Refactor debug library to use object::elf::* (#1860)

* Add GDB test * rm stray test resource * use object::elf::* structures * install gdb on CI
veracruz-project · Jun 11, 2020 · 7042403 · 7042403
1 parent b0cccf1
commit 7042403
Show file tree

Hide file tree

Showing 7 changed files with 136 additions and 44 deletions.
diff --git a/.github/actions/define-llvm-env/action.yml b/.github/actions/define-llvm-env/action.yml
@@ -1,5 +1,5 @@
 name: 'Set up a DWARFDUMP env'
-description: 'Set up a DWARFDUMP env (see tests/debug/dump.rs)'
+description: 'Set up a DWARFDUMP env (see tests/all/debug/dump.rs)'
 
 runs:
   using: node12

diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
@@ -225,7 +225,9 @@ jobs:
         RUST_BACKTRACE: 1
 
     # Test debug (DWARF) related functionality.
-    - run: cargo test test_debug_dwarf -- --ignored --test-threads 1
+    - run: |
+        sudo apt-get install -y gdb
+        cargo test test_debug_dwarf -- --ignored --test-threads 1
       if: matrix.os == 'ubuntu-latest'
       env:
         RUST_BACKTRACE: 1

diff --git a/crates/debug/src/lib.rs b/crates/debug/src/lib.rs
@@ -117,43 +117,53 @@ pub fn write_debugsections_image(
 }
 
 fn convert_object_elf_to_loadable_file(bytes: &mut Vec<u8>, code_ptr: *const u8) {
+    use object::elf::*;
+    use object::endian::LittleEndian;
     use std::ffi::CStr;
+    use std::mem::size_of;
     use std::os::raw::c_char;
 
+    let e = LittleEndian;
+    let header: &FileHeader64<LittleEndian> =
+        unsafe { &*(bytes.as_mut_ptr() as *const FileHeader64<_>) };
     assert!(
-        bytes[0x4] == 2 && bytes[0x5] == 1,
-        "bits and endianess in .ELF"
+        header.e_ident.class == ELFCLASS64 && header.e_ident.data == ELFDATA2LSB,
+        "bits and endianess in .ELF",
     );
-    let e_phoff = unsafe { *(bytes.as_ptr().offset(0x20) as *const u64) };
-    let e_phnum = unsafe { *(bytes.as_ptr().offset(0x38) as *const u16) };
     assert!(
-        e_phoff == 0 && e_phnum == 0,
+        header.e_phoff.get(e) == 0 && header.e_phnum.get(e) == 0,
         "program header table is empty"
     );
-    let e_shentsize = unsafe { *(bytes.as_ptr().offset(0x3A) as *const u16) };
-    assert_eq!(e_shentsize, 0x40, "size of sh");
+    let e_shentsize = header.e_shentsize.get(e);
+    assert_eq!(
+        e_shentsize as usize,
+        size_of::<SectionHeader64<LittleEndian>>(),
+        "size of sh"
+    );
 
