Merge branch 'master' into flock

rust-version

@@ -1 +1 @@
-9057c3ffec44926d5e149dc13ff3ce1613b69cce
+72d73cec61aa8f85901358cd5d386d5dd066fe52

src/bin/miri.rs

@@ -620,6 +620,14 @@ fn main() {
             "-Zmiri-unique-is-unique only has an effect when -Zmiri-tree-borrows is also used"
         );
     }
+    // Tree Borrows + permissive provenance does not work.
+    if miri_config.provenance_mode == ProvenanceMode::Permissive
+        && matches!(miri_config.borrow_tracker, Some(BorrowTrackerMethod::TreeBorrows))
+    {
+        show_error!(
+            "Tree Borrows does not support integer-to-pointer casts, and is hence not compatible with permissive provenance"
+        );
+    }
 
     debug!("rustc arguments: {:?}", rustc_args);
     debug!("crate arguments: {:?}", miri_config.args);

src/borrow_tracker/mod.rs

@@ -232,6 +232,10 @@ impl GlobalStateInner {
     pub fn remove_unreachable_allocs(&mut self, allocs: &LiveAllocs<'_, '_>) {
         self.root_ptr_tags.retain(|id, _| allocs.is_live(*id));
     }
+
+    pub fn borrow_tracker_method(&self) -> BorrowTrackerMethod {
+        self.borrow_tracker_method
+    }
 }
 
 /// Which borrow tracking method to use

src/borrow_tracker/stacked_borrows/mod.rs

@@ -5,6 +5,7 @@ pub mod diagnostics;
 mod item;
 mod stack;
 
+use std::cell::RefCell;
 use std::cmp;
 use std::fmt::Write;
 use std::mem;
@@ -820,7 +821,19 @@ trait EvalContextPrivExt<'tcx, 'ecx>: crate::MiriInterpCxExt<'tcx> {
         // See https://github.com/rust-lang/unsafe-code-guidelines/issues/276.
         let size = match size {
             Some(size) => size,
-            None => return Ok(place.clone()),
+            None => {
+                // The first time this happens, show a warning.
+                thread_local! { static WARNING_SHOWN: RefCell<bool> = const { RefCell::new(false) }; }
+                WARNING_SHOWN.with_borrow_mut(|shown| {
+                    if *shown {
+                        return;
+                    }
+                    // Not yet shown. Show it!
+                    *shown = true;
+                    this.emit_diagnostic(NonHaltingDiagnostic::ExternTypeReborrow);
+                });
+                return Ok(place.clone());
+            }
         };
 
         // Compute new borrow.

src/machine.rs

@@ -660,7 +660,7 @@ impl<'tcx> MiriMachine<'tcx> {
             tls: TlsData::default(),
             isolated_op: config.isolated_op,
             validate: config.validate,
-            fds: shims::FdTable::new(config.mute_stdout_stderr),
+            fds: shims::FdTable::init(config.mute_stdout_stderr),
             dirs: Default::default(),
             layouts,
             threads,

src/shims/env.rs

@@ -108,4 +108,9 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
             EnvVars::Windows(vars) => vars.get(name),
         }
     }
+
+    fn get_pid(&self) -> u32 {
+        let this = self.eval_context_ref();
+        if this.machine.communicate() { std::process::id() } else { 1000 }
+    }
 }

src/shims/unix/env.rs

@@ -274,12 +274,23 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
         let this = self.eval_context_mut();
         this.assert_target_os_is_unix("getpid");
 
