>> Still test driving. One frustration that I just diagnosed >> is that software breakpoints often don't work if icache is enabled by the >> target code (for reasons that are obvious in retrospect; cached instruction >> is >> executed instead of the BKPT command written to memory by openocd). I'd like >> to submit a patch soon that invalidates the icache after the BKPT instruction >> is written. Suggestions / corrections of course gratefully received. > > We've been fixing a lot of those cache problems with software breakpointing > over the last 6-8 months or so. It'll be good to see XScale fixed too.
The arm breakpoint implementation relies on target->type->write_memory to handle single word writes in a way that makes it "just work" for writing out breakpoint instructions. This includes i/dcache handling, handling of ram marked as read only memory and any other issues. Crude but effective. -- Meet us in Paris at rts EMBEDDED SYSTEMS 2010 - March 30. and April 1. http://www.zylin.com/events_rts.html Øyvind Harboe US toll free 1-866-980-3434 / International +47 51 63 25 00 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
