From 4271790cb50d8b390abe7d1d82ddd88f7e70c0d2 Mon Sep 17 00:00:00 2001
From: s1341 <s1341@users.noreply.github.com>
Date: Tue, 8 Jun 2021 10:54:38 +0300
Subject: [PATCH] Add unique_name() to Input. Use it to generate filename in
 OnDiskCorpus (#152)

* Add unique_name() to Input. Use unique_name to generate filename in OnDiskCorpus

* updated duplicate ahash

* nostd fixes

* fmt

* rename unique_name to generate_name

Co-authored-by: Dominik Maier <domenukk@gmail.com>
---
 fuzzers/frida_libpng/Cargo.toml |  1 -
 libafl/Cargo.toml               |  3 ++-
 libafl/src/corpus/ondisk.rs     |  8 +++++++-
 libafl/src/inputs/bytes.rs      | 12 +++++++++++-
 libafl/src/inputs/mod.rs        | 14 ++++++++++++--
 5 files changed, 32 insertions(+), 6 deletions(-)

diff --git a/fuzzers/frida_libpng/Cargo.toml b/fuzzers/frida_libpng/Cargo.toml
index c6b48feeb1..589c202717 100644
--- a/fuzzers/frida_libpng/Cargo.toml
+++ b/fuzzers/frida_libpng/Cargo.toml
@@ -30,7 +30,6 @@ libc = "0.2"
 libloading = "0.7.0"
 num-traits = "0.2.14"
 rangemap = "0.1.10"
-seahash = "4.1.0"
 clap = "2.33"
 serde = "1.0"
 
diff --git a/libafl/Cargo.toml b/libafl/Cargo.toml
index 1cfa0b0796..9a5585d503 100644
--- a/libafl/Cargo.toml
+++ b/libafl/Cargo.toml
@@ -16,7 +16,7 @@ rustc_version = "0.3.3"
 
 [dev-dependencies]
 criterion = "0.3" # Benchmarking
-ahash = "0.6.1" # another hash
+ahash = "0.7" # another hash
 fxhash = "0.2.1" # yet another hash
 xxhash-rust = { version = "0.8.2", features = ["xxh3"] } # xxh3 hashing for rust
 serde_json = "1.0.60"
@@ -68,6 +68,7 @@ core_affinity = { version = "0.5", git = "https://github.com/s1341/core_affinity
 num_enum = "0.5.1"
 hostname = "^0.3" # Is there really no gethostname in the stdlib?
 typed-builder = "0.9.0"
+ahash ="0.7" 
 
 [target.'cfg(target_os = "android")'.dependencies]
 backtrace = { version = "0.3", optional = true, default-features = false, features = ["std", "libbacktrace"] } # for llmp_debug
diff --git a/libafl/src/corpus/ondisk.rs b/libafl/src/corpus/ondisk.rs
index 15d7cbc277..8e9503846d 100644
--- a/libafl/src/corpus/ondisk.rs
+++ b/libafl/src/corpus/ondisk.rs
@@ -50,7 +50,13 @@ where
     fn add(&mut self, mut testcase: Testcase<I>) -> Result<usize, Error> {
         if testcase.filename().is_none() {
             // TODO walk entry metadata to ask for pices of filename (e.g. :havoc in AFL)
-            let filename = self.dir_path.join(format!("id_{}", &self.entries.len()));
+            let filename = self.dir_path.join(
+                testcase
+                    .input()
+                    .as_ref()
+                    .unwrap()
+                    .generate_name(self.entries.len()),
+            );
             let filename_str = filename.to_str().expect("Invalid Path");
             testcase.set_filename(filename_str.into());
         };
diff --git a/libafl/src/inputs/bytes.rs b/libafl/src/inputs/bytes.rs
index 2eb97a7e15..c07c2b509d 100644
--- a/libafl/src/inputs/bytes.rs
+++ b/libafl/src/inputs/bytes.rs
@@ -1,7 +1,10 @@
 //! The `BytesInput` is the "normal" input, a map of bytes, that can be sent directly to the client
 //! (As opposed to other, more abstract, imputs, like an Grammar-Based AST Input)
 
-use alloc::{borrow::ToOwned, rc::Rc, vec::Vec};
+use ahash::AHasher;
+use core::hash::Hasher;
+
+use alloc::{borrow::ToOwned, rc::Rc, string::String, vec::Vec};
 use core::{cell::RefCell, convert::From};
 use serde::{Deserialize, Serialize};
 #[cfg(feature = "std")]
@@ -48,6 +51,13 @@ impl Input for BytesInput {
         file.read_to_end(&mut bytes)?;
         Ok(BytesInput::new(bytes))
     }
+
+    /// Generate a name for this input
+    fn generate_name(&self, _idx: usize) -> String {
+        let mut hasher = AHasher::new_with_keys(0, 0);
+        hasher.write(self.bytes());
+        format!("{:016x}", hasher.finish())
+    }
 }
 
 /// Rc Ref-cell from Input
diff --git a/libafl/src/inputs/mod.rs b/libafl/src/inputs/mod.rs
index 831ab8fa9f..d8448f9e63 100644
--- a/libafl/src/inputs/mod.rs
+++ b/libafl/src/inputs/mod.rs
@@ -3,7 +3,10 @@
 pub mod bytes;
 pub use bytes::BytesInput;
 
-use alloc::vec::Vec;
+use alloc::{
+    string::{String, ToString},
+    vec::Vec,
+};
 use core::{clone::Clone, fmt::Debug};
 #[cfg(feature = "std")]
 use std::{
@@ -53,12 +56,19 @@ pub trait Input: Clone + serde::Serialize + serde::de::DeserializeOwned + Debug
     fn from_file<P>(_path: P) -> Result<Self, Error> {
         Err(Error::NotImplemented("Not supprted in no_std".into()))
     }
+
+    /// Generate a name for this input
+    fn generate_name(&self, idx: usize) -> String;
 }
 
 /// An input for tests, mainly. There is no real use much else.
 #[derive(Copy, Clone, Serialize, Deserialize, Debug)]
 pub struct NopInput {}
-impl Input for NopInput {}
+impl Input for NopInput {
+    fn generate_name(&self, _idx: usize) -> String {
+        "nop-input".to_string()
+    }
+}
 impl HasTargetBytes for NopInput {
     fn target_bytes(&self) -> OwnedSlice<u8> {
         OwnedSlice::Owned(vec![0])