Rollup merge of rust-lang#144221 - usamoi:versym, r=bjorn3

generate elf symbol version in raw-dylib

For link names like `aaa@bbb`, it generates a symbol named `aaa` and a version named `bbb`.

For link names like `aaa\0bbb`, `aaa@`@bbb`` or `aa@bb@cc`, it emits errors.

It adds a test that the executable is linked with glibc using raw-dylib.

cc rust-lang#135694

Author: matthiaskrgr

compiler/rustc_codegen_ssa/src/back/link/raw_dylib.rs

@@ -4,7 +4,7 @@ use std::path::{Path, PathBuf};
 
 use rustc_abi::Endian;
 use rustc_data_structures::base_n::{CASE_INSENSITIVE, ToBaseN};
-use rustc_data_structures::fx::FxIndexMap;
+use rustc_data_structures::fx::{FxHashMap, FxIndexMap};
 use rustc_data_structures::stable_hasher::StableHasher;
 use rustc_hashes::Hash128;
 use rustc_session::Session;
@@ -230,40 +230,66 @@ fn create_elf_raw_dylib_stub(sess: &Session, soname: &str, symbols: &[DllImport]
         Endian::Little => object::Endianness::Little,
         Endian::Big => object::Endianness::Big,
     };
+
     let mut stub = write::Writer::new(endianness, true, &mut stub_buf);
 
+    let mut vers = Vec::new();
+    let mut vers_map = FxHashMap::default();
+    let mut syms = Vec::new();
+
+    for symbol in symbols {
+        let symbol_name = symbol.name.as_str();
+        if let Some((name, version_name)) = symbol_name.split_once('@') {
+            assert!(!version_name.contains('@'));
+            let dynstr = stub.add_dynamic_string(name.as_bytes());
+            let ver = if let Some(&ver_id) = vers_map.get(version_name) {
+                ver_id
+            } else {
+                let id = vers.len();
+                vers_map.insert(version_name, id);
+                let dynstr = stub.add_dynamic_string(version_name.as_bytes());
+                vers.push((version_name, dynstr));
+                id
+            };
+            syms.push((name, dynstr, Some(ver)));
+        } else {
+            let dynstr = stub.add_dynamic_string(symbol_name.as_bytes());
+            syms.push((symbol_name, dynstr, None));
+        }
+    }
+
+    let soname = stub.add_dynamic_string(soname.as_bytes());
+
     // These initial reservations don't reserve any bytes in the binary yet,
     // they just allocate in the internal data structures.
 
-    // First, we crate the dynamic symbol table. It starts with a null symbol
+    // First, we create the dynamic symbol table. It starts with a null symbol
     // and then all the symbols and their dynamic strings.
     stub.reserve_null_dynamic_symbol_index();
 
-    let dynstrs = symbols
-        .iter()
-        .map(|sym| {
-            stub.reserve_dynamic_symbol_index();
-            (sym, stub.add_dynamic_string(sym.name.as_str().as_bytes()))
-        })
-        .collect::<Vec<_>>();
-
-    let soname = stub.add_dynamic_string(soname.as_bytes());
+    for _ in syms.iter() {
+        stub.reserve_dynamic_symbol_index();
+    }
 
     // Reserve the sections.
     // We have the minimal sections for a dynamic SO and .text where we point our dummy symbols to.
     stub.reserve_shstrtab_section_index();
     let text_section_name = stub.add_section_name(".text".as_bytes());
     let text_section = stub.reserve_section_index();
-    stub.reserve_dynstr_section_index();
     stub.reserve_dynsym_section_index();
+    stub.reserve_dynstr_section_index();
+    stub.reserve_gnu_versym_section_index();
+    stub.reserve_gnu_verdef_section_index();
     stub.reserve_dynamic_section_index();
 
     // These reservations now determine the actual layout order of the object file.
     stub.reserve_file_header();
     stub.reserve_shstrtab();
     stub.reserve_section_headers();
-    stub.reserve_dynstr();
     stub.reserve_dynsym();
+    stub.reserve_dynstr();
+    stub.reserve_gnu_versym();
+    stub.reserve_gnu_verdef(1 + vers.len(), 1 + vers.len());
     stub.reserve_dynamic(2); // DT_SONAME, DT_NULL
 
     // First write the ELF header with the arch information.
@@ -342,18 +368,17 @@ fn create_elf_raw_dylib_stub(sess: &Session, soname: &str, symbols: &[DllImport]
         sh_addralign: 1,
         sh_entsize: 0,
     });
-    stub.write_dynstr_section_header(0);
     stub.write_dynsym_section_header(0, 1);
+    stub.write_dynstr_section_header(0);
+    stub.write_gnu_versym_section_header(0);
+    stub.write_gnu_verdef_section_header(0);
     stub.write_dynamic_section_header(0);
 