-        this.check_no_isolation("`getpid`")?;
-
         // The reason we need to do this wacky of a conversion is because
         // `libc::getpid` returns an i32, however, `std::process::id()` return an u32.
         // So we un-do the conversion that stdlib does and turn it back into an i32.
         #[allow(clippy::cast_possible_wrap)]
-        Ok(std::process::id() as i32)
+        Ok(this.get_pid() as i32)
+    }
+
+    fn linux_gettid(&mut self) -> InterpResult<'tcx, i32> {
+        let this = self.eval_context_ref();
+        this.assert_target_os("linux", "gettid");
+
+        let index = this.machine.threads.active_thread().to_u32();
+
+        // Compute a TID for this thread, ensuring that the main thread has PID == TID.
+        let tid = this.get_pid().strict_add(index);
+
+        #[allow(clippy::cast_possible_wrap)]
+        Ok(tid as i32)
     }
 }

src/shims/unix/fd.rs

@@ -36,6 +36,30 @@ pub trait FileDescription: std::fmt::Debug + Any {
         throw_unsup_format!("cannot write to {}", self.name());
     }
 
+    /// Reads as much as possible into the given buffer from a given offset,
+    /// and returns the number of bytes read.
+    fn pread<'tcx>(
+        &mut self,
+        _communicate_allowed: bool,
+        _bytes: &mut [u8],
+        _offset: u64,
+        _ecx: &mut MiriInterpCx<'tcx>,
+    ) -> InterpResult<'tcx, io::Result<usize>> {
+        throw_unsup_format!("cannot pread from {}", self.name());
+    }
+
+    /// Writes as much as possible from the given buffer starting at a given offset,
+    /// and returns the number of bytes written.
+    fn pwrite<'tcx>(
+        &mut self,
+        _communicate_allowed: bool,
+        _bytes: &[u8],
+        _offset: u64,
+        _ecx: &mut MiriInterpCx<'tcx>,
+    ) -> InterpResult<'tcx, io::Result<usize>> {
+        throw_unsup_format!("cannot pwrite to {}", self.name());
+    }
+
     /// Seeks to the given offset (which can be relative to the beginning, end, or current position).
     /// Returns the new position from the start of the stream.
     fn seek<'tcx>(
@@ -176,10 +200,6 @@ impl FileDescription for NullOutput {
 pub struct FileDescriptor(Rc<RefCell<Box<dyn FileDescription>>>);
 
 impl FileDescriptor {
-    pub fn new<T: FileDescription>(fd: T) -> Self {
-        FileDescriptor(Rc::new(RefCell::new(Box::new(fd))))
-    }
-
     pub fn borrow(&self) -> Ref<'_, dyn FileDescription> {
         Ref::map(self.0.borrow(), |fd| fd.as_ref())
     }
@@ -211,20 +231,25 @@ impl VisitProvenance for FdTable {
 }
 
 impl FdTable {
-    pub(crate) fn new(mute_stdout_stderr: bool) -> FdTable {
-        let mut fds: BTreeMap<_, FileDescriptor> = BTreeMap::new();
-        fds.insert(0i32, FileDescriptor::new(io::stdin()));
+    fn new() -> Self {
+        FdTable { fds: BTreeMap::new() }
+    }
+    pub(crate) fn init(mute_stdout_stderr: bool) -> FdTable {
+        let mut fds = FdTable::new();
+        fds.insert_fd(io::stdin());
         if mute_stdout_stderr {
-            fds.insert(1i32, FileDescriptor::new(NullOutput));
-            fds.insert(2i32, FileDescriptor::new(NullOutput));
+            assert_eq!(fds.insert_fd(NullOutput), 1);
+            assert_eq!(fds.insert_fd(NullOutput), 2);
         } else {
-            fds.insert(1i32, FileDescriptor::new(io::stdout()));
-            fds.insert(2i32, FileDescriptor::new(io::stderr()));
+            assert_eq!(fds.insert_fd(io::stdout()), 1);
+            assert_eq!(fds.insert_fd(io::stderr()), 2);
         }
-        FdTable { fds }
+        fds
     }
 
-    pub fn insert_fd(&mut self, file_handle: FileDescriptor) -> i32 {
+    /// Insert a file descriptor to the FdTable.
+    pub fn insert_fd<T: FileDescription>(&mut self, fd: T) -> i32 {
+        let file_handle = FileDescriptor(Rc::new(RefCell::new(Box::new(fd))));
         self.insert_fd_with_min_fd(file_handle, 0)
     }
 
