ww/selfhost/cmd/w6l/main.ww

// selfhost/cmd/w6l/main.ww — port of cmd/w6l/main.c.
//
// w6l = amd64 linker. Reads relocatable ELF .o files, SysV `ar`
// archives, and shared objects (ET_DYN). Resolves symbols, applies
// relocations, writes a static or dynamic-linked ELF executable.
//
//   w6l_ww -o out [-L<dir>...] [-l<name>...] file1.o file2.o ...

package main;

import os;
import mem;
import sym;
import obj;
import dyn;
import pass;
import out;

def BASE: u64 = 4194304u64;	// 0x400000
def CODE_VA_OFF: u64 = 4096u64;	// .text starts at base + 0x1000

fn mklnk(a: *arena) *lnk = {
	let l: *lnk = amalloc(a, 144u64): *lnk;
	l.a = a;
	return l;
};

// `cstreq` lives in obj.ww — same bundle, single definition.

// `cstrlen` lives in obj.ww — same bundle, single definition.

// Build "<dir>/lib<name>.<ext>" into dst (NUL-terminated). Returns total
// length excluding NUL. dst must be large enough.
fn buildpath(dst: *u8, dir: *u8, name: *u8, ext: str) u64 = {
	let i: u64 = 0u64;
	let dn: u64 = cstrlen(dir);
	let nn: u64 = cstrlen(name);
	let k: u64 = 0u64;
	for (k < dn) { dst[i] = dir[k]; i += 1u64; k += 1u64; };
	dst[i] = 47u8;	// '/'
	i += 1u64;
	dst[i] = 108u8;	// 'l'
	i += 1u64;
	dst[i] = 105u8;	// 'i'
	i += 1u64;
	dst[i] = 98u8;	// 'b'
	i += 1u64;
	k = 0u64;
	for (k < nn) { dst[i] = name[k]; i += 1u64; k += 1u64; };
	k = 0u64;
	for (k < ext.len: u64) {
		let li: i32 = k: i32;
		dst[i] = ext[li];
		i += 1u64;
		k += 1u64;
	};
	dst[i] = 0u8;
	return i;
};

// Append decimal n to dst at offset i. Returns new offset.
fn appenddec(dst: *u8, i: u64, n: u64) u64 = {
	if (n == 0u64) {
		dst[i] = 48u8;
		return i + 1u64;
	};
	let buf: *u8 = os.alloc(16u64): *u8;
	let k: u64 = 0u64;
	let v: u64 = n;
	for (v > 0u64) {
		buf[k] = (v % 10u64): u8 + 48u8;
		v = v / 10u64;
		k += 1u64;
	};
	let oi: u64 = i;
	for (k > 0u64) {
		k -= 1u64;
		dst[oi] = buf[k];
		oi += 1u64;
	};
	return oi;
};

fn buildpathv(dst: *u8, dir: *u8, name: *u8, v: u64) u64 = {
	let i: u64 = 0u64;
	let dn: u64 = cstrlen(dir);
	let nn: u64 = cstrlen(name);
	let k: u64 = 0u64;
	for (k < dn) { dst[i] = dir[k]; i += 1u64; k += 1u64; };
	dst[i] = 47u8; i += 1u64;
	dst[i] = 108u8; i += 1u64;
	dst[i] = 105u8; i += 1u64;
	dst[i] = 98u8; i += 1u64;
	k = 0u64;
	for (k < nn) { dst[i] = name[k]; i += 1u64; k += 1u64; };
	dst[i] = 46u8; i += 1u64; dst[i] = 115u8; i += 1u64; dst[i] = 111u8; i += 1u64; dst[i] = 46u8; i += 1u64;
	i = appenddec(dst, i, v);
	dst[i] = 0u8;
	return i;
};