-    // .dynstr
-    stub.write_dynstr();
-
     // .dynsym
     stub.write_null_dynamic_symbol();
-    for (_, name) in dynstrs {
+    for (_name, dynstr, _ver) in syms.iter().copied() {
         stub.write_dynamic_symbol(&write::Sym {
-            name: Some(name),
+            name: Some(dynstr),
             st_info: (elf::STB_GLOBAL << 4) | elf::STT_NOTYPE,
             st_other: elf::STV_DEFAULT,
             section: Some(text_section),
@@ -363,10 +388,44 @@ fn create_elf_raw_dylib_stub(sess: &Session, soname: &str, symbols: &[DllImport]
         });
     }
 
+    // .dynstr
+    stub.write_dynstr();
+
+    // .gnu_version
+    stub.write_null_gnu_versym();
+    for (_name, _dynstr, ver) in syms.iter().copied() {
+        stub.write_gnu_versym(if let Some(ver) = ver {
+            assert!((2 + ver as u16) < elf::VERSYM_HIDDEN);
+            elf::VERSYM_HIDDEN | (2 + ver as u16)
+        } else {
+            1
+        });
+    }
+
+    // .gnu_version_d
+    stub.write_align_gnu_verdef();
+    stub.write_gnu_verdef(&write::Verdef {
+        version: elf::VER_DEF_CURRENT,
+        flags: elf::VER_FLG_BASE,
+        index: 1,
+        aux_count: 1,
+        name: soname,
+    });
+    for (ver, (_name, dynstr)) in vers.into_iter().enumerate() {
+        stub.write_gnu_verdef(&write::Verdef {
+            version: elf::VER_DEF_CURRENT,
+            flags: 0,
+            index: 2 + ver as u16,
+            aux_count: 1,
+            name: dynstr,
+        });
+    }
+
     // .dynamic
     // the DT_SONAME will be used by the linker to populate DT_NEEDED
     // which the loader uses to find the library.
     // DT_NULL terminates the .dynamic table.
+    stub.write_align_dynamic();
     stub.write_dynamic_string(elf::DT_SONAME, soname);
     stub.write_dynamic(elf::DT_NULL, 0);
 

compiler/rustc_metadata/messages.ftl

@@ -330,3 +330,6 @@ metadata_wasm_import_form =
 
 metadata_whole_archive_needs_static =
     linking modifier `whole-archive` is only compatible with `static` linking kind
+
+metadata_raw_dylib_malformed =
+    link name must be well-formed if link kind is `raw-dylib`

compiler/rustc_metadata/src/errors.rs

@@ -815,3 +815,10 @@ pub struct AsyncDropTypesInDependency {
     pub extern_crate: Symbol,
     pub local_crate: Symbol,
 }
+
+#[derive(Diagnostic)]
+#[diag(metadata_raw_dylib_malformed)]
+pub struct RawDylibMalformed {
+    #[primary_span]
+    pub span: Span,
+}

compiler/rustc_metadata/src/native_libs.rs

@@ -700,8 +700,21 @@ impl<'tcx> Collector<'tcx> {
             .link_ordinal
             .map_or(import_name_type, |ord| Some(PeImportNameType::Ordinal(ord)));
 
+        let name = codegen_fn_attrs.link_name.unwrap_or_else(|| self.tcx.item_name(item));
+
+        if self.tcx.sess.target.binary_format == BinaryFormat::Elf {
+            let name = name.as_str();
+            if name.contains('\0') {
+                self.tcx.dcx().emit_err(errors::RawDylibMalformed { span });
+            } else if let Some((left, right)) = name.split_once('@')
+                && (left.is_empty() || right.is_empty() || right.contains('@'))
+            {
+                self.tcx.dcx().emit_err(errors::RawDylibMalformed { span });
+            }
+        }
+
         DllImport {
-            name: codegen_fn_attrs.link_name.unwrap_or_else(|| self.tcx.item_name(item)),
+            name,
             import_name_type,
             calling_convention,
             span,

