“Can gdb segfault when checking the value of a certain address after loading a symbol file in Ubuntu?”

My environment:

• Linux My-PC 5.15.90.1-microsoft-standard-WSL2 x86_64 x86_64 x86_64 GNU/Linux
• Distributor ID: Ubuntu
• Description: Ubuntu 22.04.1 LTS

GDB-multiarch version

$ gdb-multiarch -v
GNU gdb (Ubuntu 12.1-0ubuntu1~22.04) 12.1

Selected part of Makefile (Many things are ignored):

AS  =arm-linux-gnueabihf-as
LD  =arm-linux-gnueabihf-ld
NM  =arm-linux-gnueabihf-nm
LDFLAGS =-Timx6ul.lds 
CC  =arm-linux-gnueabihf-gcc  $(RAMDISK)
CFLAGS  =-Wall -fomit-frame-pointer -nostdlib -fno-stack-protector -g

CPP =arm-linux-gnueabihf-cpp -nostdinc -Iinclude -lgcc -L/usr/lib/gcc-cross/arm-linux-gnueabihf/11 -g

.c.s:
    $(CC) $(CFLAGS) 
    -nostdinc -Iinclude -S -o $*.s $<
.s.o:
    $(AS)  -o $*.o $<
.c.o:
    $(CC) $(CFLAGS) 
    -nostdinc -Iinclude -c -o $*.o $<

all:    Image

Image: boot/start tools/system
    arm-linux-gnueabihf-objcopy -O binary -R .note -R .comment tools/system Image
    sync

tools/system:    init/main.o boot/start.o 
    $(ARCHIVES) $(DRIVERS) $(MATH) $(LIBS) 
    mkdir -p tools
    $(LD) $(LDFLAGS) init/main.o boot/start.o 
    $(ARCHIVES) 
    $(DRIVERS) 
    $(MATH) 
    $(LIBS) 
    -o tools/system -lgcc -L /usr/lib/gcc-cross/arm-linux-gnueabihf/11 
    $(NM) -n tools/system > System.map

boot/start:boot/start.S
    $(AS) -o boot/start.o boot/start.S

qemu: all
    qemu-system-arm 
        -machine virt,gic-version=3 
        -m 4G 
        -cpu cortex-a7 
        -nographic 
        -device loader,file=Image,addr=0xC0008000,cpu-num=0 
        -s -S 
        # -serial null -serial mon:stdio 

init.gdb:

target remote localhost:1234
set architecture arm
symbol-file tools/system
x/i 0xc0010f68

The problem is, when I check value at some certain address in gdb terminal, this error occurs. For thousands of instructions, gdb works as what I expect, but I’m so confused that why a segmentation fault occurs when value of certain address is fetched. In this case, I try to fetch instructions at 0xc0010f64 and 0xc0010f6c, which turns out to be no problem.

Fatal signal: Segmentation fault
----- Backtrace -----
0x55e79b1e4197 ???
0x55e79b2e6599 ???
0x55e79b2e6762 ???
0x7f66f400951f ???
        ./signal/../sysdeps/unix/sysv/linux/x86_64/libc_sigaction.c:0
0x55e79b64560c ???
0x55e79b6452c5 ???
0x55e79b646831 ???
0x55e79b6469c9 ???
0x55e79b2761b1 ???
0x55e79b27718b ???
0x55e79b4437d7 ???
0x55e79b447f4b ???
0x55e79b448a0c ???
0x55e79b219774 ???
0x55e79b5a9d94 ???
0x55e79b2e76e4 ???
0x55e79b5ab99e ???
0x55e79b228443 ???
0x55e79b2163b7 ???
0x55e79b3b1da5 ???
0x55e79b3b1e39 ???
0x55e79b3b38d3 ???
0x55e79b3b43fe ???
0x55e79b10e0ef ???
0x7f66f3ff0d8f __libc_start_call_main
        ../sysdeps/nptl/libc_start_call_main.h:58
0x7f66f3ff0e3f __libc_start_main_impl
        ../csu/libc-start.c:392
0x55e79b113e24 ???
0xffffffffffffffff ???
---------------------
A fatal error internal to GDB has been detected, further
debugging is not possible.  GDB will now terminate.

I also found out that this happens only if symbol-file tools/system executes. If symbols are not loaded in gdb, gdb works well, so I wonder what’s happening?

Thanks!