SAM4S + Atmel ICE: Debugger not working

[Physikant]

Hello,

our company is using VisualGDB to develop embedded Software for Microchip SAM devices. We are using the Atmel ICE ISP for this. VisualGDB  is working flawless with the SAME70, however debugging is almost impossible on SAM4S. It’s not a hardware issue; everything is working fine in Atmel/Microchip Studio.

Here is what usually happens:

• the chip erase takes unusually long, but works in the end
• after that, an exception is triggered at the reset handler line “_libc_init_array()”, so before the main(). The exception is: “Received a SIGTRAP: Trace/breakpoint trap”. The execution can not be continued.

Sometimes, a hard erase is needed after this. Sometimes, the execution progresses and it runs to the first breakpoint, but never hits another one after this (although the code is working in Atmel/Microchip Studio).

OpenOCD-Log:

target halted due to debug-request, current mode: Thread
xPSR: 0x01000000 pc: 0x0080004c msp: 0x20000800
Info : Erasing FLASH: 0x00400000-0x00482000...
Info : sam4 does not auto-erase while programming (Erasing relevant sectors)
Info : sam4 First: 0x00000000 Last: 0x00000040
Info : Erasing sector: 0x00000000
Info : Erasing sector: 0x00000001
Info : Erasing sector: 0x00000002
Info : Erasing sector: 0x00000003
Info : Erasing sector: 0x00000004
Info : Erasing sector: 0x00000005
Info : Erasing sector: 0x00000006
Info : Erasing sector: 0x00000007
Info : Erasing sector: 0x00000008
Info : Erasing sector: 0x00000009
Info : Erasing sector: 0x0000000a
Info : Erasing sector: 0x0000000b
Info : Erasing sector: 0x0000000c
Info : Erasing sector: 0x0000000d
Info : Erasing sector: 0x0000000e
Info : Erasing sector: 0x0000000f
Info : Erasing sector: 0x00000010
Info : Erasing sector: 0x00000011
Info : Erasing sector: 0x00000012
Info : Erasing sector: 0x00000013
Info : Erasing sector: 0x00000014
Info : Erasing sector: 0x00000015
Info : Erasing sector: 0x00000016
Info : Erasing sector: 0x00000017
Info : Erasing sector: 0x00000018
Info : Erasing sector: 0x00000019
Info : Erasing sector: 0x0000001a
Info : Erasing sector: 0x0000001b
Info : Erasing sector: 0x0000001c
Info : Erasing sector: 0x0000001d
Info : Erasing sector: 0x0000001e
Info : Erasing sector: 0x0000001f
Info : Erasing sector: 0x00000020
Info : Erasing sector: 0x00000021
Info : Erasing sector: 0x00000022
Info : Erasing sector: 0x00000023
Info : Erasing sector: 0x00000024
Info : Erasing sector: 0x00000025
Info : Erasing sector: 0x00000026
Info : Erasing sector: 0x00000027
Info : Erasing sector: 0x00000028
Info : Erasing sector: 0x00000029
Info : Erasing sector: 0x0000002a
Info : Erasing sector: 0x0000002b
Info : Erasing sector: 0x0000002c
Info : Erasing sector: 0x0000002d
Info : Erasing sector: 0x0000002e
Info : Erasing sector: 0x0000002f
Info : Erasing sector: 0x00000030
Info : Erasing sector: 0x00000031
Info : Erasing sector: 0x00000032
Info : Erasing sector: 0x00000033
Info : Erasing sector: 0x00000034
Info : Erasing sector: 0x00000035
Info : Erasing sector: 0x00000036
Info : Erasing sector: 0x00000037
Info : Erasing sector: 0x00000038
Info : Erasing sector: 0x00000039
Info : Erasing sector: 0x0000003a
Info : Erasing sector: 0x0000003b
Info : Erasing sector: 0x0000003c
Info : Erasing sector: 0x0000003d
Info : Erasing sector: 0x0000003e
Info : Erasing sector: 0x0000003f
Info : Erasing sector: 0x00000040
Warn : keep_alive() was not invoked in the 1000 ms timelimit. GDB alive packet not sent! (3248 ms). Workaround: increase "set remotetimeout" in GDB
Info : Programming FLASH (2 sections, 525720 bytes)...
Info : Programming FLASH section 1/2 (204 bytes) at 0x00400000...
Info : Padding image section 0 at 0x004000cc with 4 bytes
Warn : keep_alive() was not invoked in the 1000 ms timelimit. GDB alive packet not sent! (24901 ms). Workaround: increase "set remotetimeout" in GDB
target halted due to debug-request, current mode: Thread
xPSR: 0x01000000 pc: 0x0080004c msp: 0x20000800

Any idea what could help?

Best regards,

Nikolas Becker

• This topic was modified 2 years, 4 months ago by support. Reason: formatting

[support]

Hi,

This looks like an issue in OpenOCD rather than something VisualGDB-specific. You can try building OpenOCD from sources by following this tutorial, and then step through it to understand what is going on.

Another option would be to switch to Segger J-Link. It comes with its own fully supported replacement to OpenOCD, that generally works better for many devices.

If the Atmel IDE works well, you can also try using Process Monitor to to see if uses gdb and whether it launches its own gdb stub similar to OpenOCD. If both are true, and you can run the same stub manually (and connect GDB to it), you can configure VisualGDB to launch it as well by selecting the “Custom GDB stub” mode in VisualGDB Project Properties -> Debug Settings. Note that the last option is something to do at your own risk, as it involves running undocumented Atmel tools.

[Author]

Posts