From 3a4dd1b7b279435b4289b9b9a45db6647f1d3f9c Mon Sep 17 00:00:00 2001 From: joboet Date: Sat, 29 Mar 2025 13:08:18 +0100 Subject: std: make `cmath` functions safe --- library/std/src/sys/cmath.rs | 118 +++++++++++++++++++++---------------------- 1 file changed, 59 insertions(+), 59 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/cmath.rs b/library/std/src/sys/cmath.rs index c9969b4e376..668fd928534 100644 --- a/library/std/src/sys/cmath.rs +++ b/library/std/src/sys/cmath.rs @@ -3,70 +3,70 @@ // These symbols are all defined by `libm`, // or by `compiler-builtins` on unsupported platforms. unsafe extern "C" { - pub fn acos(n: f64) -> f64; - pub fn asin(n: f64) -> f64; - pub fn atan(n: f64) -> f64; - pub fn atan2(a: f64, b: f64) -> f64; - pub fn cbrt(n: f64) -> f64; - pub fn cbrtf(n: f32) -> f32; - pub fn cosh(n: f64) -> f64; - pub fn expm1(n: f64) -> f64; - pub fn expm1f(n: f32) -> f32; - pub fn fdim(a: f64, b: f64) -> f64; - pub fn fdimf(a: f32, b: f32) -> f32; + pub safe fn acos(n: f64) -> f64; + pub safe fn asin(n: f64) -> f64; + pub safe fn atan(n: f64) -> f64; + pub safe fn atan2(a: f64, b: f64) -> f64; + pub safe fn cbrt(n: f64) -> f64; + pub safe fn cbrtf(n: f32) -> f32; + pub safe fn cosh(n: f64) -> f64; + pub safe fn expm1(n: f64) -> f64; + pub safe fn expm1f(n: f32) -> f32; + pub safe fn fdim(a: f64, b: f64) -> f64; + pub safe fn fdimf(a: f32, b: f32) -> f32; #[cfg_attr(target_env = "msvc", link_name = "_hypot")] - pub fn hypot(x: f64, y: f64) -> f64; + pub safe fn hypot(x: f64, y: f64) -> f64; #[cfg_attr(target_env = "msvc", link_name = "_hypotf")] - pub fn hypotf(x: f32, y: f32) -> f32; - pub fn log1p(n: f64) -> f64; - pub fn log1pf(n: f32) -> f32; - pub fn sinh(n: f64) -> f64; - pub fn tan(n: f64) -> f64; - pub fn tanh(n: f64) -> f64; - pub fn tgamma(n: f64) -> f64; - pub fn tgammaf(n: f32) -> f32; - pub fn lgamma_r(n: f64, s: &mut i32) -> f64; + pub safe fn hypotf(x: f32, y: f32) -> f32; + pub safe fn log1p(n: f64) -> f64; + pub safe fn log1pf(n: f32) -> f32; + pub safe fn sinh(n: f64) -> f64; + pub safe fn tan(n: f64) -> f64; + pub safe fn tanh(n: f64) -> f64; + pub safe fn tgamma(n: f64) -> f64; + pub safe fn tgammaf(n: f32) -> f32; + pub safe fn lgamma_r(n: f64, s: &mut i32) -> f64; #[cfg(not(target_os = "aix"))] - pub fn lgammaf_r(n: f32, s: &mut i32) -> f32; - pub fn erf(n: f64) -> f64; - pub fn erff(n: f32) -> f32; - pub fn erfc(n: f64) -> f64; - pub fn erfcf(n: f32) -> f32; + pub safe fn lgammaf_r(n: f32, s: &mut i32) -> f32; + pub safe fn erf(n: f64) -> f64; + pub safe fn erff(n: f32) -> f32; + pub safe fn erfc(n: f64) -> f64; + pub safe fn erfcf(n: f32) -> f32; - pub fn acosf128(n: f128) -> f128; - pub fn asinf128(n: f128) -> f128; - pub fn atanf128(n: f128) -> f128; - pub fn atan2f128(a: f128, b: f128) -> f128; - pub fn cbrtf128(n: f128) -> f128; - pub fn coshf128(n: f128) -> f128; - pub fn expm1f128(n: f128) -> f128; - pub fn hypotf128(x: f128, y: f128) -> f128; - pub fn log1pf128(n: f128) -> f128; - pub fn sinhf128(n: f128) -> f128; - pub fn tanf128(n: f128) -> f128; - pub fn tanhf128(n: f128) -> f128; - pub fn tgammaf128(n: f128) -> f128; - pub fn lgammaf128_r(n: f128, s: &mut i32) -> f128; - pub fn erff128(n: f128) -> f128; - pub fn erfcf128(n: f128) -> f128; + pub safe fn acosf128(n: f128) -> f128; + pub safe fn asinf128(n: f128) -> f128; + pub safe fn atanf128(n: f128) -> f128; + pub safe fn atan2f128(a: f128, b: f128) -> f128; + pub safe fn cbrtf128(n: f128) -> f128; + pub safe fn coshf128(n: f128) -> f128; + pub safe fn expm1f128(n: f128) -> f128; + pub safe fn hypotf128(x: f128, y: f128) -> f128; + pub safe fn log1pf128(n: f128) -> f128; + pub safe fn sinhf128(n: f128) -> f128; + pub safe fn tanf128(n: f128) -> f128; + pub safe fn tanhf128(n: f128) -> f128; + pub safe fn tgammaf128(n: f128) -> f128; + pub safe fn lgammaf128_r(n: f128, s: &mut i32) -> f128; + pub safe fn erff128(n: f128) -> f128; + pub safe fn erfcf128(n: f128) -> f128; cfg_if::cfg_if! { if #[cfg(not(all(target_os = "windows", target_env = "msvc", target_arch = "x86")))] { - pub fn acosf(n: f32) -> f32; - pub fn asinf(n: f32) -> f32; - pub fn atan2f(a: f32, b: f32) -> f32; - pub fn atanf(n: f32) -> f32; - pub fn coshf(n: f32) -> f32; - pub fn sinhf(n: f32) -> f32; - pub fn tanf(n: f32) -> f32; - pub fn tanhf(n: f32) -> f32; + pub safe fn acosf(n: f32) -> f32; + pub safe fn asinf(n: f32) -> f32; + pub safe fn atan2f(a: f32, b: f32) -> f32; + pub safe fn atanf(n: f32) -> f32; + pub safe fn coshf(n: f32) -> f32; + pub safe fn sinhf(n: f32) -> f32; + pub safe fn tanf(n: f32) -> f32; + pub safe fn tanhf(n: f32) -> f32; }} } // On AIX, we don't have lgammaf_r only the f64 version, so we can // use the f64 version lgamma_r #[cfg(target_os = "aix")] -pub unsafe fn lgammaf_r(n: f32, s: &mut i32) -> f32 { +pub fn lgammaf_r(n: f32, s: &mut i32) -> f32 { lgamma_r(n.into(), s) as f32 } @@ -76,42 +76,42 @@ pub unsafe fn lgammaf_r(n: f32, s: &mut i32) -> f32 { cfg_if::cfg_if! { if #[cfg(all(target_os = "windows", target_env = "msvc", target_arch = "x86"))] { #[inline] - pub unsafe fn acosf(n: f32) -> f32 { + pub fn acosf(n: f32) -> f32 { f64::acos(n as f64) as f32 } #[inline] - pub unsafe fn asinf(n: f32) -> f32 { + pub fn asinf(n: f32) -> f32 { f64::asin(n as f64) as f32 } #[inline] - pub unsafe fn atan2f(n: f32, b: f32) -> f32 { + pub fn atan2f(n: f32, b: f32) -> f32 { f64::atan2(n as f64, b as f64) as f32 } #[inline] - pub unsafe fn atanf(n: f32) -> f32 { + pub fn atanf(n: f32) -> f32 { f64::atan(n as f64) as f32 } #[inline] - pub unsafe fn coshf(n: f32) -> f32 { + pub fn coshf(n: f32) -> f32 { f64::cosh(n as f64) as f32 } #[inline] - pub unsafe fn sinhf(n: f32) -> f32 { + pub fn sinhf(n: f32) -> f32 { f64::sinh(n as f64) as f32 } #[inline] - pub unsafe fn tanf(n: f32) -> f32 { + pub fn tanf(n: f32) -> f32 { f64::tan(n as f64) as f32 } #[inline] - pub unsafe fn tanhf(n: f32) -> f32 { + pub fn tanhf(n: f32) -> f32 { f64::tanh(n as f64) as f32 } }} -- cgit 1.4.1-3-g733a5 From 8b12153e3557318900675a0ab3fe1581ecf9d9e2 Mon Sep 17 00:00:00 2001 From: Chris Denton Date: Fri, 4 Apr 2025 16:41:40 +0000 Subject: Update windows-bindgen to 0.61.0 --- Cargo.lock | 6 +- library/std/src/sys/fs/windows/remove_dir_all.rs | 2 +- library/std/src/sys/pal/windows/c.rs | 4 +- library/std/src/sys/pal/windows/c/bindings.txt | 4 +- library/std/src/sys/pal/windows/c/windows_sys.rs | 303 +++++++++++++++++++++-- src/tools/generate-windows-sys/Cargo.toml | 2 +- 6 files changed, 290 insertions(+), 31 deletions(-) (limited to 'library/std/src/sys') diff --git a/Cargo.lock b/Cargo.lock index 96526f7e9e7..2ba12ea8030 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -6066,11 +6066,13 @@ dependencies = [ [[package]] name = "windows-bindgen" -version = "0.59.0" +version = "0.61.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9b7fb600834d7e868f6e5bb748a86101427330fafbf9485c331b9d5f562d54a5" +checksum = "ac1c59c20569610dd9ed784d5f003fb493ec57b4cf39d974eb03a84bb7156c90" dependencies = [ "rayon", + "serde", + "serde_json", ] [[package]] diff --git a/library/std/src/sys/fs/windows/remove_dir_all.rs b/library/std/src/sys/fs/windows/remove_dir_all.rs index f51eced8416..b213c49bcb0 100644 --- a/library/std/src/sys/fs/windows/remove_dir_all.rs +++ b/library/std/src/sys/fs/windows/remove_dir_all.rs @@ -95,7 +95,7 @@ fn open_link_no_reparse( ObjectName: &mut path_str, RootDirectory: parent.as_raw_handle(), Attributes: ATTRIBUTES.load(Ordering::Relaxed), - ..c::OBJECT_ATTRIBUTES::default() + ..c::OBJECT_ATTRIBUTES::with_length() }; let share = c::FILE_SHARE_DELETE | c::FILE_SHARE_READ | c::FILE_SHARE_WRITE; let options = c::FILE_OPEN_REPARSE_POINT | options; diff --git a/library/std/src/sys/pal/windows/c.rs b/library/std/src/sys/pal/windows/c.rs index 004cbee52f6..ac1c5e9932e 100644 --- a/library/std/src/sys/pal/windows/c.rs +++ b/library/std/src/sys/pal/windows/c.rs @@ -44,8 +44,8 @@ impl UNICODE_STRING { } } -impl Default for OBJECT_ATTRIBUTES { - fn default() -> Self { +impl OBJECT_ATTRIBUTES { + pub fn with_length() -> Self { Self { Length: size_of::() as _, RootDirectory: ptr::null_mut(), diff --git a/library/std/src/sys/pal/windows/c/bindings.txt b/library/std/src/sys/pal/windows/c/bindings.txt index e2c21633279..d5fbb453c6f 100644 --- a/library/std/src/sys/pal/windows/c/bindings.txt +++ b/library/std/src/sys/pal/windows/c/bindings.txt @@ -1,7 +1,8 @@ --out windows_sys.rs --flat --sys ---no-core +--no-deps +--link windows_targets --filter !INVALID_HANDLE_VALUE ABOVE_NORMAL_PRIORITY_CLASS @@ -19,7 +20,6 @@ ALL_PROCESSOR_GROUPS ARM64_NT_NEON128 BELOW_NORMAL_PRIORITY_CLASS bind -BOOL BY_HANDLE_FILE_INFORMATION CALLBACK_CHUNK_FINISHED CALLBACK_STREAM_SWITCH diff --git a/library/std/src/sys/pal/windows/c/windows_sys.rs b/library/std/src/sys/pal/windows/c/windows_sys.rs index 1d0e89f5d0f..eb2914b8644 100644 --- a/library/std/src/sys/pal/windows/c/windows_sys.rs +++ b/library/std/src/sys/pal/windows/c/windows_sys.rs @@ -1,4 +1,4 @@ -// Bindings generated by `windows-bindgen` 0.59.0 +// Bindings generated by `windows-bindgen` 0.61.0 #![allow(non_snake_case, non_upper_case_globals, non_camel_case_types, dead_code, clippy::all)] @@ -141,7 +141,7 @@ windows_targets::link!("ws2_32.dll" "system" fn setsockopt(s : SOCKET, level : i windows_targets::link!("ws2_32.dll" "system" fn shutdown(s : SOCKET, how : WINSOCK_SHUTDOWN_HOW) -> i32); pub const ABOVE_NORMAL_PRIORITY_CLASS: PROCESS_CREATION_FLAGS = 32768u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct ACL { pub AclRevision: u8, pub Sbz1: u8, @@ -162,6 +162,11 @@ pub struct ADDRINFOA { pub ai_addr: *mut SOCKADDR, pub ai_next: *mut ADDRINFOA, } +impl Default for ADDRINFOA { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const AF_INET: ADDRESS_FAMILY = 2u16; pub const AF_INET6: ADDRESS_FAMILY = 23u16; pub const AF_UNIX: u16 = 1u16; @@ -176,8 +181,13 @@ pub union ARM64_NT_NEON128 { pub H: [u16; 8], pub B: [u8; 16], } +impl Default for ARM64_NT_NEON128 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct ARM64_NT_NEON128_0 { pub Low: u64, pub High: i64, @@ -185,7 +195,7 @@ pub struct ARM64_NT_NEON128_0 { pub const BELOW_NORMAL_PRIORITY_CLASS: PROCESS_CREATION_FLAGS = 16384u32; pub type BOOL = i32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct BY_HANDLE_FILE_INFORMATION { pub dwFileAttributes: u32, pub ftCreationTime: FILETIME, @@ -206,9 +216,14 @@ pub type COMPARESTRING_RESULT = i32; pub struct CONDITION_VARIABLE { pub Ptr: *mut core::ffi::c_void, } +impl Default for CONDITION_VARIABLE { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type CONSOLE_MODE = u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct CONSOLE_READCONSOLE_CONTROL { pub nLength: u32, pub nInitialChars: u32, @@ -245,6 +260,12 @@ pub struct CONTEXT { pub SegSs: u32, pub ExtendedRegisters: [u8; 512], } +#[cfg(target_arch = "x86")] +impl Default for CONTEXT { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(any(target_arch = "arm64ec", target_arch = "x86_64"))] #[derive(Clone, Copy)] @@ -296,6 +317,12 @@ pub struct CONTEXT { pub LastExceptionToRip: u64, pub LastExceptionFromRip: u64, } +#[cfg(any(target_arch = "arm64ec", target_arch = "x86_64"))] +impl Default for CONTEXT { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(any(target_arch = "arm64ec", target_arch = "x86_64"))] #[derive(Clone, Copy)] @@ -303,6 +330,12 @@ pub union CONTEXT_0 { pub FltSave: XSAVE_FORMAT, pub Anonymous: CONTEXT_0_0, } +#[cfg(any(target_arch = "arm64ec", target_arch = "x86_64"))] +impl Default for CONTEXT_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(any(target_arch = "arm64ec", target_arch = "x86_64"))] #[derive(Clone, Copy)] @@ -326,6 +359,12 @@ pub struct CONTEXT_0_0 { pub Xmm14: M128A, pub Xmm15: M128A, } +#[cfg(any(target_arch = "arm64ec", target_arch = "x86_64"))] +impl Default for CONTEXT_0_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(target_arch = "aarch64")] #[derive(Clone, Copy)] @@ -343,6 +382,12 @@ pub struct CONTEXT { pub Wcr: [u32; 2], pub Wvr: [u64; 2], } +#[cfg(target_arch = "aarch64")] +impl Default for CONTEXT { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(target_arch = "aarch64")] #[derive(Clone, Copy)] @@ -350,9 +395,15 @@ pub union CONTEXT_0 { pub Anonymous: CONTEXT_0_0, pub X: [u64; 31], } +#[cfg(target_arch = "aarch64")] +impl Default for CONTEXT_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(target_arch = "aarch64")] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct CONTEXT_0_0 { pub X0: u64, pub X1: u64, @@ -2305,6 +2356,11 @@ pub struct EXCEPTION_POINTERS { pub ExceptionRecord: *mut EXCEPTION_RECORD, pub ContextRecord: *mut CONTEXT, } +impl Default for EXCEPTION_POINTERS { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub struct EXCEPTION_RECORD { @@ -2315,6 +2371,11 @@ pub struct EXCEPTION_RECORD { pub NumberParameters: u32, pub ExceptionInformation: [usize; 15], } +impl Default for EXCEPTION_RECORD { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const EXCEPTION_STACK_OVERFLOW: NTSTATUS = 0xC00000FD_u32 as _; pub const EXTENDED_STARTUPINFO_PRESENT: PROCESS_CREATION_FLAGS = 524288u32; pub const E_NOTIMPL: HRESULT = 0x80004001_u32 as _; @@ -2333,8 +2394,13 @@ pub struct FD_SET { pub fd_count: u32, pub fd_array: [SOCKET; 64], } +impl Default for FD_SET { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILETIME { pub dwLowDateTime: u32, pub dwHighDateTime: u32, @@ -2343,7 +2409,7 @@ pub type FILE_ACCESS_RIGHTS = u32; pub const FILE_ADD_FILE: FILE_ACCESS_RIGHTS = 2u32; pub const FILE_ADD_SUBDIRECTORY: FILE_ACCESS_RIGHTS = 4u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_ALLOCATION_INFO { pub AllocationSize: i64, } @@ -2369,7 +2435,7 @@ pub const FILE_ATTRIBUTE_REPARSE_POINT: FILE_FLAGS_AND_ATTRIBUTES = 1024u32; pub const FILE_ATTRIBUTE_SPARSE_FILE: FILE_FLAGS_AND_ATTRIBUTES = 512u32; pub const FILE_ATTRIBUTE_SYSTEM: FILE_FLAGS_AND_ATTRIBUTES = 4u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_ATTRIBUTE_TAG_INFO { pub FileAttributes: u32, pub ReparseTag: u32, @@ -2378,7 +2444,7 @@ pub const FILE_ATTRIBUTE_TEMPORARY: FILE_FLAGS_AND_ATTRIBUTES = 256u32; pub const FILE_ATTRIBUTE_UNPINNED: FILE_FLAGS_AND_ATTRIBUTES = 1048576u32; pub const FILE_ATTRIBUTE_VIRTUAL: FILE_FLAGS_AND_ATTRIBUTES = 65536u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_BASIC_INFO { pub CreationTime: i64, pub LastAccessTime: i64, @@ -2405,19 +2471,19 @@ pub const FILE_DISPOSITION_FLAG_IGNORE_READONLY_ATTRIBUTE: FILE_DISPOSITION_INFO pub const FILE_DISPOSITION_FLAG_ON_CLOSE: FILE_DISPOSITION_INFO_EX_FLAGS = 8u32; pub const FILE_DISPOSITION_FLAG_POSIX_SEMANTICS: FILE_DISPOSITION_INFO_EX_FLAGS = 2u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_DISPOSITION_INFO { pub DeleteFile: bool, } #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_DISPOSITION_INFO_EX { pub Flags: FILE_DISPOSITION_INFO_EX_FLAGS, } pub type FILE_DISPOSITION_INFO_EX_FLAGS = u32; pub const FILE_END: SET_FILE_POINTER_MOVE_METHOD = 2u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_END_OF_FILE_INFO { pub EndOfFile: i64, } @@ -2457,9 +2523,14 @@ pub struct FILE_ID_BOTH_DIR_INFO { pub FileId: i64, pub FileName: [u16; 1], } +impl Default for FILE_ID_BOTH_DIR_INFO { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type FILE_INFO_BY_HANDLE_CLASS = i32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_IO_PRIORITY_HINT_INFO { pub PriorityHint: PRIORITY_HINT, } @@ -2494,12 +2565,22 @@ pub struct FILE_RENAME_INFO { pub FileNameLength: u32, pub FileName: [u16; 1], } +impl Default for FILE_RENAME_INFO { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub union FILE_RENAME_INFO_0 { pub ReplaceIfExists: bool, pub Flags: u32, } +impl Default for FILE_RENAME_INFO_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const FILE_RESERVE_OPFILTER: NTCREATEFILE_CREATE_OPTIONS = 1048576u32; pub const FILE_SEQUENTIAL_ONLY: NTCREATEFILE_CREATE_OPTIONS = 4u32; pub const FILE_SESSION_AWARE: NTCREATEFILE_CREATE_OPTIONS = 262144u32; @@ -2509,7 +2590,7 @@ pub const FILE_SHARE_NONE: FILE_SHARE_MODE = 0u32; pub const FILE_SHARE_READ: FILE_SHARE_MODE = 1u32; pub const FILE_SHARE_WRITE: FILE_SHARE_MODE = 2u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct FILE_STANDARD_INFO { pub AllocationSize: i64, pub EndOfFile: i64, @@ -2549,6 +2630,12 @@ pub struct FLOATING_SAVE_AREA { pub RegisterArea: [u8; 80], pub Spare0: u32, } +#[cfg(target_arch = "x86")] +impl Default for FLOATING_SAVE_AREA { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(any(target_arch = "aarch64", target_arch = "arm64ec", target_arch = "x86_64"))] #[derive(Clone, Copy)] @@ -2563,6 +2650,12 @@ pub struct FLOATING_SAVE_AREA { pub RegisterArea: [u8; 80], pub Cr0NpxState: u32, } +#[cfg(any(target_arch = "aarch64", target_arch = "arm64ec", target_arch = "x86_64"))] +impl Default for FLOATING_SAVE_AREA { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const FORMAT_MESSAGE_ALLOCATE_BUFFER: FORMAT_MESSAGE_OPTIONS = 256u32; pub const FORMAT_MESSAGE_ARGUMENT_ARRAY: FORMAT_MESSAGE_OPTIONS = 8192u32; pub const FORMAT_MESSAGE_FROM_HMODULE: FORMAT_MESSAGE_OPTIONS = 2048u32; @@ -2639,12 +2732,22 @@ pub const IDLE_PRIORITY_CLASS: PROCESS_CREATION_FLAGS = 64u32; pub struct IN6_ADDR { pub u: IN6_ADDR_0, } +impl Default for IN6_ADDR { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub union IN6_ADDR_0 { pub Byte: [u8; 16], pub Word: [u16; 8], } +impl Default for IN6_ADDR_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const INFINITE: u32 = 4294967295u32; pub const INHERIT_CALLER_PRIORITY: PROCESS_CREATION_FLAGS = 131072u32; pub const INHERIT_PARENT_AFFINITY: PROCESS_CREATION_FLAGS = 65536u32; @@ -2653,6 +2756,11 @@ pub const INHERIT_PARENT_AFFINITY: PROCESS_CREATION_FLAGS = 65536u32; pub union INIT_ONCE { pub Ptr: *mut core::ffi::c_void, } +impl Default for INIT_ONCE { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const INIT_ONCE_INIT_FAILED: u32 = 4u32; pub const INVALID_FILE_ATTRIBUTES: u32 = 4294967295u32; pub const INVALID_SOCKET: SOCKET = -1i32 as _; @@ -2661,6 +2769,11 @@ pub const INVALID_SOCKET: SOCKET = -1i32 as _; pub struct IN_ADDR { pub S_un: IN_ADDR_0, } +impl Default for IN_ADDR { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub union IN_ADDR_0 { @@ -2668,8 +2781,13 @@ pub union IN_ADDR_0 { pub S_un_w: IN_ADDR_0_1, pub S_addr: u32, } +impl Default for IN_ADDR_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct IN_ADDR_0_0 { pub s_b1: u8, pub s_b2: u8, @@ -2677,7 +2795,7 @@ pub struct IN_ADDR_0_0 { pub s_b4: u8, } #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct IN_ADDR_0_1 { pub s_w1: u16, pub s_w2: u16, @@ -2690,12 +2808,22 @@ pub struct IO_STATUS_BLOCK { pub Anonymous: IO_STATUS_BLOCK_0, pub Information: usize, } +impl Default for IO_STATUS_BLOCK { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub union IO_STATUS_BLOCK_0 { pub Status: NTSTATUS, pub Pointer: *mut core::ffi::c_void, } +impl Default for IO_STATUS_BLOCK_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type IPPROTO = i32; pub const IPPROTO_AH: IPPROTO = 51i32; pub const IPPROTO_CBT: IPPROTO = 7i32; @@ -2742,6 +2870,11 @@ pub struct IPV6_MREQ { pub ipv6mr_multiaddr: IN6_ADDR, pub ipv6mr_interface: u32, } +impl Default for IPV6_MREQ { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const IPV6_MULTICAST_LOOP: i32 = 11i32; pub const IPV6_V6ONLY: i32 = 27i32; pub const IP_ADD_MEMBERSHIP: i32 = 12i32; @@ -2752,11 +2885,16 @@ pub struct IP_MREQ { pub imr_multiaddr: IN_ADDR, pub imr_interface: IN_ADDR, } +impl Default for IP_MREQ { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const IP_MULTICAST_LOOP: i32 = 11i32; pub const IP_MULTICAST_TTL: i32 = 10i32; pub const IP_TTL: i32 = 4i32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct LINGER { pub l_onoff: u16, pub l_linger: u16, @@ -2797,7 +2935,7 @@ pub type LPWSAOVERLAPPED_COMPLETION_ROUTINE = Option< ), >; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct M128A { pub Low: u64, pub High: i64, @@ -2838,6 +2976,11 @@ pub struct OBJECT_ATTRIBUTES { pub SecurityDescriptor: *const SECURITY_DESCRIPTOR, pub