* squash libfuzzer edits
* fixup: compat with custom mutators
* use tui flag
* add introspection support
* use libfuzzer dep now that we've merged
* force input loading
* some fixes
* begin docs, impl shrink
* make whole-archive conditional and not default
* make more copies of counters maps
* lol, remember to add the observer
* make size edge map observer an observer
* fixup: make def of run driver conditional
* add sanity checks for insertion
* revert silencing of forks
* add experimental tmin support; add default asan flags
* use default options instead of specifying our own
* implement lockless mode
* fix merge
* fixup lockless corpus
* fixup for generalisation
* remove erroneous drop_in_place
* improve error logging in the case of corpus loading failure
* ok, use lock files 😔
* fix tmin
* implement merge (again); fix rare cases with maps being too small
* implement a scheduler for removing excess
* implement a walking strategy for corpus loading for large corpora
* revert filename parameter; rename and remove duplicates
* various cleanup and clippy satisfaction
* fix no_std tests
* clang-format
* expand and satisfy the clippy gods
* fix sanitizer_ifaces bindgen for no_std
* fix wasm fuzzer
* fixup clippy script
* rename and provide a small amount of explanation for sanitizer_interfaces
* fixup: HasLastReportTime
* fix clippy oddities
* restrict clippy checks to linux-only for libafl_libfuzzer_runtime
* name the mutators
* format
* fix clippy warning
* hope docker is fixed
* fix cmin lint
* clippy pass
* more docs
* more clippy
* fix remaining clippy complaints
* fix import
* miri fixes (no constructors executed)
* exclude libafl_libfuzzer from cargo-hack
* fix clippy check for sanitizer_interfaces
* fmt
* fix CI (?)
* deduplicate sancov 8bit for improved perf on ASAN
* merge 8bit coverage regions + comment out insane deduplication
* no erroring out on free hooks
* fixup for non-forking merge
* skip the corpus dir if we use it
* fixup: recent libafl changes and feature flags
* libafl_libfuzzer: use rust-lld for whole-archive feature
* clarify cause of failure
* mark unsafe
* clippy :cursed_cowboy:
* attempt to fix wasm
* spooky unknowable bug 👻
* more clippy lints
* clippy fix for merge
* use the version pin
* add unsafe to ::register
* Serdeany autoreg fix
* make type assert actionable
* miri fixes
---------
Co-authored-by: Dominik Maier <domenukk@gmail.com>
Co-authored-by: Dominik Maier <dmnk@google.com>
Co-authored-by: Mrmaxmeier <Mrmaxmeier@gmail.com>
* Add more features to libafl_qemu to remove some of the heavy lifting from the fuzzers
* Refactor qemu_coverage
* Minor tweaks to fix other fuzzers
* Autofix
* Add CallingConvention to write_function_argument
* Replay reverted clippy fixes
---------
Co-authored-by: Your Name <you@example.com>
* Change executor trait to allow \&mut Input
* Add mut inprocess executor
* Add mut inprocess executor
* Format and fix clippy errors
* Fix more clippy errors
* Revert accidental refactoring of InMemoryCorpus
* Add mut versions of all executors that can support it
* Do not persist possible testcase mutation in stages, shadow/differential executors, or corpus minimization
* Fix missing imports
* Fix executor type for missed qemu items
* Add re-exports for mut executors
* Use InProcessForkExecutorMut in QemuForkExecutorMut
* Update BytesInput harnesses to take mutable references
* Update other-input-type-taking harnesses to take mut references
* Clippy fixes
* Feature gate TryFromIntError import
* Fix missed harness input type in baby_fuzzer
* Fix additional clippy issues
* Fix unnecessary hashes on string literal
* Even MORE clippy fixes
* Fix one more clippy issue
---------
Co-authored-by: Dongjia "toka" Zhang <tokazerkje@outlook.com>
* qemu: Return errors from Emulator::new instead of asserting
Libraries should not `assert!` except in cases of unrecoverable (library)
programmer error. These errors are all potentially recoverable, and aren't
internal errors in `libafl_qemu` itself.
* Respond to review comments
* Created macro to get the metadata form State and Testcase
* Expanded the macros for mutable, or not, State and Testcase metadata
* Created functions on traits HasMetadata and HasNamedMetadatato get, mutable or not, metadata
* Created the functions to get metadata
* Added #[inline] attribute and renamed the functions
* Renamed the functions and added #[inline] attribute
* Temporarily added testcase() function
* Added testcase() function
* Changed Ref import to core::cell:Ref
* Added testcase_mut() and renamed occurences of metadata() and metadata_mut()
* Renamed more occurences
* Renamed the metadata() on impl HasMetadata for NopState
---------
Co-authored-by: Andrea Fioraldi <andreafioraldi@gmail.com>
* scheduler replace fixes
* oops, no-std
* add
* changes on the fuzzers
* move map feedback history updates to append_metadata
* fixes for python bindings
* learn to clippy
* fix for fuzzer add_input
* clippy fixes for frida
* additional powersched differences
* corrections for bitmap_size
* off-by-one
* I live in a prison of my own creation and clippy is the warden
* clear the novelties map for the situation where is_interesting is invoked, but not append_metadata
---------
Co-authored-by: tokatoka <tokazerkje@outlook.com>
Methods read_mem and write_mem now operate on &[u8], not &[T]
The generic T slice interface was prone to various footguns:
* i32 is the default Rust integer type, but buffers are often expected
to hold u8. This means the following code writes 16 bytes to the
guest, not 4:
let buf = [0; 4];
emu.write_mem(addr, &buf);
* If a buffer of 16-bit or larger integers (&[u64] for example) is
needed to read/write, the user will need to consider host/guest
endianness. The byte array methods in std are a good, explicit
alternative.
Perhaps libafl_qemu could expose/define "to/from guest endianness"
helper functions or extension traits using the established cfg flags,
so that guest endianness is always right by default.
* emu::read_mem causes insta-UB if a user did something like:
let mut my_bool = false;
emu.read_mem(addr, &mut my_bool);
It's less surprising for users to just operate on plain-ol' bytes,
which they can explicitly transmute if they wish.
* store executor_ptr
* QemuHelpers
* working hooks and snapshot helper
* walk only the list of dirty pages on restore()
* mem hooks for snpashot
* brk snapshot
* snapshot method
* macos shit
* sugar and clippy
* starting to fix macos linker bugs
* mdetailed error prints
* start shmem service manually
* not a doc comment
* Some fixes
* only send exit msg to shmemservice when start was successful
* incorporated shmem service into provider
* removed unused imports
* trying to fix fuzzers
* fixed build
* check if join_handle is_some
* more debug prints
* fixed shmem service autolaunch
* fixed macos linker
* ignoring broken libpng testcase on macos for now (see #252)
* fmt
* try to fix missing llvm_config (see #253)
* empty issue template added
* Mmanually look for llvm-config on MacOS
* fixing CI
* fixed docs
* ignoring libmozjpg for CI, see #254
* empty libafl_qemu crate
* fuzzbench qemu fuzzer skeleton
* emu.run() works without bp
* working emu loop
* resolve elf symbols
* running Qemu fuzzer without coverage
* qemu fuzzer with edge coverage
* merge into inprocess::GLOBAL_STATE
* create QemuExecutor and remove QemuEmulator
* qemu hooks and persist edges mapping storing them in State
* windows fix
* add libafl_qemu to workspace
* windows fix
* some clippy
* clippy
* fix fuzzbench_qemu
* fix fuzzbench_qemu makefile
* fuck you macos