-    let e_shoff = unsafe { *(bytes.as_ptr().offset(0x28) as *const u64) };
-    let e_shnum = unsafe { *(bytes.as_ptr().offset(0x3C) as *const u16) };
+    let e_shoff = header.e_shoff.get(e);
+    let e_shnum = header.e_shnum.get(e);
     let mut shstrtab_off = 0;
     for i in 0..e_shnum {
         let off = e_shoff as isize + i as isize * e_shentsize as isize;
-        let sh_type = unsafe { *(bytes.as_ptr().offset(off + 0x4) as *const u32) };
-        if sh_type != /* SHT_SYMTAB */ 3 {
+        let section: &SectionHeader64<LittleEndian> =
+            unsafe { &*(bytes.as_ptr().offset(off) as *const SectionHeader64<_>) };
+        if section.sh_type.get(e) != SHT_STRTAB {
             continue;
         }
-        shstrtab_off = unsafe { *(bytes.as_ptr().offset(off + 0x18) as *const u64) };
+        shstrtab_off = section.sh_offset.get(e);
     }
     let mut segment = None;
     for i in 0..e_shnum {
         let off = e_shoff as isize + i as isize * e_shentsize as isize;
-        let sh_type = unsafe { *(bytes.as_ptr().offset(off + 0x4) as *const u32) };
-        if sh_type != /* SHT_PROGBITS */ 1 {
+        let section: &mut SectionHeader64<LittleEndian> =
+            unsafe { &mut *(bytes.as_mut_ptr().offset(off) as *mut SectionHeader64<_>) };
+        if section.sh_type.get(e) != SHT_PROGBITS {
             continue;
         }
         // It is a SHT_PROGBITS, but we need to check sh_name to ensure it is our function
+        let sh_name_off = section.sh_name.get(e);
         let sh_name = unsafe {
-            let sh_name_off = *(bytes.as_ptr().offset(off) as *const u32);
             CStr::from_ptr(
                 bytes
                     .as_ptr()
@@ -170,42 +180,36 @@ fn convert_object_elf_to_loadable_file(bytes: &mut Vec<u8>, code_ptr: *const u8)
         assert!(segment.is_none());
         // Functions was added at write_debugsections_image as .text.all.
         // Patch vaddr, and save file location and its size.
-        unsafe {
-            *(bytes.as_ptr().offset(off + 0x10) as *mut u64) = code_ptr as u64;
-        };
-        let sh_offset = unsafe { *(bytes.as_ptr().offset(off + 0x18) as *const u64) };
-        let sh_size = unsafe { *(bytes.as_ptr().offset(off + 0x20) as *const u64) };
+        section.sh_addr.set(e, code_ptr as u64);
+        let sh_offset = section.sh_offset.get(e);
+        let sh_size = section.sh_size.get(e);
         segment = Some((sh_offset, code_ptr, sh_size));
         // Fix name too: cut it to just ".text"
-        unsafe {
-            let sh_name_off = *(bytes.as_ptr().offset(off) as *const u32);
-            bytes[(shstrtab_off + sh_name_off as u64) as usize + ".text".len()] = 0;
-        }
+        bytes[(shstrtab_off + sh_name_off as u64) as usize + ".text".len()] = 0;
     }
 
     // LLDB wants segment with virtual address set, placing them at the end of ELF.
     let ph_off = bytes.len();
+    let e_phentsize = size_of::<ProgramHeader64<LittleEndian>>();
     if let Some((sh_offset, v_offset, sh_size)) = segment {
-        let segment = vec![0; 0x38];
-        unsafe {
-            *(segment.as_ptr() as *mut u32) = /* PT_LOAD */ 0x1;
-            *(segment.as_ptr().offset(0x8) as *mut u64) = sh_offset;
-            *(segment.as_ptr().offset(0x10) as *mut u64) = v_offset as u64;
-            *(segment.as_ptr().offset(0x18) as *mut u64) = v_offset as u64;
-            *(segment.as_ptr().offset(0x20) as *mut u64) = sh_size;
-            *(segment.as_ptr().offset(0x28) as *mut u64) = sh_size;
-        }
-        bytes.extend_from_slice(&segment);
+        bytes.resize(ph_off + e_phentsize, 0);
+        let program: &mut ProgramHeader64<LittleEndian> =
+            unsafe { &mut *(bytes.as_ptr().add(ph_off) as *mut ProgramHeader64<_>) };
+        program.p_type.set(e, PT_LOAD);
+        program.p_offset.set(e, sh_offset);
+        program.p_vaddr.set(e, v_offset as u64);
+        program.p_paddr.set(e, v_offset as u64);
+        program.p_filesz.set(e, sh_size as u64);
+        program.p_memsz.set(e, sh_size as u64);
     } else {
         unreachable!();
     }
 