SecurityQualityOfService: *const SECURITY_QUALITY_OF_SERVICE, } +impl Default for OBJECT_ATTRIBUTES { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type OBJECT_ATTRIBUTE_FLAGS = u32; pub const OBJ_DONT_REPARSE: OBJECT_ATTRIBUTE_FLAGS = 4096u32; pub const OPEN_ALWAYS: FILE_CREATION_DISPOSITION = 4u32; @@ -2850,14 +2993,24 @@ pub struct OVERLAPPED { pub Anonymous: OVERLAPPED_0, pub hEvent: HANDLE, } +impl Default for OVERLAPPED { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub union OVERLAPPED_0 { pub Anonymous: OVERLAPPED_0_0, pub Pointer: *mut core::ffi::c_void, } +impl Default for OVERLAPPED_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct OVERLAPPED_0_0 { pub Offset: u32, pub OffsetHigh: u32, @@ -2895,6 +3048,11 @@ pub struct PROCESS_INFORMATION { pub dwProcessId: u32, pub dwThreadId: u32, } +impl Default for PROCESS_INFORMATION { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const PROCESS_MODE_BACKGROUND_BEGIN: PROCESS_CREATION_FLAGS = 1048576u32; pub const PROCESS_MODE_BACKGROUND_END: PROCESS_CREATION_FLAGS = 2097152u32; pub const PROFILE_KERNEL: PROCESS_CREATION_FLAGS = 536870912u32; @@ -2926,6 +3084,11 @@ pub struct SECURITY_ATTRIBUTES { pub lpSecurityDescriptor: *mut core::ffi::c_void, pub bInheritHandle: BOOL, } +impl Default for SECURITY_ATTRIBUTES { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const SECURITY_CONTEXT_TRACKING: FILE_FLAGS_AND_ATTRIBUTES = 262144u32; pub const SECURITY_DELEGATION: FILE_FLAGS_AND_ATTRIBUTES = 196608u32; #[repr(C)] @@ -2939,13 +3102,18 @@ pub struct SECURITY_DESCRIPTOR { pub Sacl: *mut ACL, pub Dacl: *mut ACL, } +impl Default for SECURITY_DESCRIPTOR { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type SECURITY_DESCRIPTOR_CONTROL = u16; pub const SECURITY_EFFECTIVE_ONLY: FILE_FLAGS_AND_ATTRIBUTES = 524288u32; pub const SECURITY_IDENTIFICATION: FILE_FLAGS_AND_ATTRIBUTES = 65536u32; pub const SECURITY_IMPERSONATION: FILE_FLAGS_AND_ATTRIBUTES = 131072u32; pub type SECURITY_IMPERSONATION_LEVEL = i32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct SECURITY_QUALITY_OF_SERVICE { pub Length: u32, pub ImpersonationLevel: SECURITY_IMPERSONATION_LEVEL, @@ -2962,6 +3130,11 @@ pub struct SOCKADDR { pub sa_family: ADDRESS_FAMILY, pub sa_data: [i8; 14], } +impl Default for SOCKADDR { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub struct SOCKADDR_STORAGE { @@ -2970,12 +3143,22 @@ pub struct SOCKADDR_STORAGE { pub __ss_align: i64, pub __ss_pad2: [i8; 112], } +impl Default for SOCKADDR_STORAGE { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub struct SOCKADDR_UN { pub sun_family: ADDRESS_FAMILY, pub sun_path: [i8; 108], } +impl Default for SOCKADDR_UN { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type SOCKET = usize; pub const SOCKET_ERROR: i32 = -1i32; pub const SOCK_DGRAM: WINSOCK_SOCKET_TYPE = 2i32; @@ -2995,6 +3178,11 @@ pub const SPECIFIC_RIGHTS_ALL: FILE_ACCESS_RIGHTS = 65535u32; pub struct SRWLOCK { pub Ptr: *mut core::ffi::c_void, } +impl Default for SRWLOCK { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const STACK_SIZE_PARAM_IS_A_RESERVATION: THREAD_CREATION_FLAGS = 65536u32; pub const STANDARD_RIGHTS_ALL: FILE_ACCESS_RIGHTS = 2031616u32; pub const STANDARD_RIGHTS_EXECUTE: FILE_ACCESS_RIGHTS = 131072u32; @@ -3021,6 +3209,11 @@ pub struct STARTUPINFOEXW { pub StartupInfo: STARTUPINFOW, pub lpAttributeList: LPPROC_THREAD_ATTRIBUTE_LIST, } +impl Default for STARTUPINFOEXW { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub struct STARTUPINFOW { @@ -3043,6 +3236,11 @@ pub struct STARTUPINFOW { pub hStdOutput: HANDLE, pub hStdError: HANDLE, } +impl Default for STARTUPINFOW { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type STARTUPINFOW_FLAGS = u32; pub const STATUS_DELETE_PENDING: NTSTATUS = 0xC0000056_u32 as _; pub const STATUS_DIRECTORY_NOT_EMPTY: NTSTATUS = 0xC0000101_u32 as _; @@ -3078,14 +3276,24 @@ pub struct SYSTEM_INFO { pub wProcessorLevel: u16, pub wProcessorRevision: u16, } +impl Default for SYSTEM_INFO { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub union SYSTEM_INFO_0 { pub dwOemId: u32, pub Anonymous: SYSTEM_INFO_0_0, } +impl Default for SYSTEM_INFO_0 { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct SYSTEM_INFO_0_0 { pub wProcessorArchitecture: PROCESSOR_ARCHITECTURE, pub wReserved: u16, @@ -3097,7 +3305,7 @@ pub type THREAD_CREATION_FLAGS = u32; pub const TIMER_ALL_ACCESS: SYNCHRONIZATION_ACCESS_RIGHTS = 2031619u32; pub const TIMER_MODIFY_STATE: SYNCHRONIZATION_ACCESS_RIGHTS = 2u32; #[repr(C)] -#[derive(Clone, Copy)] +#[derive(Clone, Copy, Default)] pub struct TIMEVAL { pub tv_sec: i32, pub tv_usec: i32, @@ -3134,6 +3342,11 @@ pub struct UNICODE_STRING { pub MaximumLength: u16, pub Buffer: PWSTR, } +impl Default for UNICODE_STRING { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const VOLUME_NAME_DOS: GETFINALPATHNAMEBYHANDLE_FLAGS = 0u32; pub const VOLUME_NAME_GUID: GETFINALPATHNAMEBYHANDLE_FLAGS = 1u32; pub const VOLUME_NAME_NONE: GETFINALPATHNAMEBYHANDLE_FLAGS = 4u32; @@ -3160,6 +3373,11 @@ pub struct WIN32_FIND_DATAW { pub cFileName: [u16; 260], pub cAlternateFileName: [u16; 14], } +impl Default for WIN32_FIND_DATAW { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub type WINSOCK_SHUTDOWN_HOW = i32; pub type WINSOCK_SOCKET_TYPE = i32; pub const WRITE_DAC: FILE_ACCESS_RIGHTS = 262144u32; @@ -3171,6 +3389,11 @@ pub struct WSABUF { pub len: u32, pub buf: PSTR, } +impl Default for WSABUF { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(target_arch = "x86")] #[derive(Clone, Copy)] @@ -3183,6 +3406,12 @@ pub struct WSADATA { pub iMaxUdpDg: u16, pub lpVendorInfo: PSTR, } +#[cfg(target_arch = "x86")] +impl Default for WSADATA { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(any(target_arch = "aarch64", target_arch = "arm64ec", target_arch = "x86_64"))] #[derive(Clone, Copy)] @@ -3195,6 +3424,12 @@ pub struct WSADATA { pub szDescription: [i8; 257], pub szSystemStatus: [i8; 129], } +#[cfg(any(target_arch = "aarch64", target_arch = "arm64ec", target_arch = "x86_64"))] +impl Default for WSADATA { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const WSAEACCES: WSA_ERROR = 10013i32; pub const WSAEADDRINUSE: WSA_ERROR = 10048i32; pub const WSAEADDRNOTAVAIL: WSA_ERROR = 10049i32; @@ -3255,6 +3490,11 @@ pub struct WSAPROTOCOLCHAIN { pub ChainLen: i32, pub ChainEntries: [u32; 7], } +impl Default for WSAPROTOCOLCHAIN { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[derive(Clone, Copy)] pub struct WSAPROTOCOL_INFOW { @@ -3279,6 +3519,11 @@ pub struct WSAPROTOCOL_INFOW { pub dwProviderReserved: u32, pub szProtocol: [u16; 256], } +impl Default for WSAPROTOCOL_INFOW { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} pub const WSASERVICE_NOT_FOUND: WSA_ERROR = 10108i32; pub const WSASYSCALLFAILURE: WSA_ERROR = 10107i32; pub const WSASYSNOTREADY: WSA_ERROR = 10091i32; @@ -3348,6 +3593,12 @@ pub struct XSAVE_FORMAT { pub XmmRegisters: [M128A; 8], pub Reserved4: [u8; 224], } +#[cfg(target_arch = "x86")] +impl Default for XSAVE_FORMAT { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[repr(C)] #[cfg(any(target_arch = "aarch64", target_arch = "arm64ec", target_arch = "x86_64"))] #[derive(Clone, Copy)] @@ -3369,6 +3620,12 @@ pub struct XSAVE_FORMAT { pub XmmRegisters: [M128A; 16], pub Reserved4: [u8; 96], } +#[cfg(any(target_arch = "aarch64", target_arch = "arm64ec", target_arch = "x86_64"))] +impl Default for XSAVE_FORMAT { + fn default() -> Self { + unsafe { core::mem::zeroed() } + } +} #[cfg(target_arch = "arm")] #[repr(C)] diff --git a/src/tools/generate-windows-sys/Cargo.toml b/src/tools/generate-windows-sys/Cargo.toml index f5c0e56bb3c..152cd504abd 100644 --- a/src/tools/generate-windows-sys/Cargo.toml +++ b/src/tools/generate-windows-sys/Cargo.toml @@ -4,4 +4,4 @@ version = "0.1.0" edition = "2021" [dependencies.windows-bindgen] -version = "0.59.0" +version = "0.61.0" -- cgit 1.4.1-3-g733a5 From 0acac2cbe306b8db9bc0a7c45f529f1dfa6680c8 Mon Sep 17 00:00:00 2001 From: Thalia Archibald Date: Wed, 26 Mar 2025 00:38:28 -0700 Subject: std: Fix build for NuttX targets Fix std build for all NuttX targets. It is the single largest set of failures on . Although, ESP-IDF also requires these same gates, there are other issues for those targets. This can verified be running `x check library/std --target=` for all NuttX targets. --- library/std/src/sys/fd/unix.rs | 2 ++ library/std/src/sys/net/connection/socket/unix.rs | 1 + 2 files changed, 3 insertions(+) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/fd/unix.rs b/library/std/src/sys/fd/unix.rs index 2042ea2c73d..13fefb4031d 100644 --- a/library/std/src/sys/fd/unix.rs +++ b/library/std/src/sys/fd/unix.rs @@ -71,9 +71,11 @@ const fn max_iov() -> usize { target_os = "android", target_os = "dragonfly", target_os = "emscripten", + target_os = "espidf", target_os = "freebsd", target_os = "linux", target_os = "netbsd", + target_os = "nuttx", target_os = "nto", target_os = "openbsd", target_os = "horizon", diff --git a/library/std/src/sys/net/connection/socket/unix.rs b/library/std/src/sys/net/connection/socket/unix.rs index bbe1e038dcc..b35d5d2aa84 100644 --- a/library/std/src/sys/net/connection/socket/unix.rs +++ b/library/std/src/sys/net/connection/socket/unix.rs @@ -1,5 +1,6 @@ use libc::{MSG_PEEK, c_int, c_void, size_t, sockaddr, socklen_t}; +#[cfg(not(any(target_os = "espidf", target_os = "nuttx")))] use crate::ffi::CStr; use crate::io::{self, BorrowedBuf, BorrowedCursor, IoSlice, IoSliceMut}; use crate::net::{Shutdown, SocketAddr}; -- cgit 1.4.1-3-g733a5 From edfc74722556c659de6fa03b23af3b9c8ceb8ac2 Mon Sep 17 00:00:00 2001 From: Chris Denton Date: Mon, 24 Mar 2025 18:57:07 +0000 Subject: Avoid verbatim paths in Command::current_dir If possible, we should try not to use verbatim paths in Command::current_dir. It might work but it might also break code (including some Windows APIs) that assume the current directory isn't verbatim. --- library/std/src/sys/path/windows.rs | 43 +++++++++++++++++++++++++++++++ library/std/src/sys/path/windows/tests.rs | 12 +++++++++ library/std/src/sys/process/windows.rs | 32 ++++++++++++++++++++--- 3 files changed, 83 insertions(+), 4 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/path/windows.rs b/library/std/src/sys/path/windows.rs index 1c534721916..6547ed9aa5f 100644 --- a/library/std/src/sys/path/windows.rs +++ b/library/std/src/sys/path/windows.rs @@ -350,3 +350,46 @@ pub(crate) fn absolute(path: &Path) -> io::Result { pub(crate) fn is_absolute(path: &Path) -> bool { path.has_root() && path.prefix().is_some() } + +/// Test that the path is absolute, fully qualified and unchanged when processed by the Windows API. +/// +/// For example: +/// +/// - `C:\path\to\file` will return true. +/// - `C:\path\to\nul` returns false because the Windows API will convert it to \\.\NUL +/// - `C:\path\to\..\file` returns false because it will be resolved to `C:\path\file`. +/// +/// This is a useful property because it means the path can be converted from and to and verbatim +/// path just by changing the prefix. +pub(crate) fn is_absolute_exact(path: &[u16]) -> bool { + // This is implemented by checking that passing the path through + // GetFullPathNameW does not change the path in any way. + + // Windows paths are limited to i16::MAX length + // though the API here accepts a u32 for the length. + if path.is_empty() || path.len() > u32::MAX as usize || path.last() != Some(&0) { + return false; + } + // The path returned by `GetFullPathNameW` must be the same length as the + // given path, otherwise they're not equal. + let buffer_len = path.len(); + let mut new_path = Vec::with_capacity(buffer_len); + let result = unsafe { + c::GetFullPathNameW( + path.as_ptr(), + new_path.capacity() as u32, + new_path.as_mut_ptr(), + crate::ptr::null_mut(), + ) + }; + // Note: if non-zero, the returned result is the length of the buffer without the null termination + if result == 0 || result as usize != buffer_len - 1 { + false + } else { + // SAFETY: `GetFullPathNameW` initialized `result` bytes and does not exceed `nBufferLength - 1` (capacity). + unsafe { + new_path.set_len((result as usize) + 1); + } + path == &new_path + } +} diff --git a/library/std/src/sys/path/windows/tests.rs b/library/std/src/sys/path/windows/tests.rs index f2a60e30bc6..9eb79203dca 100644 --- a/library/std/src/sys/path/windows/tests.rs +++ b/library/std/src/sys/path/windows/tests.rs @@ -135,3 +135,15 @@ fn broken_unc_path() { assert_eq!(components.next(), Some(Component::Normal("foo".as_ref()))); assert_eq!(components.next(), Some(Component::Normal("bar".as_ref()))); } + +#[test] +fn test_is_absolute_exact() { + use crate::sys::pal::api::wide_str; + // These paths can be made verbatim by only changing their prefix. + assert!(is_absolute_exact(wide_str!(r"C:\path\to\file"))); + assert!(is_absolute_exact(wide_str!(r"\\server\share\path\to\file"))); + // These paths change more substantially + assert!(!is_absolute_exact(wide_str!(r"C:\path\to\..\file"))); + assert!(!is_absolute_exact(wide_str!(r"\\server\share\path\to\..\file"))); + assert!(!is_absolute_exact(wide_str!(r"C:\path\to\NUL"))); // Converts to \\.\NUL +} diff --git a/library/std/src/sys/process/windows.rs b/library/std/src/sys/process/windows.rs index 06c15e08f3f..4cfdf908c58 100644 --- a/library/std/src/sys/process/windows.rs +++ b/library/std/src/sys/process/windows.rs @@ -19,7 +19,7 @@ use crate::sys::args::{self, Arg}; use crate::sys::c::{self, EXIT_FAILURE, EXIT_SUCCESS}; use crate::sys::fs::{File, OpenOptions}; use crate::sys::handle::Handle; -use crate::sys::pal::api::{self, WinError}; +use crate::sys::pal::api::{self, WinError, utf16}; use crate::sys::pal::{ensure_no_nuls, fill_utf16_buf}; use crate::sys::pipe::{self, AnonPipe}; use crate::sys::{cvt, path, stdio}; @@ -880,9 +880,33 @@ fn make_envp(maybe_env: Option>) -> io::Result<(*mut fn make_dirp(d: Option<&OsString>) -> io::Result<(*const u16, Vec)> { match d { Some(dir) => { - let mut dir_str: Vec = ensure_no_nuls(dir)?.encode_wide().collect(); - dir_str.push(0); - Ok((dir_str.as_ptr(), dir_str)) + let mut dir_str: Vec = ensure_no_nuls(dir)?.encode_wide().chain([0]).collect(); + // Try to remove the `\\?\` prefix, if any. + // This is necessary because the current directory does not support verbatim paths. + // However. this can only be done if it doesn't change how the path will be resolved. + let ptr = if dir_str.starts_with(utf16!(r"\\?\UNC")) { + // Turn the `C` in `UNC` into a `\` so we can then use `\\rest\of\path`. + let start = r"\\?\UN".len(); + dir_str[start] = b'\\' as u16; + if path::is_absolute_exact(&dir_str[start..]) { + dir_str[start..].as_ptr() + } else { + // Revert the above change. + dir_str[start] = b'C' as u16; + dir_str.as_ptr() + } + } else if dir_str.starts_with(utf16!(r"\\?\")) { + // Strip the leading `\\?\` + let start = r"\\?\".len(); + if path::is_absolute_exact(&dir_str[start..]) { + dir_str[start..].as_ptr() + } else { + dir_str.as_ptr() + } + } else { + dir_str.as_ptr() + }; + Ok((ptr, dir_str)) } None => Ok((ptr::null(), Vec::new())), } -- cgit 1.4.1-3-g733a5 From 068a33245981765c3678434d5bcb71bd63d5598a Mon Sep 17 00:00:00 2001 From: Bastian Kersting Date: Fri, 11 Apr 2025 10:15:55 +0000 Subject: cfi: Remove #[no_sanitize(cfi)] for extern weak functions Previously (https://github.com/rust-lang/rust/pull/115200, https://github.com/rust-lang/rust/pull/138002), we added `#[no_sanitize(cfi)]` to all code paths that call to a weakly linked function. In https://github.com/rust-lang/rust/pull/138349 we fixed the root cause for this issue, which means we can now remove the corresponding attributes. --- library/std/src/sys/fd/unix.rs | 3 --- library/std/src/sys/fs/unix.rs | 14 -------------- library/std/src/sys/pal/unix/thread.rs | 3 --- library/std/src/sys/pal/unix/time.rs | 11 ----------- library/std/src/sys/pal/unix/weak.rs | 3 --- library/std/src/sys/process/unix/unix.rs | 3 --- library/std/src/sys/thread_local/destructors/linux_like.rs | 3 --- 7 files changed, 40 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/fd/unix.rs b/library/std/src/sys/fd/unix.rs index 2042ea2c73d..5765cb1386a 100644 --- a/library/std/src/sys/fd/unix.rs +++ b/library/std/src/sys/fd/unix.rs @@ -257,9 +257,6 @@ impl FileDesc { } #[cfg(all(target_os = "android", target_pointer_width = "32"))] - // FIXME(#115199): Rust currently omits weak function definitions - // and its metadata from LLVM IR. - #[no_sanitize(cfi)] pub fn read_vectored_at(&self, bufs: &mut [IoSliceMut<'_>], offset: u64) -> io::Result { weak!( fn preadv64( diff --git a/library/std/src/sys/fs/unix.rs b/library/std/src/sys/fs/unix.rs index 87865be0387..687fc322e59 100644 --- a/library/std/src/sys/fs/unix.rs +++ b/library/std/src/sys/fs/unix.rs @@ -1463,20 +1463,6 @@ impl File { Ok(()) } - // FIXME(#115199): Rust currently omits weak function definitions - // and its metadata from LLVM IR. - #[cfg_attr( - any( - target_os = "android", - all( - target_os = "linux", - target_env = "gnu", - target_pointer_width = "32", - not(target_arch = "riscv32") - ) - ), - no_sanitize(cfi) - )] pub fn set_times(&self, times: FileTimes) -> io::Result<()> { #[cfg(not(any( target_os = "redox", diff --git a/library/std/src/sys/pal/unix/thread.rs b/library/std/src/sys/pal/unix/thread.rs index 9078dd1c231..c0662600db1 100644 --- a/library/std/src/sys/pal/unix/thread.rs +++ b/library/std/src/sys/pal/unix/thread.rs @@ -189,9 +189,6 @@ impl Thread { } #[cfg(any(target_os = "solaris", target_os = "illumos", target_os = "nto"))] - // FIXME(#115199): Rust currently omits weak function definitions - // and its metadata from LLVM IR. - #[no_sanitize(cfi)] pub fn set_name(name: &CStr) { weak!( fn pthread_setname_np( diff --git a/library/std/src/sys/pal/unix/time.rs b/library/std/src/sys/pal/unix/time.rs index b8469b1681f..0074d767474 100644 --- a/library/std/src/sys/pal/unix/time.rs +++ b/library/std/src/sys/pal/unix/time.rs @@ -96,17 +96,6 @@ impl Timespec { } } - // FIXME(#115199): Rust currently omits weak function definitions - // and its metadata from LLVM IR. - #[cfg_attr( - all( - target_os = "linux", - target_env = "gnu", - target_pointer_width = "32", - not(target_arch = "riscv32") - ), - no_sanitize(cfi) - )] pub fn now(clock: libc::clockid_t) -> Timespec { use crate::mem::MaybeUninit; use crate::sys::cvt; diff --git a/library/std/src/sys/pal/unix/weak.rs b/library/std/src/sys/pal/unix/weak.rs index e4c814fba8c..a034995e652 100644 --- a/library/std/src/sys/pal/unix/weak.rs +++ b/library/std/src/sys/pal/unix/weak.rs @@ -155,9 +155,6 @@ unsafe fn fetch(name: &str) -> *mut libc::c_void { #[cfg(not(any(target_os = "linux", target_os = "android")))] pub(crate) macro syscall { (fn $name:ident($($param:ident : $t:ty),* $(,)?) -> $ret:ty;) => ( - // FIXME(#115199): Rust currently omits weak function definitions - // and its metadata from LLVM IR. - #[no_sanitize(cfi)] unsafe fn $name($($param: $t),*) -> $ret { weak!