Add shims for gettid-esque functions

Various platforms provide a function to return the current OS thread ID,
but they all use a slightly different name. Add shims for these
functions for Apple, FreeBSD, and Windows, with tests to account for
those and a few more platforms that are not yet supported by Miri. The
syscall and extern symbol is included as well on Linux.

These should be useful in general but will also help support printing
the OS thread ID in panic messages [1].

[1]: rust-lang#115746

Squashed commit from Ralf:

    try_from_scalar: extend comment

Author: tgross35

src/tools/miri/src/helpers.rs

@@ -1337,7 +1337,6 @@ where
 
 /// Check that the number of varargs is at least the minimum what we expect.
 /// Fixed args should not be included.
-/// Use `check_vararg_fixed_arg_count` to extract the varargs slice from full function arguments.
 pub fn check_min_vararg_count<'a, 'tcx, const N: usize>(
     name: &'a str,
     args: &'a [OpTy<'tcx>],

src/tools/miri/src/shims/env.rs

@@ -110,8 +110,30 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
         }
     }
 
+    /// Get the process identifier.
     fn get_pid(&self) -> u32 {
         let this = self.eval_context_ref();
         if this.machine.communicate() { std::process::id() } else { 1000 }
     }
+
+    /// Get an "OS" thread ID for the current thread.
+    fn get_current_tid(&self) -> u32 {
+        let this = self.eval_context_ref();
+        self.get_tid(this.machine.threads.active_thread())
+    }
+
+    /// Get an "OS" thread ID for any thread.
+    fn get_tid(&self, thread: ThreadId) -> u32 {
+        let this = self.eval_context_ref();
+        let index = thread.to_u32();
+        let target_os = &this.tcx.sess.target.os;
+        if target_os == "linux" || target_os == "netbsd" {
+            // On Linux, the main thread has PID == TID so we uphold this. NetBSD also appears
+            // to exhibit the same behavior, though I can't find a citation.
+            this.get_pid().strict_add(index)
+        } else {
+            // Other platforms do not display any relationship between PID and TID.
+            index
+        }
+    }
 }

src/tools/miri/src/shims/extern_static.rs

@@ -66,7 +66,7 @@ impl<'tcx> MiriMachine<'tcx> {
                     ecx,
                     &["__cxa_thread_atexit_impl", "__clock_gettime64"],
                 )?;
-                Self::weak_symbol_extern_statics(ecx, &["getrandom", "statx"])?;
+                Self::weak_symbol_extern_statics(ecx, &["getrandom", "gettid", "statx"])?;
             }
             "freebsd" => {
                 Self::null_ptr_extern_statics(ecx, &["__cxa_thread_atexit_impl"])?;

src/tools/miri/src/shims/unix/env.rs

@@ -274,15 +274,52 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
         interp_ok(Scalar::from_u32(this.get_pid()))
     }
 
-    fn linux_gettid(&mut self) -> InterpResult<'tcx, Scalar> {
+    /// The `gettid`-like function for Unix platforms that take no parameters and return a 32-bit
+    /// integer. It is not always named "gettid".
+    fn unix_gettid(&mut self, link_name: &str) -> InterpResult<'tcx, Scalar> {
         let this = self.eval_context_ref();
-        this.assert_target_os("linux", "gettid");
+        this.assert_target_os_is_unix(link_name);
 
-        let index = this.machine.threads.active_thread().to_u32();
+        // For most platforms the return type is an `i32`, but some are unsigned. The TID
+        // will always be positive so we don't need to differentiate.
+        interp_ok(Scalar::from_u32(this.get_current_tid()))
+    }
+
+    /// The Apple-specific `int pthread_threadid_np(pthread_t thread, uint64_t *thread_id)`, which
+    /// allows querying the ID for arbitrary threads, identified by their pthread_t.
+    ///
+    /// API documentation: <https://www.manpagez.com/man/3/pthread_threadid_np/>.
+    fn apple_pthread_threadip_np(
+        &mut self,
+        thread_op: &OpTy<'tcx>,
+        tid_op: &OpTy<'tcx>,
+    ) -> InterpResult<'tcx, Scalar> {
+        let this = self.eval_context_mut();
+        this.assert_target_os("macos", "pthread_threadip_np");
+
+        let tid_dest = this.read_pointer(tid_op)?;
+        if this.ptr_is_null(tid_dest)? {
+            // If NULL is passed, an error is immediately returned
+            return interp_ok(this.eval_libc("EINVAL"));
+        }
+
+        let thread = this.read_scalar(thread_op)?.to_int(this.libc_ty_layout("pthread_t").size)?;
+        let thread = if thread == 0 {
+            // Null thread ID indicates that we are querying the active thread.
+            this.machine.threads.active_thread()
+        } else {
+            // Our pthread_t is just the raw ThreadId.
+            let Ok(thread) = this.thread_id_try_from(thread) else {
+                return interp_ok(this.eval_libc("ESRCH"));
+            };
+            thread
+        };
 
-        // Compute a TID for this thread, ensuring that the main thread has PID == TID.
-        let tid = this.get_pid().strict_add(index);
+        let tid = this.get_tid(thread);
+        let tid_dest = this.deref_pointer_as(tid_op, this.machine.layouts.u64)?;
+        this.write_int(tid, &tid_dest)?;
 
