Auto merge of #962 - christianpoveda:file-shim, r=oli-obk

Add shims for file handling

This adds the bare minimum to be able to do `File::open` and `File::read`. I also need some feedback about how to handle certain things

src/lib.rs

@@ -34,6 +34,7 @@ pub use crate::shims::intrinsics::EvalContextExt as IntrinsicsEvalContextExt;
 pub use crate::shims::tls::{EvalContextExt as TlsEvalContextExt, TlsData};
 pub use crate::shims::dlsym::{Dlsym, EvalContextExt as DlsymEvalContextExt};
 pub use crate::shims::env::{EnvVars, EvalContextExt as EnvEvalContextExt};
+pub use crate::shims::io::{FileHandler, EvalContextExt as FileEvalContextExt};
 pub use crate::operator::EvalContextExt as OperatorEvalContextExt;
 pub use crate::range_map::RangeMap;
 pub use crate::helpers::{EvalContextExt as HelpersEvalContextExt};

src/machine.rs

@@ -96,6 +96,8 @@ pub struct Evaluator<'tcx> {
     /// If enabled, the `env_vars` field is populated with the host env vars during initialization
     /// and random number generation is delegated to the host.
     pub(crate) communicate: bool,
+
+    pub(crate) file_handler: FileHandler,
 }
 
 impl<'tcx> Evaluator<'tcx> {
@@ -110,6 +112,7 @@ impl<'tcx> Evaluator<'tcx> {
             last_error: 0,
             tls: TlsData::default(),
             communicate,
+            file_handler: Default::default(),
         }
     }
 }

src/shims/foreign_items.rs

@@ -446,6 +446,26 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriEvalContextExt<'mir, 'tcx
                 this.write_scalar(Scalar::from_int(result, dest.layout.size), dest)?;
             }
 
+            "open" | "open64" => {
+                let result = this.open(args[0], args[1])?;
+                this.write_scalar(Scalar::from_int(result, dest.layout.size), dest)?;
+            }
+
+            "fcntl" => {
+                let result = this.fcntl(args[0], args[1], args.get(2).cloned())?;
+                this.write_scalar(Scalar::from_int(result, dest.layout.size), dest)?;
+            }
+
+            "close" | "close$NOCANCEL" => {
+                let result = this.close(args[0])?;
+                this.write_scalar(Scalar::from_int(result, dest.layout.size), dest)?;
+            }
+
+            "read" => {
+                let result = this.read(args[0], args[1], args[2])?;
+                this.write_scalar(Scalar::from_int(result, dest.layout.size), dest)?;
+            }
+
             "write" => {
                 let fd = this.read_scalar(args[0])?.to_i32()?;
                 let buf = this.read_scalar(args[1])?.not_undef()?;
@@ -929,6 +949,14 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriEvalContextExt<'mir, 'tcx
         }
         return Ok(None);
     }
+
+    fn eval_libc_i32(&mut self, name: &str) -> InterpResult<'tcx, i32> {
+        self
+            .eval_context_mut()
+            .eval_path_scalar(&["libc", name])?
+            .ok_or_else(|| err_unsup_format!("Path libc::{} cannot be resolved.", name).into())
+            .and_then(|scalar| scalar.to_i32())
+    }
 }
 
 // Shims the linux 'getrandom()' syscall.