(fn $name($($param: $t),*) -> $ret;); diff --git a/library/std/src/sys/process/unix/unix.rs b/library/std/src/sys/process/unix/unix.rs index 191a09c8da9..3b04ec50db3 100644 --- a/library/std/src/sys/process/unix/unix.rs +++ b/library/std/src/sys/process/unix/unix.rs @@ -434,9 +434,6 @@ impl Command { target_os = "nto", target_vendor = "apple", ))] - // FIXME(#115199): Rust currently omits weak function definitions - // and its metadata from LLVM IR. - #[cfg_attr(target_os = "linux", no_sanitize(cfi))] fn posix_spawn( &mut self, stdio: &ChildPipes, diff --git a/library/std/src/sys/thread_local/destructors/linux_like.rs b/library/std/src/sys/thread_local/destructors/linux_like.rs index 817941229ee..d7cbaeb89f4 100644 --- a/library/std/src/sys/thread_local/destructors/linux_like.rs +++ b/library/std/src/sys/thread_local/destructors/linux_like.rs @@ -12,9 +12,6 @@ use crate::mem::transmute; -// FIXME: The Rust compiler currently omits weakly function definitions (i.e., -// __cxa_thread_atexit_impl) and its metadata from LLVM IR. -#[no_sanitize(cfi, kcfi)] pub unsafe fn register(t: *mut u8, dtor: unsafe extern "C" fn(*mut u8)) { /// This is necessary because the __cxa_thread_atexit_impl implementation /// std links to by default may be a C or C++ implementation that was not -- cgit 1.4.1-3-g733a5 From b613e9781a773232f81af9f61a8c8ac34b04873a Mon Sep 17 00:00:00 2001 From: Chris Denton Date: Fri, 11 Apr 2025 16:15:34 +0000 Subject: Use with_native_path for Windows Also add a WCStr type --- library/std/src/sys/fs/mod.rs | 30 +++++++++++----- library/std/src/sys/fs/windows.rs | 69 ++++++++++++++++--------------------- library/std/src/sys/path/windows.rs | 34 ++++++++++++++++++ 3 files changed, 86 insertions(+), 47 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/fs/mod.rs b/library/std/src/sys/fs/mod.rs index 3b176d0d16c..4c5e36ce67a 100644 --- a/library/std/src/sys/fs/mod.rs +++ b/library/std/src/sys/fs/mod.rs @@ -20,6 +20,7 @@ cfg_if::cfg_if! { mod windows; use windows as imp; pub use windows::{symlink_inner, junction_point}; + use crate::sys::path::with_native_path; } else if #[cfg(target_os = "hermit")] { mod hermit; use hermit as imp; @@ -39,7 +40,7 @@ cfg_if::cfg_if! { } // FIXME: Replace this with platform-specific path conversion functions. -#[cfg(not(target_family = "unix"))] +#[cfg(not(any(target_family = "unix", target_os = "windows")))] #[inline] pub fn with_native_path(path: &Path, f: &dyn Fn(&Path) -> io::Result) -> io::Result { f(path) @@ -51,7 +52,7 @@ pub use imp::{ }; pub fn read_dir(path: &Path) -> io::Result { - // FIXME: use with_native_path + // FIXME: use with_native_path on all platforms imp::readdir(path) } @@ -68,8 +69,11 @@ pub fn remove_dir(path: &Path) -> io::Result<()> { } pub fn remove_dir_all(path: &Path) -> io::Result<()> { - // FIXME: use with_native_path - imp::remove_dir_all(path) + // FIXME: use with_native_path on all platforms + #[cfg(not(windows))] + return imp::remove_dir_all(path); + #[cfg(windows)] + with_native_path(path, &imp::remove_dir_all) } pub fn read_link(path: &Path) -> io::Result { @@ -77,6 +81,10 @@ pub fn read_link(path: &Path) -> io::Result { } pub fn symlink(original: &Path, link: &Path) -> io::Result<()> { + // FIXME: use with_native_path on all platforms + #[cfg(windows)] + return imp::symlink(original, link); + #[cfg(not(windows))] with_native_path(original, &|original| { with_native_path(link, &|link| imp::symlink(original, link)) }) @@ -105,11 +113,17 @@ pub fn canonicalize(path: &Path) -> io::Result { } pub fn copy(from: &Path, to: &Path) -> io::Result { - // FIXME: use with_native_path - imp::copy(from, to) + // FIXME: use with_native_path on all platforms + #[cfg(not(windows))] + return imp::copy(from, to); + #[cfg(windows)] + with_native_path(from, &|from| with_native_path(to, &|to| imp::copy(from, to))) } pub fn exists(path: &Path) -> io::Result { - // FIXME: use with_native_path - imp::exists(path) + // FIXME: use with_native_path on all platforms + #[cfg(not(windows))] + return imp::exists(path); + #[cfg(windows)] + with_native_path(path, &imp::exists) } diff --git a/library/std/src/sys/fs/windows.rs b/library/std/src/sys/fs/windows.rs index 15727c99683..9215f937567 100644 --- a/library/std/src/sys/fs/windows.rs +++ b/library/std/src/sys/fs/windows.rs @@ -12,7 +12,7 @@ use crate::sync::Arc; use crate::sys::handle::Handle; use crate::sys::pal::api::{self, WinError, set_file_information_by_handle}; use crate::sys::pal::{IoResult, fill_utf16_buf, to_u16s, truncate_utf16_at_nul}; -use crate::sys::path::maybe_verbatim; +use crate::sys::path::{WCStr, maybe_verbatim}; use crate::sys::time::SystemTime; use crate::sys::{Align8, c, cvt}; use crate::sys_common::{AsInner, FromInner, IntoInner}; @@ -298,10 +298,12 @@ impl OpenOptions { impl File { pub fn open(path: &Path, opts: &OpenOptions) -> io::Result { let path = maybe_verbatim(path)?; + // SAFETY: maybe_verbatim returns null-terminated strings + let path = unsafe { WCStr::from_wchars_with_null_unchecked(&path) }; Self::open_native(&path, opts) } - fn open_native(path: &[u16], opts: &OpenOptions) -> io::Result { + fn open_native(path: &WCStr, opts: &OpenOptions) -> io::Result { let creation = opts.get_creation_mode()?; let handle = unsafe { c::CreateFileW( @@ -1212,9 +1214,8 @@ pub fn readdir(p: &Path) -> io::Result { } } -pub fn unlink(p: &Path) -> io::Result<()> { - let p_u16s = maybe_verbatim(p)?; - if unsafe { c::DeleteFileW(p_u16s.as_ptr()) } == 0 { +pub fn unlink(path: &WCStr) -> io::Result<()> { + if unsafe { c::DeleteFileW(path.as_ptr()) } == 0 { let err = api::get_last_error(); // if `DeleteFileW` fails with ERROR_ACCESS_DENIED then try to remove // the file while ignoring the readonly attribute. @@ -1223,7 +1224,7 @@ pub fn unlink(p: &Path) -> io::Result<()> { let mut opts = OpenOptions::new(); opts.access_mode(c::DELETE); opts.custom_flags(c::FILE_FLAG_OPEN_REPARSE_POINT); - if let Ok(f) = File::open_native(&p_u16s, &opts) { + if let Ok(f) = File::open_native(&path, &opts) { if f.posix_delete().is_ok() { return Ok(()); } @@ -1236,10 +1237,7 @@ pub fn unlink(p: &Path) -> io::Result<()> { } } -pub fn rename(old: &Path, new: &Path) -> io::Result<()> { - let old = maybe_verbatim(old)?; - let new = maybe_verbatim(new)?; - +pub fn rename(old: &WCStr, new: &WCStr) -> io::Result<()> { if unsafe { c::MoveFileExW(old.as_ptr(), new.as_ptr(), c::MOVEFILE_REPLACE_EXISTING) } == 0 { let err = api::get_last_error(); // if `MoveFileExW` fails with ERROR_ACCESS_DENIED then try to move @@ -1253,7 +1251,8 @@ pub fn rename(old: &Path, new: &Path) -> io::Result<()> { // Calculate the layout of the `FILE_RENAME_INFO` we pass to `SetFileInformation` // This is a dynamically sized struct so we need to get the position of the last field to calculate the actual size. - let Ok(new_len_without_nul_in_bytes): Result = ((new.len() - 1) * 2).try_into() + let Ok(new_len_without_nul_in_bytes): Result = + ((new.count_bytes() - 1) * 2).try_into() else { return Err(err).io_result(); }; @@ -1282,7 +1281,7 @@ pub fn rename(old: &Path, new: &Path) -> io::Result<()> { new.as_ptr().copy_to_nonoverlapping( (&raw mut (*file_rename_info).FileName).cast::(), - new.len(), + new.count_bytes(), ); } @@ -1309,20 +1308,19 @@ pub fn rename(old: &Path, new: &Path) -> io::Result<()> { Ok(()) } -pub fn rmdir(p: &Path) -> io::Result<()> { - let p = maybe_verbatim(p)?; +pub fn rmdir(p: &WCStr) -> io::Result<()> { cvt(unsafe { c::RemoveDirectoryW(p.as_ptr()) })?; Ok(()) } -pub fn remove_dir_all(path: &Path) -> io::Result<()> { +pub fn remove_dir_all(path: &WCStr) -> io::Result<()> { // Open a file or directory without following symlinks. let mut opts = OpenOptions::new(); opts.access_mode(c::FILE_LIST_DIRECTORY); // `FILE_FLAG_BACKUP_SEMANTICS` allows opening directories. // `FILE_FLAG_OPEN_REPARSE_POINT` opens a link instead of its target. opts.custom_flags(c::FILE_FLAG_BACKUP_SEMANTICS | c::FILE_FLAG_OPEN_REPARSE_POINT); - let file = File::open(path, &opts)?; + let file = File::open_native(path, &opts)?; // Test if the file is not a directory or a symlink to a directory. if (file.basic_info()?.FileAttributes & c::FILE_ATTRIBUTE_DIRECTORY) == 0 { @@ -1333,14 +1331,14 @@ pub fn remove_dir_all(path: &Path) -> io::Result<()> { remove_dir_all_iterative(file).io_result() } -pub fn readlink(path: &Path) -> io::Result { +pub fn readlink(path: &WCStr) -> io::Result { // Open the link with no access mode, instead of generic read. // By default FILE_LIST_DIRECTORY is denied for the junction "C:\Documents and Settings", so // this is needed for a common case. let mut opts = OpenOptions::new(); opts.access_mode(0); opts.custom_flags(c::FILE_FLAG_OPEN_REPARSE_POINT | c::FILE_FLAG_BACKUP_SEMANTICS); - let file = File::open(path, &opts)?; + let file = File::open_native(&path, &opts)?; file.readlink() } @@ -1378,19 +1376,17 @@ pub fn symlink_inner(original: &Path, link: &Path, dir: bool) -> io::Result<()> } #[cfg(not(target_vendor = "uwp"))] -pub fn link(original: &Path, link: &Path) -> io::Result<()> { - let original = maybe_verbatim(original)?; - let link = maybe_verbatim(link)?; +pub fn link(original: &WCStr, link: &WCStr) -> io::Result<()> { cvt(unsafe { c::CreateHardLinkW(link.as_ptr(), original.as_ptr(), ptr::null_mut()) })?; Ok(()) } #[cfg(target_vendor = "uwp")] -pub fn link(_original: &Path, _link: &Path) -> io::Result<()> { +pub fn link(_original: &WCStr, _link: &WCStr) -> io::Result<()> { return Err(io::const_error!(io::ErrorKind::Unsupported, "hard link are not supported on UWP")); } -pub fn stat(path: &Path) -> io::Result { +pub fn stat(path: &WCStr) -> io::Result { match metadata(path, ReparsePoint::Follow) { Err(err) if err.raw_os_error() == Some(c::ERROR_CANT_ACCESS_FILE as i32) => { if let Ok(attrs) = lstat(path) { @@ -1404,7 +1400,7 @@ pub fn stat(path: &Path) -> io::Result { } } -pub fn lstat(path: &Path) -> io::Result { +pub fn lstat(path: &WCStr) -> io::Result { metadata(path, ReparsePoint::Open) } @@ -1420,7 +1416,7 @@ impl ReparsePoint { } } -fn metadata(path: &Path, reparse: ReparsePoint) -> io::Result { +fn metadata(path: &WCStr, reparse: ReparsePoint) -> io::Result { let mut opts = OpenOptions::new(); // No read or write permissions are necessary opts.access_mode(0); @@ -1429,7 +1425,7 @@ fn metadata(path: &Path, reparse: ReparsePoint) -> io::Result { // Attempt to open the file normally. // If that fails with `ERROR_SHARING_VIOLATION` then retry using `FindFirstFileExW`. // If the fallback fails for any reason we return the original error. - match File::open(path, &opts) { + match File::open_native(&path, &opts) { Ok(file) => file.file_attr(), Err(e) if [Some(c::ERROR_SHARING_VIOLATION as _), Some(c::ERROR_ACCESS_DENIED as _)] @@ -1442,8 +1438,6 @@ fn metadata(path: &Path, reparse: ReparsePoint) -> io::Result { // However, there are special system files, such as // `C:\hiberfil.sys`, that are locked in a way that denies even that. unsafe { - let path = maybe_verbatim(path)?; - // `FindFirstFileExW` accepts wildcard file names. // Fortunately wildcards are not valid file names and // `ERROR_SHARING_VIOLATION` means the file exists (but is locked) @@ -1482,8 +1476,7 @@ fn metadata(path: &Path, reparse: ReparsePoint) -> io::Result { } } -pub fn set_perm(p: &Path, perm: FilePermissions) -> io::Result<()> { - let p = maybe_verbatim(p)?; +pub fn set_perm(p: &WCStr, perm: FilePermissions) -> io::Result<()> { unsafe { cvt(c::SetFileAttributesW(p.as_ptr(), perm.attrs))?; Ok(()) @@ -1499,17 +1492,17 @@ fn get_path(f: &File) -> io::Result { ) } -pub fn canonicalize(p: &Path) -> io::Result { +pub fn canonicalize(p: &WCStr) -> io::Result { let mut opts = OpenOptions::new(); // No read or write permissions are necessary opts.access_mode(0); // This flag is so we can open directories too opts.custom_flags(c::FILE_FLAG_BACKUP_SEMANTICS); - let f = File::open(p, &opts)?; + let f = File::open_native(p, &opts)?; get_path(&f) } -pub fn copy(from: &Path, to: &Path) -> io::Result { +pub fn copy(from: &WCStr, to: &WCStr) -> io::Result { unsafe extern "system" fn callback( _TotalFileSize: i64, _TotalBytesTransferred: i64, @@ -1528,13 +1521,11 @@ pub fn copy(from: &Path, to: &Path) -> io::Result { c::PROGRESS_CONTINUE } } - let pfrom = maybe_verbatim(from)?; - let pto = maybe_verbatim(to)?; let mut size = 0i64; cvt(unsafe { c::CopyFileExW( - pfrom.as_ptr(), - pto.as_ptr(), + from.as_ptr(), + to.as_ptr(), Some(callback), (&raw mut size) as *mut _, ptr::null_mut(), @@ -1624,14 +1615,14 @@ pub fn junction_point(original: &Path, link: &Path) -> io::Result<()> { } // Try to see if a file exists but, unlike `exists`, report I/O errors. -pub fn exists(path: &Path) -> io::Result { +pub fn exists(path: &WCStr) -> io::Result { // Open the file to ensure any symlinks are followed to their target. let mut opts = OpenOptions::new(); // No read, write, etc access rights are needed. opts.access_mode(0); // Backup semantics enables opening directories as well as files. opts.custom_flags(c::FILE_FLAG_BACKUP_SEMANTICS); - match File::open(path, &opts) { + match File::open_native(path, &opts) { Err(e) => match e.kind() { // The file definitely does not exist io::ErrorKind::NotFound => Ok(false), diff --git a/library/std/src/sys/path/windows.rs b/library/std/src/sys/path/windows.rs index 6547ed9aa5f..e0e003f6a81 100644 --- a/library/std/src/sys/path/windows.rs +++ b/library/std/src/sys/path/windows.rs @@ -10,6 +10,40 @@ mod tests; pub const MAIN_SEP_STR: &str = "\\"; pub const MAIN_SEP: char = '\\'; +/// A null terminated wide string. +#[repr(transparent)] +pub struct WCStr([u16]); + +impl WCStr { + /// Convert a slice to a WCStr without checks. + /// + /// Though it is memory safe, the slice should also not contain interior nulls + /// as this may lead to unwanted truncation. + /// + /// # Safety + /// + /// The slice must end in a null. + pub unsafe fn from_wchars_with_null_unchecked(s: &[u16]) -> &Self { + unsafe { &*(s as *const [u16] as *const Self) } + } + + pub fn as_ptr(&self) -> *const u16 { + self.0.as_ptr() + } + + pub fn count_bytes(&self) -> usize { + self.0.len() + } +} + +#[inline] +pub fn with_native_path(path: &Path, f: &dyn Fn(&WCStr) -> io::Result) -> io::Result { + let path = maybe_verbatim(path)?; + // SAFETY: maybe_verbatim returns null-terminated strings + let path = unsafe { WCStr::from_wchars_with_null_unchecked(&path) }; + f(path) +} + #[inline] pub fn is_sep_byte(b: u8) -> bool { b == b'/' || b == b'\\' -- cgit 1.4.1-3-g733a5 From 6ffebb65d6daf4fcc82cfc7fd14397bff98df1f2 Mon Sep 17 00:00:00 2001 From: Thalia Archibald Date: Sat, 12 Apr 2025 02:26:17 -0700 Subject: Move args into std::sys --- library/std/src/sys/args/hermit.rs | 66 ++++ library/std/src/sys/args/mod.rs | 34 ++ library/std/src/sys/args/sgx.rs | 61 ++++ library/std/src/sys/args/uefi.rs | 156 +++++++++ library/std/src/sys/args/unix.rs | 241 ++++++++++++++ library/std/src/sys/args/unsupported.rs | 36 +++ library/std/src/sys/args/wasi.rs | 61 ++++ library/std/src/sys/args/windows.rs | 442 +++++++++++++++++++++++++ library/std/src/sys/args/windows/tests.rs | 91 ++++++ library/std/src/sys/args/xous.rs | 53 +++ library/std/src/sys/args/zkvm.rs | 81 +++++ library/std/src/sys/mod.rs | 1 + library/std/src/sys/pal/hermit/args.rs | 66 ---- library/std/src/sys/pal/hermit/mod.rs | 3 +- library/std/src/sys/pal/sgx/args.rs | 59 ---- library/std/src/sys/pal/sgx/mod.rs | 3 +- library/std/src/sys/pal/solid/mod.rs | 2 - library/std/src/sys/pal/teeos/mod.rs | 2 - library/std/src/sys/pal/trusty/mod.rs | 2 - library/std/src/sys/pal/uefi/args.rs | 156 --------- library/std/src/sys/pal/uefi/mod.rs | 1 - library/std/src/sys/pal/unix/args.rs | 243 -------------- library/std/src/sys/pal/unix/mod.rs | 3 +- library/std/src/sys/pal/unsupported/args.rs | 36 --- library/std/src/sys/pal/unsupported/mod.rs | 1 - library/std/src/sys/pal/wasi/args.rs | 61 ---- library/std/src/sys/pal/wasi/mod.rs | 1 - library/std/src/sys/pal/wasip2/mod.rs | 2 - library/std/src/sys/pal/wasm/mod.rs | 2 - library/std/src/sys/pal/windows/args.rs | 445 -------------------------- library/std/src/sys/pal/windows/args/tests.rs | 91 ------ library/std/src/sys/pal/windows/mod.rs | 1 - library/std/src/sys/pal/xous/args.rs | 53 --- library/std/src/sys/pal/xous/mod.rs | 1 - library/std/src/sys/pal/zkvm/args.rs | 81 ----- library/std/src/sys/pal/zkvm/mod.rs | 4 +- 36 files changed, 1327 insertions(+), 1315 deletions(-) create mode 100644 library/std/src/sys/args/hermit.rs create mode 100644 library/std/src/sys/args/mod.rs create mode 100644 library/std/src/sys/args/sgx.rs create mode 100644 library/std/src/sys/args/uefi.rs create mode 100644 library/std/src/sys/args/unix.rs create mode 100644 library/std/src/sys/args/unsupported.rs create mode 100644 library/std/src/sys/args/wasi.rs create mode 100644 library/std/src/sys/args/windows.rs create mode 100644 library/std/src/sys/args/windows/tests.rs create mode 100644 library/std/src/sys/args/xous.rs create mode 100644 library/std/src/sys/args/zkvm.rs delete mode 100644 library/std/src/sys/pal/hermit/args.rs delete mode 100644 library/std/src/sys/pal/sgx/args.rs delete mode 100644 library/std/src/sys/pal/uefi/args.rs delete mode 100644 library/std/src/sys/pal/unix/args.rs delete mode 100644 library/std/src/sys/pal/unsupported/args.rs delete mode 100644 library/std/src/sys/pal/wasi/args.rs delete mode 100644 library/std/src/sys/pal/windows/args.rs delete mode 100644 library/std/src/sys/pal/windows/args/tests.rs delete mode 100644 library/std/src/sys/pal/xous/args.rs delete mode 100644 library/std/src/sys/pal/zkvm/args.rs (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/args/hermit.rs b/library/std/src/sys/args/hermit.rs new file mode 100644 index 00000000000..44024260277 --- /dev/null +++ b/library/std/src/sys/args/hermit.rs @@ -0,0 +1,66 @@ +use crate::ffi::{CStr, OsString, c_char}; +use crate::os::hermit::ffi::OsStringExt; +use crate::sync::atomic::Ordering::{Acquire, Relaxed, Release}; +use crate::sync::atomic::{AtomicIsize, AtomicPtr}; +use crate::{fmt, ptr, vec}; + +static ARGC: AtomicIsize = AtomicIsize::new(0); +static ARGV: AtomicPtr<*const u8> = AtomicPtr::new(ptr::null_mut()); + +/// One-time global initialization. +pub unsafe fn init(argc: isize, argv: *const *const u8) { + ARGC.store(argc, Relaxed); + // Use release ordering here to broadcast writes by the OS. + ARGV.store(argv as *mut *const u8, Release); +} + +/// Returns the command line arguments +pub fn args() -> Args { + // Synchronize with the store above. + let argv = ARGV.load(Acquire); + // If argv has not been initialized yet, do not return any arguments. + let argc = if argv.is_null() { 0 } else { ARGC.load(Relaxed) }; + let args: Vec = (0..argc) + .map(|i| unsafe { + let cstr = CStr::from_ptr(*argv.offset(i) as *const c_char); + OsStringExt::from_vec(cstr.to_bytes().to_vec()) + }) + .collect(); + + Args { iter: args.into_iter() } +} + +pub struct Args { + iter: vec::IntoIter, +} + +impl fmt::Debug for Args { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + self.iter.as_slice().fmt(f) + } +} + +impl !Send for Args {} +impl !Sync for Args {} + +impl Iterator for Args { + type Item = OsString; + fn next(&mut self) -> Option { + self.iter.next() + } + fn size_hint(&self) -> (usize, Option) { + self.iter.size_hint() + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.iter.len() + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + self.iter.next_back() + } +} diff --git a/library/std/src/sys/args/mod.rs b/library/std/src/sys/args/mod.rs new file mode 100644 index 00000000000..6edcafe2c4c --- /dev/null +++ b/library/std/src/sys/args/mod.rs @@ -0,0 +1,34 @@ +//! Platform-dependent command line arguments abstraction. + +#![forbid(unsafe_op_in_unsafe_fn)] + +cfg_if::cfg_if! { + if #[cfg(target_family = "unix")] { + mod unix; + pub use unix::*; + } else if #[cfg(target_family = "windows")] { + mod windows; + pub use windows::*; + } else if #[cfg(target_os = "hermit")] { + mod hermit; + pub use hermit::*; + } else if #[cfg(all(target_vendor = "fortanix", target_env = "sgx"))] { + mod sgx; + pub use sgx::*; + } else if #[cfg(target_os = "uefi")] { + mod uefi; + pub use uefi::*; + } else if #[cfg(target_os = "wasi")] { + mod wasi; + pub use wasi::*; + } else if #[cfg(target_os = "xous")] { + mod xous; + pub use xous::*; + } else if #[cfg(target_os = "zkvm")] { + mod zkvm; + pub use zkvm::*; + } else { + mod unsupported; + pub use unsupported::*; + } +} diff --git a/library/std/src/sys/args/sgx.rs b/library/std/src/sys/args/sgx.rs new file mode 100644 index 00000000000..efc4b791852 --- /dev/null +++ b/library/std/src/sys/args/sgx.rs @@ -0,0 +1,61 @@ +#![allow(fuzzy_provenance_casts)] // FIXME: this module systematically confuses pointers and integers + +use crate::ffi::OsString; +use crate::sync::atomic::{AtomicUsize, Ordering}; +use crate::sys::os_str::Buf; +use crate::sys::pal::abi::usercalls::alloc; +use crate::sys::pal::abi::usercalls::raw::ByteBuffer; +use crate::sys_common::FromInner; +use crate::{fmt, slice}; + +#[cfg_attr(test, linkage = "available_externally")] +#[unsafe(export_name = "_ZN16__rust_internals3std3sys3sgx4args4ARGSE")] +static