Unrolled build for rust-lang#116341

Rollup merge of rust-lang#116341 - Ayush1325:uefi-args, r=Mark-Simulacrum

Implement sys::args for UEFI

- Uses `EFI_LOADED_IMAGE_PROTOCOL`, which is implemented for all loaded images.

Tested on qemu with OVMF

cc ``@nicholasbishop``
cc ``@dvdhrm``

Author: rust-timer

library/std/src/sys/uefi/args.rs

@@ -0,0 +1,158 @@
+use r_efi::protocols::loaded_image;
+
+use crate::env::current_exe;
+use crate::ffi::OsString;
+use crate::fmt;
+use crate::iter::Iterator;
+use crate::mem::size_of;
+use crate::sys::uefi::helpers;
+use crate::vec;
+
+pub struct Args {
+    parsed_args_list: vec::IntoIter<OsString>,
+}
+
+pub fn args() -> Args {
+    let lazy_current_exe = || Vec::from([current_exe().map(Into::into).unwrap_or_default()]);
+
+    // Each loaded image has an image handle that supports `EFI_LOADED_IMAGE_PROTOCOL`. Thus, this
+    // will never fail.
+    let protocol =
+        helpers::image_handle_protocol::<loaded_image::Protocol>(loaded_image::PROTOCOL_GUID)
+            .unwrap();
+
+    let lp_size = unsafe { (*protocol.as_ptr()).load_options_size } as usize;
+    // Break if we are sure that it cannot be UTF-16
+    if lp_size < size_of::<u16>() || lp_size % size_of::<u16>() != 0 {
+        return Args { parsed_args_list: lazy_current_exe().into_iter() };
+    }
+    let lp_size = lp_size / size_of::<u16>();
+
+    let lp_cmd_line = unsafe { (*protocol.as_ptr()).load_options as *const u16 };
+    if !lp_cmd_line.is_aligned() {
+        return Args { parsed_args_list: lazy_current_exe().into_iter() };
+    }
+    let lp_cmd_line = unsafe { crate::slice::from_raw_parts(lp_cmd_line, lp_size) };
+
+    Args {
+        parsed_args_list: parse_lp_cmd_line(lp_cmd_line)
+            .unwrap_or_else(lazy_current_exe)
+            .into_iter(),
+    }
+}
+
+impl fmt::Debug for Args {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        self.parsed_args_list.as_slice().fmt(f)
+    }
+}
+
+impl Iterator for Args {
+    type Item = OsString;
+
+    fn next(&mut self) -> Option<OsString> {
+        self.parsed_args_list.next()
+    }
+
+    fn size_hint(&self) -> (usize, Option<usize>) {
+        self.parsed_args_list.size_hint()
+    }
+}
+
+impl ExactSizeIterator for Args {
+    fn len(&self) -> usize {
+        self.parsed_args_list.len()
+    }
+}
+
+impl DoubleEndedIterator for Args {
+    fn next_back(&mut self) -> Option<OsString> {
+        self.parsed_args_list.next_back()
+    }
+}
+
+/// Implements the UEFI command-line argument parsing algorithm.
+///
+/// This implementation is based on what is defined in Section 3.4 of
+/// [UEFI Shell Specification](https://uefi.org/sites/default/files/resources/UEFI_Shell_Spec_2_0.pdf)
+///
+/// Return None in the following cases:
+/// - Invalid UTF-16 (unpaired surrogate)
+/// - Empty/improper arguments
+fn parse_lp_cmd_line(code_units: &[u16]) -> Option<Vec<OsString>> {
+    const QUOTE: char = '"';
+    const SPACE: char = ' ';
+    const CARET: char = '^';
+    const NULL: char = '\0';
+
+    let mut ret_val = Vec::new();
+    let mut code_units_iter = char::decode_utf16(code_units.iter().cloned()).peekable();
+
+    // The executable name at the beginning is special.
+    let mut in_quotes = false;
+    let mut cur = String::new();
+    while let Some(w) = code_units_iter.next() {
+        let w = w.ok()?;
+        match w {
+            // break on NULL
+            NULL => break,
+            // A quote mark always toggles `in_quotes` no matter what because
+            // there are no escape characters when parsing the executable name.
+            QUOTE => in_quotes = !in_quotes,
+            // If not `in_quotes` then whitespace ends argv[0].
+            SPACE if !in_quotes => break,
+            // In all other cases the code unit is taken literally.
+            _ => cur.push(w),
+        }
+    }
+
+    // If exe name is missing, the cli args are invalid
+    if cur.is_empty() {
+        return None;
+    }
+
+    ret_val.push(OsString::from(cur));
+    // Skip whitespace.
+    while code_units_iter.next_if_eq(&Ok(SPACE)).is_some() {}
+
+    // Parse the arguments according to these rules:
+    // * All code units are taken literally except space, quote and caret.
+    // * When not `in_quotes`, space separate arguments. Consecutive spaces are
+    // treated as a single separator.
+    // * A space `in_quotes` is taken literally.
+    // * A quote toggles `in_quotes` mode unless it's escaped. An escaped quote is taken literally.
+    // * A quote can be escaped if preceded by caret.
+    // * A caret can be escaped if preceded by caret.
+    let mut cur = String::new();
+    let mut in_quotes = false;
+    while let Some(w) = code_units_iter.next() {
+        let w = w.ok()?;
+        match w {
+            // break on NULL
+            NULL => break,
+            // If not `in_quotes`, a space or tab ends the argument.
+            SPACE if !in_quotes => {
+                ret_val.push(OsString::from(&cur[..]));
+                cur.truncate(0);
+
+                // Skip whitespace.
+                while code_units_iter.next_if_eq(&Ok(SPACE)).is_some() {}
+            }
+            // Caret can escape quotes or carets
+            CARET if in_quotes => {
+                if let Some(x) = code_units_iter.next() {
+                    cur.push(x.ok()?);
+                }
+            }
+            // If quote then flip `in_quotes`
+            QUOTE => in_quotes = !in_quotes,
+            // Everything else is always taken literally.
+            _ => cur.push(w),
+        }
+    }
+    // Push the final argument, if any.
+    if !cur.is_empty() || in_quotes {
+        ret_val.push(OsString::from(cur));
+    }
+    Some(ret_val)
+}

