std: refactor the TLS implementation

As discovered by Mara in rust-lang#110897, our TLS implementation is a total mess. In the past months, I have simplified the actual macros and their expansions, but the majority of the complexity comes from the platform-specific support code needed to create keys and register destructors. In keeping with rust-lang#117276, I have therefore moved all of the `thread_local_key`/`thread_local_dtor` modules to the `thread_local` module in `sys` and merged them into a new structure, so that future porters of `std` can simply mix-and-match the existing code instead of having to copy the same (bad) implementation everywhere. The new structure should become obvious when looking at `sys/thread_local/mod.rs`.

Unfortunately, the documentation changes associated with the refactoring have made this PR rather large. That said, this contains no functional changes except for two small ones:
* the key-based destructor fallback now, by virtue of sharing the implementation used by macOS and others, stores its list in a `#[thread_local]` static instead of in the key, eliminating one indirection layer and drastically simplifying its code.
* I've switched over ZKVM (tier 3) to use the same implementation as WebAssembly, as the implementation was just a way worse version of that

Please let me know if I can make this easier to review! I know these large PRs aren't optimal, but I couldn't think of any good intermediate steps.

@rustbot label +A-thread-locals

library/std/src/sys/pal/hermit/mod.rs

@@ -33,9 +33,6 @@ pub mod pipe;
 pub mod process;
 pub mod stdio;
 pub mod thread;
-pub mod thread_local_dtor;
-#[path = "../unsupported/thread_local_key.rs"]
-pub mod thread_local_key;
 pub mod time;
 
 use crate::io::ErrorKind;
@@ -98,7 +95,6 @@ pub unsafe extern "C" fn runtime_entry(
     argv: *const *const c_char,
     env: *const *const c_char,
 ) -> ! {
-    use thread_local_dtor::run_dtors;
     extern "C" {
         fn main(argc: isize, argv: *const *const c_char) -> i32;
     }
@@ -108,7 +104,7 @@ pub unsafe extern "C" fn runtime_entry(
 
     let result = main(argc as isize, argv);
 
-    run_dtors();
+    crate::sys::thread_local::destructors::run();
     hermit_abi::exit(result);
 }
 

library/std/src/sys/pal/hermit/thread.rs

@@ -1,7 +1,6 @@
 #![allow(dead_code)]
 
 use super::hermit_abi;
-use super::thread_local_dtor::run_dtors;
 use crate::ffi::CStr;
 use crate::io;
 use crate::mem;
@@ -50,7 +49,7 @@ impl Thread {
                 Box::from_raw(ptr::with_exposed_provenance::<Box<dyn FnOnce()>>(main).cast_mut())();
 
                 // run all destructors
-                run_dtors();
+                crate::sys::thread_local::destructors::run();
             }
         }
     }

library/std/src/sys/pal/itron/thread.rs

@@ -14,7 +14,6 @@ use crate::{
     num::NonZero,
     ptr::NonNull,
     sync::atomic::{AtomicUsize, Ordering},
-    sys::thread_local_dtor::run_dtors,
     time::Duration,
 };
 
@@ -116,7 +115,7 @@ impl Thread {
 
             // Run TLS destructors now because they are not
             // called automatically for terminated tasks.
-            unsafe { run_dtors() };
+            unsafe { crate::sys::thread_local::destructors::run() };
 
             let old_lifecycle = inner
                 .lifecycle

library/std/src/sys/pal/sgx/mod.rs

@@ -26,7 +26,6 @@ pub mod pipe;
 pub mod process;
 pub mod stdio;
 pub mod thread;
-pub mod thread_local_key;
 pub mod thread_parking;
 pub mod time;
 pub mod waitqueue;

library/std/src/sys/pal/solid/mod.rs

@@ -33,8 +33,6 @@ pub mod pipe;
 pub mod process;
 pub mod stdio;
 pub use self::itron::thread;
-pub mod thread_local_dtor;
-pub mod thread_local_key;
 pub use self::itron::thread_parking;
 pub mod time;
 

library/std/src/sys/pal/teeos/mod.rs

@@ -27,9 +27,6 @@ pub mod process;
 mod rand;
 pub mod stdio;
 pub mod thread;
-pub mod thread_local_dtor;
-#[path = "../unix/thread_local_key.rs"]
-pub mod thread_local_key;
 #[allow(non_upper_case_globals)]
 #[path = "../unix/time.rs"]
 pub mod time;

library/std/src/sys/pal/uefi/mod.rs

@@ -28,8 +28,6 @@ pub mod pipe;
 pub mod process;
 pub mod stdio;
 pub mod thread;
-#[path = "../unsupported/thread_local_key.rs"]
-pub mod thread_local_key;
 pub mod time;
 
 mod helpers;

library/std/src/sys/pal/unix/mod.rs

@@ -31,8 +31,6 @@ pub mod rand;
 pub mod stack_overflow;
 pub mod stdio;
 pub mod thread;
-pub mod thread_local_dtor;
-pub mod thread_local_key;
 pub mod thread_parking;
 pub mod time;
 

library/std/src/sys/pal/unsupported/mod.rs

@@ -11,9 +11,6 @@ pub mod pipe;
 pub mod process;
 pub mod stdio;
 pub mod thread;
-#[cfg(target_thread_local)]
-pub mod thread_local_dtor;
-pub mod thread_local_key;
 pub mod time;
 
 mod common;