basic storage allocation

Cargo.lock

@@ -197,12 +197,19 @@ dependencies = [
  "bstr",
  "clap",
  "elven-parser",
+ "indexmap",
  "memmap2",
  "tempfile",
  "tracing",
  "tracing-subscriber",
 ]
 
+[[package]]
+name = "equivalent"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5443807d6dff69373d433ab9ef5378ad8df50ca6298caf15de6e52e24aaf54d5"
+
 [[package]]
 name = "errno"
 version = "0.3.5"
@@ -225,12 +232,28 @@ version = "1.0.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3f9eec918d3f24069decb9af1554cad7c880e2da24a9afd88aca000531ab82c1"
 
+[[package]]
+name = "hashbrown"
+version = "0.14.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f93e7192158dbcda357bdec5fb5788eebf8bbac027f3f33e719d29135ae84156"
+
 [[package]]
 name = "heck"
 version = "0.4.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "95505c38b4572b2d910cecb0281560f54b440a19336cbbcb27bf6ce6adc6f5a8"
 
+[[package]]
+name = "indexmap"
+version = "2.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8adf3ddd720272c6ea8bf59463c04e0f93d0bbf7c5439b691bca2987e0270897"
+dependencies = [
+ "equivalent",
+ "hashbrown",
+]
+
 [[package]]
 name = "lazy_static"
 version = "1.4.0"

elven-parser/src/read.rs

@@ -5,7 +5,7 @@
 use crate::{
     consts::{self as c, DynamicTag, ShType},
     idx::{define_idx, ElfIndexExt, ToIdxUsize},
-    Addr, Offset,
+    Addr, Offset, write::Section,
 };
 use bstr::BStr;
 

elven-wald/Cargo.toml

@@ -10,6 +10,7 @@ anyhow = "1.0.69"
 bstr = "1.3.0"
 clap = { version = "4.1.4", features = ["derive"] }
 elven-parser = { path = "../elven-parser" }
+indexmap = "2.0.2"
 memmap2 = "0.5.8"
 tracing = "0.1.37"
 tracing-subscriber = { version = "0.3.16", features = ["env-filter"] }

elven-wald/src/lib.rs

@@ -1,3 +1,6 @@
+mod storage;
+mod utils;
+
 #[macro_use]
 extern crate tracing;
 
@@ -26,6 +29,14 @@ pub struct Opts {
     pub objs: Vec<PathBuf>,
 }
 
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+struct FileId(usize);
+
+struct ElfFile<'a> {
+    id: FileId,
+    elf: ElfReader<'a>,
+}
+
 #[derive(Debug)]
 struct SymDef<'a> {
     _name: &'a BStr,
@@ -35,7 +46,7 @@ struct SymDef<'a> {
 }
 
 struct LinkCtxt<'a> {
-    elves: Vec<ElfReader<'a>>,
+    elves: Vec<ElfFile<'a>>,
     sym_defs: HashMap<&'a BStr, SymDef<'a>>,
 }
 
@@ -56,17 +67,18 @@ pub fn run(opts: Opts) -> Result<()> {
         bail!("you gotta supply at least one object file");
     }
 
-    if opts.objs.len() > 1 {
-        bail!("hey hey hey one stop back please. you want to link MULTIPLE files TOGETHER? im sorry i cant do that");
-    }
-
     info!(objs=?opts.objs, "Linking files");
 
     let elves = mmaps
         .iter()
         .zip(&opts.objs)
-        .map(|(mmap, path)| {
+        .enumerate()
-            ElfReader::new(mmap).with_context(|| format!("parsing ELF file {}", path.display()))
+        .map(|(idx, (mmap, path))| {
+            Ok(ElfFile {
+                id: FileId(idx),
+                elf: ElfReader::new(mmap)
+                    .with_context(|| format!("parsing ELF file {}", path.display()))?,
+            })
         })
         .collect::<Result<Vec<_>, anyhow::Error>>()?;
 
@@ -75,14 +87,19 @@ pub fn run(opts: Opts) -> Result<()> {
         sym_defs: HashMap::new(),
     };
 
