Age | Commit message (Collapse) | Author |
|
the unlocking procedure erases the flash mem (even if it wasn't locked), so it should be marked as erased after unlocking.
Change-Id: I5cc6a1e1d6cf4e1f243de532eff54111ffd66187
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/222
Tested-by: jenkins
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
if JTAG debug request fails then halting with a reset should be attempted.
the failure was ignored previously.
Change-Id: Ibec08e2e97f962d164a110c21aaa80bfc17b7f1a
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/221
Tested-by: jenkins
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
i had assumed two possible halt/debug states:
- halted + debug mode
- running + not debug mode
turns out this one also exists
- halted + NOT in debug mode
added code to handle this in an appropiate way.
Change-Id: Ia0ddcd55d1890c90d100a9e6f5e84ed8dda812a3
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/220
Tested-by: jenkins
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
now the flash algorithm running on the 568013 checks the buffer_empty bit (instead of the command_finished bit) before trying to write a new word to the flash mem.
this should speed up the flashing procedure. since it is open loop, this change may reduce the risk of failure. flashing will fail if JTAG speed is such that the flash module cannot keep up.
also, the USTAT register is only read once, as suggested in the flow chart provided by freescale (per. ref. manual @ 6-11)
the last step of the flow chart, exiting after commands are complete, is not implemented.
the algorithm will stay waiting for more data. it is up to the PC side to *not* send more data.
Change-Id: I47fe4b50de7da85f80868f5986a89a7e2152616c
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/219
Tested-by: jenkins
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
Change-Id: I36962a0ab0cc9d1eb6a29d7e577c24c38cab946b
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/218
Tested-by: jenkins
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
Change-Id: I17a0f22cbeb20e4f6ea4065236243f93d826ccf0
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/217
Tested-by: jenkins
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
moved reset_jtag to the top, since it will be needed in future changes.
also send a define to the .h
Change-Id: I53ef9c02ffa70e7d0416364525a6fbfef855e222
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/216
Tested-by: jenkins
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
added logging of target error codes to enable automatic error handling from tcl.
the plan is to use a computer to execute a series of tcl commands, the changes allow simple parsing of return messages to detect errors.
Change-Id: Ia98d3bd036e1b6065b475ffff6c1d30baeaf7417
Signed-off-by: Rodrigo L. Rosa <rodrigorosa.lg@gmail.com>
Reviewed-on: http://openocd.zylin.com/215
Tested-by: jenkins
Tested-by: Øyvind Harboe <oyvindharboe@gmail.com>
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
This fixes an issue when init is called before init_targets has been
executed.
Make sure init_targets is called before init.
Change-Id: Icd5bd4c2a8eea2e399d9de4e331a77560e9672ac
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/235
Tested-by: jenkins
Reviewed-by: Peter Stuge <peter@stuge.se>
|
|
see Trac #38
Change-Id: If86fda797391f10d745384794b68e60380ef0b21
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/204
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
Change-Id: Ia97dda6918e20fb29d6e76d43856eede8fcc60dd
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/203
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
The configuration script check for libusb-1.0 availability first and only
if not found check for libusb-0. So if both libraries are installed on the
system the build script will use libusb-1.0
It's possible to force compiling with libusb-0 with the --enable-libusb0 switch.
If the driver support only libusb0 the script check anly for it.
Change-Id: I7eb045d4e2bd553abefad53f3f4023ff46b0f5f6
Signed-off-by: Mauro Gamba <maurillo71@gmail.com>
Reviewed-on: http://openocd.zylin.com/33
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: I89efdc45bcd61ded437d67d4cbee4c53345d4d76
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/197
Tested-by: jenkins
|
|
This reverts commit 0ef5a90d93c5a026bcf70132e60e957ae339d1e1
Causes older versions of gcc to break - need to look into a better fix.
This passed through the jenkins build as we originally did not build this module - we do now.