@@ -422,7 +447,18 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
         Ok((-1).into())
     }
 
-    fn read(&mut self, fd: i32, buf: Pointer, count: u64) -> InterpResult<'tcx, i64> {
+    /// Read data from `fd` into buffer specified by `buf` and `count`.
+    ///
+    /// If `offset` is `None`, reads data from current cursor position associated with `fd`
+    /// and updates cursor position on completion. Otherwise, reads from the specified offset
+    /// and keeps the cursor unchanged.
+    fn read(
+        &mut self,
+        fd: i32,
+        buf: Pointer,
+        count: u64,
+        offset: Option<i128>,
+    ) -> InterpResult<'tcx, i64> {
         let this = self.eval_context_mut();
 
         // Isolation check is done via `FileDescriptor` trait.
@@ -440,25 +476,31 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
         let communicate = this.machine.communicate();
 
         // We temporarily dup the FD to be able to retain mutable access to `this`.
-        let Some(file_descriptor) = this.machine.fds.dup(fd) else {
+        let Some(fd) = this.machine.fds.dup(fd) else {
             trace!("read: FD not found");
             return this.fd_not_found();
         };
 
-        trace!("read: FD mapped to {:?}", file_descriptor);
+        trace!("read: FD mapped to {fd:?}");
         // We want to read at most `count` bytes. We are sure that `count` is not negative
         // because it was a target's `usize`. Also we are sure that its smaller than
         // `usize::MAX` because it is bounded by the host's `isize`.
         let mut bytes = vec![0; usize::try_from(count).unwrap()];
-        // `File::read` never returns a value larger than `count`,
-        // so this cannot fail.
-        let result = file_descriptor
-            .borrow_mut()
-            .read(communicate, &mut bytes, this)?
-            .map(|c| i64::try_from(c).unwrap());
-        drop(file_descriptor);
+        let result = match offset {
+            None => fd.borrow_mut().read(communicate, &mut bytes, this),
+            Some(offset) => {
+                let Ok(offset) = u64::try_from(offset) else {
+                    let einval = this.eval_libc("EINVAL");
+                    this.set_last_error(einval)?;
+                    return Ok(-1);
+                };
+                fd.borrow_mut().pread(communicate, &mut bytes, offset, this)
+            }
+        };
+        drop(fd);
 
-        match result {
+        // `File::read` never returns a value larger than `count`, so this cannot fail.
+        match result?.map(|c| i64::try_from(c).unwrap()) {
             Ok(read_bytes) => {
                 // If reading to `bytes` did not fail, we write those bytes to the buffer.
                 // Crucially, if fewer than `bytes.len()` bytes were read, only write
@@ -476,7 +518,13 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
         }
     }
 
-    fn write(&mut self, fd: i32, buf: Pointer, count: u64) -> InterpResult<'tcx, i64> {
+    fn write(
+        &mut self,
+        fd: i32,
+        buf: Pointer,
+        count: u64,
+        offset: Option<i128>,
+    ) -> InterpResult<'tcx, i64> {
         let this = self.eval_context_mut();
 
         // Isolation check is done via `FileDescriptor` trait.
@@ -493,16 +541,24 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
 
         let bytes = this.read_bytes_ptr_strip_provenance(buf, Size::from_bytes(count))?.to_owned();
         // We temporarily dup the FD to be able to retain mutable access to `this`.
-        let Some(file_descriptor) = this.machine.fds.dup(fd) else {
+        let Some(fd) = this.machine.fds.dup(fd) else {
             return this.fd_not_found();
         };
 
-        let result = file_descriptor
-            .borrow_mut()
-            .write(communicate, &bytes, this)?
-            .map(|c| i64::try_from(c).unwrap());
-        drop(file_descriptor);
+        let result = match offset {
+            None => fd.borrow_mut().write(communicate, &bytes, this),
+            Some(offset) => {
+                let Ok(offset) = u64::try_from(offset) else {
+                    let einval = this.eval_libc("EINVAL");
+                    this.set_last_error(einval)?;
+                    return Ok(-1);
+                };
+                fd.borrow_mut().pwrite(communicate, &bytes, offset, this)
+            }
+        };
+        drop(fd);
 