+    let storage =
+        storage::allocate_storage(BASE_EXEC_ADDR, &cx.elves).context("while allocating storage")?;
+
+    dbg!(storage);
+
     cx.resolve()?;
 
     dbg!(cx.sym_defs);
 
-    let text_sh = cx.elves[0].section_header_by_name(b".text")?;
+    let text_sh = cx.elves[0].elf.section_header_by_name(b".text")?;
-    let text_content = cx.elves[0].section_content(text_sh)?;
+    let text_content = cx.elves[0].elf.section_content(text_sh)?;
 
-    let _start_sym = cx.elves[0].symbol_by_name(b"_start")?;
+    let _start_sym = cx.elves[0].elf.symbol_by_name(b"_start")?;
 
     write_output(&opts, text_content, _start_sym.value)?;
 
@@ -90,12 +107,12 @@ pub fn run(opts: Opts) -> Result<()> {
 }
 
 pub const BASE_EXEC_ADDR: Addr = Addr(0x400000); // whatever ld does
-pub const DEFAULT_PROGRAM_HEADER_ALIGN_THAT_LD_USES_HERE: u64 = 0x1000;
+pub const DEFAULT_PAGE_ALIGN: u64 = 0x1000;
 
 impl<'a> LinkCtxt<'a> {
     fn resolve(&mut self) -> Result<()> {
         for (elf_idx, elf) in self.elves.iter().enumerate() {
-            for e_sym in elf.symbols()? {
+            for e_sym in elf.elf.symbols()? {
                 let ty = e_sym.info.r#type();
 
                 // Undefined symbols are not a definition.
@@ -104,8 +121,10 @@ impl<'a> LinkCtxt<'a> {
                 }
 
                 let name = match ty.0 {
-                    c::STT_SECTION => elf.sh_string(elf.section_header(e_sym.shndx)?.name)?,
+                    c::STT_SECTION => elf
-                    _ => elf.string(e_sym.name)?,
+                        .elf
+                        .sh_string(elf.elf.section_header(e_sym.shndx)?.name)?,
+                    _ => elf.elf.string(e_sym.name)?,
                 };
 
                 match self.sym_defs.entry(name) {
@@ -154,7 +173,7 @@ fn write_output(opts: &Opts, text: &[u8], entry_offset_from_text: Addr) -> Resul
         fixed_entsize: None,
         content: text.to_vec(),
         // align nicely
-        addr_align: Some(NonZeroU64::new(DEFAULT_PROGRAM_HEADER_ALIGN_THAT_LD_USES_HERE).unwrap()),
+        addr_align: Some(NonZeroU64::new(DEFAULT_PAGE_ALIGN).unwrap()),
     })?;
 
     let elf_header_and_program_headers = ProgramHeader {
@@ -168,13 +187,12 @@ fn write_output(opts: &Opts, text: &[u8], entry_offset_from_text: Addr) -> Resul
         paddr: BASE_EXEC_ADDR,
         filesz: 176, // FIXME: Do not hardocde this lol
         memsz: 176,
-        align: DEFAULT_PROGRAM_HEADER_ALIGN_THAT_LD_USES_HERE,
+        align: DEFAULT_PAGE_ALIGN,
     };
 
     write.add_program_header(elf_header_and_program_headers);
 
-    let entry_addr =
+    let entry_addr = BASE_EXEC_ADDR + DEFAULT_PAGE_ALIGN + entry_offset_from_text;
-        BASE_EXEC_ADDR + DEFAULT_PROGRAM_HEADER_ALIGN_THAT_LD_USES_HERE + entry_offset_from_text;
 
     let text_program_header = ProgramHeader {
         r#type: PT_LOAD.into(),
@@ -187,7 +205,7 @@ fn write_output(opts: &Opts, text: &[u8], entry_offset_from_text: Addr) -> Resul
         paddr: entry_addr,
         filesz: text.len() as u64,
         memsz: text.len() as u64,
-        align: DEFAULT_PROGRAM_HEADER_ALIGN_THAT_LD_USES_HERE,
+        align: DEFAULT_PAGE_ALIGN,
     };
 
