[cvrtky]

It’s working now. Excellent 🙂 !!!
Thank you.

Cheers
Tim

[cvrtky]

Hello again

I tried following your tutorial. I wasn’t success full. I might have found a new bug.
I can flash the lpc1768 and pyocd-gdbserver connects w.o. complaining to much. See below.

C:\VisualGDB\lpc1678\lpc1678\Debug
(PyOCD) λ C:\VirtualEnvs\PyOCD\Scripts\pyocd-gdbserver.exe -t lpc1768
INFO:root:DAP SWD MODE initialised
INFO:root:ROM table #0 @ 0xe00ff000 cidr=b105100d pidr=0
INFO:root:[0]<e000e000: cidr=b105e00d, pidr=4002bb000, class=14>
WARNING:root:Invalid coresight component, cidr=0x0
INFO:root:[1]<e0001000: cidr=0, pidr=0, component invalid>
INFO:root:[2]<e0002000:FPB cidr=b105e00d, pidr=4002bb003, class=14>
WARNING:root:Invalid coresight component, cidr=0x0
INFO:root:[3]<e0000000: cidr=0, pidr=0, component invalid>
WARNING:root:Invalid coresight component, cidr=0x0
INFO:root:[4]<e0040000: cidr=0, pidr=0, component invalid>
INFO:root:[5]<e0041000: cidr=b105900d, pidr=4002bb924, class=9, devtype=13, devid=0>
INFO:root:CPU core is Cortex-M3
INFO:root:6 hardware breakpoints, 4 literal comparators
INFO:root:4 hardware watchpoints
INFO:root:Telnet: server started on port 4444
INFO:root:GDB server started at port:3333
INFO:root:Telnet: server stopped

When I try to connect from VisualGDB the error as shown in the attachment. As shown in the other attachment the path to pyocd-gdbserver.exe is correct.
It looks to me as if the macro holding the port number isn’t expanded correctly.

Please advice.

Sincerely
Tim Børresen

Attachments:

You must be logged in to view attached files.

[cvrtky]

Hello again.

Just in case this is the board in question: https://os.mbed.com/platforms/mbed-LPC1768/.

On the SEGGER site, I find J-link firmware for the LPC-Link2/3 and OpenSDA. Neither are suited for the above board. At least not directly. The board has a chip only labeled “MBED interface” which I presume is the MBED DAP-LINK debugger. The only other chips on the board are the LPC1768 itself, an Atmel EEPROM and the Ethernet PHY.

Please advice.

Cheers
Tim

[cvrtky]

SWDIO and SWCLK aren’t broken out to a connector on the original LPC1768 based MBED board. Those pins (pins 3 & 5) are connected to an onboard debugger IC.

Are you suggesting I solder wires directly onto the MuC?

/Tim

 

[cvrtky]

6 Dupond wires could be used as an interim solution.

/Tim

[cvrtky]

Hello again.
Thank you for your support. I went along and purchased a license for the Custom Edition.
As I understand it, I’m entitled to do 2 installations when purchasing a personal non commercial license.
Do I need a second license key for this additional install or is my key good for 2 installations?

Sincerely
Tim Boerresen

[cvrtky]

Bummer. Failed again.
I’ve uploaded the project to google drive. It is located in the following folder:
https://drive.google.com/drive/folders/0B3ZpZQ81OjS5NDRsUmlTT1JiQ1U/
File name is EmbeddedProject4.7z

Cheers
Tim

[cvrtky]

My attachment was apparently to big. Here we try again.

[cvrtky]

Hello again
I’ve been busy again testing your feed back … I also have to do my normal job 😉

1)
I’m fully aware of the volatility of the the mbed library and that “My mileage will vary”. Yes I’m willing to “hax” around in your BSPgenerator. I already have had a taste of what I’m signing up for while building and installing a BSP for mbed-os-5.6.6 (current head of the repro).
You have done a fine job writing that program. The import facility also works fine 🙂 This will server as a good starting point. I hope you’ll keep the GitHub repro updated as you progress. I’m willing to share my work in return. That is I hope we can work together on this. It is quite a job to stay current with that library and roll BSP’s for older versions as required.