ARGS: AtomicUsize = AtomicUsize::new(0); +type ArgsStore = Vec; + +#[cfg_attr(test, allow(dead_code))] +pub unsafe fn init(argc: isize, argv: *const *const u8) { + if argc != 0 { + let args = unsafe { alloc::User::<[ByteBuffer]>::from_raw_parts(argv as _, argc as _) }; + let args = args + .iter() + .map(|a| OsString::from_inner(Buf { inner: a.copy_user_buffer() })) + .collect::(); + ARGS.store(Box::into_raw(Box::new(args)) as _, Ordering::Relaxed); + } +} + +pub fn args() -> Args { + let args = unsafe { (ARGS.load(Ordering::Relaxed) as *const ArgsStore).as_ref() }; + if let Some(args) = args { Args(args.iter()) } else { Args([].iter()) } +} + +pub struct Args(slice::Iter<'static, OsString>); + +impl fmt::Debug for Args { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + self.0.as_slice().fmt(f) + } +} + +impl Iterator for Args { + type Item = OsString; + fn next(&mut self) -> Option { + self.0.next().cloned() + } + fn size_hint(&self) -> (usize, Option) { + self.0.size_hint() + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.0.len() + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + self.0.next_back().cloned() + } +} diff --git a/library/std/src/sys/args/uefi.rs b/library/std/src/sys/args/uefi.rs new file mode 100644 index 00000000000..4e7412937bd --- /dev/null +++ b/library/std/src/sys/args/uefi.rs @@ -0,0 +1,156 @@ +use r_efi::protocols::loaded_image; + +use crate::env::current_exe; +use crate::ffi::OsString; +use crate::iter::Iterator; +use crate::sys::pal::helpers; +use crate::{fmt, vec}; + +pub struct Args { + parsed_args_list: vec::IntoIter, +} + +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_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::() || lp_size % size_of::() != 0 { + return Args { parsed_args_list: lazy_current_exe().into_iter() }; + } + let lp_size = lp_size / size_of::(); + + 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 { + self.parsed_args_list.next() + } + + fn size_hint(&self) -> (usize, Option) { + 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 { + 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) +/// +/// Returns None in the following cases: +/// - Invalid UTF-16 (unpaired surrogate) +/// - Empty/improper arguments +fn parse_lp_cmd_line(code_units: &[u16]) -> Option> { + 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) +} diff --git a/library/std/src/sys/args/unix.rs b/library/std/src/sys/args/unix.rs new file mode 100644 index 00000000000..4c30609c099 --- /dev/null +++ b/library/std/src/sys/args/unix.rs @@ -0,0 +1,241 @@ +//! Global initialization and retrieval of command line arguments. +//! +//! On some platforms these are stored during runtime startup, +//! and on some they are retrieved from the system on demand. + +#![allow(dead_code)] // runtime init functions not used during testing + +use crate::ffi::{CStr, OsString}; +use crate::os::unix::ffi::OsStringExt; +use crate::{fmt, vec}; + +/// One-time global initialization. +pub unsafe fn init(argc: isize, argv: *const *const u8) { + unsafe { imp::init(argc, argv) } +} + +/// Returns the command line arguments +pub fn args() -> Args { + let (argc, argv) = imp::argc_argv(); + + let mut vec = Vec::with_capacity(argc as usize); + + for i in 0..argc { + // SAFETY: `argv` is non-null if `argc` is positive, and it is + // guaranteed to be at least as long as `argc`, so reading from it + // should be safe. + let ptr = unsafe { argv.offset(i).read() }; + + // Some C commandline parsers (e.g. GLib and Qt) are replacing already + // handled arguments in `argv` with `NULL` and move them to the end. + // + // Since they can't directly ensure updates to `argc` as well, this + // means that `argc` might be bigger than the actual number of + // non-`NULL` pointers in `argv` at this point. + // + // To handle this we simply stop iterating at the first `NULL` + // argument. `argv` is also guaranteed to be `NULL`-terminated so any + // non-`NULL` arguments after the first `NULL` can safely be ignored. + if ptr.is_null() { + // NOTE: On Apple platforms, `-[NSProcessInfo arguments]` does not + // stop iterating here, but instead `continue`, always iterating + // up until it reached `argc`. + // + // This difference will only matter in very specific circumstances + // where `argc`/`argv` have been modified, but in unexpected ways, + // so it likely doesn't really matter which option we choose. + // See the following PR for further discussion: + // + break; + } + + // SAFETY: Just checked that the pointer is not NULL, and arguments + // are otherwise guaranteed to be valid C strings. + let cstr = unsafe { CStr::from_ptr(ptr) }; + vec.push(OsStringExt::from_vec(cstr.to_bytes().to_vec())); + } + + Args { iter: vec.into_iter() } +} + +pub struct Args { + iter: vec::IntoIter, +} + +impl !Send for Args {} +impl !Sync for Args {} + +impl fmt::Debug for Args { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + self.iter.as_slice().fmt(f) + } +} + +impl Iterator for Args { + type Item = OsString; + fn next(&mut self) -> Option { + self.iter.next() + } + fn size_hint(&self) -> (usize, Option) { + self.iter.size_hint() + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.iter.len() + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + self.iter.next_back() + } +} + +#[cfg(any( + target_os = "linux", + target_os = "android", + target_os = "freebsd", + target_os = "dragonfly", + target_os = "netbsd", + target_os = "openbsd", + target_os = "cygwin", + target_os = "solaris", + target_os = "illumos", + target_os = "emscripten", + target_os = "haiku", + target_os = "l4re", + target_os = "fuchsia", + target_os = "redox", + target_os = "vxworks", + target_os = "horizon", + target_os = "aix", + target_os = "nto", + target_os = "hurd", + target_os = "rtems", + target_os = "nuttx", +))] +mod imp { + use crate::ffi::c_char; + use crate::ptr; + use crate::sync::atomic::{AtomicIsize, AtomicPtr, Ordering}; + + // The system-provided argc and argv, which we store in static memory + // here so that we can defer the work of parsing them until its actually + // needed. + // + // Note that we never mutate argv/argc, the argv array, or the argv + // strings, which allows the code in this file to be very simple. + static ARGC: AtomicIsize = AtomicIsize::new(0); + static ARGV: AtomicPtr<*const u8> = AtomicPtr::new(ptr::null_mut()); + + unsafe fn really_init(argc: isize, argv: *const *const u8) { + // These don't need to be ordered with each other or other stores, + // because they only hold the unmodified system-provide argv/argc. + ARGC.store(argc, Ordering::Relaxed); + ARGV.store(argv as *mut _, Ordering::Relaxed); + } + + #[inline(always)] + pub unsafe fn init(argc: isize, argv: *const *const u8) { + // on GNU/Linux if we are main then we will init argv and argc twice, it "duplicates work" + // BUT edge-cases are real: only using .init_array can break most emulators, dlopen, etc. + unsafe { really_init(argc, argv) }; + } + + /// glibc passes argc, argv, and envp to functions in .init_array, as a non-standard extension. + /// This allows `std::env::args` to work even in a `cdylib`, as it does on macOS and Windows. + #[cfg(all(target_os = "linux", target_env = "gnu"))] + #[used] + #[unsafe(link_section = ".init_array.00099")] + static ARGV_INIT_ARRAY: extern "C" fn( + crate::os::raw::c_int, + *const *const u8, + *const *const u8, + ) = { + extern "C" fn init_wrapper( + argc: crate::os::raw::c_int, + argv: *const *const u8, + _envp: *const *const u8, + ) { + unsafe { really_init(argc as isize, argv) }; + } + init_wrapper + }; + + pub fn argc_argv() -> (isize, *const *const c_char) { + // Load ARGC and ARGV, which hold the unmodified system-provided + // argc/argv, so we can read the pointed-to memory without atomics or + // synchronization. + // + // If either ARGC or ARGV is still zero or null, then either there + // really are no arguments, or someone is asking for `args()` before + // initialization has completed, and we return an empty list. + let argv = ARGV.load(Ordering::Relaxed); + let argc = if argv.is_null() { 0 } else { ARGC.load(Ordering::Relaxed) }; + + // Cast from `*mut *const u8` to `*const *const c_char` + (argc, argv.cast()) + } +} + +// Use `_NSGetArgc` and `_NSGetArgv` on Apple platforms. +// +// Even though these have underscores in their names, they've been available +// since the first versions of both macOS and iOS, and are declared in +// the header `crt_externs.h`. +// +// NOTE: This header was added to the iOS 13.0 SDK, which has been the source +// of a great deal of confusion in the past about the availability of these +// APIs. +// +// NOTE(madsmtm): This has not strictly been verified to not cause App Store +// rejections; if this is found to be the case, the previous implementation +// of this used `[[NSProcessInfo processInfo] arguments]`. +#[cfg(target_vendor = "apple")] +mod imp { + use crate::ffi::{c_char, c_int}; + + pub unsafe fn init(_argc: isize, _argv: *const *const u8) { + // No need to initialize anything in here, `libdyld.dylib` has already + // done the work for us. + } + + pub fn argc_argv() -> (isize, *const *const c_char) { + unsafe extern "C" { + // These functions are in crt_externs.h. + fn _NSGetArgc() -> *mut c_int; + fn _NSGetArgv() -> *mut *mut *mut c_char; + } + + // SAFETY: The returned pointer points to a static initialized early + // in the program lifetime by `libdyld.dylib`, and as such is always + // valid. + // + // NOTE: Similar to `_NSGetEnviron`, there technically isn't anything + // protecting us against concurrent modifications to this, and there + // doesn't exist a lock that we can take. Instead, it is generally + // expected that it's only modified in `main` / before other code + // runs, so reading this here should be fine. + let argc = unsafe { _NSGetArgc().read() }; + // SAFETY: Same as above. + let argv = unsafe { _NSGetArgv().read() }; + + // Cast from `*mut *mut c_char` to `*const *const c_char` + (argc as isize, argv.cast()) + } +} + +#[cfg(any(target_os = "espidf", target_os = "vita"))] +mod imp { + use crate::ffi::c_char; + use crate::ptr; + + #[inline(always)] + pub unsafe fn init(_argc: isize, _argv: *const *const u8) {} + + pub fn argc_argv() -> (isize, *const *const c_char) { + (0, ptr::null()) + } +} diff --git a/library/std/src/sys/args/unsupported.rs b/library/std/src/sys/args/unsupported.rs new file mode 100644 index 00000000000..a2d75a61976 --- /dev/null +++ b/library/std/src/sys/args/unsupported.rs @@ -0,0 +1,36 @@ +use crate::ffi::OsString; +use crate::fmt; + +pub struct Args {} + +pub fn args() -> Args { + Args {} +} + +impl fmt::Debug for Args { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().finish() + } +} + +impl Iterator for Args { + type Item = OsString; + fn next(&mut self) -> Option { + None + } + fn size_hint(&self) -> (usize, Option) { + (0, Some(0)) + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + 0 + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + None + } +} diff --git a/library/std/src/sys/args/wasi.rs b/library/std/src/sys/args/wasi.rs new file mode 100644 index 00000000000..52cfa202af8 --- /dev/null +++ b/library/std/src/sys/args/wasi.rs @@ -0,0 +1,61 @@ +#![forbid(unsafe_op_in_unsafe_fn)] + +use crate::ffi::{CStr, OsStr, OsString}; +use crate::os::wasi::ffi::OsStrExt; +use crate::{fmt, vec}; + +pub struct Args { + iter: vec::IntoIter, +} + +impl !Send for Args {} +impl !Sync for Args {} + +/// Returns the command line arguments +pub fn args() -> Args { + Args { iter: maybe_args().unwrap_or(Vec::new()).into_iter() } +} + +fn maybe_args() -> Option> { + unsafe { + let (argc, buf_size) = wasi::args_sizes_get().ok()?; + let mut argv = Vec::with_capacity(argc); + let mut buf = Vec::with_capacity(buf_size); + wasi::args_get(argv.as_mut_ptr(), buf.as_mut_ptr()).ok()?; + argv.set_len(argc); + let mut ret = Vec::with_capacity(argc); + for ptr in argv { + let s = CStr::from_ptr(ptr.cast()); + ret.push(OsStr::from_bytes(s.to_bytes()).to_owned()); + } + Some(ret) + } +} + +impl fmt::Debug for Args { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + self.iter.as_slice().fmt(f) + } +} + +impl Iterator for Args { + type Item = OsString; + fn next(&mut self) -> Option { + self.iter.next() + } + fn size_hint(&self) -> (usize, Option) { + self.iter.size_hint() + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.iter.len() + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + self.iter.next_back() + } +} diff --git a/library/std/src/sys/args/windows.rs b/library/std/src/sys/args/windows.rs new file mode 100644 index 00000000000..55bd6864cae --- /dev/null +++ b/library/std/src/sys/args/windows.rs @@ -0,0 +1,442 @@ +//! The Windows command line is just a string +//! +//! +//! This module implements the parsing necessary to turn that string into a list of arguments. + +#[cfg(test)] +mod tests; + +use crate::ffi::{OsStr, OsString}; +use crate::num::NonZero; +use crate::os::windows::prelude::*; +use crate::path::{Path, PathBuf}; +use crate::sys::pal::os::current_exe; +use crate::sys::pal::{ensure_no_nuls, fill_utf16_buf}; +use crate::sys::path::get_long_path; +use crate::sys::{c, to_u16s}; +use crate::sys_common::AsInner; +use crate::sys_common::wstr::WStrUnits; +use crate::{fmt, io, iter, ptr, vec}; + +pub fn args() -> Args { + // SAFETY: `GetCommandLineW` returns a pointer to a null terminated UTF-16 + // string so it's safe for `WStrUnits` to use. + unsafe { + let lp_cmd_line = c::GetCommandLineW(); + let parsed_args_list = parse_lp_cmd_line(WStrUnits::new(lp_cmd_line), || { + current_exe().map(PathBuf::into_os_string).unwrap_or_else(|_| OsString::new()) + }); + + Args { parsed_args_list: parsed_args_list.into_iter() } + } +} + +/// Implements the Windows command-line argument parsing algorithm. +/// +/// Microsoft's documentation for the Windows CLI argument format can be found at +/// +/// +/// A more in-depth explanation is here: +/// +/// +/// Windows includes a function to do command line parsing in shell32.dll. +/// However, this is not used for two reasons: +/// +/// 1. Linking with that DLL causes the process to be registered as a GUI application. +/// GUI applications add a bunch of overhead, even if no windows are drawn. See +/// . +/// +/// 2. It does not follow the modern C/C++ argv rules outlined in the first two links above. +/// +/// This function was tested for equivalence to the C/C++ parsing rules using an +/// extensive test suite available at +/// . +fn parse_lp_cmd_line<'a, F: Fn() -> OsString>( + lp_cmd_line: Option>, + exe_name: F, +) -> Vec { + const BACKSLASH: NonZero = NonZero::new(b'\\' as u16).unwrap(); + const QUOTE: NonZero = NonZero::new(b'"' as u16).unwrap(); + const TAB: NonZero = NonZero::new(b'\t' as u16).unwrap(); + const SPACE: NonZero = NonZero::new(b' ' as u16).unwrap(); + + let mut ret_val = Vec::new(); + // If the cmd line pointer is null or it points to an empty string then + // return the name of the executable as argv[0]. + if lp_cmd_line.as_ref().and_then(|cmd| cmd.peek()).is_none() { + ret_val.push(exe_name()); + return ret_val; + } + let mut code_units = lp_cmd_line.unwrap(); + + // The executable name at the beginning is special. + let mut in_quotes = false; + let mut cur = Vec::new(); + for w in &mut code_units { + match w { + // 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 | TAB if !in_quotes => break, + // In all other cases the code unit is taken literally. + _ => cur.push(w.get()), + } + } + // Skip whitespace. + code_units.advance_while(|w| w == SPACE || w == TAB); + ret_val.push(OsString::from_wide(&cur)); + + // Parse the arguments according to these rules: + // * All code units are taken literally except space, tab, quote and backslash. + // * When not `in_quotes`, space and tab separate arguments. Consecutive spaces and tabs are + // treated as a single separator. + // * A space or tab `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 an odd number of backslashes. + // * If any number of backslashes is immediately followed by a quote then the number of + // backslashes is halved (rounding down). + // * Backslashes not followed by a quote are all taken literally. + // * If `in_quotes` then a quote can also be escaped using another quote + // (i.e. two consecutive quotes become one literal quote). + let mut cur = Vec::new(); + let mut in_quotes = false; + while let Some(w) = code_units.next() { + match w { + // If not `in_quotes`, a space or tab ends the argument. + SPACE | TAB if !in_quotes => { + ret_val.push(OsString::from_wide(&cur[..])); + cur.truncate(0); + + // Skip whitespace. + code_units.advance_while(|w| w == SPACE || w == TAB); + } + // Backslashes can escape quotes or backslashes but only if consecutive backslashes are followed by a quote. + BACKSLASH => { + let backslash_count = code_units.advance_while(|w| w == BACKSLASH) + 1; + if code_units.peek() == Some(QUOTE) { + cur.extend(iter::repeat(BACKSLASH.get()).take(backslash_count / 2)); + // The quote is escaped if there are an odd number of backslashes. + if backslash_count % 2 == 1 { + code_units.next(); + cur.push(QUOTE.get()); + } + } else { + // If there is no quote on the end then there is no escaping. + cur.extend(iter::repeat(BACKSLASH.get()).take(backslash_count)); + } + } + // If `in_quotes` and not backslash escaped (see above) then a quote either + // unsets `in_quote` or is escaped by another quote. + QUOTE if in_quotes => match code_units.peek() { + // Two consecutive quotes when `in_quotes` produces one literal quote. + Some(QUOTE) => { + cur.push(QUOTE.get()); + code_units.next(); + } + // Otherwise set `in_quotes`. + Some(_) => in_quotes = false, + // The end of the command line. + // Push `cur` even if empty, which we do by breaking while `in_quotes` is still set. + None => break, + }, + // If not `in_quotes` and not BACKSLASH escaped (see above) then a quote sets `in_quote`. + QUOTE => in_quotes = true, + // Everything else is always taken literally. + _ => cur.push(w.get()), + } + } + // Push the final argument, if any. + if !cur.is_empty() || in_quotes { + ret_val.push(OsString::from_wide(&cur[..])); + } + ret_val +} + +pub struct Args { + parsed_args_list: vec::IntoIter, +} + +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 { + self.parsed_args_list.next() + } + fn size_hint(&self) -> (usize, Option) { + self.parsed_args_list.size_hint() + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + self.parsed_args_list.next_back() + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.parsed_args_list.len() + } +} + +#[derive(Debug)] +pub(crate) enum Arg { + /// Add quotes (if needed) + Regular(OsString), + /// Append raw string without quoting + Raw(OsString), +} + +enum Quote { + // Every arg is quoted + Always, + // Whitespace and empty args are quoted + Auto, + // Arg appended without any changes (#29494) + Never, +} + +pub(crate) fn append_arg(cmd: &mut Vec, arg: &Arg, force_quotes: bool) -> io::Result<()> { + let (arg, quote) = match arg { + Arg::Regular(arg) => (arg, if force_quotes { Quote::Always } else { Quote::Auto }), + Arg::Raw(arg) => (arg, Quote::Never), + }; + + // If an argument has 0 characters then we need to quote it to ensure + // that it actually gets passed through on the command line or otherwise + // it will be dropped entirely when parsed on the other end. + ensure_no_nuls(arg)?; + let arg_bytes = arg.as_encoded_bytes(); + let (quote, escape) = match quote { + Quote::Always => (true, true), + Quote::Auto => { + (arg_bytes.iter().any(|c| *c == b' ' || *c == b'\t') || arg_bytes.is_empty(), true) + } + Quote::Never => (false, false), + }; + if quote { + cmd.push('"' as u16); + } + + let mut backslashes: usize = 0; + for x in arg.encode_wide() { + if escape { + if x == '\\' as u16 { + backslashes += 1; + } else { + if x == '"' as u16 { + // Add n+1 backslashes to total 2n+1 before internal '"'. + cmd.extend((0..