940ef05371
Previously, we would reject inbound channels if the funder wasn't able to meet our channel reserve on their first commitment transaction only if they also failed to push enough to us for us to not meet their initial channel reserve as well. There's not a lot of reason to care about us meeting their reserve, however - its largely expected that they may not push enough to us in the initial open to meet it, and its not actually our problem if they don't. Further, we used our own fee, instead of the channel's actual fee, to calculate fee affordability of the initial commitment transaction. We resolve both issues here, rewriting the combined affordability check conditionals in inbound channel open handling and adding a fee affordability check for outbound channels as well. The prior code may have allowed a counterparty to start the channel with "no punishment" states - violating the reason for the reserve threshold. |
||
|---|---|---|
| .. | ||
| src | ||
| .gitignore | ||
| Cargo.toml | ||
| ci-fuzz.sh | ||
| README.md | ||
| targets.h | ||
Fuzzing
Fuzz tests generate a ton of random parameter arguments to the program and then validate that none cause it to crash.
How does it work?
Typically, Travis CI will run travis-fuzz.sh on one of the environments the automated tests are configured for.
This is the most time-consuming component of the continuous integration workflow, so it is recommended that you detect
issues locally, and Travis merely acts as a sanity check. Fuzzing is further only effective with
a lot of CPU time, indicating that if crash scenarios are discovered on Travis with its low
runtime constraints, the crash is caused relatively easily.
How do I run fuzz tests locally?
You typically won't need to run the entire combination of different fuzzing tools. For local execution, honggfuzz
should be more than sufficient.
Setup
To install honggfuzz, simply run
cargo update
cargo install --force honggfuzz
Execution
To run the Hongg fuzzer, do
export CPU_COUNT=1 # replace as needed
export HFUZZ_BUILD_ARGS="--features honggfuzz_fuzz"
export HFUZZ_RUN_ARGS="-n $CPU_COUNT --exit_upon_crash"
export TARGET="msg_ping_target" # replace with the target to be fuzzed
cargo hfuzz run $TARGET
To see a list of available fuzzing targets, run:
ls ./src/bin/
A fuzz test failed on Travis, what do I do?
You're trying to create a PR, but need to find the underlying cause of that pesky fuzz failure blocking the merge?
Worry not, for this is easily traced.
If your Travis output log looks like this:
Size:639 (i,b,hw,ed,ip,cmp): 0/0/0/0/0/1, Tot:0/0/0/2036/5/28604
Seen a crash. Terminating all fuzzing threads
… # a lot of lines in between
<0x0000555555565559> [func:UNKNOWN file: line:0 module:/home/travis/build/rust-bitcoin/rust-lightning/fuzz/hfuzz_target/x86_64-unknown-linux-gnu/release/full_stack_target]
<0x0000000000000000> [func:UNKNOWN file: line:0 module:UNKNOWN]
=====================================================================
2d3136383734090101010101010101010101010101010101010101010101
010101010100040101010101010101010101010103010101010100010101
0069d07c319a4961
The command "if [ "$(rustup show | grep default | grep stable)" != "" ]; then cd fuzz && cargo test --verbose && ./travis-fuzz.sh; fi" exited with 1.
Note that the penultimate stack trace line ends in release/full_stack_target]. That indicates that
the failing target was full_stack. To reproduce the error locally, simply copy the hex,
and run the following from the fuzz directory:
export TARGET="full_stack" # adjust for your output
export HEX="2d3136383734090101010101010101010101010101010101010101010101\
010101010100040101010101010101010101010103010101010100010101\
0069d07c319a4961" # adjust for your output
mkdir -p ./test_cases/$TARGET
echo $HEX | xxd -r -p > ./test_cases/$TARGET/any_filename_works
export RUST_BACKTRACE=1
export RUSTFLAGS="--cfg=fuzzing"
cargo test
This will reproduce the failing fuzz input and yield a usable stack trace.