src/shims/io.rs

@@ -0,0 +1,213 @@
+use std::collections::HashMap;
+use std::fs::File;
+use std::io::Read;
+
+use rustc::ty::layout::Size;
+
+use crate::stacked_borrows::Tag;
+use crate::*;
+
+pub struct FileHandle {
+    file: File,
+    flag: i32,
+}
+
+pub struct FileHandler {
+    handles: HashMap<i32, FileHandle>,
+    low: i32,
+}
+
+impl Default for FileHandler {
+    fn default() -> Self {
+        FileHandler {
+            handles: Default::default(),
+            // 0, 1 and 2 are reserved for stdin, stdout and stderr
+            low: 3,
+        }
+    }
+}
+
+impl<'mir, 'tcx> EvalContextExt<'mir, 'tcx> for crate::MiriEvalContext<'mir, 'tcx> {}
+pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriEvalContextExt<'mir, 'tcx> {
+    fn open(
+        &mut self,
+        path_op: OpTy<'tcx, Tag>,
+        flag_op: OpTy<'tcx, Tag>,
+    ) -> InterpResult<'tcx, i32> {
+        let this = self.eval_context_mut();
+
+        if !this.machine.communicate {
+            throw_unsup_format!("`open` not available when isolation is enabled")
+        }
+
+        let flag = this.read_scalar(flag_op)?.to_i32()?;
+
+        if flag != this.eval_libc_i32("O_RDONLY")? && flag != this.eval_libc_i32("O_CLOEXEC")? {
+            throw_unsup_format!("Unsupported flag {:#x}", flag);
+        }
+
+        let path_bytes = this
+            .memory()
+            .read_c_str(this.read_scalar(path_op)?.not_undef()?)?;
+        let path = std::str::from_utf8(path_bytes)
+            .map_err(|_| err_unsup_format!("{:?} is not a valid utf-8 string", path_bytes))?;
+        let fd = File::open(path).map(|file| {
+            let mut fh = &mut this.machine.file_handler;
+            fh.low += 1;
+            fh.handles.insert(fh.low, FileHandle { file, flag });
+            fh.low
+        });
+
+        this.consume_result(fd)
+    }
+
+    fn fcntl(
+        &mut self,
+        fd_op: OpTy<'tcx, Tag>,
+        cmd_op: OpTy<'tcx, Tag>,
+        arg_op: Option<OpTy<'tcx, Tag>>,
+    ) -> InterpResult<'tcx, i32> {
+        let this = self.eval_context_mut();
+
+        if !this.machine.communicate {
+            throw_unsup_format!("`open` not available when isolation is enabled")
+        }
+
+        let fd = this.read_scalar(fd_op)?.to_i32()?;
+        let cmd = this.read_scalar(cmd_op)?.to_i32()?;
+
+        if cmd == this.eval_libc_i32("F_SETFD")? {
+            // This does not affect the file itself. Certain flags might require changing the file
+            // or the way it is accessed somehow.
+            let flag = this.read_scalar(arg_op.unwrap())?.to_i32()?;
+            // The only usage of this in stdlib at the moment is to enable the `FD_CLOEXEC` flag.
+            let fd_cloexec = this.eval_libc_i32("FD_CLOEXEC")?;
+            if let Some(FileHandle { flag: old_flag, .. }) =
+                this.machine.file_handler.handles.get_mut(&fd)
+            {
+                if flag ^ *old_flag == fd_cloexec {
+                    *old_flag = flag;
+                } else {
+                    throw_unsup_format!("Unsupported arg {:#x} for `F_SETFD`", flag);
+                }
+            }
+            Ok(0)
+        } else if cmd == this.eval_libc_i32("F_GETFD")? {
+            this.get_handle_and(fd, |handle| Ok(handle.flag))
+        } else {
+            throw_unsup_format!("Unsupported command {:#x}", cmd);
+        }
+    }
+
+    fn close(&mut self, fd_op: OpTy<'tcx, Tag>) -> InterpResult<'tcx, i32> {
+        let this = self.eval_context_mut();
+
+        if !this.machine.communicate {
+            throw_unsup_format!("`open` not available when isolation is enabled")
+        }
+
+        let fd = this.read_scalar(fd_op)?.to_i32()?;
+
+        this.remove_handle_and(
+            fd,
+            |handle, this| this.consume_result(handle.file.sync_all().map(|_| 0i32)),
+        )
+    }
+
+    fn read(
+        &mut self,
+        fd_op: OpTy<'tcx, Tag>,
+        buf_op: OpTy<'tcx, Tag>,
+        count_op: OpTy<'tcx, Tag>,
+    ) -> InterpResult<'tcx, i64> {
+        let this = self.eval_context_mut();
+
+        if !this.machine.communicate {
+            throw_unsup_format!("`open` not available when isolation is enabled")
+        }
+
+        let tcx = &{ this.tcx.tcx };
+
+        let fd = this.read_scalar(fd_op)?.to_i32()?;
+        let buf = this.force_ptr(this.read_scalar(buf_op)?.not_undef()?)?;
