msp430-gdbproxy loses connection with target
"Andrew McLaren" <[email protected]>
| Newsgroups | gmane.comp.hardware.texas-instruments.msp430.gcc.user |
|---|---|
| Organization | Aratika NZ Ltd |
| Message-ID | <A8A53B36A3924A45A6D531D18C4385F2@Jupiter> |
I've been struggling with an (almost) repeatable issue with gdpproxy losing connection with the target msp. After operating correctly for a period, gdbproxy will suddenly report "error: msp430: Could not determine device state (18)" in response to an MSP430_State() request by gdbproxy. then following this a stream of "error: msp430: No error (0)" in response to a sequence of MSP430_Memory(READ) requests. At this point I've only been able to get this working again by repowering the FET and the target msp (actually the target msp looks like its enough, but its typically powered via the FET). >From this I'm guessing that something in the logic being debugged is corrupting the target msp's JTAG interface in some way, which effectively disrupts communication with the FET. Repowering the target msp resets all the interfaces to a known state, which gets things back in line. However, even if this is the case, I'm not having much luck isolating where this is happening, and hoping someone can give me some ideas or pointers which may help in isolating this. When I said 'almost' above, often this seems to trigger in the same sequence in the logic, but not always, so its not totally predictable. I've been trying to 'binary chop' the logic from where it is known good to where it is often bad, but the lack of predicatability is running me in circles. I've also forced the code to link in a different order (theory being that if it was a relative addressing rather than absolute addressing issue, the problem would move), but its remained much the same (but of course the lack of predicatability means I can't be certain!) Where it is occuring it should be running logic that has all previously been exercised a number of times. - Does my diagnosis sound right, or at least reasonable? - From what I can glean, the target msp (a 439) has a dedicated JTAG interface, and can't see that any of this is configurable, even inadvertently. Am I missing something here, or is there some other way the JTAG interface can be impacted? - Any other ideas for what could cause this? - Any ideas or techniques for isolating this sort of problem. The setup is Eclipse talking to the target MSP (a 439) via msp430-gdbproxy (version 0.7.1), and a TI USB FET. Thanks for any thoughts. Andrew ------------------------------------------------------------------------------ Introducing Performance Central, a new site from SourceForge and AppDynamics. Performance Central is your source for news, insights, analysis and resources for efficient Application Performance Management. Visit us today! http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk _______________________________________________ Mspgcc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mspgcc-users