tests/ui/linkage-attr/raw-dylib/elf/glibc-x86_64.rs

@@ -0,0 +1,74 @@
+//@ only-x86_64-unknown-linux-gnu
+//@ needs-dynamic-linking
+//@ run-pass
+//@ compile-flags: -Cpanic=abort
+//@ edition: 2024
+
+#![allow(incomplete_features)]
+#![feature(raw_dylib_elf)]
+#![no_std]
+#![no_main]
+
+use core::ffi::{c_char, c_int};
+
+extern "C" fn callback(
+    _fpath: *const c_char,
+    _sb: *const (),
+    _tflag: c_int,
+    _ftwbuf: *const (),
+) -> c_int {
+    0
+}
+
+#[link(name = "libc.so.6", kind = "raw-dylib", modifiers = "+verbatim")]
+unsafe extern "C" {
+    #[link_name = "nftw@GLIBC_2.2.5"]
+    unsafe fn nftw_2_2_5(
+        dirpath: *const c_char,
+        f: extern "C" fn(*const c_char, *const (), c_int, *const ()) -> c_int,
+        nopenfd: c_int,
+        flags: c_int,
+    ) -> c_int;
+    #[link_name = "nftw@GLIBC_2.3.3"]
+    unsafe fn nftw_2_3_3(
+        dirpath: *const c_char,
+        f: extern "C" fn(*const c_char, *const (), c_int, *const ()) -> c_int,
+        nopenfd: c_int,
+        flags: c_int,
+    ) -> c_int;
+    #[link_name = "exit@GLIBC_2.2.5"]
+    safe fn exit(status: i32) -> !;
+    unsafe fn __libc_start_main() -> c_int;
+}
+
+#[unsafe(no_mangle)]
+extern "C" fn main() -> ! {
+    unsafe {
+        // The old `nftw` does not check whether unknown flags are set.
+        let res = nftw_2_2_5(c".".as_ptr(), callback, 20, 1 << 30);
+        assert_eq!(res, 0);
+    }
+    unsafe {
+        // The new `nftw` does.
+        let res = nftw_2_3_3(c".".as_ptr(), callback, 20, 1 << 30);
+        assert_eq!(res, -1);
+    }
+    exit(0);
+}
+
+#[cfg(not(test))]
+#[panic_handler]
+fn panic_handler(_: &core::panic::PanicInfo<'_>) -> ! {
+    exit(1);
+}
+
+#[unsafe(no_mangle)]
+extern "C" fn rust_eh_personality(
+    _version: i32,
+    _actions: i32,
+    _exception_class: u64,
+    _exception_object: *mut (),
+    _context: *mut (),
+) -> i32 {
+    exit(1);
+}

tests/ui/linkage-attr/raw-dylib/elf/malformed-link-name.rs

@@ -0,0 +1,20 @@
+//@ only-elf
+//@ needs-dynamic-linking
+//@ check-fail
+
+#![feature(raw_dylib_elf)]
+#![allow(incomplete_features)]
+
+#[link(name = "libc.so.6", kind = "raw-dylib", modifiers = "+verbatim")]
+unsafe extern "C" {
+    #[link_name = "exit@"]
+    pub safe fn exit_0(status: i32) -> !; //~ ERROR link name must be well-formed if link kind is `raw-dylib`
+    #[link_name = "@GLIBC_2.2.5"]
+    pub safe fn exit_1(status: i32) -> !; //~ ERROR link name must be well-formed if link kind is `raw-dylib`
+    #[link_name = "ex\0it@GLIBC_2.2.5"]
+    pub safe fn exit_2(status: i32) -> !; //~ ERROR link name must be well-formed if link kind is `raw-dylib`
+    #[link_name = "exit@@GLIBC_2.2.5"]
+    pub safe fn exit_3(status: i32) -> !; //~ ERROR link name must be well-formed if link kind is `raw-dylib`
+}
+
+fn main() {}

tests/ui/linkage-attr/raw-dylib/elf/malformed-link-name.stderr

@@ -0,0 +1,26 @@
+error: link name must be well-formed if link kind is `raw-dylib`
+  --> $DIR/malformed-link-name.rs:11:5
+   |
+LL |     pub safe fn exit_0(status: i32) -> !;
+   |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+error: link name must be well-formed if link kind is `raw-dylib`
+  --> $DIR/malformed-link-name.rs:13:5
+   |
+LL |     pub safe fn exit_1(status: i32) -> !;
+   |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+error: link name must be well-formed if link kind is `raw-dylib`
+  --> $DIR/malformed-link-name.rs:15:5
+   |
+LL |     pub safe fn exit_2(status: i32) -> !;
+   |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+error: link name must be well-formed if link kind is `raw-dylib`
+  --> $DIR/malformed-link-name.rs:17:5
+   |
+LL |     pub safe fn exit_3(status: i32) -> !;
+   |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+error: aborting due to 4 previous errors
+