     write.add_program_header(text_program_header);

elven-wald/src/main.rs

@@ -8,7 +8,7 @@ fn main() -> anyhow::Result<()> {
     tracing_subscriber::fmt()
         .with_env_filter(
             EnvFilter::builder()
-                .with_default_directive(LevelFilter::INFO.into())
+                .with_default_directive(LevelFilter::DEBUG.into())
                 .from_env_lossy(),
         )
         .init();

elven-wald/src/storage.rs

@@ -0,0 +1,76 @@
+use std::ops::Add;
+
+use anyhow::Result;
+use bstr::{BStr, BString};
+use elven_parser::{read::ElfReadError, Addr};
+use indexmap::IndexMap;
+
+use crate::{utils::AlignExt, ElfFile, FileId, DEFAULT_PAGE_ALIGN};
+
+#[derive(Debug)]
+pub struct Allocation {
+    file: FileId,
+    section: BString,
+    size: u64,
+    align: u64,
+}
+
+#[derive(Debug)]
+pub struct SegmentPart {
+    base: Addr,
+    file: FileId,
+    section: BString,
+    size: u64,
+}
+
+#[derive(Debug)]
+pub struct StorageAllocation {
+    segment_parts: Vec<SegmentPart>,
+}
+
+pub fn allocate_storage<'a>(base_addr: Addr, files: &[ElfFile<'a>]) -> Result<StorageAllocation> {
+    let mut allocs = IndexMap::<_, Vec<Allocation>>::new();
+
+    for file in files {
+        let elf = file.elf;
+
+        for name in [b".text".as_slice(), b".data", b".bss"] {
+            let section = elf.section_header_by_name(name);
+            match section {
+                Ok(section) => {
+                    allocs.entry(BStr::new(name)).or_default().push(Allocation {
+                        file: file.id,
+                        section: name.into(),
+                        size: section.size,
+                        align: section.addralign,
+                    });
+                }
+                Err(ElfReadError::NotFoundByName(_, _)) => {}
+                Err(e) => return Err(e.into()),
+            }
+        }
+    }
+
+    debug!(?allocs, "Allocation pass one completed");
+
+    let mut current_addr = base_addr;
+    let mut segment_parts = Vec::new();
+    for section in allocs {
+        current_addr = current_addr.align_up(DEFAULT_PAGE_ALIGN);
+        for alloc in section.1 {
+            let align = alloc.align;
+            let addr = current_addr.align_up(align);
+
+            current_addr = addr + alloc.size;
+
+            segment_parts.push(SegmentPart {
+                base: addr,
+                file: alloc.file,
+                size: alloc.size,
+                section: section.0.to_owned(),
+            });
+        }
+    }
+
+    Ok(StorageAllocation { segment_parts })
+}

elven-wald/src/utils.rs

@@ -0,0 +1,32 @@
+use elven_parser::Addr;
+
+pub trait AlignExt<T>: Copy {
+    fn align_down(self, align: T) -> Self;
+    fn align_up(self, align: T) -> Self;
+}
+
+impl AlignExt<u64> for u64 {
+    fn align_down(self, align: Self) -> Self {
+        assert!(align.is_power_of_two() && align > 0);
+        // We want to set all the aligment bits to zero.
+        // 0b0101 aligned to 0b0100 => 0b0100
+        // mask is !0b0011 = 0b1100
+        let mask = !(align - 1);
+        self & mask
+    }
+    fn align_up(self, align: Self) -> Self {
+        assert!(align.is_power_of_two() && align > 0);
+        // 0b0101 aligned to 0b0100 => 0b1000
+        (self + align - 1) & !(align - 1)
+    }
+}
+
+impl AlignExt<u64> for Addr {
+    fn align_down(self, align: u64) -> Self {
+        Addr(self.u64().align_down(align))
+    }
+
+    fn align_up(self, align: u64) -> Self {
+        Addr(self.u64().align_up(align))
+    }
+}