Magnus Lundin wrote: > Øyvind Harboe wrote: > >>> I have found it to, IR in bypass returns 0x01 if I recall correctly, so >>> bypass taps should not be checked, or checked with a correct value. >>> >>> The flag used to signal no IR check was to set in_handler = NULL; /* >>> disable verification by default */ >>> >>> Really ugly but you broke it anyway :), we need a cleaner implementation of >>> return check logic. >>> >>> >> Second try to create a patch. I can't test right now, but I believe this is >> it. >> >> >> >> >> > It works for the STM32, and going on the the flyswatter+OMAP3530 that > workes in the old rev, we find > > IR scan gives no problems, but the first idcode scan fails. But OpenOCD > is alive, and connecting with telnet works. > Trying to do another initialisation fails with > > openocd: jtag.c:1407: jtag_check_value_mask: Assertion `field->in_value > != ((void *)0)' failed. > > We have as yet not found any problem revealing hidden bugs in the > previous codebase. > > Magnus > > _______________________________________________ > Openocd-development mailing list > Openocd-development@lists.berlios.de > https://lists.berlios.de/mailman/listinfo/openocd-development > This is IMHO not an efficient way of developing software, it is rater unplanned trial and error, looks like amatuer work to me. Where is the structural analysis or the overall picture ? Fixing one error at a time by banging the head against the wall to see if it hurts.
Friday evening , I need some red wine. Have a good weekend everyone. Magnus _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