library/std/src/sys/uefi/helpers.rs

@@ -139,3 +139,10 @@ pub(crate) unsafe fn close_event(evt: NonNull<crate::ffi::c_void>) -> io::Result
 
     if r.is_error() { Err(crate::io::Error::from_raw_os_error(r.as_usize())) } else { Ok(()) }
 }
+
+/// Get the Protocol for current system handle.
+/// Note: Some protocols need to be manually freed. It is the callers responsibility to do so.
+pub(crate) fn image_handle_protocol<T>(protocol_guid: Guid) -> Option<NonNull<T>> {
+    let system_handle = uefi::env::try_image_handle()?;
+    open_protocol(system_handle, protocol_guid).ok()
+}

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

@@ -13,7 +13,6 @@
 //! [`OsString`]: crate::ffi::OsString
 
 pub mod alloc;
-#[path = "../unsupported/args.rs"]
 pub mod args;
 #[path = "../unix/cmath.rs"]
 pub mod cmath;

src/doc/rustc/src/platform-support/unknown-uefi.md

@@ -268,6 +268,8 @@ cargo build --target x86_64-unknown-uefi -Zbuild-std=std,panic_abort
 #### stdio
 - Uses `Simple Text Input Protocol` and `Simple Text Output Protocol`.
 - Note: UEFI uses CRLF for new line. This means Enter key is registered as CR instead of LF.
+#### args
+- Uses `EFI_LOADED_IMAGE_PROTOCOL->LoadOptions`
 
 ## Example: Hello World With std
 The following code features a valid UEFI application, including `stdio` and `alloc` (`OsString` and `Vec`):