+        let count = this.read_scalar(count_op)?.to_usize(&*this.tcx)?;
+
+        // Remove the file handle to avoid borrowing issues
+        this.remove_handle_and(
+            fd,
+            |mut handle, this| {
+                let bytes = handle
+                    .file
+                    .read(this.memory_mut().get_mut(buf.alloc_id)?.get_bytes_mut(
+                        tcx,
+                        buf,
+                        Size::from_bytes(count),
+                    )?)
+                    .map(|bytes| bytes as i64);
+                // Reinsert the file handle
+                this.machine.file_handler.handles.insert(fd, handle);
+                this.consume_result(bytes)
+            },
+        )
+    }
+
+    /// Helper function that gets a `FileHandle` immutable reference and allows to manipulate it
+    /// using `f`.
+    ///
+    /// If the `fd` file descriptor does not corresponds to a file, this functions returns `Ok(-1)`
+    /// and sets `Evaluator::last_error` to `libc::EBADF` (invalid file descriptor).
+    ///
+    /// This function uses `T: From<i32>` instead of `i32` directly because some IO related
+    /// functions return different integer types (like `read`, that returns an `i64`)
+    fn get_handle_and<F, T: From<i32>>(&mut self, fd: i32, f: F) -> InterpResult<'tcx, T>
+    where
+        F: Fn(&FileHandle) -> InterpResult<'tcx, T>,
+    {
+        let this = self.eval_context_mut();
+        if let Some(handle) = this.machine.file_handler.handles.get(&fd) {
+            f(handle)
+        } else {
+            this.machine.last_error = this.eval_libc_i32("EBADF")? as u32;
+            Ok((-1).into())
+        }
+    }
+
+    /// Helper function that removes a `FileHandle` and allows to manipulate it using the `f`
+    /// closure. This function is quite useful when you need to modify a `FileHandle` but you need
+    /// to modify `MiriEvalContext` at the same time, so you can modify the handle and reinsert it
+    /// using `f`.
+    ///
+    /// If the `fd` file descriptor does not corresponds to a file, this functions returns `Ok(-1)`
+    /// and sets `Evaluator::last_error` to `libc::EBADF` (invalid file descriptor).
+    ///
+    /// This function uses `T: From<i32>` instead of `i32` directly because some IO related
+    /// functions return different integer types (like `read`, that returns an `i64`)
+    fn remove_handle_and<F, T: From<i32>>(&mut self, fd: i32, mut f: F) -> InterpResult<'tcx, T>
+    where
+        F: FnMut(FileHandle, &mut MiriEvalContext<'mir, 'tcx>) -> InterpResult<'tcx, T>,
+    {
+        let this = self.eval_context_mut();
+        if let Some(handle) = this.machine.file_handler.handles.remove(&fd) {
+            f(handle, this)
+        } else {
+            this.machine.last_error = this.eval_libc_i32("EBADF")? as u32;
+            Ok((-1).into())
+        }
+    }
+
+    /// Helper function that consumes an `std::io::Result<T>` and returns an
+    /// `InterpResult<'tcx,T>::Ok` instead. It is expected that the result can be converted to an
+    /// OS error using `std::io::Error::raw_os_error`.
+    ///
+    /// This function uses `T: From<i32>` instead of `i32` directly because some IO related
+    /// functions return different integer types (like `read`, that returns an `i64`)
+    fn consume_result<T: From<i32>>(&mut self, result: std::io::Result<T>) -> InterpResult<'tcx, T> {
+        match result {
+            Ok(ok) => Ok(ok),
+            Err(e) => {
+                self.eval_context_mut().machine.last_error = e.raw_os_error().unwrap() as u32;
+                Ok((-1).into())
+            }
+        }
+    }
+}

src/shims/mod.rs

@@ -3,6 +3,7 @@ pub mod env;
 pub mod foreign_items;
 pub mod intrinsics;
 pub mod tls;
+pub mod io;
 
 use rustc::{mir, ty};
 

tests/hello.txt

@@ -0,0 +1 @@
+Hello, World!

tests/run-pass/file_read.rs

@@ -0,0 +1,13 @@
+// ignore-windows: File handling is not implemented yet
+// compile-flags: -Zmiri-disable-isolation
+
+use std::fs::File;
+use std::io::Read;
+
+fn main() {
+    // FIXME: create the file and delete it when `rm` is implemented.
+    let mut file = File::open("./tests/hello.txt").unwrap();
+    let mut contents = String::new();
+    file.read_to_string(&mut contents).unwrap();
+    assert_eq!("Hello, World!\n", contents);
+}