6l: dynamic linking with symbol versioning
Teach the linker to consume ET_DYN shared objects and emit a dynamically-linked ELF executable. Snake et al. can now link against libncurses + libc through the system dynamic loader. Pipeline additions: - dyn.c: read ET_DYN, parse .dynsym + DT_SONAME, walk .gnu.version_d / .gnu.version to learn each export's default version (skip hidden entries). - pass.c: when an undefined sym is provided by some Lso, promote it to dynamic, assign a PLT slot, record the matched version on the Lsym. - dynout.c: emit PT_INTERP + PT_DYNAMIC, .dynsym/.dynstr/.hash, .plt + .got.plt + .rela.plt, .gnu.version + .gnu.version_r, and the full DT_* set with DT_BIND_NOW. Patch PC32/PLT32 references against dyn syms to point at their PLT stubs. - main.c: -L<dir> and -l<name> flag parsing; resolve <name> via .so / .so.<N> / .a in libdir order, skipping GNU ld linker scripts (libc.so on most distros). - ww driver: collect -l/-L (joined and split forms) and pass through to 6l. Design choices: - DT_BIND_NOW so the loader resolves all PLT slots at startup; no PLT0 lazy resolver stub. - SysV .hash, not .gnu.hash. One bucket; loader scans the chain. Slow at scale, fine for snake-class binaries. - Non-PIE at fixed 0x400000. - No section headers — loader uses program headers, but readelf -V/-S won't display anything. Symbol versioning is the only correctness item beyond the basic PLT/GOT machinery: glibc symbols default to versions later than GLIBC_2.2.5 (e.g. clock_gettime → GLIBC_2.17 for the vDSO impl), and the loader rejects unversioned references to those without a matching Vernaux entry. test/wwc/810_dyn covers four cases: bare libc dyn call, multi-PLT, clock_gettime versioning, and fn-pointer to FFI binding (which exercises the codegen fixes from the parent commit alongside the new linker path).
This commit is contained in:
152
test/wwc/810_dyn.c
Normal file
152
test/wwc/810_dyn.c
Normal file
@@ -0,0 +1,152 @@
|
||||
/*
|
||||
* 810_dyn — end-to-end test of 6l's dynamic linker.
|
||||
*
|
||||
* Drives `ww build -L /usr/lib -l c` over a small ww program that
|
||||
* binds libc symbols via @symbol, then runs the produced binary and
|
||||
* checks the exit status. Verifies:
|
||||
*
|
||||
* - the .so loader (dyn.c) reads ET_DYN and extracts exports
|
||||
* - PLT/GOT generation and the JUMP_SLOT relocation
|
||||
* - PT_INTERP/PT_DYNAMIC emission
|
||||
* - symbol versioning (.gnu.version + .gnu.version_r) — clock_gettime
|
||||
* has both GLIBC_2.2.5 (compat) and GLIBC_2.17 (default) on glibc;
|
||||
* the loader requires the right .gnu.version_r entry to bind
|
||||
* the default vDSO-aware impl.
|
||||
*
|
||||
* Skipped (passing trivially) on systems without /usr/lib/libc.so.6.
|
||||
*/
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include <unistd.h>
|
||||
#include <sys/stat.h>
|
||||
#include <sys/wait.h>
|
||||
|
||||
static int
|
||||
runwait(const char *cmd)
|
||||
{
|
||||
int rc = system(cmd);
|
||||
if (rc == -1) return -1;
|
||||
if (WIFEXITED(rc)) return WEXITSTATUS(rc);
|
||||
return 1;
|
||||
}
|
||||
|
||||
struct row {
|
||||
const char *src;
|
||||
int want_exit;
|
||||
};
|
||||
|
||||
static const struct row rows[] = {
|
||||
/* 1. _exit via dynamically-linked libc. Exit code is the plumbing
|
||||
* of choice — proves the PLT entry calls into libc.so.6. */
|
||||
{ "@symbol(\"_exit\") fn libc_exit(c: i32) void;\n"
|
||||
"export fn main() i32 = { libc_exit(42); return 0; };", 42 },
|
||||
|
||||
/* 2. Multiple dyn syms in one binary: write + _exit. */
|
||||
{ "@symbol(\"write\") fn libc_write(fd: i32, b: *u8, n: u64) i64;\n"
|
||||
"@symbol(\"_exit\") fn libc_exit(c: i32) void;\n"
|
||||
"export fn main() i32 = {\n"
|
||||
" libc_write(1, \"x\".ptr, 1u64);\n"
|
||||
" libc_exit(7);\n"
|
||||
" return 0;\n"
|
||||
"};", 7 },
|
||||
|
||||
/* 3. Symbol versioning: clock_gettime. Without DT_VERSYM/VERNEED
|
||||
* pointing at a Vernaux entry for GLIBC_2.17, the loader either
|
||||
* picks the wrong impl or fails outright on modern glibc. */
|
||||
{ "@symbol(\"clock_gettime\") fn clock_gettime(c: i32, ts: *u8) i32;\n"
|
||||
"@symbol(\"_exit\") fn libc_exit(c: i32) void;\n"
|
||||
"export fn main() i32 = {\n"
|
||||
" let ts: [16]u8;\n"
|
||||
" let r: i32 = clock_gettime(1, ts.ptr);\n" /* CLOCK_MONOTONIC */
|
||||
" libc_exit(r);\n"
|
||||
" return 0;\n"
|
||||
"};", 0 },
|
||||
|
||||
/* 4. Take the address of an FFI binding and call it indirectly.
