0.2 (#2)

* this is a continuation of the pullrequest for pdb_wrapper (#1)

* update dependency versions

* load rva from fakepdb json

* add arg that allows passing clang args

* link the new arg into the code

* make bao compilable by providing an env var

* crash and burn only if we detect an actual error

* pull in the data from the pdb section if it exists

* rewrote parts to be more concise.

* add features to bao and wrap the pdb generation inside of a feature controlled function

* have a default return value if it fails to access the needed debug information

* implement some clippy recomendations

* correctly implement the feature

* start_rva is usize

* correctly setup rng rand

* Make the code more idiomatic, TODO: fix dependency link

* add code to get symbol location from rva_start

* revert the lib change

* update git with working link

* update git with working link

Co-authored-by: noah the goodra <peterpan0413@live.com>
Co-authored-by: jan <not-wlan@protonmail.com>

* 0.2 release, rename binary to bao-pdb for publishing

Co-authored-by: Jordan Francis Moran-Meyers <69991101+jfm535@users.noreply.github.com>
Co-authored-by: noah the goodra <peterpan0413@live.com>

Cargo.toml

@@ -1,8 +1,14 @@
 [package]
-name = "bao"
-version = "0.1.0"
+name = "bao-pdb"
+version = "0.2.0"
 authors = ["jan"]
 edition = "2018"
+build = "build.rs"
+license = "Unlicense"
+description = "bao-pdb can be used to generate PDB files from existing applications."
+homepage = "https://github.com/not-wlan/bao"
+repository = "https://github.com/not-wlan/bao"
+readme = "README.md"
 
 [dependencies]
 snafu = "0.6.9"
@@ -13,18 +19,25 @@ serde_derive = "1.0"
 simplelog = "0.8.0"
 log = "~0.4"
 clap = "2.33"
+uuid = { version = "0.8", features = ["v4"] }
 
 [dependencies.clang-sys]
-version = "1.0.1"
-features = ["clang_10_0"]
+version = "1.2.0"
+features = ["clang_11_0"]
 
 [dependencies.pdb_wrapper]
-git = "https://github.com/not-wlan/pdb_wrapper"
+version = "0.2"
+default-features = false
 
 [dependencies.clang]
-version = "~1.0"
+version = "1.0.3"
 features = ["clang_10_0"]
 
 [dependencies.goblin]
 version = "~0.2"
-features = ["std"]
+features = ["std"]
+
+[features]
+default = ["llvm_13"]
+llvm_10 = ["pdb_wrapper/llvm_10"]
+llvm_13 = ["pdb_wrapper/llvm_13"]

Dockerfile

@@ -7,12 +7,12 @@ RUN printf "deb http://apt.llvm.org/buster/ llvm-toolchain-buster-10 main\ndeb-s
 # Install dependencies
 RUN apt-get update && apt-get install -y curl git make build-essential libllvm10 llvm-10 llvm-10-dev llvm-10-runtime \
     cmake clang-10 clang-tools-10 libclang-common-10-dev libclang-10-dev libclang1-10 clang-format-10 gcc-multilib
-
+RUN ln -s /usr/bin/llvm-config-10 /usr/bin/llvm-config
 WORKDIR /app
 COPY ./Cargo.toml .
 COPY ./Cargo.lock .
 COPY ./src/ /app/src/
-
+COPY build.rs .
 RUN cargo update -p pdb_wrapper && cargo install --path . --target-dir /app/bin
 
 RUN ls -alh /app/bin/release
@@ -26,4 +26,4 @@ RUN bash -c "wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key|apt-key add -"
 RUN printf "deb http://apt.llvm.org/buster/ llvm-toolchain-buster-10 main\ndeb-src http://apt.llvm.org/buster/ llvm-toolchain-buster-10 main" > /etc/apt/sources.list.d/backports.list
 # Install dependencies
 RUN apt-get update && apt-get install -y libllvm10 llvm-10-runtime libclang1-10 gcc-multilib
-COPY --from=0 /app/bin/release/bao /usr/bin/
+COPY --from=0 /app/bin/release/bao-pdb /usr/bin/

README.md

@@ -96,7 +96,7 @@ The easiest way to run _bao_ is to use [Docker](https://www.docker.com/):
 ~$ docker build . -t bao:latest
 ~$ docker run -v /path/to/project:/project -it bao:latest
 #$ cd /project
-#$ bao -o vac.pdb -c vac.json vac.dll structs.c
+#$ bao-pdb -o vac.pdb -c vac.json vac.dll structs.c
 ```
 
 The first three commands are only necessary on your first run or after an update of _bao_.

build.rs

@@ -0,0 +1,6 @@
+fn main() {
+    println!("If the build fails pass the llvm lib folder to env LLVMLIB_DIR");
+    if let Ok(name) = std::env::var("LLVMLIB_DIR") {
+        println!("cargo:rustc-link-search={}", name);
+    }
+}

src/main.rs

@@ -9,6 +9,7 @@ use crate::{
     parsing::{BaoTU, BaoType},
     pe::BaoPE,
 };
+use clang::diagnostic::Severity;
 use clap::{App, Arg};
 use log::{error, info, warn};
 use simplelog::{CombinedLogger, Config, LevelFilter, TermLogger, TerminalMode};
@@ -58,6 +59,14 @@ pub fn main() -> Result<(), Box<dyn Error>> {
                 .required(true)
                 .index(2),
         )
+        .arg(
+            Arg::with_name("coptions")
+                .short("d")
+                .long("coptions")
+                .required(false)
+                .multiple(true)
+                .takes_value(true),
+        )
         .get_matches();
 
     // Unwrapping these is fine since they're marked as required.
@@ -95,25 +104,50 @@ pub fn main() -> Result<(), Box<dyn Error>> {
     let clang = Clang::new()?;
     let index = Index::new(&clang, false, false);
 
-    let tu = BaoTU::from(
-        index
-            .parser(source)
-            .arguments(if pe.is_64 {
-                &["-Werror"]
-            } else {
-                &["-m32", "-Werror"]
-            })
-            .parse()?,
-    );
+    let mut args = matches
+        .values_of("coptions")
+        .map(|values| {
+            values
+                .map(|value| value.replace('\"', ""))
+                .collect::<Vec<_>>()
+        })
+        .unwrap_or_default();
+
+    if !pe.is_64 {
+        args.push(String::from("-m32"));
+    }
+
+    let tu = BaoTU::from(index.parser(source).arguments(&args).parse()?);
 
-    let mut generated = pdb_wrapper::PDB::new(false)?;
+    #[cfg(not(feature = "llvm_13"))]
+    let mut generated = pdb_wrapper::PDB::new(pe.is_64)?;
+
+    #[cfg(feature = "llvm_13")]
+    let mut generated = {
+        let guid = pe
+            .debug_data
+            .and_then(|dbg| dbg.codeview_pdb70_debug_info)
+            .map(|code_view| code_view.signature)
+            .unwrap_or_else(|| *uuid::Uuid::new_v4().as_bytes());
+
+        pdb_wrapper::PDB::new(pe.is_64, 1, 0, guid)
+    }?;
 
     if tu.has_errors() {
-        for error in tu.get_diagnostics() {
-            error!("{}", error);
-        }
+        let has_errors = tu
+            .get_diagnostics()
+            .iter()
+            .any(|diag| diag.get_severity() > Severity::Warning);
+
+        tu.get_diagnostics()
+            .iter()
+            .for_each(|err| error!("{}", err));
+
         info!("Please fix these errors before continuing!");
-        return Ok(());
+
+        if has_errors {
+            return Ok(());
+        }
     }
 
     let funcs = tu.get_entities(EntityKind::FunctionDecl);
@@ -139,7 +173,7 @@ pub fn main() -> Result<(), Box<dyn Error>> {
     // aren't included in a pattern.
     let func_types = funcs
         .into_iter()
-        .map(|func| BaoFunc::try_from(func))
+        .map(BaoFunc::try_from)
         .collect::<Result<Vec<_>, BaoError>>()?
         .into_iter()
         .map(|func| {
@@ -158,7 +192,7 @@ pub fn main() -> Result<(), Box<dyn Error>> {
         .map(|(ty, result)| {
             generated
                 .insert_function(result.index, result.offset, &result.name, ty.cloned())
-                .map_err(|e| BaoError::from(e))
+                .map_err(BaoError::from)
         })
         .collect::<Result<_, BaoError>>()?;
 
@@ -178,7 +212,7 @@ pub fn main() -> Result<(), Box<dyn Error>> {
         .map(|(ty, result)| {
             generated
                 .insert_global(&result.name, result.index, result.offset, ty)
-                .map_err(|e| BaoError::from(e))
+                .map_err(BaoError::from)
         })
         .collect::<Result<_, BaoError>>()?;
 
@@ -188,6 +222,6 @@ pub fn main() -> Result<(), Box<dyn Error>> {
     warnings.into_iter().for_each(|err| warn!("{}", err));
 
     // Finally, save the generated PDB to the path we calculated in the beginning
-    generated.commit(&path, &output)?;
+    generated.commit(path, &output)?;
     Ok(())
 }

src/matching.rs

@@ -6,8 +6,11 @@ use std::convert::TryInto;
 #[derive(Serialize, Deserialize, Debug)]
 pub(crate) struct BaoSymbol {
     pub(crate) name: String,
+    #[serde(default)]
     pub(crate) pattern: String,
     #[serde(default)]
+    pub(crate) start_rva:usize,
+    #[serde(default)]
     extra: isize,
     #[serde(default)]
     pub offsets: Vec<isize>,
@@ -30,7 +33,9 @@ fn add_signed(first: usize, second: isize) -> Option<usize> {
 impl BaoSymbol {
     pub fn find(&self, data: &[u8], imagebase: usize) -> Result<(usize, bool), BaoError> {
         let mut va = false;
-
+        if self.start_rva != 0 {
+            return Ok((self.start_rva + imagebase, true));
+        }
         let peek_bytes = |offset| -> Result<u32, BaoError> {
             let bfr: &[u8; 4] =
                 &data[offset..offset + 4]

src/parsing.rs

@@ -10,18 +10,18 @@ pub(crate) struct BaoFunc {
     pub(crate) cconv: CConv,
 }
 
-impl Into<PDBFunction> for BaoFunc {
-    fn into(self) -> PDBFunction {
-        PDBFunction::new(self.retn, &self.args, self.cconv)
+impl From<BaoFunc> for PDBFunction {
+    fn from(s: BaoFunc) -> PDBFunction {
+        PDBFunction::new(s.retn, &s.args, s.cconv)
     }
 }
 
 #[derive(Debug)]
 pub(crate) struct BaoType(PDBType);
 
-impl Into<PDBType> for BaoType {
-    fn into(self) -> PDBType {
-        self.0
+impl From<BaoType> for PDBType {
+    fn from(s:BaoType) -> PDBType {
+        s.0
     }
 }
 
@@ -195,7 +195,7 @@ impl<'tu> TryFrom<Type<'tu>> for BaoFunc {
                 function: value.get_display_name(),
             })?
             .into_iter()
-            .map(|ty| BaoType::try_from(ty))
+            .map(BaoType::try_from)
             .map(|ty| ty.map(|ty| ty.into()))
             .collect::<Result<Vec<PDBType>, BaoError>>()?;