e- segfaults and crashes

[divyaranjan1905]

Work environment

Questions	Answers
OS/arch/bits (mandatory)	GNU Guix, AMD64
File format of the file you reverse (mandatory)	ELF
Architecture/bits of the file (mandatory)	8051
rizin -v full output, not truncated (mandatory)	rizin 0.7.2 @ linux-x86-64, package: 0.7.2 (guix)

Expected behavior

As listed by e?, e- should reset the config variables.

Actual behavior

Doing e- crashes with a segmentation fault and no further error or backtrace is provided.

Steps to reproduce the behavior

• The bug is irrelevant of the binary, opening rizin without any binary also suffers from the same issue.
• Recording: https://asciinema.org/a/ruOqyaRdG5T1oBgjxLjzc1oSM

[divyaranjan1905]

Even though GNU Guix has only packaged upto 0.7.2, the same issue can be reproduced with the latest build of rizin from the dev branch.

[Rot127]

Stack trace:

[0x00000000]> e-
../librz/arch/platform_profile.c:94:18: runtime error: null pointer passed as argument 2, which is declared to never be null
    #0 0x7115cd819b1a in cpu_reload_needed ../librz/arch/platform_profile.c:94
    #1 0x7115cd81ae91 in rz_platform_profiles_init ../librz/arch/platform_profile.c:219
    #2 0x7115c6db6635 in cb_asmcpu ../librz/core/cconfig.c:429
    #3 0x7115d303a7b5 in rz_config_set_cb ../librz/config/config.c:173
    #4 0x7115c6de6f63 in rz_core_config_init ../librz/core/cconfig.c:3281
    #5 0x7115c71eda79 in rz_eval_reset_handler ../librz/core/cmd/cmd_eval.c:452
    #6 0x7115c71c9ba8 in argv_call_cb ../librz/core/cmd/cmd_api.c:742
    #7 0x7115c71ca9e4 in call_cd ../librz/core/cmd/cmd_api.c:801
    #8 0x7115c71cac5c in rz_cmd_call_parsed_args ../librz/core/cmd/cmd_api.c:819
    #9 0x7115c7197c56 in handle_ts_arged_stmt_internal ../librz/core/cmd/cmd.c:3216
    #10 0x7115c7196a38 in handle_ts_arged_stmt ../librz/core/cmd/cmd.c:3164
    #11 0x7115c71b81d5 in handle_ts_stmt ../librz/core/cmd/cmd.c:4803
    #12 0x7115c71b9a43 in handle_ts_statements_internal ../librz/core/cmd/cmd.c:4860
    #13 0x7115c71b8af8 in handle_ts_statements ../librz/core/cmd/cmd.c:4825
    #14 0x7115c71baf40 in core_cmd_tsrzcmd ../librz/core/cmd/cmd.c:4972
    #15 0x7115c71bb7a1 in rz_core_cmd ../librz/core/cmd/cmd.c:5020
    #16 0x7115c6e83ecd in rz_core_prompt_exec ../librz/core/core.c:1931
    #17 0x7115c6e81d8c in rz_core_prompt_loop ../librz/core/core.c:1800
    #18 0x7115d32e3c1a in rz_main_rizin ../librz/main/rizin.c:1469
    #19 0x401a5f in main ../binrz/rizin/rizin.c:57
    #20 0x7115d2838087 in __libc_start_call_main (/lib64/libc.so.6+0x2a087) (BuildId: c8c3fa52aaee3f5d73b6fd862e39e9d4c010b6ba)
    #21 0x7115d283814a in __libc_start_main_impl (/lib64/libc.so.6+0x2a14a) (BuildId: c8c3fa52aaee3f5d73b6fd862e39e9d4c010b6ba)
    #22 0x4011d4 in _start (/usr/local/bin/rizin+0x4011d4) (BuildId: f4f2bfc3b96322cc69c294647c791a386cb77ce9)

SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior ../librz/arch/platform_profile.c:94:18 in 
Aborted (core dumped)

[divyaranjan1905]

Since I'm interested in getting to work with rizin better, so I'll take this as an opportunity to debug and work on this. But from @Rot127 's stack trace, it looks like we have got a classic null-pointer problem?

[Rot127]

Yeah, there is an rz_return_val_if_fail() missing in rz_platform_profiles_init. But it also should never pass a NULL pointer there. Just adding the assert is not enough. Feel free to debug why it passes a NULL pointer in the first place :)