|
||||
* Codegen must apply @symbol resolution at the LEAQ site (so the
|
||||
* stub address is `write`, not the ww-side ident `libc_write`),
|
||||
* and the local-variable call must be CALL *AX, not CALL fp(SB). */
|
||||
{ "@symbol(\"write\") fn libc_write(fd: i32, b: *u8, n: u64) i64;\n"
|
||||
"@symbol(\"_exit\") fn libc_exit(c: i32) void;\n"
|
||||
"export fn main() i32 = {\n"
|
||||
" let fp: fn(fd: i32, b: *u8, n: u64) i64 = libc_write;\n"
|
||||
" fp(1, \"X\".ptr, 1u64);\n"
|
||||
" libc_exit(5);\n"
|
||||
" return 0;\n"
|
||||
"};", 5 },
|
||||
|
||||
{ NULL, 0 }
|
||||
};
|
||||
|
||||
int
|
||||
main(void)
|
||||
{
|
||||
if (access("/usr/lib/libc.so.6", 0) != 0
|
||||
&& access("/lib/x86_64-linux-gnu/libc.so.6", 0) != 0
|
||||
&& access("/lib64/libc.so.6", 0) != 0) {
|
||||
puts("dyn: no libc.so.6 on this system — skipping");
|
||||
return 0;
|
||||
}
|
||||
|
||||
const char *bin = getenv("BIN");
|
||||
if (!bin) bin = "out/bin";
|
||||
char absbin[1024];
|
||||
if (bin[0] != '/') {
|
||||
char cwd[1024];
|
||||
if (getcwd(cwd, sizeof cwd) == NULL) return 1;
|
||||
snprintf(absbin, sizeof absbin, "%s/%s", cwd, bin);
|
||||
bin = absbin;
|
||||
}
|
||||
|
||||
/* Probe each common libc dir to find the right -L. */
|
||||
const char *libdir = NULL;
|
||||
if (access("/usr/lib/libc.so.6", 0) == 0) libdir = "/usr/lib";
|
||||
else if (access("/lib/x86_64-linux-gnu/libc.so.6", 0) == 0) libdir = "/lib/x86_64-linux-gnu";
|
||||
else if (access("/lib64/libc.so.6", 0) == 0) libdir = "/lib64";
|
||||
|
||||
int n = 0, fail = 0;
|
||||
for (int i = 0; rows[i].src; i++, n++) {
|
||||
char src[80], exe[80], tmpdir[80];
|
||||
snprintf(src, sizeof src, "/tmp/wwdyn_%d_%d.ww", getpid(), i);
|
||||
snprintf(exe, sizeof exe, "/tmp/wwdyn_%d_%d", getpid(), i);
|
||||
snprintf(tmpdir, sizeof tmpdir, "/tmp/wwdyn_%d_d_%d", getpid(), i);
|
||||
mkdir(tmpdir, 0755);
|
||||
|
||||
FILE *f = fopen(src, "wb");
|
||||
fputs(rows[i].src, f);
|
||||
fclose(f);
|
||||
|
||||
char cmd[1024];
|
||||
snprintf(cmd, sizeof cmd,
|
||||
"cd %s && %s/ww build %s -L %s -l c",
|
||||
tmpdir, bin, src, libdir);
|
||||
if (runwait(cmd) != 0) {
|
||||
fprintf(stderr, "dyn row %d: build failed\n", i);
|
||||
fail++;
|
||||
unlink(src); rmdir(tmpdir);
|
||||
continue;
|
||||
}
|
||||
|
||||
char outbin[160];
|
||||
const char *base = strrchr(src, '/');
|
||||
base = base ? base + 1 : src;
|
||||
snprintf(outbin, sizeof outbin, "%s/%s", tmpdir, base);
|
||||
char *dot = strrchr(outbin, '.');
|
||||
if (dot && strcmp(dot, ".ww") == 0) *dot = '\0';
|
||||
|
||||
int got = runwait(outbin);
|
||||
if (got != rows[i].want_exit) {
|
||||
fprintf(stderr, "dyn row %d: exit %d, want %d\n",
|
||||
i, got, rows[i].want_exit);
|
||||
fail++;
|
||||
}
|
||||
unlink(src); unlink(outbin); rmdir(tmpdir);
|
||||
}
|
||||
if (fail) {
|
||||
fprintf(stderr, "%d/%d dyn tests failed\n", fail, n);
|
||||
return 1;
|
||||
}
|
||||
printf("dyn: %d/%d ok\n", n, n);
|
||||
return 0;
|
||||
}
|
||||
Reference in New Issue
Block a user