JTAG Debugging "Failed to verify app image @0x10000 <8194>"

[aaquilina]

Hi,

I'm debugging using JTAG and recently, more often than not, I get the error "Failed to verify app image @0x10000 <8194>" when trying to reprogram for debug. I then try to flash the ESP32 with the hello world program over serial and it succeeds without fail. After a while of a series of reprogramming, unplugging, plugging and resetting, I manage to get it to debug for a short while until this happens again. Full disclosure: I'm using VisualGDB with VisualStudio for debugging but this has been working for the past month and a half. I've also emailed them to see if the issue might be from their side. Any thought?

[WiFive]

Have you considered setting flash voltage in efuse to bypass mtdi strapping pin?

[aaquilina]

Not yet ! How would I do this?

[WiFive]

https://github.com/espressif/esp-idf/co ... 65410c6b5c

[ESP_Angus]

To expand on WiFive's answer, the exact command is:

Code: Select all

components/esptool_py/esptool/espefuse.py set_flash_voltage 3.3V

(MTDI pin will be ignored on reset when setting flash voltage, it will always be 3.3V.)

[aaquilina]

I've managed to set the flash voltage. However, I still getting the same errors while using VisualGDB JTAG debugger. I need to identify where the issue is coming from. Whether it's a VisualGDB issue, JTAG issue or a code issue. Luckily, I have another ESP32 Board with the same issue if anyone has any other suggestion or maybe a follow up?


The ESP32 also outputs this feed when I try debug sometimes. The issue is that I cant really identify a pattern in the different outputs.

[f.h-f.s.]

I've had a similar problem using VisualGDB, They responded on their forum saying to edit the Debug setting Flash timeout in VisualGDB project properties. I set it to 10000.

Or

open with notepad: "YourProjectNameHere-Debug.vgdbsettings" located in your project folder and add

Code: Select all

        <KeyValue>
          <Key>com.sysprogs.esp32.openocd.alg_timeout</Key>
          <Value>10000</Value>
        </KeyValue>

inside the </Entries> </DebugMethodProperties> tags.

My problem was, the debugger moved on to the next bit of programming before it was done. Atleast thats what I think it did. Hope it helps.

[aaquilina]

Thanks for your input. Unfortunately, I already tried this. Its strange because I used the board for a month before it started doing this. I ordered some new boards to test whether its caused by the boards at all. Are you still using VisualGDB?

[aaquilina]

It turns out the second serial output that I posted is caused by something else, so it seems to be two different errors. The second I believe is caused by me hard coding a rather large website (815KB) into a header file. I will start a separate post on this issue. Thanks for your help so far. This also means that the efuse burning may have in fact fixed my issue. I will confirm later today if it occurs again.