DAVE™ Forum Discussions

Hi,
I am trying to link some files from another project to my current project. The problem is that linking folders were perfectly fine. But when I am trying to link individual files, "No such file or directory" error occurring. For example, I want to link a header file say header.h, after linking it, during build the error occurs at point where other files have"include header.h" statement. Is this allowed in eclipse to link a resource file which is also included in other files of project? Show Less

Hello,
maybe I understand it not, but I have the following problem.
I have to build the frequency (~50 Hz) and the RMS-value of an electrical 3-phase system. For that I use the PWM App. From each wave I take 32 measurements with the ADC-MEASUREMENT App.
So, at the end of the wave the period duration is calculated (and also the frequency).
Included in this calculation for the period is the calcuation for the next PWM-frequency. The ADC-MEASUREMENT-Conversion is started with the PWM-compare-period Interrupt.
So far, so good, but it seems to me, that, after changing the PWM-frequency, something happens, so that I have the same ADC-MEASUREMENT-Values for exact 2 PWM-Interrupts.
I am not sure, if I have explained it in a way, that can be understood. But if, please ask for more details.

Best regards,
Oliver Show Less

Hi,

I'm looking to commit our software to subversion, for the first time.
Infact, I've got two versions to check-in, a V1, and a V2.
They're sitting on my PC presently, in C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1\, and C:\Workspaces\DAVE-4.3-64Bit\Foobar\V2\
Within each, is my main.c, and the usual subfolders, Dave, Libraries, Startup, and so forth.
And both build, where they are sitting, 0 errors, 0 warnings.
So far, so good.

I learned recently, that DAVE stores the *fullpath*, within several project settings.
Specifically, on the C/C++ Project or Project Explorer tab, if you right-click Project,
then Properties, then C/C++ Build > Logging > Log file location, you'll find there eg:

C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1\.metadata\.plugins\org.eclipse.cdt.ui\Project.build.log

Also, for C/C++ Build > Environment > CWD, you've got eg:

C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1\Project\Debug

And the same, for C/C++ Build > Environment > PWD.

And maybe others, too, I don't know. I tried grepping for 'C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1',
to see if I could find it any places, but none were found, so I think DAVE must store that, in some kind of
'coded' fashion.

Anyway, I'm wondering, whether the C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1\ bit,
can be replaced with something like %Foobars_place% or similar, if you are with me,
so as to make a project, entirely independent, of the folder path, it's living in.

Presently, if in the future, if someone checks-out my project, to a different folder path, and builds,
DAVE will generate the log file as follows, including recreating the complete path leading up to it:

C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1\.metadata\.plugins\org.eclipse.cdt.ui\Project.build.log

On the other hand, if using something like %Foobars_place%, could upset some other behaviour or
operation of DAVE, or cause incompatibilities with other parts of the DAVE system, or those other parts
make certain assumptions, about the contents of the above settings, and could be upset by finding
%Foobars_place% in those settings, or if those other parts of DAVE could overwrite those settings,
losing the %Foobars_place% bit, them I'll probably decide not to try making the project independent
of the folder path, it's living in.

I guess what I'm asking, is whether DAVE has built-in, a proper way, and proper support,
for making a project, independent of the folder path, it's living in.

I know, that using the DAVE export and import facility, avoids the need to change the mentioned settings,
to avoid having to do so, manually. But I don't really want to check-into subversion, an export. I really want
to check-in just the C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1\ contents, as-is..

..although perhaps, you might advise otherwise.

Of course, an alternative tack, would be to export C:\Workspaces\DAVE-4.3-64Bit\Foobar\V1\,
create a C:\Workspaces\DAVE-4.3-64Bit\Foobar\Current\ folder, say, then import to there,
and then check-in to subversion.

Then, do the same, for C:\Workspaces\DAVE-4.3-64Bit\Foobar\V2\

If you are with me.

But the project is still tied to a particular folder. It's just named instead, C:\Workspaces\DAVE-4.3-64Bit\Foobar\Current\

I've had a read through all the forum posts, about version control systems, but they seem to focus on
what can safely be 'left out', from the check-in, rather than the matter I describe.

Hoping someone can enlighten me..

