| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
by libswd internals. Will change and might not work as expected, feel free to test and feedback results :-) The goal is to have enqueued by openocd operations executed with no need for error handling (which is impossible on multiple operations enqueued and also not supported by openocd anyway) just as they were executed correctly for the first time. Error handling, retries, communication verification, etc should be done by underlying transport library (libswd in this case). swdctx->config.trunccmdqonerror=0 will result in untouched queue on error response from target (before the queue was truncated to make room for manual error handling). work in progres...
|
|
|
|
| |
Target read functions should now work with double pointers in order to return data on flush and not produce flush on read.
|
| |
|
|
|
|
| |
and all sticky error flags (do we need to clear them anyway even after ABORT?).
|
|
|
|
| |
SWD and JTAG work again.
|
|
|
|
| |
not return data where is should, 2. calling reads with NULL pointer caused program crash.
|
|\ |
|
| |
| |
| |
| | |
ubuntu 11.10.
|
| |
| |
| |
| | |
is set in OpenOCD.
|
|/
|
|
| |
cannot be queued up and fulshed at later time because OpenOCD use single pointers for data return from function and it is impossible to read out dynamically allocated memory this way in C.
|
|
|
|
| |
data. ACK and DATA packets are always transmitted LSb-First, no matter how ARM draws its fuckin timing diagrams!
|
|
|
|
| |
11.04. Thanks to Przemyslaw Wegrzyn for pointing this out :-)
|
|
|
|
| |
to Przemysław Węgrzyn for pointing this out :-)
|
|
|
|
| |
startup with other commands.
|
|
|
|
| |
on 64bit machines, now use %p.
|
|
|
|
| |
swd_log() does not produce endline anymore.
|
| |
|
|
|
|
| |
transport initialization and only if context is created. 2. swd_log() use vprintf as using openocd logging mechanisms caused program crash by va recurrence. 3. Fixed tcl interface to use/update loglevel configuration parameter from swdctx stored within dap.
|
| |
|
|
|
|
| |
after the DAP is created, DAP also holds the swdctx. transport_select() now only marks transport type and registers swd transport commands. Target system now consists of TAP (that is the test access port of the device), DAP (that is the debug access port of the arm system) and the TARGET+FLASH itself that allow to perform operations on the target architecture + memory. Interface and Transport connects to the TAP and allows operations on the TARGET using DAP.
|
|
|
|
| |
registered at transport select. Available commands are 'newdap' currently pointing to jtag_newtap, 'loglevel' that can show/set/inherit loglevel for easier log analysis during development.
|
| |
|
|
|
|
| |
loglevel inherited from openocd.
|
| |
|
| |
|
|
|
|
| |
and verified. Transport select does not operate on interface anymore, just prepares memory and configures transport, because signals are not yet definet at this point and interface is not initialized. Transport init operates on interface and reads out IDCODE by default.
|
| |
|
|
|
|
| |
src/transport.h and defined by src/transport.c. Multiple definitions of 'jtag_only' were removed and code was updated to use 'oocd_transport_jtag_only' (also updated headers if necessary) as defined by src/transport.
|
|
|
|
| |
and libswd. Not yet functional!
|
|
|
|
| |
code updates.
|
|
as we introduce swd and jtag as two transports, we want
to start up with a new transport folder to organize the
code a bit.
|