// islinkable: read first 8 bytes; require !<arch>\n or \x7fELF.
fn islinkable(path: *u8) bool = {
	let fd: i32 = os.open(pathstr(path), os.flag.RDONLY, 0i32);
	if (fd < 0) { return false; };
	let mp: *u8 = os.alloc(8u64): *u8;
	let n: i64 = os.read(fd, mp, 8u64);
	os.close(fd);
	if (n < 4i64) { return false; };
	// archive: "!<arch>\n"
	if (n >= 8i64) {
		if (mp[0u64] == 33u8) { if (mp[1u64] == 60u8) {
		if (mp[2u64] == 97u8) { if (mp[3u64] == 114u8) {
		if (mp[4u64] == 99u8) { if (mp[5u64] == 104u8) {
		if (mp[6u64] == 62u8) { if (mp[7u64] == 10u8) {
			return true;
		}; }; }; }; }; }; }; };
	};
	// ELF: "\x7fELF"
	if (mp[0u64] == 127u8) {
		if (mp[1u64] == 69u8) {
			if (mp[2u64] == 76u8) {
				if (mp[3u64] == 70u8) { return true; };
			};
		};
	};
	return false;
};

// Walk libdirs[0..n) trying lib<name>.so, then lib<name>.so.{0..8},
// then lib<name>.a. Return arena-owned NUL-terminated path on success,
// nil on miss.
fn resolvelib(a: *arena, name: *u8, libdirs: **u8, nlibdirs: i32) *u8 = {
	let bufp: *u8 = os.alloc(1024u64): *u8;
	let i: i32 = 0;
	for (i < nlibdirs) {
		let dir: *u8 = libdirs[i];
		let _l1: u64 = buildpath(bufp, dir, name, ".so");
		if (islinkable(bufp)) {
			let pl: u64 = cstrlen(bufp);
			let p: *u8 = amalloc(a, pl + 1u64): *u8;
			let k: u64 = 0u64;
			for (k <= pl) { p[k] = bufp[k]; k += 1u64; };
			return p;
		};
		let v: u64 = 0u64;
		for (v <= 8u64) {
			let _l2: u64 = buildpathv(bufp, dir, name, v);
			if (islinkable(bufp)) {
				let pl2: u64 = cstrlen(bufp);
				let p2: *u8 = amalloc(a, pl2 + 1u64): *u8;
				let k2: u64 = 0u64;
				for (k2 <= pl2) { p2[k2] = bufp[k2]; k2 += 1u64; };
				return p2;
			};
			v += 1u64;
		};
		let _l3: u64 = buildpath(bufp, dir, name, ".a");
		if (islinkable(bufp)) {
			let pl3: u64 = cstrlen(bufp);
			let p3: *u8 = amalloc(a, pl3 + 1u64): *u8;
			let k3: u64 = 0u64;
			for (k3 <= pl3) { p3[k3] = bufp[k3]; k3 += 1u64; };
			return p3;
		};
		i += 1;
	};
	return nil;
};

// Read first 20 bytes; return 1 for ET_DYN .so, 0 for ar/.o.
fn isso(path: *u8) i32 = {
	let fd: i32 = os.open(pathstr(path), os.flag.RDONLY, 0i32);
	if (fd < 0) { return 0; };
	let mp: *u8 = os.alloc(20u64): *u8;
	let n: i64 = os.read(fd, mp, 20u64);
	os.close(fd);
	if (n < 20i64) { return 0; };
	if (mp[0u64] != 127u8) { return 0; };
	if (mp[1u64] != 69u8) { return 0; };
	if (mp[2u64] != 76u8) { return 0; };
	if (mp[3u64] != 70u8) { return 0; };
	// e_type at offset 16, u16 little-endian
	let t: u16 = (mp[16u64]: u16) | ((mp[17u64]: u16) << 8u16);
	if (t == 3u16) { return 1; };
	return 0;
};

export fn main(argc: i32, argv: **u8) i32 = {
	let outpath: *u8 = nil;
	let maxinputs: i32 = 64;
	let inputs:  **u8 = os.alloc((maxinputs: u64) * 8u64): **u8;
	let ninputs: i32 = 0;
	let libdirs: **u8 = os.alloc((maxinputs: u64) * 8u64): **u8;
	let nlibdirs: i32 = 0;
	let lflags:  **u8 = os.alloc((maxinputs: u64) * 8u64): **u8;
	let nlflags: i32 = 0;