     // It is somewhat loadable ELF file at this moment.
-    // Update e_flags, e_phoff, e_phentsize and e_phnum.
-    unsafe {
-        *(bytes.as_ptr().offset(0x10) as *mut u16) = /* ET_DYN */ 3;
-        *(bytes.as_ptr().offset(0x20) as *mut u64) = ph_off as u64;
-        *(bytes.as_ptr().offset(0x36) as *mut u16) = 0x38 as u16;
-        *(bytes.as_ptr().offset(0x38) as *mut u16) = 1 as u16;
-    }
+    let header: &mut FileHeader64<LittleEndian> =
+        unsafe { &mut *(bytes.as_mut_ptr() as *mut FileHeader64<_>) };
+    header.e_type.set(e, ET_DYN);
+    header.e_phoff.set(e, ph_off as u64);
+    header.e_phentsize.set(e, e_phentsize as u16);
+    header.e_phnum.set(e, 1u16);
 }
diff --git a/tests/all/debug/gdb.rs b/tests/all/debug/gdb.rs
@@ -0,0 +1,85 @@
+#![allow(dead_code)]
+
+use anyhow::{bail, format_err, Result};
+use filecheck::{CheckerBuilder, NO_VARIABLES};
+use std::env;
+use std::io::Write;
+use std::process::Command;
+use tempfile::NamedTempFile;
+
+fn gdb_with_script(args: &[&str], script: &str) -> Result<String> {
+    let lldb_path = env::var("GDB").unwrap_or("gdb".to_string());
+    let mut cmd = Command::new(&lldb_path);
+
+    cmd.arg("--batch");
+    let mut script_file = NamedTempFile::new()?;
+    script_file.write(script.as_bytes())?;
+    let script_path = script_file.path().to_str().unwrap();
+    cmd.args(&["-x", &script_path]);
+
+    cmd.arg("--args");
+
+    let mut me = std::env::current_exe().expect("current_exe specified");
+    me.pop(); // chop off the file name
+    me.pop(); // chop off `deps`
+    me.push("wasmtime");
+    cmd.arg(me);
+
+    cmd.args(args);
+
+    let output = cmd.output().expect("success");
+    if !output.status.success() {
+        bail!(
+            "failed to execute {:?}: {}",
+            cmd,
+            String::from_utf8_lossy(&output.stderr),
+        );
+    }
+    Ok(String::from_utf8(output.stdout)?)
+}
+
+fn check_gdb_output(output: &str, directives: &str) -> Result<()> {
+    let mut builder = CheckerBuilder::new();
+    builder
+        .text(directives)
+        .map_err(|e| format_err!("unable to build checker: {:?}", e))?;
+    let checker = builder.finish();
+    let check = checker
+        .explain(output, NO_VARIABLES)
+        .map_err(|e| format_err!("{:?}", e))?;
+    assert!(check.0, "didn't pass check {}", check.1);
+    Ok(())
+}
+
+#[test]
+#[ignore]
+#[cfg(all(target_os = "linux", target_pointer_width = "64"))]
+pub fn test_debug_dwarf_gdb() -> Result<()> {
+    let output = gdb_with_script(
+        &[
+            "-g",
+            "tests/all/debug/testsuite/fib-wasm.wasm",
+            "--invoke",
+            "fib",
+            "3",
+        ],
+        r#"set breakpoint pending on
+b fib
+r
+info locals
+c"#,
+    )?;
+
+    check_gdb_output(
+        &output,
+        r#"
+check: Breakpoint 1 (fib) pending
+check: hit Breakpoint 1
+sameln: fib (n=3)
+check: a = 0
+check: b = 0
+check: exited normally
+"#,
+    )?;
+    Ok(())
+}
diff --git a/tests/all/debug/mod.rs b/tests/all/debug/mod.rs
@@ -1,4 +1,5 @@
 mod dump;
+mod gdb;
 mod lldb;
 mod obj;
 mod simulate;

diff --git a/tests/all/debug/translate.rs b/tests/all/debug/translate.rs
@@ -65,7 +65,7 @@ check:        DW_AT_decl_line	(10)
 ))]
 fn test_debug_dwarf5_translate() -> Result<()> {
     check_wasm(
-        "tests/debug/testsuite/fib-wasm-dwarf5.wasm",
+        "tests/all/debug/testsuite/fib-wasm-dwarf5.wasm",
         r##"
 check: DW_TAG_compile_unit
 # We have "fib" function

diff --git a/tests/debug/testsuite/fib-wasm-dwarf5.wasm b/tests/debug/testsuite/fib-wasm-dwarf5.wasm