Hello, Ok, for the fix without goto. I think we can merge it right now, because the orginal implementation creates an infinite loop in following case : For tap that are dynamically enabled in the chain : - jtag_examine_chain is not executed at tap insertion, (jtag_examine_chain is executed following jtag arp_init) - at target examine hasidcode is still false and infinite loop is entered.
Later on, we have to make it work with dynamic enable tap. Best regards -----Original Message----- From: Mathias K. [mailto:[email protected]] Sent: Monday, September 19, 2011 7:01 PM To: Øyvind Harboe Cc: Michel JAOUEN; [email protected] Subject: Re: [Openocd-development] [PATCH] kinetis cpu flash driver Hello, i prefer something like this (i don't want to start a goto discussion): if ( tap->hasidcode && (dap_syssec_filter_data[i].idcode == tap->idcode) ) but we should be sure that idcode is initialized with an invalid value like zero but independent of the hasidcode flag. Regards, Mathias Am 19.09.2011 18:50, schrieb Øyvind Harboe: > I'll let it cool off for a few days. Please give word if further work > is required and > we should submit an updated patch. >
0001-kinetis-fix-deadlock-on-device-having-hasidcode-fals.patch
Description: 0001-kinetis-fix-deadlock-on-device-having-hasidcode-fals.patch
_______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