Best regards,

David King Show Less

Hi,
We are working on a project that displays lots of different text in different languages. The supporting characters fit within UTF16 space and I am using wchar_t as my character type. I set the gcc and g++ command line option to -fshort-wchar to conserve memory (we have a lot of stored text strings in flash) but I get a lot of linker warnings regarding libraries using 4-byte wchar_t when my code is using 2-byte wchar_t (libg.a, libstdc++.a, etc.). Is there a set of DAVE libraries that support this?

I am going to guess the answer is 'no' and I would need to roll my own functions that use wchar_t. I am hoping for an answer before I go down this path....

Thanks,
Paul Show Less

Hi folks,

I'm getting started on a second phase of our project, that will see us add CANopen comms.
A question, if I may, about sending byte strings over CANopen, of 7 to 261 bytes..
..if there's someone reading that's clued up on these matters.

From what I've gathered so far, SDO is the right vehicle for this.
Specifically, as 'segmented' or 'block' transfer, rather than the 4-byte limited 'expedited'.
Although in many respects, that's down to the CANopen stack, to handle.

My question, concerns the fact that the strings I wish to send, are variable in length, anywhere between 7 and 261 bytes.
And I don't want to always have to send 261 bytes, and swamp the CAN bus with a ton of frames, each carrying 8 bytes of the 261 !

My first thought, was create an SDO object/index, comprising just a single 262-byte OCTET_STRING
262, to allowing one extra byte, for terminating null, per the Port CANopen stack (demo version) documentation, at least.
And start sending as if I plan to send the full 262 bytes..
.. but bail, with an abort frame, when done.

But I didn't much like that idea. Especially as abort, from what I gather, is intended really for errors,
and diagnostics might record the aborts as such - doesn't look great!

My second thought, was to create an SDO array object, comprising 255 strings, the first being 8 bytes, the next 9 bytes, through to the last, of 262 bytes.
In other words, an SDO object/index, with subindex 0 being the array size, namely 255. Subindex 1 would be a 8-byte OCTET_STRING, subindex 2 would be an 9-byte OCTET_STRING, through to subindex 255 being a 262-byte OCTET_STRING. And use whichever of the strings matches the length of my variable string (including its terminating null).

Nice idea! But the Port stack at least, requires that with an array of OCTET_STRING's, they all have to be the same size. The manual says so. And when I tried the above, in the Port Design Tool, it errors with: Error: 'line 0, index 2000 - Different 'Size' values for sub-indices of this index. This is not allowed for Arrays.'. OK, fair enough.

My third idea then, is to create 255 SDO objects, each comprising just a different length OCTET_STRING, from 8 bytes, through to 262 bytes. And likewise, use whichever string matches.

But, my question is, is this really necessary? Do CANopen stacks, generally, accept an SDO write request, of say a 123-byte string, to an OCTET_STRING object configured to be size 262-bytes? And if so, do they infact only send sufficient frames to communicate the 124 bytes (string and terminating null) ?

And does the receiving stack accept the string? Or reject on the basis of it not equalling 262 bytes ?

I see in the great description, at http://www.byteme.org.uk/canopenparent/canopen/sdo-service-data-objects-canopen/, if you scroll down to 'Write Dictionary Object (segmented)', that the first frame includes unsigned 32-bit 'Len'.

The CANopen standards, unfortunately don't have anything to say on the above. Infact, CiA301 doesn't even indicate the maximum length of an OCTET_STRING, or the other extended data types.

Which leads me to my last question: Anyone know whether the various CANopen stacks / tools, can support a 262 byte OCTET_STRING ?

With the Port Design Tool, I tried entering various lengths, including 9999999999, above the 'Len' limit, but it didn't complain, on entry or code generate, and simply generated the below. So that, unfortunately, didn't inform.

:
extern CO_DT_INIT_VAR OCT_STRING_T xxx[9999999999];
:

Looking at the below, I was concerned to read, 'Please note that the size of an octet string is limited to 127 characters.'.

http://www.port.de/fileadmin/user_upload/Dateien_IST_fuer_Migration/DeviceMonitor/UserMan_CANopen_DevMon_english.pdf

So we might find that tool limiting, for our project..

..anyone have a handle, on which stacks / tools can support OCTET_STRING's of 262 bytes?