=backslashes).map(|_| '\\' as u16)); + } + backslashes = 0; + } + } + cmd.push(x); + } + + if quote { + // Add n backslashes to total 2n before ending '"'. + cmd.extend((0..backslashes).map(|_| '\\' as u16)); + cmd.push('"' as u16); + } + Ok(()) +} + +fn append_bat_arg(cmd: &mut Vec, arg: &OsStr, mut quote: bool) -> io::Result<()> { + ensure_no_nuls(arg)?; + // If an argument has 0 characters then we need to quote it to ensure + // that it actually gets passed through on the command line or otherwise + // it will be dropped entirely when parsed on the other end. + // + // We also need to quote the argument if it ends with `\` to guard against + // bat usage such as `"%~2"` (i.e. force quote arguments) otherwise a + // trailing slash will escape the closing quote. + if arg.is_empty() || arg.as_encoded_bytes().last() == Some(&b'\\') { + quote = true; + } + for cp in arg.as_inner().inner.code_points() { + if let Some(cp) = cp.to_char() { + // Rather than trying to find every ascii symbol that must be quoted, + // we assume that all ascii symbols must be quoted unless they're known to be good. + // We also quote Unicode control blocks for good measure. + // Note an unquoted `\` is fine so long as the argument isn't otherwise quoted. + static UNQUOTED: &str = r"#$*+-./:?@\_"; + let ascii_needs_quotes = + cp.is_ascii() && !(cp.is_ascii_alphanumeric() || UNQUOTED.contains(cp)); + if ascii_needs_quotes || cp.is_control() { + quote = true; + } + } + } + + if quote { + cmd.push('"' as u16); + } + // Loop through the string, escaping `\` only if followed by `"`. + // And escaping `"` by doubling them. + let mut backslashes: usize = 0; + for x in arg.encode_wide() { + if x == '\\' as u16 { + backslashes += 1; + } else { + if x == '"' as u16 { + // Add n backslashes to total 2n before internal `"`. + cmd.extend((0..backslashes).map(|_| '\\' as u16)); + // Appending an additional double-quote acts as an escape. + cmd.push(b'"' as u16) + } else if x == '%' as u16 || x == '\r' as u16 { + // yt-dlp hack: replaces `%` with `%%cd:~,%` to stop %VAR% being expanded as an environment variable. + // + // # Explanation + // + // cmd supports extracting a substring from a variable using the following syntax: + // %variable:~start_index,end_index% + // + // In the above command `cd` is used as the variable and the start_index and end_index are left blank. + // `cd` is a built-in variable that dynamically expands to the current directory so it's always available. + // Explicitly omitting both the start and end index creates a zero-length substring. + // + // Therefore it all resolves to nothing. However, by doing this no-op we distract cmd.exe + // from potentially expanding %variables% in the argument. + cmd.extend_from_slice(&[ + '%' as u16, '%' as u16, 'c' as u16, 'd' as u16, ':' as u16, '~' as u16, + ',' as u16, + ]); + } + backslashes = 0; + } + cmd.push(x); + } + if quote { + // Add n backslashes to total 2n before ending `"`. + cmd.extend((0..backslashes).map(|_| '\\' as u16)); + cmd.push('"' as u16); + } + Ok(()) +} + +pub(crate) fn make_bat_command_line( + script: &[u16], + args: &[Arg], + force_quotes: bool, +) -> io::Result> { + const INVALID_ARGUMENT_ERROR: io::Error = + io::const_error!(io::ErrorKind::InvalidInput, r#"batch file arguments are invalid"#); + // Set the start of the command line to `cmd.exe /c "` + // It is necessary to surround the command in an extra pair of quotes, + // hence the trailing quote here. It will be closed after all arguments + // have been added. + // Using /e:ON enables "command extensions" which is essential for the `%` hack to work. + let mut cmd: Vec = "cmd.exe /e:ON /v:OFF /d /c \"".encode_utf16().collect(); + + // Push the script name surrounded by its quote pair. + cmd.push(b'"' as u16); + // Windows file names cannot contain a `"` character or end with `\\`. + // If the script name does then return an error. + if script.contains(&(b'"' as u16)) || script.last() == Some(&(b'\\' as u16)) { + return Err(io::const_error!( + io::ErrorKind::InvalidInput, + "Windows file names may not contain `\"` or end with `\\`" + )); + } + cmd.extend_from_slice(script.strip_suffix(&[0]).unwrap_or(script)); + cmd.push(b'"' as u16); + + // Append the arguments. + // FIXME: This needs tests to ensure that the arguments are properly + // reconstructed by the batch script by default. + for arg in args { + cmd.push(' ' as u16); + match arg { + Arg::Regular(arg_os) => { + let arg_bytes = arg_os.as_encoded_bytes(); + // Disallow \r and \n as they may truncate the arguments. + const DISALLOWED: &[u8] = b"\r\n"; + if arg_bytes.iter().any(|c| DISALLOWED.contains(c)) { + return Err(INVALID_ARGUMENT_ERROR); + } + append_bat_arg(&mut cmd, arg_os, force_quotes)?; + } + _ => { + // Raw arguments are passed on as-is. + // It's the user's responsibility to properly handle arguments in this case. + append_arg(&mut cmd, arg, force_quotes)?; + } + }; + } + + // Close the quote we left opened earlier. + cmd.push(b'"' as u16); + + Ok(cmd) +} + +/// Takes a path and tries to return a non-verbatim path. +/// +/// This is necessary because cmd.exe does not support verbatim paths. +pub(crate) fn to_user_path(path: &Path) -> io::Result> { + from_wide_to_user_path(to_u16s(path)?) +} +pub(crate) fn from_wide_to_user_path(mut path: Vec) -> io::Result> { + // UTF-16 encoded code points, used in parsing and building UTF-16 paths. + // All of these are in the ASCII range so they can be cast directly to `u16`. + const SEP: u16 = b'\\' as _; + const QUERY: u16 = b'?' as _; + const COLON: u16 = b':' as _; + const U: u16 = b'U' as _; + const N: u16 = b'N' as _; + const C: u16 = b'C' as _; + + // Early return if the path is too long to remove the verbatim prefix. + const LEGACY_MAX_PATH: usize = 260; + if path.len() > LEGACY_MAX_PATH { + return Ok(path); + } + + match &path[..] { + // `\\?\C:\...` => `C:\...` + [SEP, SEP, QUERY, SEP, _, COLON, SEP, ..] => unsafe { + let lpfilename = path[4..].as_ptr(); + fill_utf16_buf( + |buffer, size| c::GetFullPathNameW(lpfilename, size, buffer, ptr::null_mut()), + |full_path: &[u16]| { + if full_path == &path[4..path.len() - 1] { + let mut path: Vec = full_path.into(); + path.push(0); + path + } else { + path + } + }, + ) + }, + // `\\?\UNC\...` => `\\...` + [SEP, SEP, QUERY, SEP, U, N, C, SEP, ..] => unsafe { + // Change the `C` in `UNC\` to `\` so we can get a slice that starts with `\\`. + path[6] = b'\\' as u16; + let lpfilename = path[6..].as_ptr(); + fill_utf16_buf( + |buffer, size| c::GetFullPathNameW(lpfilename, size, buffer, ptr::null_mut()), + |full_path: &[u16]| { + if full_path == &path[6..path.len() - 1] { + let mut path: Vec = full_path.into(); + path.push(0); + path + } else { + // Restore the 'C' in "UNC". + path[6] = b'C' as u16; + path + } + }, + ) + }, + // For everything else, leave the path unchanged. + _ => get_long_path(path, false), + } +} diff --git a/library/std/src/sys/args/windows/tests.rs b/library/std/src/sys/args/windows/tests.rs new file mode 100644 index 00000000000..484a90ab056 --- /dev/null +++ b/library/std/src/sys/args/windows/tests.rs @@ -0,0 +1,91 @@ +use super::*; +use crate::ffi::OsString; + +fn chk(string: &str, parts: &[&str]) { + let mut wide: Vec = OsString::from(string).encode_wide().collect(); + wide.push(0); + let parsed = + unsafe { parse_lp_cmd_line(WStrUnits::new(wide.as_ptr()), || OsString::from("TEST.EXE")) }; + let expected: Vec = parts.iter().map(|k| OsString::from(k)).collect(); + assert_eq!(parsed.as_slice(), expected.as_slice(), "{:?}", string); +} + +#[test] +fn empty() { + chk("", &["TEST.EXE"]); + chk("\0", &["TEST.EXE"]); +} + +#[test] +fn single_words() { + chk("EXE one_word", &["EXE", "one_word"]); + chk("EXE a", &["EXE", "a"]); + chk("EXE 😅", &["EXE", "😅"]); + chk("EXE 😅🤦", &["EXE", "😅🤦"]); +} + +#[test] +fn official_examples() { + chk(r#"EXE "abc" d e"#, &["EXE", "abc", "d", "e"]); + chk(r#"EXE a\\\b d"e f"g h"#, &["EXE", r"a\\\b", "de fg", "h"]); + chk(r#"EXE a\\\"b c d"#, &["EXE", r#"a\"b"#, "c", "d"]); + chk(r#"EXE a\\\\"b c" d e"#, &["EXE", r"a\\b c", "d", "e"]); +} + +#[test] +fn whitespace_behavior() { + chk(" test", &["", "test"]); + chk(" test", &["", "test"]); + chk(" test test2", &["", "test", "test2"]); + chk(" test test2", &["", "test", "test2"]); + chk("test test2 ", &["test", "test2"]); + chk("test test2 ", &["test", "test2"]); + chk("test ", &["test"]); +} + +#[test] +fn genius_quotes() { + chk(r#"EXE "" """#, &["EXE", "", ""]); + chk(r#"EXE "" """"#, &["EXE", "", r#"""#]); + chk( + r#"EXE "this is """all""" in the same argument""#, + &["EXE", r#"this is "all" in the same argument"#], + ); + chk(r#"EXE "a"""#, &["EXE", r#"a""#]); + chk(r#"EXE "a"" a"#, &["EXE", r#"a" a"#]); + // quotes cannot be escaped in command names + chk(r#""EXE" check"#, &["EXE", "check"]); + chk(r#""EXE check""#, &["EXE check"]); + chk(r#""EXE """for""" check"#, &["EXE for check"]); + chk(r#""EXE \"for\" check"#, &[r"EXE \for\ check"]); + chk(r#""EXE \" for \" check"#, &[r"EXE \", "for", r#"""#, "check"]); + chk(r#"E"X"E test"#, &["EXE", "test"]); + chk(r#"EX""E test"#, &["EXE", "test"]); +} + +// from https://daviddeley.com/autohotkey/parameters/parameters.htm#WINCRULESEX +#[test] +fn post_2008() { + chk("EXE CallMeIshmael", &["EXE", "CallMeIshmael"]); + chk(r#"EXE "Call Me Ishmael""#, &["EXE", "Call Me Ishmael"]); + chk(r#"EXE Cal"l Me I"shmael"#, &["EXE", "Call Me Ishmael"]); + chk(r#"EXE CallMe\"Ishmael"#, &["EXE", r#"CallMe"Ishmael"#]); + chk(r#"EXE "CallMe\"Ishmael""#, &["EXE", r#"CallMe"Ishmael"#]); + chk(r#"EXE "Call Me Ishmael\\""#, &["EXE", r"Call Me Ishmael\"]); + chk(r#"EXE "CallMe\\\"Ishmael""#, &["EXE", r#"CallMe\"Ishmael"#]); + chk(r#"EXE a\\\b"#, &["EXE", r"a\\\b"]); + chk(r#"EXE "a\\\b""#, &["EXE", r"a\\\b"]); + chk(r#"EXE "\"Call Me Ishmael\"""#, &["EXE", r#""Call Me Ishmael""#]); + chk(r#"EXE "C:\TEST A\\""#, &["EXE", r"C:\TEST A\"]); + chk(r#"EXE "\"C:\TEST A\\\"""#, &["EXE", r#""C:\TEST A\""#]); + chk(r#"EXE "a b c" d e"#, &["EXE", "a b c", "d", "e"]); + chk(r#"EXE "ab\"c" "\\" d"#, &["EXE", r#"ab"c"#, r"\", "d"]); + chk(r#"EXE a\\\b d"e f"g h"#, &["EXE", r"a\\\b", "de fg", "h"]); + chk(r#"EXE a\\\"b c d"#, &["EXE", r#"a\"b"#, "c", "d"]); + chk(r#"EXE a\\\\"b c" d e"#, &["EXE", r"a\\b c", "d", "e"]); + // Double Double Quotes + chk(r#"EXE "a b c"""#, &["EXE", r#"a b c""#]); + chk(r#"EXE """CallMeIshmael""" b c"#, &["EXE", r#""CallMeIshmael""#, "b", "c"]); + chk(r#"EXE """Call Me Ishmael""""#, &["EXE", r#""Call Me Ishmael""#]); + chk(r#"EXE """"Call Me Ishmael"" b c"#, &["EXE", r#""Call"#, "Me", "Ishmael", "b", "c"]); +} diff --git a/library/std/src/sys/args/xous.rs b/library/std/src/sys/args/xous.rs new file mode 100644 index 00000000000..2dd37d53422 --- /dev/null +++ b/library/std/src/sys/args/xous.rs @@ -0,0 +1,53 @@ +use crate::ffi::OsString; +use crate::sys::pal::os::get_application_parameters; +use crate::sys::pal::os::params::ArgumentList; +use crate::{fmt, vec}; + +pub struct Args { + parsed_args_list: vec::IntoIter, +} + +pub fn args() -> Args { + let Some(params) = get_application_parameters() else { + return Args { parsed_args_list: vec![].into_iter() }; + }; + + for param in params { + if let Ok(args) = ArgumentList::try_from(¶m) { + let mut parsed_args = vec![]; + for arg in args { + parsed_args.push(arg.into()); + } + return Args { parsed_args_list: parsed_args.into_iter() }; + } + } + Args { parsed_args_list: vec![].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 { + self.parsed_args_list.next() + } + fn size_hint(&self) -> (usize, Option) { + self.parsed_args_list.size_hint() + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + self.parsed_args_list.next_back() + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.parsed_args_list.len() + } +} diff --git a/library/std/src/sys/args/zkvm.rs b/library/std/src/sys/args/zkvm.rs new file mode 100644 index 00000000000..194ba7159d4 --- /dev/null +++ b/library/std/src/sys/args/zkvm.rs @@ -0,0 +1,81 @@ +use crate::ffi::OsString; +use crate::fmt; +use crate::sys::os_str; +use crate::sys::pal::{WORD_SIZE, abi}; +use crate::sys_common::FromInner; + +pub struct Args { + i_forward: usize, + i_back: usize, + count: usize, +} + +pub fn args() -> Args { + let count = unsafe { abi::sys_argc() }; + Args { i_forward: 0, i_back: 0, count } +} + +impl Args { + /// Use sys_argv to get the arg at the requested index. Does not check that i is less than argc + /// and will not return if the index is out of bounds. + fn argv(i: usize) -> OsString { + let arg_len = unsafe { abi::sys_argv(crate::ptr::null_mut(), 0, i) }; + + let arg_len_words = (arg_len + WORD_SIZE - 1) / WORD_SIZE; + let words = unsafe { abi::sys_alloc_words(arg_len_words) }; + + let arg_len2 = unsafe { abi::sys_argv(words, arg_len_words, i) }; + debug_assert_eq!(arg_len, arg_len2); + + // Convert to OsString. + // + // FIXME: We can probably get rid of the extra copy here if we + // reimplement "os_str" instead of just using the generic unix + // "os_str". + let arg_bytes: &[u8] = + unsafe { crate::slice::from_raw_parts(words.cast() as *const u8, arg_len) }; + OsString::from_inner(os_str::Buf { inner: arg_bytes.to_vec() }) + } +} + +impl fmt::Debug for Args { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().finish() + } +} + +impl Iterator for Args { + type Item = OsString; + + fn next(&mut self) -> Option { + if self.i_forward >= self.count - self.i_back { + None + } else { + let arg = Self::argv(self.i_forward); + self.i_forward += 1; + Some(arg) + } + } + + fn size_hint(&self) -> (usize, Option) { + (self.count, Some(self.count)) + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.count + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + if self.i_back >= self.count - self.i_forward { + None + } else { + let arg = Self::argv(self.count - 1 - self.i_back); + self.i_back += 1; + Some(arg) + } + } +} diff --git a/library/std/src/sys/mod.rs b/library/std/src/sys/mod.rs index f8f220fafd1..bc4bf11cb74 100644 --- a/library/std/src/sys/mod.rs +++ b/library/std/src/sys/mod.rs @@ -9,6 +9,7 @@ mod alloc; mod personality; pub mod anonymous_pipe; +pub mod args; pub mod backtrace; pub mod cmath; pub mod exit_guard; diff --git a/library/std/src/sys/pal/hermit/args.rs b/library/std/src/sys/pal/hermit/args.rs deleted file mode 100644 index 44024260277..00000000000 --- a/library/std/src/sys/pal/hermit/args.rs +++ /dev/null @@ -1,66 +0,0 @@ -use crate::ffi::{CStr, OsString, c_char}; -use crate::os::hermit::ffi::OsStringExt; -use crate::sync::atomic::Ordering::{Acquire, Relaxed, Release}; -use crate::sync::atomic::{AtomicIsize, AtomicPtr}; -use crate::{fmt, ptr, vec}; - -static ARGC: AtomicIsize = AtomicIsize::new(0); -static ARGV: AtomicPtr<*const u8> = AtomicPtr::new(ptr::null_mut()); - -/// One-time global initialization. -pub unsafe fn init(argc: isize, argv: *const *const u8) { - ARGC.store(argc, Relaxed); - // Use release ordering here to broadcast writes by the OS. - ARGV.store(argv as *mut *const u8, Release); -} - -/// Returns the command line arguments -pub fn args() -> Args { - // Synchronize with the store above. - let argv = ARGV.load(Acquire); - // If argv has not been initialized yet, do not return any arguments. - let argc = if argv.is_null() { 0 } else { ARGC.load(Relaxed) }; - let args: Vec = (0..argc) - .map(|i| unsafe { - let cstr = CStr::from_ptr(*argv.offset(i) as *const c_char); - OsStringExt::from_vec(cstr.to_bytes().to_vec()) - }) - .collect(); - - Args { iter: args.into_iter() } -} - -pub struct Args { - iter: vec::IntoIter, -} - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - self.iter.as_slice().fmt(f) - } -} - -impl !Send for Args {} -impl !Sync for Args {} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - self.iter.next() - } - fn size_hint(&self) -> (usize, Option) { - self.iter.size_hint() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.iter.len() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.iter.next_back() - } -} diff --git a/library/std/src/sys/pal/hermit/mod.rs b/library/std/src/sys/pal/hermit/mod.rs index 26211bcb152..821836824e2 100644 --- a/library/std/src/sys/pal/hermit/mod.rs +++ b/library/std/src/sys/pal/hermit/mod.rs @@ -18,7 +18,6 @@ use crate::os::raw::c_char; -pub mod args; pub mod env; pub mod futex; pub mod os; @@ -58,7 +57,7 @@ pub extern "C" fn __rust_abort() { // NOTE: this is not guaranteed to run, for example when Rust code is called externally. pub unsafe fn init(argc: isize, argv: *const *const u8, _sigpipe: u8) { unsafe { - args::init(argc, argv); + crate::sys::args::init(argc, argv); } } diff --git a/library/std/src/sys/pal/sgx/args.rs b/library/std/src/sys/pal/sgx/args.rs deleted file mode 100644 index e62bf383954..00000000000 --- a/library/std/src/sys/pal/sgx/args.rs +++ /dev/null @@ -1,59 +0,0 @@ -use super::abi::usercalls::alloc; -use super::abi::usercalls::raw::ByteBuffer; -use crate::ffi::OsString; -use crate::sync::atomic::{AtomicUsize, Ordering}; -use crate::sys::os_str::Buf; -use crate::sys_common::FromInner; -use crate::{fmt, slice}; - -#[cfg_attr(test, linkage = "available_externally")] -#[unsafe(export_name = "_ZN16__rust_internals3std3sys3sgx4args4ARGSE")] -static ARGS: AtomicUsize = AtomicUsize::new(0); -type ArgsStore = Vec; - -#[cfg_attr(test, allow(dead_code))] -pub unsafe fn init(argc: isize, argv: *const *const u8) { - if argc != 0 { - let args = unsafe { alloc::User::<[ByteBuffer]>::from_raw_parts(argv as _, argc as _) }; - let args = args - .iter() - .map(|a| OsString::from_inner(Buf { inner: a.copy_user_buffer() })) - .collect::(); - ARGS.store(Box::into_raw(Box::new(args)) as _, Ordering::Relaxed); - } -} - -pub fn args() -> Args { - let args = unsafe { (ARGS.load(Ordering::Relaxed) as *const ArgsStore).as_ref() }; - if let Some(args) = args { Args(args.iter()) } else { Args([].iter()) } -} - -pub struct Args(slice::Iter<'static, OsString>); - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - self.0.as_slice().fmt(f) - } -} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - self.0.next().cloned() - } - fn size_hint(&self) -> (usize, Option) { - self.0.size_hint() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.0.len() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.0.next_back().cloned() - } -} diff --git a/library/std/src/sys/pal/sgx/mod.rs b/library/std/src/sys/pal/sgx/mod.rs index 52684e18ac2..8a87e7a7ae1 100644 --- a/library/std/src/sys/pal/sgx/mod.rs +++ b/library/std/src/sys/pal/sgx/mod.rs @@ -9,7 +9,6 @@ use crate::io::ErrorKind; use crate::sync::atomic::{AtomicBool, Ordering}; pub mod abi; -pub mod args; pub mod env; mod libunwind_integration; pub mod os; @@ -24,7 +23,7 @@ pub mod waitqueue; // NOTE: this is not guaranteed to run, for example when Rust code is called externally. pub unsafe fn init(argc: isize, argv: *const *const u8, _sigpipe: u8) { unsafe { - args::init(argc, argv); + crate::sys::args::init(argc, argv); } } diff --git a/library/std/src/sys/pal/solid/mod.rs b/library/std/src/sys/pal/solid/mod.rs index 22052a168fd..c41dc848a1b 100644 --- a/library/std/src/sys/pal/solid/mod.rs +++ b/library/std/src/sys/pal/solid/mod.rs @@ -16,8 +16,6 @@ pub mod itron { use super::unsupported; } -#[path = "../unsupported/args.rs"] -pub mod args; pub mod env; // `error` is `pub(crate)` so that it can be accessed by `itron/error.rs` as // `crate::sys::error` diff --git a/library/std/src/sys/pal/teeos/mod.rs b/library/std/src/sys/pal/teeos/mod.rs index c1921a2f40d..b8095cec3e9 100644 --- a/library/std/src/sys/pal/teeos/mod.rs +++ b/library/std/src/sys/pal/teeos/mod.rs @@ -6,8 +6,6 @@ #![allow(unused_variables)] #![allow(dead_code)] -#[path = "../unsupported/args.rs"] -pub mod args; #[path = "../unsupported/env.rs"] pub mod env; //pub mod fd; diff --git a/library/std/src/sys/pal/trusty/mod.rs b/library/std/src/sys/pal/trusty/mod.rs index 5295d3fdc91..04e6b4c8186 100644 --- a/library/std/src/sys/pal/trusty/mod.rs +++ b/library/std/src/sys/pal/trusty/mod.rs @@ -1,7 +1,5 @@ //! System bindings for the Trusty OS. -#[path = "../unsupported/args.rs"] -pub mod args; #[path = "../unsupported/common.rs"] #[deny(unsafe_op_in_unsafe_fn)] mod common; diff --git