[gdb] target state locked in "running" after step instruction with st-util and gdb #831
Closed
6 tasks done
Labels
Milestone
Dear developers and maintainers of
stlink
,I am trying to debug a simple program for the STM32F769I-DISCO board with Stlink/v2. I am using an Ubuntu machine, st-util for the gdb server and gdb-multiarch for the client. The connection to the target and loading of the program are successful, as well as the detection of the target in st-util (at least it seems so). However, when I try to step through the first instruction (
stepi
in gdb), it hangs, while the target seems to enter into continuous execution mode. Verbose output ofst-util
reveals that indeed, the core status isrunning
.Attached version information and output logs.
Ubuntu:
st-util:
gdb-multiarch:
Target identification:
STLink Firmware:
st-info
,st-flash
,st-util
Reproduction steps:
st-util -v99
gdb-multiarch program.elf
and connect to target, plus load the program into memoryat this point GDB cli looks like this (additional disassemble run)
nexti
GDB commandexpected result: the target executes one instruction and gdb regains control afterwards
observed result: the target enters some kind of continuous execution mode, GDB never gains control again.
Moreover, if I stop the execution with ^C in GDB,
PC
is shown to point to0x00000000
or, in other cases to random addresses.Output from st-util (clipped):
Please note that at line
2019-09-14T11:53:12 DEBUG common.c: *** stlink_force_debug_mode ***
of output, I pressed ^C to stop execution in GDB.This seems to be in connection to issue #627 however, premises differ since now we are using a Stlinkv2 with a STM32F7x. I also managed to reproduce the same issue on a STM32VLDISCOVERY. This started happening after updating the firmware to the specified version above. Before the update, the step instruction command worked fine.
The text was updated successfully, but these errors were encountered: