DWARF Import : >450,000% Performance Improvement

1 files changed, 13 insertions, 156 deletions

diff --git a/rust/examples/dwarf/shared/src/lib.rs b/rust/examples/dwarf/shared/src/lib.rs
index fd141f1f..0bf11771 100644
--- a/rust/examples/dwarf/shared/src/lib.rs
+++ b/rust/examples/dwarf/shared/src/lib.rs
@@ -1,4 +1,4 @@
-// Copyright 2021-2022 Vector 35 Inc.
+// Copyright 2021-2023 Vector 35 Inc.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -12,7 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use gimli::{Endianity, Error, Reader, ReaderOffsetId, RunTimeEndian, SectionId};
+use gimli::{EndianRcSlice, Endianity, Error, RunTimeEndian, SectionId};
 
 use binaryninja::binaryninjacore_sys::*;
 
@@ -22,10 +22,7 @@ use binaryninja::{
     Endianness,
 };
 
-use std::borrow::Cow;
-use std::convert::TryInto;
-use std::ffi::CString;
-use std::{fmt, str};
+use std::{ffi::CString, rc::Rc};
 
 //////////////////////
 // Dwarf Validation
@@ -69,155 +66,12 @@ pub fn get_endian(view: &BinaryView) -> RunTimeEndian {
     }
 }
 
-#[derive(Clone)]
-pub struct DWARFReader<Endian: Endianity> {
-    data: Vec<u8>,
-    endian: Endian,
-    data_offset: usize,
-    section_offset: usize,
-}
-
-impl<Endian: Endianity> DWARFReader<Endian> {
-    pub fn new(data: Vec<u8>, endian: Endian) -> Self {
-        Self {
-            data,
-            endian,
-            data_offset: 0,
-            section_offset: 0,
-        }
-    }
-}
-
-impl<Endian: Endianity> fmt::Debug for DWARFReader<Endian> {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        let data = if self.data.len() < 6 {
-            self.data.clone()
-        } else {
-            let mut vec = vec![0; 6];
-            vec.clone_from_slice(&self.data[0..6]);
-            vec
-        };
-        f.debug_struct("DWARFReader")
-            .field("data", &data)
-            .field("endian", &self.endian)
-            .field("data_offset", &self.data_offset)
-            .field("section_offset", &self.section_offset)
-            .finish()
-    }
-}
-
-impl<Endian: Endianity> Reader for DWARFReader<Endian> {
-    type Endian = Endian;
-    type Offset = usize;
-
-    fn endian(&self) -> Endian {
-        self.endian
-    }
-
-    fn len(&self) -> usize {
-        self.data.len() - self.data_offset
-    }
-
-    fn empty(&mut self) {
-        self.data.clear();
-        self.data_offset = 0;
-    }
-
-    fn truncate(&mut self, len: usize) -> Result<(), Error> {
-        self.data.truncate(self.data_offset + len);
-        Ok(())
-    }
-
-    fn offset_from(&self, base: &Self) -> usize {
-        (self.section_offset + self.data_offset) - (base.section_offset + base.data_offset)
-    }
-
-    fn offset_id(&self) -> ReaderOffsetId {
-        ReaderOffsetId(self.data_offset.try_into().unwrap())
-    }
-
-    fn lookup_offset_id(&self, id: ReaderOffsetId) -> Option<usize> {
-        Some(id.0.try_into().unwrap())
-    }
-
-    fn find(&self, byte: u8) -> Result<usize, Error> {
-        match self
-            .data
-            .iter()
-            .skip(self.data_offset)
-            .position(|&b| b == byte)
-        {
-            Some(value) => Ok(value),
-            _ => Err(Error::UnexpectedEof(self.offset_id())),
-        }
-    }
-
-    fn skip(&mut self, len: usize) -> Result<(), Error> {
-        if self.data.len() < self.data_offset + len {
-            Err(Error::UnexpectedEof(self.offset_id()))
-        } else {
-            self.data_offset += len;
-            Ok(())
-        }
-    }
-
-    fn split(&mut self, len: usize) -> Result<Self, Error> {
-        if self.data.len() < self.data_offset + len {
-            unreachable!();
-            // Err(Error::UnexpectedEof(self.offset_id()))
-        } else {
-            self.data_offset += len;
-
-            Ok(Self {
-                data: self.data[(self.data_offset - len)..self.data_offset].to_vec(),
-                endian: self.endian,
-                data_offset: 0,
-                section_offset: self.section_offset + self.data_offset - len,
-            })
-        }
-    }
-
-    fn to_slice(&self) -> Result<Cow<'_, [u8]>, Error> {
-        Ok(self.data[self.data_offset..].into())
-    }
-
-    fn to_string(&self) -> Result<Cow<'_, str>, Error> {
-        Ok(str::from_utf8(&self.data[self.data_offset..])
-            .unwrap()
-            .into())
-    }
-
-    fn to_string_lossy(&self) -> Result<Cow<'_, str>, Error> {
-        Ok(str::from_utf8(&self.data[self.data_offset..])
-            .unwrap()
-            .into())
-    }
-
-    fn read_slice(&mut self, buf: &mut [u8]) -> Result<(), Error> {
-        if self.len() >= 4 {
-            let mut vec = vec![0; 4];
-            vec.clone_from_slice(&self.data[self.data_offset..self.data_offset + 4]);
-        }
-
-        if self.data.len() < self.data_offset + buf.len() {
-            Err(Error::UnexpectedEof(self.offset_id()))
-        } else {
-            for b in buf {
-                *b = self.data[self.data_offset];
-                self.data_offset += 1;
-            }
-
-            Ok(())
-        }
-    }
-}
-
 pub fn create_section_reader<'a, Endian: 'a + Endianity>(
     section_id: SectionId,
     view: &'a BinaryView,
     endian: Endian,
     dwo_file: bool,
-) -> Result<DWARFReader<Endian>, Error> {
+) -> Result<EndianRcSlice<Endian>, Error> {
     let section_name = if dwo_file && section_id.dwo_name().is_some() {
         section_id.dwo_name().unwrap()
     } else {
@@ -276,7 +130,7 @@ pub fn create_section_reader<'a, Endian: 'a + Endianity>(
                                 let output_buffer: DataBuffer =
                                     unsafe { std::mem::transmute(raw_buf) };
 
-                                return Ok(DWARFReader::new(
+                                return Ok(EndianRcSlice::new(
                                     output_buffer.get_data().into(),
                                     endian,
                                 ));
@@ -289,16 +143,19 @@ pub fn create_section_reader<'a, Endian: 'a + Endianity>(
         let offset = section.start();
         let len = section.len();
         if len == 0 {
-            Ok(DWARFReader::new(vec![], endian))
+            Ok(EndianRcSlice::new(Rc::from([]), endian))
         } else {
-            Ok(DWARFReader::new(view.read_vec(offset, len), endian))
+            Ok(EndianRcSlice::new(
+                Rc::from(view.read_vec(offset, len).as_slice()),
+                endian,
+            ))
         }
     } else if let Ok(section) = view.section_by_name("__".to_string() + &section_name[1..]) {
-        Ok(DWARFReader::new(
-            view.read_vec(section.start(), section.len()),
+        Ok(EndianRcSlice::new(
+            Rc::from(view.read_vec(section.start(), section.len()).as_slice()),
             endian,
         ))
     } else {
-        Ok(DWARFReader::new(vec![], endian))
+        Ok(EndianRcSlice::new(Rc::from([]), endian))
     }
 }