W dniu 2012-09-28 16:25, Freddie Chopin pisze: > Here is a link to logs for telnet and gdb, for good and bad version of > OpenOCD. > > https://www.dropbox.com/s/50m5xtoeexqwl1d/logs.7z
Comparing this the first significant difference I see (for telnet logs) is that "good" version does allocate a 180b buffer for the second flashing, and the "bad" one doesn't. They both allocate it for the first flashing. This is present for "good" and absent for "bad": > Debug: 885 27701 target.c:1523 target_alloc_working_area_try(): allocated new > working area of 180 bytes at address 0x10000000 > Debug: 886 27701 target.c:1386 print_wa_layout(): * 0x10000000-0x100000b3 > (180 bytes) > Debug: 887 27701 target.c:1386 print_wa_layout(): 0x100000b4-0x10007fff > (32588 bytes) A little lower for "good" version: > Debug: 1039 28208 target.c:1386 print_wa_layout(): * 0x10000000-0x100000b3 > (180 bytes) > Debug: 1040 28208 target.c:1386 print_wa_layout(): * 0x100000b4-0x100010b3 > (4096 bytes) > Debug: 1041 28208 target.c:1386 print_wa_layout(): 0x100010b4-0x10007fff > (28492 bytes) For the "bad" version: > Debug: 1042 37309 target.c:1386 print_wa_layout(): * 0x10000000-0x10000fff > (4096 bytes) > Debug: 1043 37309 target.c:1386 print_wa_layout(): 0x10001000-0x10007fff > (28672 bytes) Next differences are clearing lockups after double fault and HardFaults vs. "normal" restoring of operation after flashing. 4\/3!! ------------------------------------------------------------------------------ Got visibility? Most devs has no idea what their production app looks like. Find out how fast your code is with AppDynamics Lite. http://ad.doubleclick.net/clk;262219671;13503038;y? http://info.appdynamics.com/FreeJavaPerformanceDownload.html _______________________________________________ OpenOCD-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/openocd-devel
