gdbserver, Failed to read memory at 0xfffffffe

[manoj153]

Any idea what is going wrong?

• Development board : stm32h7a3zi-q

openocd.cfg :

source [find interface/stlink-dap.cfg]
transport select dapdirect_swd

set WORKAREASIZE 0x3000
set CHIPNAME STM32H7A3ZI
set BOARDNAME NUCLEO-H7A3ZI_Q

source [find target/stm32h7x.cfg]

# Use connect_assert_srst here to be able to program
# even when core is in sleep mode
reset_config srst_only srst_nogate connect_assert_srst

$_CHIPNAME.cpu0 configure -event gdb-attach {
        echo "Debugger attaching: halting execution"
        gdb_breakpoint_override hard
}

$_CHIPNAME.cpu0 configure -event gdb-detach {
        echo "Debugger detaching: resuming execution"
        resume
}

# Due to the use of connect_assert_srst, running gdb requires
# to reset halt just after openocd init.
rename init old_init
proc init {} {
        old_init
        reset halt
}

openocd runtime info :

Open On-Chip Debugger 0.11.0+dev-00242-g7036ed509-dirty (2021-09-20-14:06)
Licensed under GNU GPL v2
For bug reports, read
    http://openocd.org/doc/doxygen/bugs.html
init
Info : STLINK V3J8M3 (API v3) VID:PID 0483:374E
Info : stlink_dap_op_connect(connect)
Info : SWD DPIDR 0x6ba02477
Info : STM32H7A3ZI.cpu0: hardware has 8 breakpoints, 4 watchpoints
Info : STM32H7A3ZI.cpu0: external reset detected
Info : gdb port disabled
Info : starting gdb server for STM32H7A3ZI.cpu0 on 3333
Info : Listening on port 3333 for gdb connections
target halted due to debug-request, current mode: Thread 
xPSR: 0x01000000 pc: 0x08001530 msp: 0x24000700
    TargetName         Type       Endian TapName            State       
--  ------------------ ---------- ------ ------------------ ------------
 0  STM32H7A3ZI.ap2    mem_ap     little STM32H7A3ZI.cpu    halted
 1* STM32H7A3ZI.cpu0   cortex_m   little STM32H7A3ZI.cpu    halted

target halted due to debug-request, current mode: Thread 
xPSR: 0x01000000 pc: 0x08001530 msp: 0x24000700
Info : Listening on port 6333 for tcl connections
Info : Listening on port 4444 for telnet connections
target halted due to debug-request, current mode: Thread 
xPSR: 0x01000000 pc: 0x08001530 msp: 0x24000700
Info : accepting 'gdb' connection on tcp/3333
Debugger attaching: halting execution
force hard breakpoints
Info : Device: STM32H7Ax/7Bx
Info : flash size probed value 2048
Info : STM32H7 flash has dual banks
Info : Bank (0) size is 1024 kb, base address is 0x08000000
Info : New GDB Connection: 1, Target STM32H7A3ZI.cpu0, state: halted
Error: Failed to read memory at 0xfffffffe

[tarek-bochkati]

Hi,

I acknowledge the issue you have noticed, and confirm that we can reproduce.

We have identified that this issue is mainly caused by bad call stack management at gdb level.
Thus we reported the issue to gdb community (please refer to https://sourceware.org/bugzilla/show_bug.cgi?id=28549 )

Several tentative have been done to fix the gdb stack unwinder, but this required much effort.
Currently there is a promising patch series (at gdb mailing list https://sourceware.org/pipermail/gdb-patches/2022-January/185131.html).

I will keep you posted whenever this is merged upstream.
Once done, you will need to rebuild gdb binary ( if you are using your own toolchain ).
At STM32CubeIDE level, we will try to include the fixed ASAP.

Thank you for reporting this.
Regards,
Tarek