a/library/std/src/sys/pal/uefi/args.rs b/library/std/src/sys/pal/uefi/args.rs deleted file mode 100644 index 0c29caf2db6..00000000000 --- a/library/std/src/sys/pal/uefi/args.rs +++ /dev/null @@ -1,156 +0,0 @@ -use r_efi::protocols::loaded_image; - -use super::helpers; -use crate::env::current_exe; -use crate::ffi::OsString; -use crate::iter::Iterator; -use crate::{fmt, vec}; - -pub struct Args { - parsed_args_list: vec::IntoIter, -} - -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_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::() || lp_size % size_of::() != 0 { - return Args { parsed_args_list: lazy_current_exe().into_iter() }; - } - let lp_size = lp_size / size_of::(); - - 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 { - self.parsed_args_list.next() - } - - fn size_hint(&self) -> (usize, Option) { - 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 { - 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) -/// -/// Returns None in the following cases: -/// - Invalid UTF-16 (unpaired surrogate) -/// - Empty/improper arguments -fn parse_lp_cmd_line(code_units: &[u16]) -> Option> { - 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) -} diff --git a/library/std/src/sys/pal/uefi/mod.rs b/library/std/src/sys/pal/uefi/mod.rs index 9760a23084a..cd901f48b76 100644 --- a/library/std/src/sys/pal/uefi/mod.rs +++ b/library/std/src/sys/pal/uefi/mod.rs @@ -13,7 +13,6 @@ //! [`OsString`]: crate::ffi::OsString #![forbid(unsafe_op_in_unsafe_fn)] -pub mod args; pub mod env; pub mod helpers; pub mod os; diff --git a/library/std/src/sys/pal/unix/args.rs b/library/std/src/sys/pal/unix/args.rs deleted file mode 100644 index 0bb7b64007a..00000000000 --- a/library/std/src/sys/pal/unix/args.rs +++ /dev/null @@ -1,243 +0,0 @@ -//! Global initialization and retrieval of command line arguments. -//! -//! On some platforms these are stored during runtime startup, -//! and on some they are retrieved from the system on demand. - -#![allow(dead_code)] // runtime init functions not used during testing - -use crate::ffi::{CStr, OsString}; -use crate::os::unix::ffi::OsStringExt; -use crate::{fmt, vec}; - -/// One-time global initialization. -pub unsafe fn init(argc: isize, argv: *const *const u8) { - imp::init(argc, argv) -} - -/// Returns the command line arguments -pub fn args() -> Args { - let (argc, argv) = imp::argc_argv(); - - let mut vec = Vec::with_capacity(argc as usize); - - for i in 0..argc { - // SAFETY: `argv` is non-null if `argc` is positive, and it is - // guaranteed to be at least as long as `argc`, so reading from it - // should be safe. - let ptr = unsafe { argv.offset(i).read() }; - - // Some C commandline parsers (e.g. GLib and Qt) are replacing already - // handled arguments in `argv` with `NULL` and move them to the end. - // - // Since they can't directly ensure updates to `argc` as well, this - // means that `argc` might be bigger than the actual number of - // non-`NULL` pointers in `argv` at this point. - // - // To handle this we simply stop iterating at the first `NULL` - // argument. `argv` is also guaranteed to be `NULL`-terminated so any - // non-`NULL` arguments after the first `NULL` can safely be ignored. - if ptr.is_null() { - // NOTE: On Apple platforms, `-[NSProcessInfo arguments]` does not - // stop iterating here, but instead `continue`, always iterating - // up until it reached `argc`. - // - // This difference will only matter in very specific circumstances - // where `argc`/`argv` have been modified, but in unexpected ways, - // so it likely doesn't really matter which option we choose. - // See the following PR for further discussion: - // - break; - } - - // SAFETY: Just checked that the pointer is not NULL, and arguments - // are otherwise guaranteed to be valid C strings. - let cstr = unsafe { CStr::from_ptr(ptr) }; - vec.push(OsStringExt::from_vec(cstr.to_bytes().to_vec())); - } - - Args { iter: vec.into_iter() } -} - -pub struct Args { - iter: vec::IntoIter, -} - -impl !Send for Args {} -impl !Sync for Args {} - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - self.iter.as_slice().fmt(f) - } -} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - self.iter.next() - } - fn size_hint(&self) -> (usize, Option) { - self.iter.size_hint() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.iter.len() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.iter.next_back() - } -} - -#[cfg(any( - target_os = "linux", - target_os = "android", - target_os = "freebsd", - target_os = "dragonfly", - target_os = "netbsd", - target_os = "openbsd", - target_os = "cygwin", - target_os = "solaris", - target_os = "illumos", - target_os = "emscripten", - target_os = "haiku", - target_os = "l4re", - target_os = "fuchsia", - target_os = "redox", - target_os = "vxworks", - target_os = "horizon", - target_os = "aix", - target_os = "nto", - target_os = "hurd", - target_os = "rtems", - target_os = "nuttx", -))] -mod imp { - use crate::ffi::c_char; - use crate::ptr; - use crate::sync::atomic::{AtomicIsize, AtomicPtr, Ordering}; - - // The system-provided argc and argv, which we store in static memory - // here so that we can defer the work of parsing them until its actually - // needed. - // - // Note that we never mutate argv/argc, the argv array, or the argv - // strings, which allows the code in this file to be very simple. - static ARGC: AtomicIsize = AtomicIsize::new(0); - static ARGV: AtomicPtr<*const u8> = AtomicPtr::new(ptr::null_mut()); - - unsafe fn really_init(argc: isize, argv: *const *const u8) { - // These don't need to be ordered with each other or other stores, - // because they only hold the unmodified system-provide argv/argc. - ARGC.store(argc, Ordering::Relaxed); - ARGV.store(argv as *mut _, Ordering::Relaxed); - } - - #[inline(always)] - pub unsafe fn init(argc: isize, argv: *const *const u8) { - // on GNU/Linux if we are main then we will init argv and argc twice, it "duplicates work" - // BUT edge-cases are real: only using .init_array can break most emulators, dlopen, etc. - really_init(argc, argv); - } - - /// glibc passes argc, argv, and envp to functions in .init_array, as a non-standard extension. - /// This allows `std::env::args` to work even in a `cdylib`, as it does on macOS and Windows. - #[cfg(all(target_os = "linux", target_env = "gnu"))] - #[used] - #[unsafe(link_section = ".init_array.00099")] - static ARGV_INIT_ARRAY: extern "C" fn( - crate::os::raw::c_int, - *const *const u8, - *const *const u8, - ) = { - extern "C" fn init_wrapper( - argc: crate::os::raw::c_int, - argv: *const *const u8, - _envp: *const *const u8, - ) { - unsafe { - really_init(argc as isize, argv); - } - } - init_wrapper - }; - - pub fn argc_argv() -> (isize, *const *const c_char) { - // Load ARGC and ARGV, which hold the unmodified system-provided - // argc/argv, so we can read the pointed-to memory without atomics or - // synchronization. - // - // If either ARGC or ARGV is still zero or null, then either there - // really are no arguments, or someone is asking for `args()` before - // initialization has completed, and we return an empty list. - let argv = ARGV.load(Ordering::Relaxed); - let argc = if argv.is_null() { 0 } else { ARGC.load(Ordering::Relaxed) }; - - // Cast from `*mut *const u8` to `*const *const c_char` - (argc, argv.cast()) - } -} - -// Use `_NSGetArgc` and `_NSGetArgv` on Apple platforms. -// -// Even though these have underscores in their names, they've been available -// since the first versions of both macOS and iOS, and are declared in -// the header `crt_externs.h`. -// -// NOTE: This header was added to the iOS 13.0 SDK, which has been the source -// of a great deal of confusion in the past about the availability of these -// APIs. -// -// NOTE(madsmtm): This has not strictly been verified to not cause App Store -// rejections; if this is found to be the case, the previous implementation -// of this used `[[NSProcessInfo processInfo] arguments]`. -#[cfg(target_vendor = "apple")] -mod imp { - use crate::ffi::{c_char, c_int}; - - pub unsafe fn init(_argc: isize, _argv: *const *const u8) { - // No need to initialize anything in here, `libdyld.dylib` has already - // done the work for us. - } - - pub fn argc_argv() -> (isize, *const *const c_char) { - unsafe extern "C" { - // These functions are in crt_externs.h. - fn _NSGetArgc() -> *mut c_int; - fn _NSGetArgv() -> *mut *mut *mut c_char; - } - - // SAFETY: The returned pointer points to a static initialized early - // in the program lifetime by `libdyld.dylib`, and as such is always - // valid. - // - // NOTE: Similar to `_NSGetEnviron`, there technically isn't anything - // protecting us against concurrent modifications to this, and there - // doesn't exist a lock that we can take. Instead, it is generally - // expected that it's only modified in `main` / before other code - // runs, so reading this here should be fine. - let argc = unsafe { _NSGetArgc().read() }; - // SAFETY: Same as above. - let argv = unsafe { _NSGetArgv().read() }; - - // Cast from `*mut *mut c_char` to `*const *const c_char` - (argc as isize, argv.cast()) - } -} - -#[cfg(any(target_os = "espidf", target_os = "vita"))] -mod imp { - use crate::ffi::c_char; - use crate::ptr; - - #[inline(always)] - pub unsafe fn init(_argc: isize, _argv: *const *const u8) {} - - pub fn argc_argv() -> (isize, *const *const c_char) { - (0, ptr::null()) - } -} diff --git a/library/std/src/sys/pal/unix/mod.rs b/library/std/src/sys/pal/unix/mod.rs index d7106c33974..20078839a14 100644 --- a/library/std/src/sys/pal/unix/mod.rs +++ b/library/std/src/sys/pal/unix/mod.rs @@ -6,7 +6,6 @@ use crate::io::ErrorKind; #[macro_use] pub mod weak; -pub mod args; pub mod env; #[cfg(target_os = "fuchsia")] pub mod fuchsia; @@ -47,7 +46,7 @@ pub unsafe fn init(argc: isize, argv: *const *const u8, sigpipe: u8) { reset_sigpipe(sigpipe); stack_overflow::init(); - args::init(argc, argv); + crate::sys::args::init(argc, argv); // Normally, `thread::spawn` will call `Thread::set_name` but since this thread // already exists, we have to call it ourselves. We only do this on Apple targets diff --git a/library/std/src/sys/pal/unsupported/args.rs b/library/std/src/sys/pal/unsupported/args.rs deleted file mode 100644 index a2d75a61976..00000000000 --- a/library/std/src/sys/pal/unsupported/args.rs +++ /dev/null @@ -1,36 +0,0 @@ -use crate::ffi::OsString; -use crate::fmt; - -pub struct Args {} - -pub fn args() -> Args { - Args {} -} - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - f.debug_list().finish() - } -} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - None - } - fn size_hint(&self) -> (usize, Option) { - (0, Some(0)) - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - 0 - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - None - } -} diff --git a/library/std/src/sys/pal/unsupported/mod.rs b/library/std/src/sys/pal/unsupported/mod.rs index 38838b915b5..dea42a95dcc 100644 --- a/library/std/src/sys/pal/unsupported/mod.rs +++ b/library/std/src/sys/pal/unsupported/mod.rs @@ -1,6 +1,5 @@ #![deny(unsafe_op_in_unsafe_fn)] -pub mod args; pub mod env; pub mod os; pub mod pipe; diff --git a/library/std/src/sys/pal/wasi/args.rs b/library/std/src/sys/pal/wasi/args.rs deleted file mode 100644 index 52cfa202af8..00000000000 --- a/library/std/src/sys/pal/wasi/args.rs +++ /dev/null @@ -1,61 +0,0 @@ -#![forbid(unsafe_op_in_unsafe_fn)] - -use crate::ffi::{CStr, OsStr, OsString}; -use crate::os::wasi::ffi::OsStrExt; -use crate::{fmt, vec}; - -pub struct Args { - iter: vec::IntoIter, -} - -impl !Send for Args {} -impl !Sync for Args {} - -/// Returns the command line arguments -pub fn args() -> Args { - Args { iter: maybe_args().unwrap_or(Vec::new()).into_iter() } -} - -fn maybe_args() -> Option> { - unsafe { - let (argc, buf_size) = wasi::args_sizes_get().ok()?; - let mut argv = Vec::with_capacity(argc); - let mut buf = Vec::with_capacity(buf_size); - wasi::args_get(argv.as_mut_ptr(), buf.as_mut_ptr()).ok()?; - argv.set_len(argc); - let mut ret = Vec::with_capacity(argc); - for ptr in argv { - let s = CStr::from_ptr(ptr.cast()); - ret.push(OsStr::from_bytes(s.to_bytes()).to_owned()); - } - Some(ret) - } -} - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - self.iter.as_slice().fmt(f) - } -} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - self.iter.next() - } - fn size_hint(&self) -> (usize, Option) { - self.iter.size_hint() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.iter.len() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.iter.next_back() - } -} diff --git a/library/std/src/sys/pal/wasi/mod.rs b/library/std/src/sys/pal/wasi/mod.rs index 80853e7b5a2..4ea42b1082b 100644 --- a/library/std/src/sys/pal/wasi/mod.rs +++ b/library/std/src/sys/pal/wasi/mod.rs @@ -13,7 +13,6 @@ //! compiling for wasm. That way it's a compile time error for something that's //! guaranteed to be a runtime error! -pub mod args; pub mod env; #[allow(unused)] #[path = "../wasm/atomics/futex.rs"] diff --git a/library/std/src/sys/pal/wasip2/mod.rs b/library/std/src/sys/pal/wasip2/mod.rs index 504b947d09e..6445bf2cc0d 100644 --- a/library/std/src/sys/pal/wasip2/mod.rs +++ b/library/std/src/sys/pal/wasip2/mod.rs @@ -6,8 +6,6 @@ //! To begin with, this target mirrors the wasi target 1 to 1, but over //! time this will change significantly. -#[path = "../wasi/args.rs"] -pub mod args; #[path = "../wasi/env.rs"] pub mod env; #[allow(unused)] diff --git a/library/std/src/sys/pal/wasm/mod.rs b/library/std/src/sys/pal/wasm/mod.rs index 8d39b70d039..af370020d96 100644 --- a/library/std/src/sys/pal/wasm/mod.rs +++ b/library/std/src/sys/pal/wasm/mod.rs @@ -16,8 +16,6 @@ #![deny(unsafe_op_in_unsafe_fn)] -#[path = "../unsupported/args.rs"] -pub mod args; pub mod env; #[path = "../unsupported/os.rs"] pub mod os; diff --git a/library/std/src/sys/pal/windows/args.rs b/library/std/src/sys/pal/windows/args.rs deleted file mode 100644 index d973743639a..00000000000 --- a/library/std/src/sys/pal/windows/args.rs +++ /dev/null @@ -1,445 +0,0 @@ -//! The Windows command line is just a string -//! -//! -//! This module implements the parsing necessary to turn that string into a list of arguments. - -#[cfg(test)] -mod tests; - -use super::ensure_no_nuls; -use super::os::current_exe; -use crate::ffi::{OsStr, OsString}; -use crate::num::NonZero; -use crate::os::windows::prelude::*; -use crate::path::{Path, PathBuf}; -use crate::sys::path::get_long_path; -use crate::sys::{c, to_u16s}; -use crate::sys_common::AsInner; -use crate::sys_common::wstr::WStrUnits; -use crate::{fmt, io, iter, vec}; - -pub fn args() -> Args { - // SAFETY: `GetCommandLineW` returns a pointer to a null terminated UTF-16 - // string so it's safe for `WStrUnits` to use. - unsafe { - let lp_cmd_line = c::GetCommandLineW(); - let parsed_args_list = parse_lp_cmd_line(WStrUnits::new(lp_cmd_line), || { - current_exe().map(PathBuf::into_os_string).unwrap_or_else(|_| OsString::new()) - }); - - Args { parsed_args_list: parsed_args_list.into_iter() } - } -} - -/// Implements the Windows command-line argument parsing algorithm. -/// -/// Microsoft's documentation for the Windows CLI argument format can be found at -/// -/// -/// A more in-depth explanation is here: -/// -/// -/// Windows includes a function to do command line parsing in shell32.dll. -/// However, this is not used for two reasons: -/// -/// 1. Linking with that DLL causes the process to be registered as a GUI application. -/// GUI applications add a bunch of overhead, even if no windows are drawn. See -/// . -/// -/// 2. It does not follow the modern C/C++ argv rules outlined in the first two links above. -/// -/// This function was tested for equivalence to the C/C++ parsing rules using an -/// extensive test suite available at -/// . -fn parse_lp_cmd_line<'a, F: Fn() -> OsString>( - lp_cmd_line: Option>, - exe_name: F, -) -> Vec { - const BACKSLASH: NonZero = NonZero::new(b'\\' as u16).unwrap(); - const QUOTE: NonZero = NonZero::new(b'"' as u16).unwrap(); - const TAB: NonZero = NonZero::new(b'\t' as u16).unwrap(); - const SPACE: NonZero = NonZero::new(b' ' as u16).unwrap(); - - let mut ret_val = Vec::new(); - // If the cmd line pointer is null or it points to an empty string then - // return the name of the executable as argv[0]. - if lp_cmd_line.as_ref().and_then(|cmd| cmd.peek()).is_none() { - ret_val.push(exe_name()); - return ret_val; - } - let mut code_units = lp_cmd_line.unwrap(); - - // The executable name at the beginning is special. - let mut in_quotes = false; - let mut cur = Vec::new(); - for w in &mut code_units { - match w { - // 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 | TAB if !in_quotes => break, - // In all other cases the code unit is taken literally. - _ => cur.push(w.get()), - } - } - // Skip whitespace. - code_units.advance_while(|w| w == SPACE || w == TAB); - ret_val.push(OsString::from_wide(&cur)); - - // Parse the arguments according to these rules: - // * All code units are taken literally except space, tab, quote and backslash. - // * When not `in_quotes`, space and tab separate arguments. Consecutive spaces and tabs are - // treated as a single separator. - // * A space or tab `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 an odd number of backslashes. - // * If any number of backslashes is immediately followed by a quote then the number of - // backslashes is halved (rounding down). - // * Backslashes not followed by a quote are all taken literally. - // * If `in_quotes` then a quote can also be escaped using another quote - // (i.e. two consecutive quotes become one literal quote). - let mut cur = Vec::new(); - let mut in_quotes = false; - while let Some(w) = code_units.next() { - match w { - // If not `in_quotes`, a space or tab ends the argument. - SPACE | TAB if !in_quotes => { - ret_val.push(OsString::from_wide(&cur[..])); - cur.truncate(0); - - // Skip whitespace. - code_units.advance_while(|w| w == SPACE || w == TAB); - } - // Backslashes can escape quotes or backslashes but only if consecutive backslashes are followed by a quote. - BACKSLASH => { - let backslash_count = code_units.advance_while(|w| w == BACKSLASH) + 1; - if code_units.peek() == Some(QUOTE) { - cur.extend(iter::repeat(BACKSLASH.get()).take(backslash_count / 2)); - // The quote is escaped if there are an odd number of backslashes. - if backslash_count % 2 == 1 { - code_units.next(); - cur.push(QUOTE.get()); - } - } else { - // If there is no quote on the end then there is no escaping. - cur.extend(iter::repeat(BACKSLASH.get()).take(backslash_count)); - } - } - // If `in_quotes` and not backslash escaped (see above) then a quote either - // unsets `in_quote` or is escaped by another quote. - QUOTE if in_quotes => match code_units.peek() { - // Two consecutive quotes when `in_quotes` produces one literal quote. - Some(QUOTE) => { - cur.push(QUOTE.get()); - code_units.next(); - } - // Otherwise set `in_quotes`. - Some(_) => in_quotes = false, - // The end of the command line. - // Push `cur` even if empty, which we do by breaking while `in_quotes` is still set. - None => break, - }, - // If not `in_quotes` and not BACKSLASH escaped (see above) then a quote sets `in_quote`. - QUOTE => in_quotes = true, - // Everything else is always taken literally. - _ => cur.push(w.get()), - } - } - // Push the final argument, if any. - if !cur.is_empty() || in_quotes { - ret_val.push(OsString::from_wide(&cur[..])); - } - ret_val -} - -pub struct Args { - parsed_args_list: vec::IntoIter, -} - -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 { - self.parsed_args_list.next() - } - fn size_hint(&self) -> (usize, Option) { - self.parsed_args_list.size_hint() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.parsed_args_list.next_back() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.parsed_args_list.len() - } -} - -#[derive(Debug)] -pub(crate) enum Arg { - /// Add quotes (if needed) - Regular(OsString), - /// Append raw string without quoting - Raw(OsString), -} - -enum Quote { - // Every arg is quoted - Always, - // Whitespace and empty args are quoted - Auto, - // Arg appended without any changes (#29494) - Never, -} - -pub(crate) fn append_arg(cmd: &mut Vec, arg: &Arg, force_quotes: bool) -> io::Result<()> { - let (arg, quote) = match arg { - Arg::Regular(arg) => (arg, if force_quotes { Quote::Always } else { Quote::Auto }), - Arg::Raw(arg) => (arg, Quote::Never), - }; - - // If an argument has 0 characters then we need to quote it to ensure - // that it actually gets passed through on the command line or otherwise - // it will be dropped entirely when parsed on the other end. - ensure_no_nuls(arg)?; - let arg_bytes = arg.as_encoded_bytes(); - let (quote, escape) = match quote { - Quote::Always => (true, true), - Quote::Auto => { - (arg_bytes.iter().any(|c| *c == b' ' || *c == b'\t') || arg_bytes.is_empty(), true) - } - Quote::Never => (false, false), - }; - if quote { - cmd.push('"' as u16); - } - - let mut backslashes: usize = 0; - for x in arg.encode_wide() { - if escape { - if x == '\\' as u16 { - backslashes += 1; - } else { - if x == '"' as u16 { - // Add n+1 backslashes to total 2n+1 before internal '"'. - cmd.extend((0..