Changed reset handling: save off polling state, disable
polling and restore after reset.
With timer based polling enabled a drscan/irscan sequence
in a reset event script could have a poll run inbetween commands,
which would cause havoc, occasionally.
Passed smoketests.
Note that reset may explicitly wait for the target to enter the halted
state and here polling is called directly instead of via a timer, this
polling still happens and was never a problem.
Comments?
Objections?
--
Øyvind Harboe
Embedded software and hardware consulting services
http://consulting.zylin.com
### Eclipse Workspace Patch 1.0
#P openocd
Index: src/target/target.c
===================================================================
--- src/target/target.c (revision 2179)
+++ src/target/target.c (working copy)
@@ -434,9 +434,14 @@
return ERROR_FAIL;
}
+ int save_poll = target_continous_poll;
+ target_continous_poll = 0;
+
sprintf( buf, "ocd_process_reset %s", n->name );
retval = Jim_Eval( interp, buf );
+ target_continous_poll = save_poll;
+
if(retval != JIM_OK) {
Jim_PrintErrorMessage(interp);
return ERROR_FAIL;
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