	let i: i32 = 1;
	for (i < argc) {
		let a: *u8 = argv[i];
		if (cstreq(a, "-o")) {
			i += 1;
			if (i >= argc) {
				os.write(2, "w6l: -o requires argument\n".ptr, 25u64);
				return 2;
			};
			outpath = argv[i];
		} else { if (cstreq(a, "-L")) {
			i += 1;
			if (i >= argc) {
				os.write(2, "w6l: -L requires argument\n".ptr, 25u64);
				return 2;
			};
			libdirs[nlibdirs] = argv[i];
			nlibdirs += 1;
		} else { if (cstreq(a, "-l")) {
			i += 1;
			if (i >= argc) {
				os.write(2, "w6l: -l requires argument\n".ptr, 25u64);
				return 2;
			};
			lflags[nlflags] = argv[i];
			nlflags += 1;
		} else { if (a[0u64] == 45u8) {
			// -L<dir> joined form.
			if (a[1u64] == 76u8) {
				if (a[2u64] != 0u8) {
					libdirs[nlibdirs] = a + 2u64;
					nlibdirs += 1;
				} else {
					os.write(2, "w6l: bare -L\n".ptr, 12u64);
					return 2;
				};
			} else { if (a[1u64] == 108u8) {
				if (a[2u64] != 0u8) {
					lflags[nlflags] = a + 2u64;
					nlflags += 1;
				} else {
					os.write(2, "w6l: bare -l\n".ptr, 12u64);
					return 2;
				};
			} else {
				os.write(2, "w6l: unknown flag\n".ptr, 17u64);
				return 2;
			};};
		} else {
			if (ninputs >= maxinputs) {
				os.write(2, "w6l: too many inputs\n".ptr, 20u64);
				return 2;
			};
			inputs[ninputs] = a;
			ninputs += 1;
		};};};};
		i += 1;
	};

	if (outpath == nil) {
		os.write(2, "usage: w6l_ww -o exe [-L<dir>...] [-l<name>...] file1.o [file2.o...]\n".ptr, 68u64);
		return 2;
	};
	if (ninputs == 0) {
		os.write(2, "w6l: no inputs\n".ptr, 14u64);
		return 2;
	};

	let a: *arena = newarena();
	let l: *lnk = mklnk(a);

	// Seed _start so libwwrt-style start.o is recognised as wanted.
	intern(l, "_start");

	// Load positional inputs first (preserving order).
	let k: i32 = 0;
	for (k < ninputs) {
		if (load(l, inputs[k]) != 0) {
			return 1;
		};
		k += 1;
	};

	// Then resolve -l flags and load each. Archives append; shared
	// objects register their exports.
	let lf: i32 = 0;
	for (lf < nlflags) {
		let p: *u8 = resolvelib(a, lflags[lf], libdirs, nlibdirs);
		if (p == nil) {
			os.write(2, "w6l: cannot find -l".ptr, 18u64);
			let nm: *u8 = lflags[lf];
			os.write(2, nm, cstrlen(nm));
			os.write(2, "\n".ptr, 1u64);
			return 1;
		};
		if (isso(p) != 0) {
			if (loadso(l, p) != 0) { return 1; };
		} else {
			if (load(l, p) != 0) { return 1; };
		};
		lf += 1;
	};

	if (resolve(l) != 0) { return 1; };
	// Relocation is deferred to the emit functions — each path
	// knows its own layout (textva, datava); the static and dyn
	// paths place .data at different VAs.

	let entrysym: *lsym = lookup(l, "_start");
	if (entrysym == nil) { entrysym = lookup(l, "main"); }
	else { if (entrysym.defined == 0) { entrysym = lookup(l, "main"); }; };
	if (entrysym == nil) {
		os.write(2, "w6l: no _start or main symbol\n".ptr, 29u64);
		return 1;
	};
	if (entrysym.defined == 0) {
		os.write(2, "w6l: no _start or main symbol\n".ptr, 29u64);
		return 1;
	};

	let flags: os.flag = os.flag.WRONLY | os.flag.CREATE | os.flag.TRUNC;
	let fd: i32 = os.open(pathstr(outpath), flags, 493i32);	// 0o755
	if (fd < 0) {
		os.write(2, "w6l: cannot open output\n".ptr, 23u64);
		return 1;
	};

	let entryva: u64 = BASE + CODE_VA_OFF + entrysym.val;
	let rc: i32 = emitelf(l, fd, BASE, entryva);
	os.close(fd);
	return rc;
};