Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix symbol resolver issues #562

Open
wants to merge 4 commits into
base: master
Choose a base branch
from

Conversation

endofunky
Copy link

@endofunky endofunky commented Sep 24, 2023

I've been having some frustrating hours of woes with the sym_resolver - mainly it forcing a radix change and symbol offsets to be calculated correctly.

This is mostly based on suggestions from other issues and a cherry-picked commit from another PR. I've added a few more examples to bindings/python/sample.py to illustrate it seemingly behaving correctly now. I've also added regression tests for anything mentioned below. All the examples in that file now also run for me again - on master they appear broken. Details below.

1. Invalid symbol offsets

Issues: #244 #271 #351

Given a symbol table:

SYM   ADDR
_l1   0x1000

On current master with base 0x1000:

jmp _l1
nop

Assembles to eb ff 90. That offset is incorrect and looks like it's calculated from the start of the instruction rather than the end of it, missing the offset for it.

The suggestion from #351 fixes that and symbol offsets now appear to be calculated correctly. With the suggested change the instructions correctly assemble to eb fe 90.

I've fixed the Python sym_resolver examples. The x64_sym_resolver.py regression test, too, which also passes.

2. Radix set from default (10) to 16 when setting sym_resolver

Issues: #481 #436 #538

Cherry-picked 5c7ed87 from #528

Another issue is present in the setting of ks_option:

KEYSTONE_EXPORT
ks_err ks_option(ks_engine *ks, ks_opt_type type, size_t value)
{
    ks->MAI->setRadix(16);
    switch(type) {
        case KS_OPT_SYNTAX:
            if (ks->arch != KS_ARCH_X86)
                return KS_ERR_OPT_INVALID;
            switch(value) {
                default:
                    return KS_ERR_OPT_INVALID;
                case KS_OPT_SYNTAX_RADIX16: // default syntax is Intel
                case KS_OPT_SYNTAX_NASM | KS_OPT_SYNTAX_RADIX16:
                case KS_OPT_SYNTAX_INTEL | KS_OPT_SYNTAX_RADIX16:
                    ks->MAI->setRadix(16);
                case KS_OPT_SYNTAX_NASM:
                case KS_OPT_SYNTAX_INTEL:
                    ks->syntax = (ks_opt_value)value;
                    ks->MAI->setAssemblerDialect(1);
                    break;
                case KS_OPT_SYNTAX_GAS | KS_OPT_SYNTAX_RADIX16:
                case KS_OPT_SYNTAX_ATT | KS_OPT_SYNTAX_RADIX16:
                    ks->MAI->setRadix(16);
                case KS_OPT_SYNTAX_GAS:
                case KS_OPT_SYNTAX_ATT:
                    ks->syntax = (ks_opt_value)value;
                    ks->MAI->setAssemblerDialect(0);
                    break;
            }

            return KS_ERR_OK;
        case KS_OPT_SYM_RESOLVER:
            ks->sym_resolver = (ks_sym_resolver)value;
            return KS_ERR_OK;
    }

    return KS_ERR_OPT_INVALID;
}

This currently forces a radix of 16 every time any ks_option call is made. So when a symbol resolver is set without explicitly setting a radix of 16, one still ends up with a radix of 16.

This change for this was made for #382 to address https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=10437 but unless ks_option is called at least once, the issue still remains.

This is now also now covered by a new regression test.

In #382 was also a short discussion if the default should be 16 or 10 and it was set to 16. This also appears to be incorrect because the actual default in keystone unless explicitly changed is a radix of 10, which is also NASM's default.

Notes

I hope we can get this (or a similar fix) merged a new version released. I'm happy to write additional regression tests or make any necessary needed. My current project has some code for incremental multi-pass assembling on top of keystone where I handle the symbol table myself and this is currently blocking my work :(

@endofunky
Copy link
Author

I can't see what's failing on the semaphoreci CI step. The appveyor task is failing on master with the same error. So I assume they're both the same failures that are occurring on master as well.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants