Auto merge of #141860 - RalfJung:miri-sync, r=RalfJung

Miri subtree update

r? `@ghost`

Includes rust-lang/miri#4370 to unbreak PR CI. (So we're committing to having bda28aa38 in the Miri history by landing this, whether or not that Miri PR lands.)

Cc `@Noratrieb` `@tiif`

Author: bors

src/tools/miri/cargo-miri/src/phases.rs

@@ -90,7 +90,7 @@ pub fn phase_cargo_miri(mut args: impl Iterator<Item = String>) {
                 "`cargo miri` supports the following subcommands: `run`, `test`, `nextest`, `clean`, and `setup`."
             ),
     };
-    let verbose = num_arg_flag("-v");
+    let verbose = num_arg_flag("-v") + num_arg_flag("--verbose");
     let quiet = has_arg_flag("-q") || has_arg_flag("--quiet");
 
     // Determine the involved architectures.

src/tools/miri/ci/ci.sh

@@ -142,12 +142,12 @@ case $HOST_TARGET in
     # Host
     GC_STRESS=1 MIR_OPT=1 MANY_SEEDS=64 TEST_BENCH=1 CARGO_MIRI_ENV=1 run_tests
     # Extra tier 1
-    # With reduced many-seed count to avoid spending too much time on that.
-    # (All OSes and ABIs are run with 64 seeds at least once though via the macOS runner.)
-    MANY_SEEDS=16 TEST_TARGET=i686-unknown-linux-gnu run_tests
-    MANY_SEEDS=16 TEST_TARGET=aarch64-unknown-linux-gnu run_tests
-    MANY_SEEDS=16 TEST_TARGET=x86_64-apple-darwin run_tests
-    MANY_SEEDS=16 TEST_TARGET=x86_64-pc-windows-gnu run_tests
+    MANY_SEEDS=64 TEST_TARGET=i686-unknown-linux-gnu run_tests
+    MANY_SEEDS=64 TEST_TARGET=aarch64-unknown-linux-gnu run_tests
+    MANY_SEEDS=64 TEST_TARGET=x86_64-apple-darwin run_tests
+    MANY_SEEDS=64 TEST_TARGET=x86_64-pc-windows-gnu run_tests
+    # Extra tier 1 candidate
+    MANY_SEEDS=64 TEST_TARGET=aarch64-pc-windows-msvc run_tests
     ;;
   aarch64-apple-darwin)
     # Host
@@ -156,7 +156,8 @@ case $HOST_TARGET in
     MANY_SEEDS=64 TEST_TARGET=i686-pc-windows-gnu run_tests
     MANY_SEEDS=64 TEST_TARGET=x86_64-pc-windows-msvc CARGO_MIRI_ENV=1 run_tests
     # Extra tier 2
-    MANY_SEEDS=16 TEST_TARGET=arm-unknown-linux-gnueabi run_tests
+    MANY_SEEDS=16 TEST_TARGET=arm-unknown-linux-gnueabi run_tests # 32bit ARM
+    MANY_SEEDS=16 TEST_TARGET=aarch64-pc-windows-gnullvm run_tests # gnullvm ABI
     MANY_SEEDS=16 TEST_TARGET=s390x-unknown-linux-gnu run_tests # big-endian architecture of choice
     # Not officially supported tier 2
     MANY_SEEDS=16 TEST_TARGET=mips-unknown-linux-gnu run_tests # a 32bit big-endian target, and also a target without 64bit atomics
@@ -178,7 +179,7 @@ case $HOST_TARGET in
     # Host
     # Without GC_STRESS and with reduced many-seeds count as this is the slowest runner.
     # (The macOS runner checks windows-msvc with full many-seeds count.)
-    MIR_OPT=1 MANY_SEEDS=16 TEST_BENCH=1 run_tests
+    MIR_OPT=1 MANY_SEEDS=64 TEST_BENCH=1 run_tests
     # Extra tier 1
     # We really want to ensure a Linux target works on a Windows host,
     # and a 64bit target works on a 32bit host.

src/tools/miri/miri-script/src/commands.rs

@@ -1,4 +1,4 @@
-use std::collections::HashMap;
+use std::collections::BTreeMap;
 use std::ffi::{OsStr, OsString};
 use std::fmt::Write as _;
 use std::fs::{self, File};
@@ -404,7 +404,28 @@ impl Command {
         // We want to forward the host stdin so apparently we cannot use `cmd!`.
         let mut cmd = process::Command::new("git");
         cmd.arg("rebase").arg(&base).arg("--interactive");
-        cmd.env("GIT_SEQUENCE_EDITOR", env::current_exe()?);
+        let current_exe = {
+            if cfg!(windows) {
+                // Apparently git-for-Windows gets confused by backslashes if we just use
+                // `current_exe()` here. So replace them by forward slashes if this is not a "magic"
+                // path starting with "\\". This is clearly a git bug but we work around it here.
+                // Also see <https://github.com/rust-lang/miri/issues/4340>.
+                let bin = env::current_exe()?;
+                match bin.into_os_string().into_string() {
+                    Err(not_utf8) => not_utf8.into(), // :shrug:
+                    Ok(str) => {
+                        if str.starts_with(r"\\") {
+                            str.into() // don't touch these magic paths, they must use backslashes
+                        } else {
+                            str.replace('\\', "/").into()
+                        }
+                    }
+                }
+            } else {
+                env::current_exe()?
+            }
+        };
+        cmd.env("GIT_SEQUENCE_EDITOR", current_exe);
         cmd.env("MIRI_SCRIPT_IS_GIT_SEQUENCE_EDITOR", "1");
         cmd.current_dir(sh.current_dir());
         let result = cmd.status()?;
@@ -489,7 +510,9 @@ impl Command {
             sh.read_dir(benches_dir)?
                 .into_iter()
                 .filter(|path| path.is_dir())
-                .map(|path| path.into_os_string().into_string().unwrap())
+                // Only keep the basename: that matches the usage with a manual bench list,
+                // and it ensure the path concatenations below work as intended.
+                .map(|path| path.file_name().unwrap().to_owned().into_string().unwrap())
                 .collect()
         } else {
             benches.into_iter().collect()
@@ -530,14 +553,16 @@ impl Command {
             stddev: f64,
         }
 
-        let gather_results = || -> Result<HashMap<&str, BenchResult>> {
+        let gather_results = || -> Result<BTreeMap<&str, BenchResult>> {
             let baseline_temp_dir = results_json_dir.unwrap();
-            let mut results = HashMap::new();
+            let mut results = BTreeMap::new();
             for bench in &benches {
-                let result = File::open(path!(baseline_temp_dir / format!("{bench}.bench.json")))?;
-                let mut result: serde_json::Value =
-                    serde_json::from_reader(BufReader::new(result))?;
-                let result: BenchResult = serde_json::from_value(result["results"][0].take())?;
+                let result = File::open(path!(baseline_temp_dir / format!("{bench}.bench.json")))
+                    .context("failed to read hyperfine JSON")?;
+                let mut result: serde_json::Value = serde_json::from_reader(BufReader::new(result))
+                    .context("failed to parse hyperfine JSON")?;
+                let result: BenchResult = serde_json::from_value(result["results"][0].take())
+                    .context("failed to interpret hyperfine JSON")?;
                 results.insert(bench as &str, result);
             }
             Ok(results)
@@ -549,15 +574,15 @@ impl Command {
             serde_json::to_writer_pretty(BufWriter::new(baseline), &results)?;
         } else if let Some(baseline_file) = load_baseline {
             let new_results = gather_results()?;
-            let baseline_results: HashMap<String, BenchResult> = {
+            let baseline_results: BTreeMap<String, BenchResult> = {
                 let f = File::open(baseline_file)?;
                 serde_json::from_reader(BufReader::new(f))?
             };
             println!(
                 "Comparison with baseline (relative speed, lower is better for the new results):"
             );
-            for (bench, new_result) in new_results.iter() {
-                let Some(baseline_result) = baseline_results.get(*bench) else { continue };
+            for (bench, new_result) in new_results {
+                let Some(baseline_result) = baseline_results.get(bench) else { continue };
 
                 // Compare results (inspired by hyperfine)
                 let ratio = new_result.mean / baseline_result.mean;

src/tools/miri/rust-version

@@ -1 +1 @@
-2b96ddca1272960623e41829439df8dae82d20af
+337c11e5932275e7d450c1f2e26f289f0ddfa717

src/tools/miri/src/alloc_bytes.rs renamed to src/tools/miri/src/alloc/alloc_bytes.rs

@@ -1,12 +1,23 @@
 use std::alloc::Layout;
 use std::borrow::Cow;
 use std::{alloc, slice};
+#[cfg(target_os = "linux")]
+use std::{cell::RefCell, rc::Rc};
 
 use rustc_abi::{Align, Size};
 use rustc_middle::mir::interpret::AllocBytes;
 
+#[cfg(target_os = "linux")]
+use crate::alloc::isolated_alloc::IsolatedAlloc;
 use crate::helpers::ToU64 as _;
 
+#[derive(Clone, Debug)]
+pub enum MiriAllocParams {
+    Global,
+    #[cfg(target_os = "linux")]
+    Isolated(Rc<RefCell<IsolatedAlloc>>),
+}
+
 /// Allocation bytes that explicitly handle the layout of the data they're storing.
 /// This is necessary to interface with native code that accesses the program store in Miri.
 #[derive(Debug)]
@@ -18,13 +29,16 @@ pub struct MiriAllocBytes {
     /// * If `self.layout.size() == 0`, then `self.ptr` was allocated with the equivalent layout with size 1.
     /// * Otherwise, `self.ptr` points to memory allocated with `self.layout`.
     ptr: *mut u8,
+    /// Whether this instance of `MiriAllocBytes` had its allocation created by calling `alloc::alloc()`
+    /// (`Global`) or the discrete allocator (`Isolated`)
+    params: MiriAllocParams,
 }
 
 impl Clone for MiriAllocBytes {
     fn clone(&self) -> Self {
         let bytes: Cow<'_, [u8]> = Cow::Borrowed(self);
         let align = Align::from_bytes(self.layout.align().to_u64()).unwrap();
-        MiriAllocBytes::from_bytes(bytes, align, ())
+        MiriAllocBytes::from_bytes(bytes, align, self.params.clone())
     }
 }
 
