[openocd-svn] Main OpenOCD repository branch, master, updated. 0.2.0-371-g36564d7

[Øyvind H. <go...@us...>]

This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "Main OpenOCD repository".

The branch, master has been updated
       via  36564d74505598c328fab91d0cfe415b0fc61676 (commit)
      from  77ca2f3a702c7a9cfceba14166042a0ed09947fe (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit 36564d74505598c328fab91d0cfe415b0fc61676
Author: Øyvind Harboe <oyv...@zy...>
Date:   Wed Oct 14 11:14:04 2009 +0200

    Work in progress on arm11 reset. Assert srst.

diff --git a/src/target/arm11.c b/src/target/arm11.c
index f46e424..94d5fb1 100644
--- a/src/target/arm11.c
+++ b/src/target/arm11.c
@@ -1200,18 +1200,62 @@ int arm11_step(struct target_s *target, int current, uint32_t address, int handl
 int arm11_assert_reset(target_t *target)
 {
 	FNC_INFO;
+	int retval;
 
-	/* FIX! we really should assert srst here, but
-	 * how do we reset the target into the halted state?
-	 *
-	 * Also arm11 behaves "funny" when srst is asserted
-	 * (as of writing the rules are not understood).
-	 */
+	arm11_common_t * arm11 = target->arch_info;
+	retval = arm11_check_init(arm11, NULL);
+	if (retval != ERROR_OK)
+		return retval;
+
+	target->state = TARGET_UNKNOWN;
+
+	/* we would very much like to reset into the halted, state,
+	 * but resetting and halting is second best... */
 	if (target->reset_halt)
 	{
 		CHECK_RETVAL(target_halt(target));
 	}
 
+
+	/* srst is funny. We can not do *anything* else while it's asserted
+	 * and it has unkonwn side effects. Make sure no other code runs
+	 * meanwhile.
+	 *
+	 * Code below assumes srst:
+	 *
+	 * - Causes power-on-reset (but of what parts of the system?). Bug
+	 * in arm11?
+	 *
+	 * - Messes us TAP state without asserting trst.
+	 *
+	 * - There is another bug in the arm11 core. When you generate an access to
+	 * external logic (for example ddr controller via AHB bus) and that block
+	 * is not configured (perhaps it is still held in reset), that transaction
+	 * will never complete. This will hang arm11 core but it will also hang
+	 * JTAG controller. Nothing, short of srst assertion will bring it out of
+	 * this.
+	 *
+	 * Mysteries:
+	 *
+	 * - What should the PC be after an srst reset when starting in the halted
+	 * state?
+	 */
+
+	jtag_add_reset(0, 1);
+	jtag_add_reset(0, 0);
+
+	/* How long do we have to wait? */
+	jtag_add_sleep(5000);
+
+	/* un-mess up TAP state */
+	jtag_add_tlr();
+
+	retval = jtag_execute_queue();
+	if (retval != ERROR_OK)
+	{
+		return retval;
+	}
+
 	return ERROR_OK;
 }
 

-----------------------------------------------------------------------

Summary of changes:
 src/target/arm11.c |   56 ++++++++++++++++++++++++++++++++++++++++++++++-----
 1 files changed, 50 insertions(+), 6 deletions(-)


hooks/post-receive
-- 
Main OpenOCD repository