+        let result = result?.map(|c| i64::try_from(c).unwrap());
         this.try_unwrap_io_result(result)
     }
 }

src/shims/unix/foreign_items.rs

@@ -92,7 +92,7 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
                 let fd = this.read_scalar(fd)?.to_i32()?;
                 let buf = this.read_pointer(buf)?;
                 let count = this.read_target_usize(count)?;
-                let result = this.read(fd, buf, count)?;
+                let result = this.read(fd, buf, count, None)?;
                 this.write_scalar(Scalar::from_target_isize(result, this), dest)?;
             }
             "write" => {
@@ -101,7 +101,47 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
                 let buf = this.read_pointer(buf)?;
                 let count = this.read_target_usize(n)?;
                 trace!("Called write({:?}, {:?}, {:?})", fd, buf, count);
-                let result = this.write(fd, buf, count)?;
+                let result = this.write(fd, buf, count, None)?;
+                // Now, `result` is the value we return back to the program.
+                this.write_scalar(Scalar::from_target_isize(result, this), dest)?;
+            }
+            "pread" => {
+                let [fd, buf, count, offset] = this.check_shim(abi, Abi::C { unwind: false }, link_name, args)?;
+                let fd = this.read_scalar(fd)?.to_i32()?;
+                let buf = this.read_pointer(buf)?;
+                let count = this.read_target_usize(count)?;
+                let offset = this.read_scalar(offset)?.to_int(this.libc_ty_layout("off_t").size)?;
+                let result = this.read(fd, buf, count, Some(offset))?;
+                this.write_scalar(Scalar::from_target_isize(result, this), dest)?;
+            }
+            "pwrite" => {
+                let [fd, buf, n, offset] = this.check_shim(abi, Abi::C { unwind: false }, link_name, args)?;
+                let fd = this.read_scalar(fd)?.to_i32()?;
+                let buf = this.read_pointer(buf)?;
+                let count = this.read_target_usize(n)?;
+                let offset = this.read_scalar(offset)?.to_int(this.libc_ty_layout("off_t").size)?;
+                trace!("Called pwrite({:?}, {:?}, {:?}, {:?})", fd, buf, count, offset);
+                let result = this.write(fd, buf, count, Some(offset))?;
+                // Now, `result` is the value we return back to the program.
+                this.write_scalar(Scalar::from_target_isize(result, this), dest)?;
+            }
+            "pread64" => {
+                let [fd, buf, count, offset] = this.check_shim(abi, Abi::C { unwind: false }, link_name, args)?;
+                let fd = this.read_scalar(fd)?.to_i32()?;
+                let buf = this.read_pointer(buf)?;
+                let count = this.read_target_usize(count)?;
+                let offset = this.read_scalar(offset)?.to_int(this.libc_ty_layout("off64_t").size)?;
+                let result = this.read(fd, buf, count, Some(offset))?;
+                this.write_scalar(Scalar::from_target_isize(result, this), dest)?;
+            }
+            "pwrite64" => {
+                let [fd, buf, n, offset] = this.check_shim(abi, Abi::C { unwind: false }, link_name, args)?;
+                let fd = this.read_scalar(fd)?.to_i32()?;
+                let buf = this.read_pointer(buf)?;
+                let count = this.read_target_usize(n)?;
+                let offset = this.read_scalar(offset)?.to_int(this.libc_ty_layout("off64_t").size)?;
+                trace!("Called pwrite64({:?}, {:?}, {:?}, {:?})", fd, buf, count, offset);
+                let result = this.write(fd, buf, count, Some(offset))?;
                 // Now, `result` is the value we return back to the program.
                 this.write_scalar(Scalar::from_target_isize(result, this), dest)?;
             }