=backslashes).map(|_| '\\' as u16)); - } - backslashes = 0; - } - } - cmd.push(x); - } - - if quote { - // Add n backslashes to total 2n before ending '"'. - cmd.extend((0..backslashes).map(|_| '\\' as u16)); - cmd.push('"' as u16); - } - Ok(()) -} - -fn append_bat_arg(cmd: &mut Vec, arg: &OsStr, mut quote: bool) -> io::Result<()> { - ensure_no_nuls(arg)?; - // If an argument has 0 characters then we need to quote it to ensure - // that it actually gets passed through on the command line or otherwise - // it will be dropped entirely when parsed on the other end. - // - // We also need to quote the argument if it ends with `\` to guard against - // bat usage such as `"%~2"` (i.e. force quote arguments) otherwise a - // trailing slash will escape the closing quote. - if arg.is_empty() || arg.as_encoded_bytes().last() == Some(&b'\\') { - quote = true; - } - for cp in arg.as_inner().inner.code_points() { - if let Some(cp) = cp.to_char() { - // Rather than trying to find every ascii symbol that must be quoted, - // we assume that all ascii symbols must be quoted unless they're known to be good. - // We also quote Unicode control blocks for good measure. - // Note an unquoted `\` is fine so long as the argument isn't otherwise quoted. - static UNQUOTED: &str = r"#$*+-./:?@\_"; - let ascii_needs_quotes = - cp.is_ascii() && !(cp.is_ascii_alphanumeric() || UNQUOTED.contains(cp)); - if ascii_needs_quotes || cp.is_control() { - quote = true; - } - } - } - - if quote { - cmd.push('"' as u16); - } - // Loop through the string, escaping `\` only if followed by `"`. - // And escaping `"` by doubling them. - let mut backslashes: usize = 0; - for x in arg.encode_wide() { - if x == '\\' as u16 { - backslashes += 1; - } else { - if x == '"' as u16 { - // Add n backslashes to total 2n before internal `"`. - cmd.extend((0..backslashes).map(|_| '\\' as u16)); - // Appending an additional double-quote acts as an escape. - cmd.push(b'"' as u16) - } else if x == '%' as u16 || x == '\r' as u16 { - // yt-dlp hack: replaces `%` with `%%cd:~,%` to stop %VAR% being expanded as an environment variable. - // - // # Explanation - // - // cmd supports extracting a substring from a variable using the following syntax: - // %variable:~start_index,end_index% - // - // In the above command `cd` is used as the variable and the start_index and end_index are left blank. - // `cd` is a built-in variable that dynamically expands to the current directory so it's always available. - // Explicitly omitting both the start and end index creates a zero-length substring. - // - // Therefore it all resolves to nothing. However, by doing this no-op we distract cmd.exe - // from potentially expanding %variables% in the argument. - cmd.extend_from_slice(&[ - '%' as u16, '%' as u16, 'c' as u16, 'd' as u16, ':' as u16, '~' as u16, - ',' as u16, - ]); - } - backslashes = 0; - } - cmd.push(x); - } - if quote { - // Add n backslashes to total 2n before ending `"`. - cmd.extend((0..backslashes).map(|_| '\\' as u16)); - cmd.push('"' as u16); - } - Ok(()) -} - -pub(crate) fn make_bat_command_line( - script: &[u16], - args: &[Arg], - force_quotes: bool, -) -> io::Result> { - const INVALID_ARGUMENT_ERROR: io::Error = - io::const_error!(io::ErrorKind::InvalidInput, r#"batch file arguments are invalid"#); - // Set the start of the command line to `cmd.exe /c "` - // It is necessary to surround the command in an extra pair of quotes, - // hence the trailing quote here. It will be closed after all arguments - // have been added. - // Using /e:ON enables "command extensions" which is essential for the `%` hack to work. - let mut cmd: Vec = "cmd.exe /e:ON /v:OFF /d /c \"".encode_utf16().collect(); - - // Push the script name surrounded by its quote pair. - cmd.push(b'"' as u16); - // Windows file names cannot contain a `"` character or end with `\\`. - // If the script name does then return an error. - if script.contains(&(b'"' as u16)) || script.last() == Some(&(b'\\' as u16)) { - return Err(io::const_error!( - io::ErrorKind::InvalidInput, - "Windows file names may not contain `\"` or end with `\\`" - )); - } - cmd.extend_from_slice(script.strip_suffix(&[0]).unwrap_or(script)); - cmd.push(b'"' as u16); - - // Append the arguments. - // FIXME: This needs tests to ensure that the arguments are properly - // reconstructed by the batch script by default. - for arg in args { - cmd.push(' ' as u16); - match arg { - Arg::Regular(arg_os) => { - let arg_bytes = arg_os.as_encoded_bytes(); - // Disallow \r and \n as they may truncate the arguments. - const DISALLOWED: &[u8] = b"\r\n"; - if arg_bytes.iter().any(|c| DISALLOWED.contains(c)) { - return Err(INVALID_ARGUMENT_ERROR); - } - append_bat_arg(&mut cmd, arg_os, force_quotes)?; - } - _ => { - // Raw arguments are passed on as-is. - // It's the user's responsibility to properly handle arguments in this case. - append_arg(&mut cmd, arg, force_quotes)?; - } - }; - } - - // Close the quote we left opened earlier. - cmd.push(b'"' as u16); - - Ok(cmd) -} - -/// Takes a path and tries to return a non-verbatim path. -/// -/// This is necessary because cmd.exe does not support verbatim paths. -pub(crate) fn to_user_path(path: &Path) -> io::Result> { - from_wide_to_user_path(to_u16s(path)?) -} -pub(crate) fn from_wide_to_user_path(mut path: Vec) -> io::Result> { - use super::fill_utf16_buf; - use crate::ptr; - - // UTF-16 encoded code points, used in parsing and building UTF-16 paths. - // All of these are in the ASCII range so they can be cast directly to `u16`. - const SEP: u16 = b'\\' as _; - const QUERY: u16 = b'?' as _; - const COLON: u16 = b':' as _; - const U: u16 = b'U' as _; - const N: u16 = b'N' as _; - const C: u16 = b'C' as _; - - // Early return if the path is too long to remove the verbatim prefix. - const LEGACY_MAX_PATH: usize = 260; - if path.len() > LEGACY_MAX_PATH { - return Ok(path); - } - - match &path[..] { - // `\\?\C:\...` => `C:\...` - [SEP, SEP, QUERY, SEP, _, COLON, SEP, ..] => unsafe { - let lpfilename = path[4..].as_ptr(); - fill_utf16_buf( - |buffer, size| c::GetFullPathNameW(lpfilename, size, buffer, ptr::null_mut()), - |full_path: &[u16]| { - if full_path == &path[4..path.len() - 1] { - let mut path: Vec = full_path.into(); - path.push(0); - path - } else { - path - } - }, - ) - }, - // `\\?\UNC\...` => `\\...` - [SEP, SEP, QUERY, SEP, U, N, C, SEP, ..] => unsafe { - // Change the `C` in `UNC\` to `\` so we can get a slice that starts with `\\`. - path[6] = b'\\' as u16; - let lpfilename = path[6..].as_ptr(); - fill_utf16_buf( - |buffer, size| c::GetFullPathNameW(lpfilename, size, buffer, ptr::null_mut()), - |full_path: &[u16]| { - if full_path == &path[6..path.len() - 1] { - let mut path: Vec = full_path.into(); - path.push(0); - path - } else { - // Restore the 'C' in "UNC". - path[6] = b'C' as u16; - path - } - }, - ) - }, - // For everything else, leave the path unchanged. - _ => get_long_path(path, false), - } -} diff --git a/library/std/src/sys/pal/windows/args/tests.rs b/library/std/src/sys/pal/windows/args/tests.rs deleted file mode 100644 index 484a90ab056..00000000000 --- a/library/std/src/sys/pal/windows/args/tests.rs +++ /dev/null @@ -1,91 +0,0 @@ -use super::*; -use crate::ffi::OsString; - -fn chk(string: &str, parts: &[&str]) { - let mut wide: Vec = OsString::from(string).encode_wide().collect(); - wide.push(0); - let parsed = - unsafe { parse_lp_cmd_line(WStrUnits::new(wide.as_ptr()), || OsString::from("TEST.EXE")) }; - let expected: Vec = parts.iter().map(|k| OsString::from(k)).collect(); - assert_eq!(parsed.as_slice(), expected.as_slice(), "{:?}", string); -} - -#[test] -fn empty() { - chk("", &["TEST.EXE"]); - chk("\0", &["TEST.EXE"]); -} - -#[test] -fn single_words() { - chk("EXE one_word", &["EXE", "one_word"]); - chk("EXE a", &["EXE", "a"]); - chk("EXE 😅", &["EXE", "😅"]); - chk("EXE 😅🤦", &["EXE", "😅🤦"]); -} - -#[test] -fn official_examples() { - chk(r#"EXE "abc" d e"#, &["EXE", "abc", "d", "e"]); - chk(r#"EXE a\\\b d"e f"g h"#, &["EXE", r"a\\\b", "de fg", "h"]); - chk(r#"EXE a\\\"b c d"#, &["EXE", r#"a\"b"#, "c", "d"]); - chk(r#"EXE a\\\\"b c" d e"#, &["EXE", r"a\\b c", "d", "e"]); -} - -#[test] -fn whitespace_behavior() { - chk(" test", &["", "test"]); - chk(" test", &["", "test"]); - chk(" test test2", &["", "test", "test2"]); - chk(" test test2", &["", "test", "test2"]); - chk("test test2 ", &["test", "test2"]); - chk("test test2 ", &["test", "test2"]); - chk("test ", &["test"]); -} - -#[test] -fn genius_quotes() { - chk(r#"EXE "" """#, &["EXE", "", ""]); - chk(r#"EXE "" """"#, &["EXE", "", r#"""#]); - chk( - r#"EXE "this is """all""" in the same argument""#, - &["EXE", r#"this is "all" in the same argument"#], - ); - chk(r#"EXE "a"""#, &["EXE", r#"a""#]); - chk(r#"EXE "a"" a"#, &["EXE", r#"a" a"#]); - // quotes cannot be escaped in command names - chk(r#""EXE" check"#, &["EXE", "check"]); - chk(r#""EXE check""#, &["EXE check"]); - chk(r#""EXE """for""" check"#, &["EXE for check"]); - chk(r#""EXE \"for\" check"#, &[r"EXE \for\ check"]); - chk(r#""EXE \" for \" check"#, &[r"EXE \", "for", r#"""#, "check"]); - chk(r#"E"X"E test"#, &["EXE", "test"]); - chk(r#"EX""E test"#, &["EXE", "test"]); -} - -// from https://daviddeley.com/autohotkey/parameters/parameters.htm#WINCRULESEX -#[test] -fn post_2008() { - chk("EXE CallMeIshmael", &["EXE", "CallMeIshmael"]); - chk(r#"EXE "Call Me Ishmael""#, &["EXE", "Call Me Ishmael"]); - chk(r#"EXE Cal"l Me I"shmael"#, &["EXE", "Call Me Ishmael"]); - chk(r#"EXE CallMe\"Ishmael"#, &["EXE", r#"CallMe"Ishmael"#]); - chk(r#"EXE "CallMe\"Ishmael""#, &["EXE", r#"CallMe"Ishmael"#]); - chk(r#"EXE "Call Me Ishmael\\""#, &["EXE", r"Call Me Ishmael\"]); - chk(r#"EXE "CallMe\\\"Ishmael""#, &["EXE", r#"CallMe\"Ishmael"#]); - chk(r#"EXE a\\\b"#, &["EXE", r"a\\\b"]); - chk(r#"EXE "a\\\b""#, &["EXE", r"a\\\b"]); - chk(r#"EXE "\"Call Me Ishmael\"""#, &["EXE", r#""Call Me Ishmael""#]); - chk(r#"EXE "C:\TEST A\\""#, &["EXE", r"C:\TEST A\"]); - chk(r#"EXE "\"C:\TEST A\\\"""#, &["EXE", r#""C:\TEST A\""#]); - chk(r#"EXE "a b c" d e"#, &["EXE", "a b c", "d", "e"]); - chk(r#"EXE "ab\"c" "\\" d"#, &["EXE", r#"ab"c"#, r"\", "d"]); - chk(r#"EXE a\\\b d"e f"g h"#, &["EXE", r"a\\\b", "de fg", "h"]); - chk(r#"EXE a\\\"b c d"#, &["EXE", r#"a\"b"#, "c", "d"]); - chk(r#"EXE a\\\\"b c" d e"#, &["EXE", r"a\\b c", "d", "e"]); - // Double Double Quotes - chk(r#"EXE "a b c"""#, &["EXE", r#"a b c""#]); - chk(r#"EXE """CallMeIshmael""" b c"#, &["EXE", r#""CallMeIshmael""#, "b", "c"]); - chk(r#"EXE """Call Me Ishmael""""#, &["EXE", r#""Call Me Ishmael""#]); - chk(r#"EXE """"Call Me Ishmael"" b c"#, &["EXE", r#""Call"#, "Me", "Ishmael", "b", "c"]); -} diff --git a/library/std/src/sys/pal/windows/mod.rs b/library/std/src/sys/pal/windows/mod.rs index bdf0cc2c59c..3c0a5c2de26 100644 --- a/library/std/src/sys/pal/windows/mod.rs +++ b/library/std/src/sys/pal/windows/mod.rs @@ -14,7 +14,6 @@ pub mod compat; pub mod api; -pub mod args; pub mod c; pub mod env; #[cfg(not(target_vendor = "win7"))] diff --git a/library/std/src/sys/pal/xous/args.rs b/library/std/src/sys/pal/xous/args.rs deleted file mode 100644 index 00c44ca220a..00000000000 --- a/library/std/src/sys/pal/xous/args.rs +++ /dev/null @@ -1,53 +0,0 @@ -use crate::ffi::OsString; -use crate::sys::pal::xous::os::get_application_parameters; -use crate::sys::pal::xous::os::params::ArgumentList; -use crate::{fmt, vec}; - -pub struct Args { - parsed_args_list: vec::IntoIter, -} - -pub fn args() -> Args { - let Some(params) = get_application_parameters() else { - return Args { parsed_args_list: vec![].into_iter() }; - }; - - for param in params { - if let Ok(args) = ArgumentList::try_from(¶m) { - let mut parsed_args = vec![]; - for arg in args { - parsed_args.push(arg.into()); - } - return Args { parsed_args_list: parsed_args.into_iter() }; - } - } - Args { parsed_args_list: vec![].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 { - self.parsed_args_list.next() - } - fn size_hint(&self) -> (usize, Option) { - self.parsed_args_list.size_hint() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.parsed_args_list.next_back() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.parsed_args_list.len() - } -} diff --git a/library/std/src/sys/pal/xous/mod.rs b/library/std/src/sys/pal/xous/mod.rs index 58926e2beb1..4f652d3f130 100644 --- a/library/std/src/sys/pal/xous/mod.rs +++ b/library/std/src/sys/pal/xous/mod.rs @@ -1,6 +1,5 @@ #![forbid(unsafe_op_in_unsafe_fn)] -pub mod args; #[path = "../unsupported/env.rs"] pub mod env; pub mod os; diff --git a/library/std/src/sys/pal/zkvm/args.rs b/library/std/src/sys/pal/zkvm/args.rs deleted file mode 100644 index 47857f6c448..00000000000 --- a/library/std/src/sys/pal/zkvm/args.rs +++ /dev/null @@ -1,81 +0,0 @@ -use super::{WORD_SIZE, abi}; -use crate::ffi::OsString; -use crate::fmt; -use crate::sys::os_str; -use crate::sys_common::FromInner; - -pub struct Args { - i_forward: usize, - i_back: usize, - count: usize, -} - -pub fn args() -> Args { - let count = unsafe { abi::sys_argc() }; - Args { i_forward: 0, i_back: 0, count } -} - -impl Args { - /// Use sys_argv to get the arg at the requested index. Does not check that i is less than argc - /// and will not return if the index is out of bounds. - fn argv(i: usize) -> OsString { - let arg_len = unsafe { abi::sys_argv(crate::ptr::null_mut(), 0, i) }; - - let arg_len_words = (arg_len + WORD_SIZE - 1) / WORD_SIZE; - let words = unsafe { abi::sys_alloc_words(arg_len_words) }; - - let arg_len2 = unsafe { abi::sys_argv(words, arg_len_words, i) }; - debug_assert_eq!(arg_len, arg_len2); - - // Convert to OsString. - // - // FIXME: We can probably get rid of the extra copy here if we - // reimplement "os_str" instead of just using the generic unix - // "os_str". - let arg_bytes: &[u8] = - unsafe { crate::slice::from_raw_parts(words.cast() as *const u8, arg_len) }; - OsString::from_inner(os_str::Buf { inner: arg_bytes.to_vec() }) - } -} - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - f.debug_list().finish() - } -} - -impl Iterator for Args { - type Item = OsString; - - fn next(&mut self) -> Option { - if self.i_forward >= self.count - self.i_back { - None - } else { - let arg = Self::argv(self.i_forward); - self.i_forward += 1; - Some(arg) - } - } - - fn size_hint(&self) -> (usize, Option) { - (self.count, Some(self.count)) - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.count - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - if self.i_back >= self.count - self.i_forward { - None - } else { - let arg = Self::argv(self.count - 1 - self.i_back); - self.i_back += 1; - Some(arg) - } - } -} diff --git a/library/std/src/sys/pal/zkvm/mod.rs b/library/std/src/sys/pal/zkvm/mod.rs index 4659dad16e8..ebd7b036779 100644 --- a/library/std/src/sys/pal/zkvm/mod.rs +++ b/library/std/src/sys/pal/zkvm/mod.rs @@ -8,11 +8,9 @@ //! will likely change over time. #![forbid(unsafe_op_in_unsafe_fn)] -const WORD_SIZE: usize = size_of::(); +pub const WORD_SIZE: usize = size_of::(); pub mod abi; -#[path = "../zkvm/args.rs"] -pub mod args; pub mod env; pub mod os; #[path = "../unsupported/pipe.rs"] -- cgit 1.4.1-3-g733a5 From e014fd6b872edd1af42b00d33a963632107618f9 Mon Sep 17 00:00:00 2001 From: Thalia Archibald Date: Sat, 12 Apr 2025 02:37:24 -0700 Subject: Use unsupported args for espidf and vita --- library/std/src/sys/args/mod.rs | 2 +- library/std/src/sys/args/unix.rs | 13 ------------- library/std/src/sys/pal/unix/mod.rs | 2 ++ 3 files changed, 3 insertions(+), 14 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/args/mod.rs b/library/std/src/sys/args/mod.rs index 6edcafe2c4c..f24d6eb123e 100644 --- a/library/std/src/sys/args/mod.rs +++ b/library/std/src/sys/args/mod.rs @@ -3,7 +3,7 @@ #![forbid(unsafe_op_in_unsafe_fn)] cfg_if::cfg_if! { - if #[cfg(target_family = "unix")] { + if #[cfg(all(target_family = "unix", not(any(target_os = "espidf", target_os = "vita"))))] { mod unix; pub use unix::*; } else if #[cfg(target_family = "windows")] { diff --git a/library/std/src/sys/args/unix.rs b/library/std/src/sys/args/unix.rs index 4c30609c099..13c459f682f 100644 --- a/library/std/src/sys/args/unix.rs +++ b/library/std/src/sys/args/unix.rs @@ -226,16 +226,3 @@ mod imp { (argc as isize, argv.cast()) } } - -#[cfg(any(target_os = "espidf", target_os = "vita"))] -mod imp { - use crate::ffi::c_char; - use crate::ptr; - - #[inline(always)] - pub unsafe fn init(_argc: isize, _argv: *const *const u8) {} - - pub fn argc_argv() -> (isize, *const *const c_char) { - (0, ptr::null()) - } -} diff --git a/library/std/src/sys/pal/unix/mod.rs b/library/std/src/sys/pal/unix/mod.rs index 20078839a14..f8733eb6119 100644 --- a/library/std/src/sys/pal/unix/mod.rs +++ b/library/std/src/sys/pal/unix/mod.rs @@ -26,6 +26,7 @@ pub mod time; pub fn init(_argc: isize, _argv: *const *const u8, _sigpipe: u8) {} #[cfg(not(target_os = "espidf"))] +#[cfg_attr(target_os = "vita", allow(unused_variables))] // SAFETY: must be called only once during runtime initialization. // NOTE: this is not guaranteed to run, for example when Rust code is called externally. // See `fn init()` in `library/std/src/rt.rs` for docs on `sigpipe`. @@ -46,6 +47,7 @@ pub unsafe fn init(argc: isize, argv: *const *const u8, sigpipe: u8) { reset_sigpipe(sigpipe); stack_overflow::init(); + #[cfg(not(target_os = "vita"))] crate::sys::args::init(argc, argv); // Normally, `thread::spawn` will call `Thread::set_name` but since this thread -- cgit 1.4.1-3-g733a5 From bea202253ec03e2fdc4c97915a24ebe530277edf Mon Sep 17 00:00:00 2001 From: Thalia Archibald Date: Sat, 12 Apr 2025 04:29:12 -0700 Subject: Unify owned Args types between platforms --- library/std/src/sys/args/common.rs | 43 +++++++++++++++++++++++++++++++++ library/std/src/sys/args/hermit.rs | 43 +++++---------------------------- library/std/src/sys/args/uefi.rs | 47 +++++-------------------------------- library/std/src/sys/args/unix.rs | 44 +++++----------------------------- library/std/src/sys/args/wasi.rs | 40 ++++--------------------------- library/std/src/sys/args/windows.rs | 40 +++++-------------------------- library/std/src/sys/args/xous.rs | 42 +++++---------------------------- 7 files changed, 77 insertions(+), 222 deletions(-) create mode 100644 library/std/src/sys/args/common.rs (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/args/common.rs b/library/std/src/sys/args/common.rs new file mode 100644 index 00000000000..43ac5e95923 --- /dev/null +++ b/library/std/src/sys/args/common.rs @@ -0,0 +1,43 @@ +use crate::ffi::OsString; +use crate::{fmt, vec}; + +pub struct Args { + iter: vec::IntoIter, +} + +impl !Send for Args {} +impl !Sync for Args {} + +impl Args { + pub(super) fn new(args: Vec) -> Self { + Args { iter: args.into_iter() } + } +} + +impl fmt::Debug for Args { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + self.iter.as_slice().fmt(f) + } +} + +impl Iterator for Args { + type Item = OsString; + fn next(&mut self) -> Option { + self.iter.next() + } + fn size_hint(&self) -> (usize, Option) { + self.iter.size_hint() + } +} + +impl ExactSizeIterator for Args { + fn len(&self) -> usize { + self.iter.len() + } +} + +impl DoubleEndedIterator for Args { + fn next_back(&mut self) -> Option { + self.iter.next_back() + } +} diff --git a/library/std/src/sys/args/hermit.rs b/library/std/src/sys/args/hermit.rs index 44024260277..ddd644a5540 100644 --- a/library/std/src/sys/args/hermit.rs +++ b/library/std/src/sys/args/hermit.rs @@ -1,8 +1,12 @@ use crate::ffi::{CStr, OsString, c_char}; use crate::os::hermit::ffi::OsStringExt; +use crate::ptr; use crate::sync::atomic::Ordering::{Acquire, Relaxed, Release}; use crate::sync::atomic::{AtomicIsize, AtomicPtr}; -use crate::{fmt, ptr, vec}; + +#[path = "common.rs"] +mod common; +pub use common::Args; static ARGC: AtomicIsize = AtomicIsize::new(0); static ARGV: AtomicPtr<*const u8> = AtomicPtr::new(ptr::null_mut()); @@ -27,40 +31,5 @@ pub fn args() -> Args { }) .collect(); - Args { iter: args.into_iter() } -} - -pub struct Args { - iter: vec::IntoIter, -} - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - self.iter.as_slice().fmt(f) - } -} - -impl !Send for Args {} -impl !Sync for Args {} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - self.iter.next() - } - fn size_hint(&self) -> (usize, Option) { - self.iter.size_hint() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.iter.len() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.iter.next_back() - } + Args::new(args) } diff --git a/library/std/src/sys/args/uefi.rs b/library/std/src/sys/args/uefi.rs index 4e7412937bd..84406c7f69d 100644 --- a/library/std/src/sys/args/uefi.rs +++ b/library/std/src/sys/args/uefi.rs @@ -4,11 +4,10 @@ use crate::env::current_exe; use crate::ffi::OsString; use crate::iter::Iterator; use crate::sys::pal::helpers; -use crate::{fmt, vec}; -pub struct Args { - parsed_args_list: vec::IntoIter, -} +#[path = "common.rs"] +mod common; +pub use common::Args; pub fn args() -> Args { let lazy_current_exe = || Vec::from([current_exe().map(Into::into).unwrap_or_default()]); @@ -22,51 +21,17 @@ pub fn args() -> Args { 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::() || lp_size % size_of::() != 0 { - return Args { parsed_args_list: lazy_current_exe().into_iter() }; + return Args::new(lazy_current_exe()); } let lp_size = lp_size / size_of::(); 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() }; + return Args::new(lazy_current_exe()); } 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 { - self.parsed_args_list.next() - } - - fn size_hint(&self) -> (usize, Option) { - 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 { - self.parsed_args_list.next_back() - } + Args::new(parse_lp_cmd_line(lp_cmd_line).unwrap_or_else(lazy_current_exe)) } /// Implements the UEFI command-line argument parsing algorithm. diff --git a/library/std/src/sys/args/unix.rs b/library/std/src/sys/args/unix.rs index 13c459f682f..7d7815c6dff 100644 --- a/library/std/src/sys/args/unix.rs +++ b/library/std/src/sys/args/unix.rs @@ -5,9 +5,12 @@ #![allow(dead_code)] // runtime init functions not used during testing -use crate::ffi::{CStr, OsString}; +use crate::ffi::CStr; use crate::os::unix::ffi::OsStringExt; -use crate::{fmt, vec}; + +#[path = "common.rs"] +mod common; +pub use common::Args; /// One-time global initialization. pub unsafe fn init(argc: isize, argv: *const *const u8) { @@ -55,42 +58,7 @@ pub fn args() -> Args { vec.push(OsStringExt::from_vec(cstr.to_bytes().to_vec())); } - Args { iter: vec.into_iter() } -} - -pub struct Args { - iter: vec::IntoIter, -} - -impl !Send for Args {} -impl !Sync for Args {} - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - self.iter.as_slice().fmt(f) - } -} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - self.iter.next() - } - fn size_hint(&self) -> (usize, Option) { - self.iter.size_hint() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.iter.len() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.iter.next_back() - } + Args::new(vec) } #[cfg(any( diff --git a/library/std/src/sys/args/wasi.rs b/library/std/src/sys/args/wasi.rs index 52cfa202af8..4795789e4c7 100644 --- a/library/std/src/sys/args/wasi.rs +++ b/library/std/src/sys/args/wasi.rs @@ -2,18 +2,14 @@ use crate::ffi::{CStr, OsStr, OsString}; use crate::os::wasi::ffi::OsStrExt; -use crate::{fmt, vec}; -pub struct Args { - iter: vec::IntoIter, -} - -impl !Send for Args {} -impl !Sync for Args {} +#[path = "common.rs"] +mod common; +pub use common::Args; /// Returns the command line arguments pub fn args() -> Args { - Args { iter: maybe_args().unwrap_or(Vec::new()).into_iter() } + Args::new(maybe_args().unwrap_or(Vec::new())) } fn maybe_args() -> Option> { @@ -31,31 +27,3 @@ fn maybe_args() -> Option> { Some(ret) } } - -impl fmt::Debug for Args { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - self.iter.as_slice().fmt(f) - } -} - -impl Iterator for Args { - type Item = OsString; - fn next(&mut self) -> Option { - self.iter.next() - } - fn size_hint(&self) -> (usize, Option) { - self.iter.size_hint() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.iter.len() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.iter.next_back() - } -} diff --git a/library/std/src/sys/args/windows.rs b/library/std/src/sys/args/windows.rs index 55bd6864cae..47f0e5f2d05 100644 --- a/library/std/src/sys/args/windows.rs +++ b/library/std/src/sys/args/windows.rs @@ -16,7 +16,11 @@ use crate::sys::path::get_long_path; use crate::sys::{c, to_u16s}; use crate::sys_common::AsInner; use crate::sys_common::wstr::WStrUnits; -use crate::{fmt, io, iter, ptr, vec}; +use crate::{io, iter, ptr}; + +#[path = "common.rs"] +mod common; +pub use common::Args; pub fn args() -> Args { // SAFETY: `GetCommandLineW` returns a pointer to a null terminated UTF-16 @@ -27,7 +31,7 @@ pub fn args() -> Args { current_exe().map(PathBuf::into_os_string).unwrap_or_else(|_| OsString::new()) }); - Args { parsed_args_list: parsed_args_list.into_iter() } + Args::new(parsed_args_list) } } @@ -153,38 +157,6 @@ fn parse_lp_cmd_line<'a, F: Fn() -> OsString>( ret_val } -pub struct Args { - parsed_args_list: vec::IntoIter, -} - -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 { - self.parsed_args_list.next() - } - fn size_hint(&self) -> (usize, Option) { - self.parsed_args_list.size_hint() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.parsed_args_list.next_back() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.parsed_args_list.len() - } -} - #[derive(Debug)] pub(crate) enum Arg { /// Add quotes (if needed) diff --git a/library/std/src/sys/args/xous.rs b/library/std/src/sys/args/xous.rs index 2dd37d53422..09a47283d65 100644 --- a/library/std/src/sys/args/xous.rs +++ b/library/std/src/sys/args/xous.rs @@ -1,15 +1,13 @@ -use crate::ffi::OsString; use crate::sys::pal::os::get_application_parameters; use crate::sys::pal::os::params::ArgumentList; -use crate::{fmt, vec}; -pub struct Args { - parsed_args_list: vec::IntoIter, -} +#[path = "common.rs"] +mod common; +pub use common::Args; pub fn args() -> Args { let Some(params) = get_application_parameters() else { - return Args { parsed_args_list: vec![].into_iter() }; + return Args::new(vec![]); }; for param in params { @@ -18,36 +16,8 @@ pub fn args() -> Args { for arg in args { parsed_args.push(arg.into()); } - return Args { parsed_args_list: parsed_args.into_iter() }; + return Args::new(parsed_args); } } - Args { parsed_args_list: vec![].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 { - self.parsed_args_list.next() - } - fn size_hint(&self) -> (usize, Option) { - self.parsed_args_list.size_hint() - } -} - -impl DoubleEndedIterator for Args { - fn next_back(&mut self) -> Option { - self.parsed_args_list.next_back() - } -} - -impl ExactSizeIterator for Args { - fn len(&self) -> usize { - self.parsed_args_list.len() - } + Args::new(vec![]) } -- cgit 1.4.1-3-g733a5 From 15a93c5683fa4a5c7a10b3e3385178ff0fd0fed6 Mon Sep 17 00:00:00 2001 From: Huang Qi Date: Sun, 13 Apr 2025 21:15:15 +0800 Subject: std/thread: Use default stack size from menuconfig for NuttX * Update comments to clarify the usage of zero as an indication for default stack size configuration * Adjust conditional compilation to reflect the changes in stack size handling for the NuttX platform This change improves clarity and consistency in stack size configuration across platforms. Signed-off-by: Huang Qi --- library/std/src/sys/pal/unix/thread.rs | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/pal/unix/thread.rs b/library/std/src/sys/pal/unix/thread.rs index 9078dd1c231..4f1517b7ef4 100644 --- a/library/std/src/sys/pal/unix/thread.rs +++ b/library/std/src/sys/pal/unix/thread.rs @@ -8,14 +8,19 @@ use crate::sys::weak::weak; use crate::sys::{os, stack_overflow}; use crate::time::Duration; use crate::{cmp, io, ptr}; -#[cfg(not(any(target_os = "l4re", target_os = "vxworks", target_os = "espidf")))] +#[cfg(not(any( + target_os = "l4re", + target_os = "vxworks", + target_os = "espidf", + target_os = "nuttx" +)))] pub const DEFAULT_MIN_STACK_SIZE: usize = 2 * 1024 * 1024; #[cfg(target_os = "l4re")] pub const DEFAULT_MIN_STACK_SIZE: usize = 1024 * 1024; #[cfg(target_os = "vxworks")] pub const DEFAULT_MIN_STACK_SIZE: usize = 256 * 1024; -#[cfg(target_os = "espidf")] -pub const DEFAULT_MIN_STACK_SIZE: usize = 0; // 0 indicates that the stack size configured in the ESP-IDF menuconfig system should be used +#[cfg(any(target_os = "espidf", target_os = "nuttx"))] +pub const DEFAULT_MIN_STACK_SIZE: usize = 0; // 0 indicates that the stack size configured in the ESP-IDF/NuttX menuconfig system should be used #[cfg(target_os = "fuchsia")] mod zircon { @@ -52,10 +57,10 @@ impl Thread { let mut attr: mem::MaybeUninit = mem::MaybeUninit::uninit(); assert_eq!(libc::pthread_attr_init(attr.as_mut_ptr()), 0); - #[cfg(target_os = "espidf")] + #[cfg(any(target_os = "espidf", target_os = "nuttx"))] if stack > 0 { // Only set the stack if a non-zero value is passed - // 0 is used as an indication that the default stack size configured in the ESP-IDF menuconfig system should be used + // 0 is used as an indication that the default stack size configured in the ESP-IDF/NuttX menuconfig system should be used assert_eq!( libc::pthread_attr_setstacksize( attr.as_mut_ptr(), @@ -65,7 +70,7 @@ impl Thread { ); } - #[cfg(not(target_os = "espidf"))] + #[cfg(not(any(target_os = "espidf", target_os = "nuttx")))] { let stack_size = cmp::max(stack, min_stack_size(attr.as_ptr())); -- cgit 1.4.1-3-g733a5 From a404015775276f2e21e16c69fab36cc8228d51f6 Mon Sep 17 00:00:00 2001 From: Ayush Singh Date: Tue, 8 Apr 2025 12:28:42 +0530 Subject: std: sys: process: uefi: Use NULL stdin by default According to the docs in `Command::output`: > By default, stdout and stderr are captured (and used to provide the resulting output). Stdin is not inherited from the parent and any attempt by the child process to read from the stdin stream will result in the stream immediately closing. This was being violated by UEFI which was inheriting stdin by default. While the docs don't explicitly state that the default should be NULL, the behaviour seems like reading from NULL. UEFI however, has a bit of a problem. The `EFI_SIMPLE_TEXT_INPUT_PROTOCOL` only provides support for reading 1 key press. This means that you either get an error, or it is assumed that the keypress was read successfully. So there is no way to have a successful read of length 0. Currently, I am returning UNSUPPORTED error when trying to read from NULL stdin. On linux however, you will get a read of length 0 for Null stdin. One possible way to get around this is to translate one of the UEFI errors to a read 0 (Maybe unsupported?). It is also possible to have a non-standard error code, but well, not sure if we go that route. Alternatively, if meaning of Stdio::Null is platform dependent, it should be fine to keep the current behaviour of returning an error. Signed-off-by: Ayush Singh --- library/std/src/sys/process/uefi.rs | 112 +++++++++++++++++++++++++++++++++++- 1 file changed, 109 insertions(+), 3 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/process/uefi.rs b/library/std/src/sys/process/uefi.rs index b46418ae9bb..fb5aa5ca266 100644 --- a/library/std/src/sys/process/uefi.rs +++ b/library/std/src/sys/process/uefi.rs @@ -1,4 +1,4 @@ -use r_efi::protocols::simple_text_output; +use r_efi::protocols::{simple_text_input, simple_text_output}; use crate::collections::BTreeMap; pub use crate::ffi::OsString as EnvKey; @@ -122,6 +122,22 @@ impl Command { } } + fn create_stdin( + s: Stdio, + ) -> io::Result>> { + match s { + Stdio::Null => unsafe { + helpers::OwnedProtocol::create( + uefi_command_internal::InputProtocol::null(), + simple_text_input::PROTOCOL_GUID, + ) + } + .map(Some), + Stdio::Inherit => Ok(None), + Stdio::MakePipe => unsupported(), + } + } + pub fn output(&mut self) -> io::Result<(ExitStatus, Vec, Vec)> { let mut cmd = uefi_command_internal::Image::load_image(&self.prog)?; @@ -149,6 +165,14 @@ impl Command { cmd.stderr_inherit() }; + // Setup Stdin + let stdin = Self::create_stdin(Stdio::Null)?; + if let Some(con) = stdin { + cmd.stdin_init(con) + } else { + cmd.stdin_inherit() + }; + let env = env_changes(&self.env); // Set any new vars @@ -334,7 +358,7 @@ impl<'a> fmt::Debug for CommandArgs<'a> { #[allow(dead_code)] mod uefi_command_internal { - use r_efi::protocols::{loaded_image, simple_text_output}; + use r_efi::protocols::{loaded_image, simple_text_input, simple_text_output}; use crate::ffi::{OsStr, OsString}; use crate::io::{self, const_error}; @@ -350,6 +374,7 @@ mod uefi_command_internal { handle: NonNull, stdout: Option>, stderr: Option>, + stdin: Option>, st: OwnedTable, args: Option<(*mut u16, usize)>, } @@ -384,7 +409,14 @@ mod uefi_command_internal { helpers::open_protocol(child_handle, loaded_image::PROTOCOL_GUID).unwrap(); let st = OwnedTable::from_table(unsafe { (*loaded_image.as_ptr()).system_table }); - Ok(Self { handle: child_handle, stdout: None, stderr: None, st, args: None }) + Ok(Self { + handle: child_handle, + stdout: None, + stderr: None, + stdin: None, + st, + args: None, + }) } } @@ -445,6 +477,17 @@ mod uefi_command_internal { } } + fn set_stdin( + &mut self, + handle: r_efi::efi::Handle, + protocol: *mut simple_text_input::Protocol, + ) { + unsafe { + (*self.st.as_mut_ptr()).console_in_handle = handle; + (*self.st.as_mut_ptr()).con_in = protocol; + } + } + pub fn stdout_init(&mut self, protocol: helpers::OwnedProtocol) { self.set_stdout( protocol.handle().as_ptr(), @@ -471,6 +514,19 @@ mod uefi_command_internal { unsafe { self.set_stderr((*st.as_ptr()).standard_error_handle, (*st.as_ptr()).std_err) } } + pub(crate) fn stdin_init(&mut self, protocol: helpers::OwnedProtocol) { + self.set_stdin( + protocol.handle().as_ptr(), + protocol.as_ref() as *const InputProtocol as *mut simple_text_input::Protocol, + ); + self.stdin = Some(protocol); + } + + pub(crate) fn stdin_inherit(&mut self) { + let st: NonNull = system_table().cast(); + unsafe { self.set_stdin((*st.as_ptr()).console_in_handle, (*st.as_ptr()).con_in) } + } + pub fn stderr(&self) -> io::Result> { match &self.stderr { Some(stderr) => stderr.as_ref().utf8(), @@ -722,6 +778,56 @@ mod uefi_command_internal { } } + #[repr(C)] + pub(crate) struct InputProtocol { + reset: simple_text_input::ProtocolReset, + read_key_stroke: simple_text_input::ProtocolReadKeyStroke, + wait_for_key: r_efi::efi::Event, + } + + impl InputProtocol { + pub(crate) fn null() -> Self { + let evt = helpers::OwnedEvent::new( + r_efi::efi::EVT_NOTIFY_WAIT, + r_efi::efi::TPL_CALLBACK, + Some(Self::empty_notify), + None, + ) + .unwrap(); + + Self { + reset: Self::null_reset, + read_key_stroke: Self::null_read_key, + wait_for_key: evt.into_raw(), + } + } + + extern "efiapi" fn null_reset( + _: *mut simple_text_input::Protocol, + _: r_efi::efi::Boolean, + ) -> r_efi::efi::Status { + r_efi::efi::Status::SUCCESS + } + + extern "efiapi" fn null_read_key( + _: *mut simple_text_input::Protocol, + _: *mut simple_text_input::InputKey, + ) -> r_efi::efi::Status { + r_efi::efi::Status::UNSUPPORTED + } + + extern "efiapi" fn empty_notify(_: r_efi::efi::Event, _: *mut crate::ffi::c_void) {} + } + + impl Drop for InputProtocol { + fn drop(&mut self) { + // Close wait_for_key + unsafe { + let _ = helpers::OwnedEvent::from_raw(self.wait_for_key); + } + } + } + pub fn create_args(prog: &OsStr, args: &[OsString]) -> Box<[u16]> { const QUOTE: u16 = 0x0022; const SPACE: u16 = 0x0020; -- cgit 1.4.1-3-g733a5 From af25995d11f8630a071a7f4776f0798d22942a9f Mon Sep 17 00:00:00 2001 From: Ayush Singh Date: Sun, 13 Apr 2025 20:11:44 +0530 Subject: std: sys: stdio: uefi: Tread UNSUPPORTED Status as read(0) Allows implementing Stdio::Null for Command in a deterministic manner. Signed-off-by: Ayush Singh --- library/std/src/sys/stdio/uefi.rs | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/stdio/uefi.rs b/library/std/src/sys/stdio/uefi.rs index 257e321dd03..ccd6bf658b0 100644 --- a/library/std/src/sys/stdio/uefi.rs +++ b/library/std/src/sys/stdio/uefi.rs @@ -142,8 +142,12 @@ impl io::Write for Stderr { // UTF-16 character should occupy 4 bytes at most in UTF-8 pub const STDIN_BUF_SIZE: usize = 4; -pub fn is_ebadf(_err: &io::Error) -> bool { - false +pub fn is_ebadf(err: &io::Error) -> bool { + if let Some(x) = err.raw_os_error() { + r_efi::efi::Status::UNSUPPORTED.as_usize() == x + } else { + false + } } pub fn panic_output() -> Option { -- cgit 1.4.1-3-g733a5 From d994fef7495ecbc1e18c95588253869f3a765edf Mon Sep 17 00:00:00 2001 From: Ayush Singh Date: Thu, 10 Apr 2025 10:40:50 +0530 Subject: std: sys: process: uefi: Allow specifying Stdin Stdio::MakePipe is not supported. For Stdio::Null, return UNSUPPORTED. This is treated as read(0). Additionally, have infinte loop on the notify function to prevent wait_for_key from returning. Signed-off-by: Ayush Singh --- library/std/src/sys/process/uefi.rs | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/process/uefi.rs b/library/std/src/sys/process/uefi.rs index fb5aa5ca266..5f922292d05 100644 --- a/library/std/src/sys/process/uefi.rs +++ b/library/std/src/sys/process/uefi.rs @@ -23,6 +23,7 @@ pub struct Command { args: Vec, stdout: Option, stderr: Option, + stdin: Option, env: CommandEnv, } @@ -48,6 +49,7 @@ impl Command { args: Vec::new(), stdout: None, stderr: None, + stdin: None, env: Default::default(), } } @@ -64,8 +66,8 @@ impl Command { panic!("unsupported") } - pub fn stdin(&mut self, _stdin: Stdio) { - panic!("unsupported") + pub fn stdin(&mut self, stdin: Stdio) { + self.stdin = Some(stdin); } pub fn stdout(&mut self, stdout: Stdio) { @@ -166,7 +168,8 @@ impl Command { }; // Setup Stdin - let stdin = Self::create_stdin(Stdio::Null)?; + let stdin = self.stdin.unwrap_or(Stdio::Null); + let stdin = Self::create_stdin(stdin)?; if let Some(con) = stdin { cmd.stdin_init(con) } else { -- cgit 1.4.1-3-g733a5 From 188d44dd6e4949606e31c8009dffa4773ea25214 Mon Sep 17 00:00:00 2001 From: 0x79de <0x79de@gmail.com> Date: Mon, 14 Apr 2025 21:25:48 +0300 Subject: Fix: Map EOPNOTSUPP to ErrorKind::Unsupported on Unix This change maps the EOPNOTSUPP errno value (95) to std::io::ErrorKind::Unsupported in the decode_error_kind function for Unix platforms. Previously, it was incorrectly mapped to ErrorKind::Uncategorized. Fixes #139803 --- library/std/src/sys/pal/unix/mod.rs | 1 + 1 file changed, 1 insertion(+) (limited to 'library/std/src/sys') diff --git a/library/std/src/sys/pal/unix/mod.rs b/library/std/src/sys/pal/unix/mod.rs index f8733eb6119..3a790d9c868 100644 --- a/library/std/src/sys/pal/unix/mod.rs +++ b/library/std/src/sys/pal/unix/mod.rs @@ -274,6 +274,7 @@ pub fn decode_error_kind(errno: i32) -> ErrorKind { libc::ETXTBSY => ExecutableFileBusy, libc::EXDEV => CrossesDevices, libc::EINPROGRESS => InProgress, + libc::EOPNOTSUPP => Unsupported, libc::EACCES | libc::EPERM => PermissionDenied, -- cgit 1.4.1-3-g733a5