OK your .xml files are more detailed compared to usual .SVD files. Could you please tell me how to convert .SVD’s to your .XML format using VisualGDB ?

2)
I dropped using the tools chains from the EmBitz IDE by Gerard Zagema and fetched the equivalent installation packages from Launchpad. It’s working as expected now. The correct specs are being read and the correct libraries are being linked. I renamed the libraries temporarily to make sure that this was the case.

5)
Got this sorted to. I get the stats printed out even when building using make. It’s quite trivial: Don’t use a source tree w. code for several MuCs, that will confuse the import. Insert preprocessor symbols by hand. The same goes for include directories. Remove the generated startup code and include the startup code from the project. Switch the scatter file to the one comming w. the project. Done …. Almost, since I found another bug in VGDB
If I insert preprocessor symbol containing pipe – / logical or symbol like the (im)famous example from previous posts: CLOCK_SOURCE=’USE_PLL_HSE_EXTC|USE_PLL_HSI’ into the input in the GUI, the last apostrophe ie. the one after USE_PLL_HSI is removed when saving . Pt. I have to put it into stm32.mak I’ve added the project so you may inspect it. It’s just a trivial mbed blinky.

I’m starting to get a good feeling with yo product. 🙂
Sincerely
Tim Børresen

[cvrtky]

Hmm upload blocked. Trying again using a 7zip archive.

Attachments:

You must be logged in to view attached files.

[cvrtky]

Hello again Sysprogs
you put me to work wit your last reply 😉

1)
Thank you for providing link to your mbed BSPgenerator. Having installed the required Python modules in a virtual environment and correcting the path to your toolChain from e:\SysGCC\arm-eabi to C:\…, I manged to create BSP for mbed-os version 5.6.3 & 5.6.4 / mbed-dev library veriosn 154 & 155.
I get a lot of errors about missing register definitions and missing linker scripts. You are probably aware of this.
See log file attached.
I wanted to try adding a register file for the target: STM32L072CZ_LRWAN1. My partner in the USA are using this board. Secondly adding support for the STM32L072CZ should be fairly easy since one just have to remove the LCD Registers sets from the STM32L073CZ 😉
Looking at the the stm32l072xx.xml it resembles a ordinary .SVD file from ARM / ST.com. Please confirm this. Or is is specially tailored for your register view? I tried to remove the LCD register set and save a copy as stm32l072xx.xml.gz, but I still get complaints about missing register definitions for STM32L072CZ_LRWAN1. Should the file be named differently? Please advice. I would be nice to get rid of these errors. Where should I look for away to patch for the linker scripts? Please advice.

The RunTests function in Program.cs fail miserable because it rely on reading the location of you toolchain from the registry. One could of cause override that, but the test takes a long time, so I didn’t care.

For mbed-os-5.6.5 and 5.6.6 ( mbed-dev library version 156 & 157) I had to extend the search for RAM segment in the scatter files, because the Analog Device Mucs ADuMC3029 & ADuMC4050 were added in mbed-os-5.6.5.
in the function DetectAndApplyMemorySizes in MbedBSPGenerator.cs I added the match for the DSRAM_B segment as follows.

var ram = memories.First(m => m.Name.ToUpper() == “RAM” || m.Name == “m_data” || m.Name == “RAM_INTERN” || m.Name == “SRAM1” || m.Name == “RAM0” || m.Name.StartsWith(“Ram0_”) || m.Name == “DSRAM_B” );

2)
Please see the attached files. I copied the sections for handling nano lib and semihosting from your tool chain.

3)
This patch solved the problem. Thank you.

