shell cmd: "devmem load" only works for addresses below 0x80000000

[AaronFontaine-DojoFive]

Describe the bug

Shell command "devmem load" works properly only if the address parameter is less than 0x80000000. If it is equal to or above 0x80000000, then a bus fault will occur after the first byte (i.e. the first two hex characters) have been entered into the shell.

The problem is that cmd_load() is using strtol() when assigning to bytes and data. Everywhere else in zephyr/subsys/shell/modules/devmem_service.c uses strtoul(). This seems to be an oversight in the cmd_load() function since addresses can never be negative.

If the value is outside the range of a signed long, then strtol() returns LONG_MAX, which in my case is 0x7FFFFFFF. This then gets used as the address to start writing bytes to. As soon as the first byte is entered in the shell, bypass_cb() tries to write it and bus faults.

To Reproduce

Steps to reproduce the behavior:

1. Choose any MCU that places a memory region above 0x80000000. In my case, the STM32H7 maps the external SDRAM starting at 0xC0000000.
2. Ensure that the shell is enabled. CONFIG_SHELL=y in prj.conf.
3. Build and flash the board and connect a terminal program to the shell
4. Issue command devmem load 0xC0000000
5. Start typing hexadecimal characters
6. After 2 characters, a bus fault occurs

Expected behavior

"devmem load" works for addresses above 0x80000000.

Impact

Minimal. I can work around this, by changing the strotol() calls to strtoul(), but it requires managing and applying a Zephyr patch file in the West workspace.

Logs and console output

[00:00:21.883,000] <err> os: ***** BUS FAULT *****
[00:00:21.883,000] <err> os:   Imprecise data bus error
[00:00:21.883,000] <err> os: r0/a1:  0x00000023  r1/a2:  0x00000000  r2/a3:  0x2400cc70
[00:00:21.883,000] <err> os: r3/a4:  0x00000000 r12/ip:  0x0000000f r14/lr:  0x0800884d
[00:00:21.883,000] <err> os:  xpsr:  0x41000000
[00:00:21.883,000] <err> os: Faulting instruction address (r15/pc): 0x08009004
[00:00:21.883,000] <err> os: >>> ZEPHYR FATAL ERROR 26: Unknown error on CPU 0
[00:00:21.883,000] <err> os: Current thread: 0x24009608 (shell_uart)

0x08009004 is in shell_process().
0x0800884d is in bypass_cb().

Environment (please complete the following information):

• OS: WSL/Ubuntu 22.04 on Windows 10 22H2
• Zephyr SDK sdk0.16.5-1
• v3.6.0, but problem is not fixed in v3.7.0

Additional context

None