[support]

Hi,

Please try updating to VisualGDB 5.3 Preview 8 and installing the latest ESP32 toolchain. It includes a FLASH programming mechanism provided by Espressif that might be more tolerant to non-typical configurations (our FLASH programming algorithm was based on undocumented API). If it doesn’t work either, please attach both OpenOCD logs and the full gdb log to see which exact command fails and what error is related to the failing command.

[support]

Hi,

This looks like a gdb bug actually. Can you confirm that you also get the wrong value when running “print <expression>” manually via the GDB Session window?

[support]

Hi,

It looks like you are trying to create a live variable for an expression that is not valid in the current context. Is it a global variable visible from all the source files or a local variable inside a function/method?

[support]

Hi,

Please check the OpenOCD window for errors. The GDB log only shows that OpenOCD refused the connection, while the real reason for the error is normally shown in the OpenOCD log.

[support]

Hi,

Please try using the “Test OpenOCD Settings” button (or installing VisualGDB 5.3 Preview 7). This should automatically fix the driver issues.

[support]

Hi,

Yes, but you would need to contact our support in order to get your previous license deactivated from our side.

[support]

Hi,

Thanks for confirming this. This could be caused by IntelliSense and the compiler using different sets of headers or different build parameters. Please let us know the type of the project (Embedded/Linux/Android) and the machine on which it is built (Windows or Linux via SSH). If you could also attach a screenshot of the problem, we might be able to diagnose it by checking for known irregularities.

[support]

Hi,

This feature hasn’t been officially announced yet and will be included in the upcoming Preview 8.

In the meanwhile you can try the following build that already includes it: http://sysprogs.com/files/tmp/VisualGDB-5.3.7.1769.msi

[support]

OK, we have finally updated our KSDK importer to work with the latest v2.1 and v2.2 KSDK releases. You can download a build supporting it here: http://sysprogs.com/files/tmp/VisualGDB-5.3.7.1769.msi

[support]

Hi,

Thanks for confirming this. We have retested it and confirmed that the exception would happen before VisualGDB asks to download the matching sources. We have fixed it in this build: http://sysprogs.com/files/tmp/VisualGDB-5.3.7.1767.msi

[support]

Hi,

This error happens because VisualGDB cannot automatically download the source code for the system library you stepped in. If your PC does not have an Internet connection, simply choose “no” when VisualGDB suggests downloading the source code and it will show the disassembly instead.

[support]

Hi,

The virtual drive could be using a different FLASH programming mechanism (e.g. different frequency) that does not trigger the bug, so it’s hard to say what exactly happens without an in-depth analysis of what is going on with the board. As the problem looks specific to one board instance, it could be easier to simply get another one.

[support]

Hi,

Thanks for the detailed log. If the FLASH programming fails, it could be an indication of a damaged board, so simply getting a new one could be the easiest fix.

Another option would be to try using J-FLASH or other Segger-specific tools for programming FLASH and see if they report any errors when you try to program the same ELF file.

[support]

Hi,

It looks like VisualGDB sent the “flushregs” command before the last “evaluate-expression” was complete:

^done,value=”0”
-data-evaluate-expression “\$s30″
^done,value=”0”
-data-evaluate-expression “\$s31″
-interpreter-exec console flushregs
-gdb-set $r0=0x1
~”The target is not responding to GDB commands.\nStop debugging it? ”
~”(y or n) [answered Y; input not from terminal]\n”
=thread-group-exited,id=”i1″
^error,msg=”Disconnected from target.”
-gdb-set $r1=0x20003448
~”Register cache flushed.\n”

Did you have to cancel the “-data-evaluate-expression” command via the “GDB Command Taking Too Long” window? If not, it could be caused by some strange race condition and we can add soem extra logging on VisualGDB side to help you diagnose this.

[support]

Hi,

Please ensure you are using Clang-based IntelliSense. It supports GCC-specific code much better than the regular VS IntelliSense.

[Author]

Posts