Change-Id: Iafeac8442b2249269ff45a52ccd3e2870920f635
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/214
Tested-by: jenkins
|
|
see trac #47
Change-Id: I48a3e963354dfc82209477672c2508c96fb737d6
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/198
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
fix issue when using tms_sequence short
see Trac #31
Change-Id: I22a9cd2af59eae4d8a276dae60b6a99d05af53bb
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/201
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
The calculation for reg_packet_size in gdb_get_registers_packet() could
generate a wrong result in the case of multiple registers whose size is
not a multiple of 8.
The current calculation sums the sizes for all registers then rounds the result
up to the next multiple of 8.
Instead it should round each register size up individually and sum the results for all registers.
Change-Id: Idfb5e5eeee0e69a6889dbe9769c0bf17feacb63b
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/200
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
gdb_memory_map() correctly calculates the target specific number of flash banks, but then uses the total number (all targets) instead of the target specific number to construct its GDB response, causing a crash.
Change-Id: I3f8639b3e90303a59753ebe140ce4fff96fd5db0
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/199
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
Change-Id: I60523f809f2d9ec9c9283e0456746ce9a63576a7
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/196
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
Don't call a variable num_pages if it holds the flash size. Also rearrange
flash size to num_pages calculation to avoid divide-by-zero if there will
be a device with < 1024 byte pages someday.
Change-Id: I2febea39694a2f9750de141f52ec88ae1599c086
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/211
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: I3fabbdbda4ba8ba6557d79b97444fe06f1710b58
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/209
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: Ie1049b9d8ed5e44aee038e9181e423c35b4263c4
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/208
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
in_mask and in_idx are just another encoding of the same state information
that is already kept in bitq_in_state.bit_pos so derive them from that
instead of maintaining them separately.
Change-Id: I4ac6bbe923698a8c1090a785b8babcbb90f82931
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/207
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
this at least checks the post conditions after convoluted code.
Change-Id: Idfa8cbedce5288d8bae5743687949f141dfb07b2
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/187
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
in_buff is only ever set to field->in_value and that pointer is still
available when the parsing is restarted so it could just as well be used
directly, removing the need for the static variable.
Change-Id: I3dd7a8315ed5c5bdc3bfb74044f89492bca9816c
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/206
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
field->in_value is already checked so it must be non-null here, which means
the else clause can never execute so the entire buffer allocation and
handling code is completely dead.
Change-Id: Id465012a7e607349401d554fc7a8e5db7e967998
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/205
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Check whether the destination is in the program flash or NVM regions,
in the former case, use the normal longword mechanism, not the fast NVM
write.
Change-Id: I7366b7c8919928ee690252df83b99701776aee82
Signed-off-by: Tomas Frydrych <tomas@sleepfive.com>
Reviewed-on: http://openocd.zylin.com/194
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
This is a successor to the Flyswatter cable and is very close to the original.
The new revision is based on FT2232H.
Change-Id: Icc6efcf0e4f9d8a10b65df8679b4973f6b375a9f
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: David Anders <danders@tincantools.com>
Reviewed-on: http://openocd.zylin.com/193
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Before this commit, openocd used to segfault when probing flash
of an at91sam7x512 (which contains 2 banks of flash). This was due
to the way it systematically insert a new flash bank without setting
its name.
Then, when get_flash_bank_by_name_noprobe() is called, it is doing
a strcmp() on the non-initialized bank->name.
This commit prevents allocation of second probed bank if it is already
allocated (for example, if it is set in a target config file). If a
new bank really needs to be allocated, it ensures that a default name
is set.
Change-Id: I38d15bef1fda2ec746efad37171975136cf7b371
Signed-off-by: Aurelien Jacobs <aurel@gnuage.org>
Reviewed-on: http://openocd.zylin.com/171
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
there are no comments about what the rules w.r.t. null pointers
are and it is inconsistent.