5)
Please elaborate on how to use this manged Makefile. I’ve had no luck attempting to do that.
I followed the tutorial on the generic import. that works fine, but as you wrote won’t provide me with memory usage stats. the size command from Atmel Studios shell utils can provide some information. the size reported this way for text – + data segment = total size of .bin file

Sincerely
Tim Børresen

[cvrtky]

Hello again Sysprogs.
1)
I believe we misunderstood each other on this issue. From my point of view, you refer to mbed-os as you BSP, that is the mbed-dev core library plus all the extras including RTOS, events and feature subsystems, while I was talking solely about mbed-dev. This subsystem consist of the files mbed.h, mbed.conf and the 4 directories drivers, hal, platform and targets. The mbed-dev libray is distributed separately through a Mercurial repro, which is updated each time the Library version number is updated
You mbed BSP 5.6.3r2 contain mbed-os version 5.6.3 and mbed-dev library version 154. The current version of mbed-dev is 157. Following your instructions, I was given the option to install the old mbed BSP version 5.4.2R2, which gave me mbed-os versopn 5.4.1 and mbed-dev library version 123.
This is not what I’m asking. As mbed contributor I’m working with bug fixes and updates to the mbed-dev library. What I’m looking for is either to construct my own BSP containing basically any version of the mbed-dev sub module plus the ability to either switch the version of the subtree forth and back at my discretion, or the ability to do the above by updating mbed BSP 5.6.3r2
The mbed-dev submodule changes often / is in constant flux. Files may be added to, removed from or moved around in the source tree. Because of these frequent changes, I need a way to generate the BSP.xml file since it has a description of the location of the files in the tree and the compile options per supported boards. I presume you generate this file somehow “automagic”. Please advice whether this is possible and if you are willing to share this program.
Albeit we have a project buid with mbed-dev version 95. Until we migrated this to using HAL instead of SPL we’ll handle this using CrossWorks. We mainly are interested in version 140 and later. This probably makes it easier to accommodate our requirement.

2)
This is a bug. The nano.specs in C:/SysGCC/EmBitz/1.11/arm-none-eabi/lib is NOT being read. The only way I can make this work as expected is to edit mbed.mak in my project directory changing the name of the specs files. If I change the name of nano.specs to sat nanonano.specs then everything works as expected. I tried without any success to specify nanonano.specs in the tag in toolchain.xml under the default tag. This problem arises with every external tool chain I’m attempting to use even if I don’t install your arm-eabi toolchain.

3)
This is a bug, but there is a very simple work around.
The problem turned out not to be related with the version of make being used and neither the -IC:/… syntax. The culprit is the “-DCLOCK_SOURCE=USE_PLL_HSE_EXTC|USE_PLL_HSI” command line option. The “|” is being interpreted in the shell as a pipe command. This problem occurs independent of which make and tool chain being used. The solution straight forward: One have to specify the option as follows:
“-DCLOCK_SOURCE=’USE_PLL_HSE_EXTC|USE_PLL_HSI’” That is in the GUI enter an apostrophe before and after “USE_HSE_EXTC|USE_PLL_HSI” in the GUI. I haven’t tried VisualGDB on Linux nor OSX. I’m using Windows 10.

4)
OK. This is NOT a major problem, we can live with this limitation ?

5)
Hmm. I’m not sure I can follow you on this. I am using the GNU make utility packaged with VGDB. Isn’t this what you refer to as “The Embedded Make?” Could I get this information using a custom post make command?

6)
Jup. Problem solved.

7)
I haven’t had time to look further into this. I could probably use atbackend from Atmell Studio. We can discus this later. Issue 1 .. 3 are the important ones.

[cvrtky]

Hello again,
seems like I found the answer to question 6.
Navigating from the menu bar to Tools -> Options -> Project and Solutions -> Locations.
Then I’m allowed to change the Project location. Nice 🙂

Cheers
Tim

[Author]

Posts