Best regards,

David Show Less

Is there anyone know whether the DAVE IDE is capable to display the variable value in the Expression window real time during DEBUG, or not? TI’s code composer studio IDE is able to do it. This is a very useful feature for the code development and test. Show Less

Hi,

Trying to install the DAVE device library, but when I go to the "help" section, there's no option to Install DAVE APP?Example/Device Library...

I'm on DAVE 4.3.2

Was this option removed/integrated with something else or is there a problem with my install?

Thanks,

Richard Li Show Less

The DAVE resource solver is used to map required chip resources to the available resources of the selected MCU considering all relevant constraints defined by the user (e.g. manual pin assignment), or given by the HW design (connectivity options). It is also used for chip relevant plausibility checks when adding a new DAVE App or performing resource relevant configurations.
During a solver run, DAVE does quite a few checks that may result in an error message. Below lists all related error messages, the possible root cause and the suggestion to the users.

Error Message1: "Failed to update one or more resources in the application model, please see the error log for more information”
1. Solver is executed successfully and returned the valid results, but processing of solver results failed
- This could be due to absence or partial installation of DD (Device Description)
2. Suggestion to users: install the Device Descriptions Library (delete or un-install if already installed, details see below)

Error Message2: Error occurred while connecting to the solver process
1. Connection to solver process failed
- communication problem, TCP port may be already in use by other application
2. Already implemented functionality in DAVE 3.1.10: if connection failed, 4 attempts are made to connect to solver process using different port numbers (increment of port number in preferences)
3. Suggestion to users: change the TCP port number in the DAVE CE preference page (->Window ->Preferences ->Dave ->DAVE CE Preferences), restart DAVE

Error Message3: Execution of solver query failed
1. Solver process is launched & connected successfully
2. No response from the Solver (java timeout of 60 seconds)
- Although connected successfully this indicates a communication problem
3. Suggestion to users: Change the TCP port number in DAVE CE preferences & restart DAVE

Error Message4: Resource mapping failed, please review the user defined constraints
1. Solver process is launched & connected successfully
2. Solver returns the result indicating the timeout (by default 15 seconds is the solver timeout) & it is interpreted as “no possible solution” because the chip doesn’t provide the required resources. In very rare cases a solution with a higher time out may be found, therefore it might help to increase the time in the DAVE CE preferences to 30 or 60 sec.
3. Suggestion to users: remove required resources (DAVE Apps) or signal connection or pin assignment constraint and then run the solver again.

Error Message5: Error occurred while creating the solver process
1. Creation of Solver process failed
- This could be due to lack of permission to create new process
2. Suggestion to users: Access rights of user needs to be checked (admin rights missing, suggestion to start DAVE with admin rights: right mouse click on DAVE-3.1.10.exe in the eclipse folder and select “run as administrator”)

Error Message 6: Error occurred while executing solver
This is a general error message and analysis of the log files is required.
The log files (.log and solver.log) are located under the workspace .metadata folder (eg. C:\DAVE3_Workspace\.metadata).

If none of the above hints helps, please provide us the log files so that we can do a more detailed analysis with these files.

Delete (rename) and re-install the Device Descriptions libraries:

The device description libraries are located together with the DAVE Apps libraries and example project Libraries in the local repository: C:\Users\\Infineon\D3LibraryStore_3.1.10

Instead of deletion we recommend to rename the existing library store and then install it again. In this cases the renamed library store can still be used.
The used library store is defined in the Library Manager preferences (Go to: >Window>Preferences>Dave>Library Manager).

Instructions:
Close DAVE
Rename the folder D3LibaryStore_3.1.10 to a different name
Start DAVE

Install the actual DAVE Apps and Device Description Library:
Go to: >Help>Install DAVE Apps/Example Library --> Select "Work with: DAVE Apps Library Manager" --> Press Select All and continue… Show Less

hi
i have A project with 100 boards. I am doing udp packet length 512 and 40 frames per sec. in average every 10 hours one boards hangs an is no longer pingable. is this what you would expect?
Unfortunatly I have no way to connect a debugger.
So how reliable is lwip and freertos? Show Less

Hi Everyone!

does anyone know how to configure the DAVE IDE to use the DAP miniWiggler?

I've installed the DAS software running the UDAS server. My target board is recognised and the memory contents can be read using the Memtool software.
So the target device and cable are working properly. I just now need to configure the DAVE IDE to start debugging.

Any help would be great. Thanks in advance

Kfw Show Less