@@ -37,8 +51,16 @@ impl Drop for MiriAllocBytes {
         } else {
             self.layout
         };
+
         // SAFETY: Invariant, `self.ptr` points to memory allocated with `self.layout`.
-        unsafe { alloc::dealloc(self.ptr, alloc_layout) }
+        unsafe {
+            match self.params.clone() {
+                MiriAllocParams::Global => alloc::dealloc(self.ptr, alloc_layout),
+                #[cfg(target_os = "linux")]
+                MiriAllocParams::Isolated(alloc) =>
+                    alloc.borrow_mut().dealloc(self.ptr, alloc_layout),
+            }
+        }
     }
 }
 
@@ -67,35 +89,45 @@ impl MiriAllocBytes {
     fn alloc_with(
         size: u64,
         align: u64,
-        alloc_fn: impl FnOnce(Layout) -> *mut u8,
+        params: MiriAllocParams,
+        alloc_fn: impl FnOnce(Layout, &MiriAllocParams) -> *mut u8,
     ) -> Result<MiriAllocBytes, ()> {
         let size = usize::try_from(size).map_err(|_| ())?;
         let align = usize::try_from(align).map_err(|_| ())?;
         let layout = Layout::from_size_align(size, align).map_err(|_| ())?;
         // When size is 0 we allocate 1 byte anyway, to ensure each allocation has a unique address.
         let alloc_layout =
             if size == 0 { Layout::from_size_align(1, align).unwrap() } else { layout };
-        let ptr = alloc_fn(alloc_layout);
+        let ptr = alloc_fn(alloc_layout, &params);
         if ptr.is_null() {
             Err(())
         } else {
             // SAFETY: All `MiriAllocBytes` invariants are fulfilled.
-            Ok(Self { ptr, layout })
+            Ok(Self { ptr, layout, params })
         }
     }
 }
 
 impl AllocBytes for MiriAllocBytes {
-    /// Placeholder!
-    type AllocParams = ();
+    type AllocParams = MiriAllocParams;
 
-    fn from_bytes<'a>(slice: impl Into<Cow<'a, [u8]>>, align: Align, _params: ()) -> Self {
+    fn from_bytes<'a>(
+        slice: impl Into<Cow<'a, [u8]>>,
+        align: Align,
+        params: MiriAllocParams,
+    ) -> Self {
         let slice = slice.into();
         let size = slice.len();
         let align = align.bytes();
         // SAFETY: `alloc_fn` will only be used with `size != 0`.
-        let alloc_fn = |layout| unsafe { alloc::alloc(layout) };
-        let alloc_bytes = MiriAllocBytes::alloc_with(size.to_u64(), align, alloc_fn)
+        let alloc_fn = |layout, params: &MiriAllocParams| unsafe {
+            match params {
+                MiriAllocParams::Global => alloc::alloc(layout),
+                #[cfg(target_os = "linux")]
+                MiriAllocParams::Isolated(alloc) => alloc.borrow_mut().alloc(layout),
+            }
+        };
+        let alloc_bytes = MiriAllocBytes::alloc_with(size.to_u64(), align, params, alloc_fn)
             .unwrap_or_else(|()| {
                 panic!("Miri ran out of memory: cannot create allocation of {size} bytes")
             });
@@ -105,12 +137,18 @@ impl AllocBytes for MiriAllocBytes {
         alloc_bytes
     }
 
-    fn zeroed(size: Size, align: Align, _params: ()) -> Option<Self> {
+    fn zeroed(size: Size, align: Align, params: MiriAllocParams) -> Option<Self> {
         let size = size.bytes();
         let align = align.bytes();
         // SAFETY: `alloc_fn` will only be used with `size != 0`.
-        let alloc_fn = |layout| unsafe { alloc::alloc_zeroed(layout) };
-        MiriAllocBytes::alloc_with(size, align, alloc_fn).ok()
+        let alloc_fn = |layout, params: &MiriAllocParams| unsafe {
+            match params {
+                MiriAllocParams::Global => alloc::alloc_zeroed(layout),
+                #[cfg(target_os = "linux")]
+                MiriAllocParams::Isolated(alloc) => alloc.borrow_mut().alloc_zeroed(layout),
+            }
+        };
+        MiriAllocBytes::alloc_with(size, align, params, alloc_fn).ok()
     }
 
     fn as_mut_ptr(&mut self) -> *mut u8 {