-        interp_ok(Scalar::from_u32(tid))
+        // Possible errors have been handled, return success.
+        interp_ok(Scalar::from_u32(0))
     }
 }

src/tools/miri/src/shims/unix/freebsd/foreign_items.rs

@@ -56,6 +56,11 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
                 };
                 this.write_scalar(res, dest)?;
             }
+            "pthread_getthreadid_np" => {
+                let [] = this.check_shim(abi, CanonAbi::C, link_name, args)?;
+                let result = this.unix_gettid(link_name.as_str())?;
+                this.write_scalar(result, dest)?;
+            }
 
             "cpuset_getaffinity" => {
                 // The "same" kind of api as `sched_getaffinity` but more fine grained control for FreeBSD specifically.

src/tools/miri/src/shims/unix/linux/foreign_items.rs

@@ -18,7 +18,7 @@ use crate::*;
 const TASK_COMM_LEN: u64 = 16;
 
 pub fn is_dyn_sym(name: &str) -> bool {
-    matches!(name, "statx")
+    matches!(name, "gettid" | "statx")
 }
 
 impl<'tcx> EvalContextExt<'tcx> for crate::MiriInterpCx<'tcx> {}
@@ -117,7 +117,7 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
             }
             "gettid" => {
                 let [] = this.check_shim(abi, CanonAbi::C, link_name, args)?;
-                let result = this.linux_gettid()?;
+                let result = this.unix_gettid(link_name.as_str())?;
                 this.write_scalar(result, dest)?;
             }
 

src/tools/miri/src/shims/unix/linux_like/syscall.rs

@@ -4,6 +4,7 @@ use rustc_span::Symbol;
 use rustc_target::callconv::FnAbi;
 
 use crate::helpers::check_min_vararg_count;
+use crate::shims::unix::env::EvalContextExt;
 use crate::shims::unix::linux_like::eventfd::EvalContextExt as _;
 use crate::shims::unix::linux_like::sync::futex;
 use crate::*;
@@ -24,6 +25,7 @@ pub fn syscall<'tcx>(
     let sys_getrandom = ecx.eval_libc("SYS_getrandom").to_target_usize(ecx)?;
     let sys_futex = ecx.eval_libc("SYS_futex").to_target_usize(ecx)?;
     let sys_eventfd2 = ecx.eval_libc("SYS_eventfd2").to_target_usize(ecx)?;
+    let sys_gettid = ecx.eval_libc("SYS_gettid").to_target_usize(ecx)?;
 
     match ecx.read_target_usize(op)? {
         // `libc::syscall(NR_GETRANDOM, buf.as_mut_ptr(), buf.len(), GRND_NONBLOCK)`
@@ -53,6 +55,10 @@ pub fn syscall<'tcx>(
             let result = ecx.eventfd(initval, flags)?;
             ecx.write_int(result.to_i32()?, dest)?;
         }
+        num if num == sys_gettid => {
+            let result = ecx.unix_gettid("SYS_gettid")?;
+            ecx.write_int(result.to_u32()?, dest)?;
+        }
         num => {
             throw_unsup_format!("syscall: unsupported syscall number {num}");
         }

src/tools/miri/src/shims/unix/macos/foreign_items.rs

@@ -222,6 +222,11 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
                 };
                 this.write_scalar(res, dest)?;
             }
+            "pthread_threadid_np" => {
+                let [thread, tid_ptr] = this.check_shim(abi, CanonAbi::C, link_name, args)?;
+                let res = this.apple_pthread_threadip_np(thread, tid_ptr)?;
+                this.write_scalar(res, dest)?;
+            }
 
             // Synchronization primitives
             "os_sync_wait_on_address" => {

src/tools/miri/src/shims/windows/foreign_items.rs

@@ -629,6 +629,23 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
                 this.write_scalar(name, &name_ptr)?;
                 this.write_scalar(res, dest)?;
             }
+            "GetThreadId" => {
+                let [handle] = this.check_shim(abi, sys_conv, link_name, args)?;
+                let handle = this.read_handle(handle, "GetThreadId")?;
+                let thread = match handle {
+                    Handle::Thread(thread) => thread,
+                    Handle::Pseudo(PseudoHandle::CurrentThread) => this.active_thread(),
+                    _ => this.invalid_handle("GetThreadDescription")?,
+                };
+                let tid = this.get_tid(thread);
+                this.write_scalar(Scalar::from_u32(tid), dest)?;
+            }
+            "GetCurrentThreadId" => {
+                let [] = this.check_shim(abi, sys_conv, link_name, args)?;
+                let thread = this.active_thread();
+                let tid = this.get_tid(thread);
+                this.write_scalar(Scalar::from_u32(tid), dest)?;
+            }
 
             // Miscellaneous
             "ExitProcess" => {

src/tools/miri/src/shims/windows/handle.rs

@@ -166,6 +166,10 @@ impl Handle {
     /// Structurally invalid handles return [`HandleError::InvalidHandle`].
     /// If the handle is structurally valid but semantically invalid, e.g. a for non-existent thread
     /// ID, returns [`HandleError::ThreadNotFound`].
+    ///
+    /// This function is deliberately private; shims should always use `read_handle`.
+    /// That enforces handle validity even when Windows does not: for now, we argue invalid
+    /// handles are always a bug and programmers likely want to know about them.
     fn try_from_scalar<'tcx>(
         handle: Scalar,
         cx: &MiriInterpCx<'tcx>,