It's simply a bug in the app if we ask about the properties of a null
pointer w.r.t. what kind of target it is. It's equally wrong to say
that it is an arm target as that it isn't an arm target.
Change-Id: I0925a6a5c8b38e594ffa7c3ca4390487b5e9b718
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/168
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
|
|
Change-Id: I1a5e968f165e060fd4aa7c023ad870a9e21bb5dd
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/191
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
variable is not read afterwards.
Change-Id: I905bbb10c596190f75494e6c6ad400a3e51843f6
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/192
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: Ib7f67612db3a865f9acc5ae349455da7ddcd3348
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/177
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
The CPLD on the xds100v2 expects to see a rising edge on PWR_RST to
enable the outputs. This patch creates that transition correctly by
fixing the direction register for PWR_RST.
THe CPLD will also loop back the data if the LOOPBACK signal is
asserted. Set this signal to an output and keep it clear.
This was tested with a TI DM3730 Beagleboard xM.
Change-Id: I4ea216bef6ae5c40e935741af5c69dc844d5d494
Signed-off-by: Kyle Manna <kyle.manna@fuel7.com>
Reviewed-on: http://openocd.zylin.com/189
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
* Add flash waitstate support for Atmel SAM3 chips.
* Set default waitstates to 6, to workaround a silicon bug in the SAM3 family
This code has been tested on SAM3U4, SAM3N4 and SAM3N1
based on Change-Id: I477446f9bfb3e910ea3e2414a6e9a75beb14a214
by Jim Norris <u17263@att.net>
Change-Id: I8d360080f6968979ca5e197ad638282cadd18fb7
Signed-off-by: Attila Kinali <attila@kinali.ch>
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/128
Tested-by: jenkins
|
|
Change-Id: I836038b3518e617291ac7d5d255a388d9486f67f
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/180
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
found by clang.
Change-Id: I80ea8e6afc8dcc1aa7edb6f63af0d94f6781b81c
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/182
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: I8c73b0f9a6a10734e539eedcedd79e1a34122f60
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/185
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: I9def63d36ed4fa8bf9cdeeedc18b1b25d0e487d6
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/184
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
clang found silly code that was trivially fixed.
Change-Id: Ied6c1b254c1823cd111140cbe0c8a03d2ede65fb
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/186
Tested-by: jenkins
Reviewed-by: Peter Stuge <peter@stuge.se>
|
|
add support for checking target against the device CLASS rather
then just the PARTNO.
This change also adds the new LM4F family (Blizzard).
Change-Id: Ia9d1e33f1f1c2817c0039a2232ecf932fae072f9
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/161
Reviewed-by: Øyvind Harboe <oyvindharboe@gmail.com>
|
|
Change-Id: Idc384b733056a72108680b073da7c327a8eeedc3
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/179
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
return error message instead. Found by clang.
Change-Id: Ica109d077206236a12d007e77cc78061ffd05834
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/169
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: Ibc70deb980d6d18ceb376b72d9104e6180b16acf
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/176
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
- propagate error
- remove dead assignment
Change-Id: I0d7078f531d96e421e95f08bfd908e818500d74c
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/170
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: I9ddb2793b4cdbf6acea6f69973531491e4ebcc5b
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/145
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
- reduce scope of variables
- assert on post condition to fix clang warning
Change-Id: Id91038f73a632f9688ba52c9c34eae923626e770
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/160
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
stick to convention of "retval" being used as error value to
be propagated and use "flash_flags" local variable for flash
flags read from how.
Change-Id: I63f1f2248b4f4538d6cd7634ae277f7c0aadc346
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/178
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
Change-Id: Ib34815c9cf654517f22486a7c8001fdb7471338c
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/174
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|
|
- fix error propagation
- add some debug output for "unused" variables
- reduce scope of variable to avoid "unread" variables.
Change-Id: I136995f2a043102e64b8fff3017502fae564013b
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Reviewed-on: http://openocd.zylin.com/175
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
|