This is an automated email from Gerrit. Luis de Arquer ([email protected]) just uploaded a new patch set to Gerrit, which you can find at http://openocd.zylin.com/6126
-- gerrit commit 87a0f35abca5bc9d2dd527b4f3850ebc1d3b254e Author: Luis de Arquer <[email protected]> Date: Wed Mar 24 17:20:09 2021 +0100 COVER LETTER: Initial support for ST On Chip Emulator (OnCE) This patchset provides basic support for debug control of the Power Architecture e200 core. All of the SPC56x/RPC56x devices from ST (and possibly MPC56x from NXP) should be compatible with this implementation, based on ST Application Note AN4035. The following functionality is provided: - Read/Write CPU registers - Read/Write arbitrary memory location (32-bit) - Execute single instruction - Load binary file to memory location - Execute code at given memory location, until breakpoint is reached Tested on ST SPC560D. Even though OnCE can be implemented mostly by regular IRSCAN and DRSCAN instructions, there are some nasty particularities which required low level changes, and which may not be to everyone's taste in its current form. 1) OnCE Status Register (OSR): This register is read in a very special way: it is shifted out serially during an IRSCAN. To do so, Commit 1 changes the IRSCAN tcl implementation so that it can return a value. Maybe this could be implemented by providing an extra option to IRSCAN. 2) If DR-PAUSE is entered before DR-UPDATE, then OnCE TAP gets disabled. Unfortunately, this is the default behaviour in OpenOCD. Commit 2 proposes a way to adapt for this (currently only implemented on the FTDI driver), with the naive hope of not breaking anything. As with the change on IRSCAN, maybe this could be enabled by an extra option to the TCL command. Commit 3 creates a new TCL file, once.cfg. This is the core of the patchset, and feedback is welcome on how to shape it. Commit 4 provides a board target file, for the ST SPC560D discovery board, providing an example on how to initialize the OnCE interface. The TCL code is not meant to be exhaustive yet (e.g. error handling and documentation are still missing). I am looking for feedback at this stage, rather than providing a polished implementation which may not fit OpenOCD practices (for example, maybe it makes more sense to code once.cfg in C) Change-Id: Id821afb1a57276aa911331ef0d75e956d4c305a4 Signed-off-by: Luis de Arquer <[email protected]> -- _______________________________________________ OpenOCD-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/openocd-devel
