[support]

Hi,

Sorry, we tried reproducing it on our hardware (ST-Link v1 and v2) and could not get the problem. It looks like OpenOCD may not fully support st-link v2.1 at this point. Does the Texane/STLink tool work for your hardware?

[support]

Hi,

We use a special build of OpenOCD that contains a workaround to the LibUSB problem on Windows 8.1 (the source code modifications are included in the libusb_workaround.patch file together with the package). You mentioned that running the x64 OpenOCD manually works. Does running our build manually from %LOCALAPPDATA%VisualGDBEmbeddedDebugPackagescom.sysprogs.arm.openocd work as well?

The KARL – should read back as 0x03, not 60 02 00 00 message comes from the texane/stlink tool and it looks like it gets printed unconditionally:

static void init_code_breakpoints(stlink_t *sl) {
memset(sl->q_buf, 0, 4);

stlink_write_debug32(sl, CM3_REG_FP_CTRL, 0x03 /*KEY | ENABLE4*/);

printf("KARL - should read back as 0x03, not 60 02 00 00n");

stlink_read_debug32(sl, CM3_REG_FP_CTRL);



for(int i = 0; i < CODE_BREAK_NUM; i++) {

code_breaks.type = 0;

stlink_write_debug32(sl, CM3_REG_FP_COMP0 + i * 4, 0);

}

}

Most likely, it is safe to ignore this message.

[support]

Hi,

VisualGDB supports IntelliSense by providing configuration and compatibility definitions to the Visual Studio IntelliSense engine. However, some GCC-specific constructs still confuse it and cause false positives. We are currently developing an independent clang-based IntelliSense engine that will solve that problem. As a temporary workaround, you can right-click on the errors window and disable IntelliSense errors completely.

[support]

Hi,

LIBUSB_ERROR_ACCESS looks like a permission problem. Does running Visual Studio as Administrator help?

[support]

Hi,

You can also check out our UART tutorial explaining how to fix a similar issue: http://visualgdb.com/tutorials/arm/stm32/uart/

[support]

Hi,

Is uploading the same file with SmarTTY also slow? It is using the same SSH engine.

[support]

Hi,

Can you try starting the Driver tool (select ‘manual mode’ in OpenOCD settings) and trying to manually install the WinUSB-LibUSB driver for the ST-Link device?

[support]

Hi,

Most likely some of your symbols refer to an incorrect path. Can you try running arm-eabi-objdump -g and see if the file tables mentioned in the output contain some suspicious names? If not, can you try commenting out parts of your project to see which source file causes it?

[support]

Hi,

What is the final size of the binary you’re debugging? If it’s several tens of megabytes, it could explain why deployment is slow. You can speed it up by adding a custom action that will strip it and deploy a stripped version (don’t forget to disable auto-deployment and don’t strip the version used by GDB to read symbols). We could also add a deployment progress window if it turns out that the deployment is the bottleneck.

[support]

Hi,
Based on the log, the clean action fails because there is no project built yet (it should be normally ignored), but the build succeeds. Does the normal build command (not rebuild) work?

[support]

Hi,

The disassembly window in Visual Studio sometimes behaves strangely is some of the requested code ranges cannot be disassembled. Can you provide a screenshot of the Disassembly window, the address you’re trying to see and the corresponding GDB log showing what gdb reports when asked to disassemble those addresses?

[support]

Hi,

Looks like 374b is mentioned in the stlink-v2-1.cfg configuration file in the OpenOCD directory. You can edit the %LOCALAPPDATA%VisualGDBEmbeddedDebugPackagescom.sysprogs.arm.openocdQuickSetupinterfaces.xml file to make VisualGDB recognize it automatically:

	

stlink-v2-1

ST-Link v2

interface/stlink-v2-1.cfg





0483

374B





true

NotSupported

com.sysprogs.libusb.mini

[support]

Hi,

Sorry for the delay. Can you send us an archive with your test application (or attach it here) so that we could experiment with it on our side to figure out the root cause of the problem?

[support]

Hi,

Thanks, the deployment was indeed skipped. We have added a default option to deploy it to the upcoming maintenance release.

[support]

Hi,

Your Windows itself may not have the limit, but the GNU make may have been compiled with an old version of the cygwin/MinGW runtime that does. That would explain why the shell started from the Make sees the variable, but the Make itself ignores it.
If this is a critical issue, you could rebuild the Make binary from sources and step through it to see why the variables are lost. Let us know if you need instructions on that.

[Author]

Posts