + $some_rtos system? > > I'm not sure porting it would be the right solution. > > For starters, OpenOCD is demonstrably too bloated to > fit on a typical Cortex-M3. I did a test build using > a "Cortex-A8 Thumb2" target, and saw over 640 KBytes > of object code, on top of needing libc etc. It ought > to be possible to trim that down ... but that'd be a > lot of work, and it'd surely look VERY different before > it could fit into a sane-sized microcontroller. > > Also, there are more CM3 chips than just STM32. ;) > > I think the more-likely-to-succeed-soon model would > be to aim at a "smart dongle" that can just offload > more of the work than an FT2232 ("dumb") does.
Which is what jlink does. jlink is supported by OpenOCD. Those Segger guys know what they are doing. There are plenty of open source dongle projects out there(schematics & PCB & even samples you can buy). The cost of a dongle is $1 in volume when you add a chip to an eval board. That same bug-by-bug compatible dongle can be sold standalone. This is done all the time. Look at e.g. LuminaryMicro, Altera Byteblaster, etc. -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 ARM11 XScale Cortex JTAG debugger and flash programmer _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
