From c99224536152cff14639b64a3fa4f7a215fd037c Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sat, 9 Aug 2025 23:52:11 +0100 Subject: refactor: Include table sizes in comment at top of `unicode_data.rs` To make changes in table size obvious from git diffs --- src/tools/unicode-table-generator/src/main.rs | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-) (limited to 'src') diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index 6cdb82a87bd..c1017142097 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -239,6 +239,11 @@ fn main() { std::fs::write(&path, generate_tests(&write_location, ranges_by_property)).unwrap(); } + let mut table_file = String::new(); + table_file.push_str( + "///! This file is generated by `./x run src/tools/unicode-table-generator`; do not edit manually!\n", + ); + let mut total_bytes = 0; let mut modules = Vec::new(); for (property, ranges) in ranges_by_property { @@ -252,8 +257,8 @@ fn main() { } modules.push((property.to_lowercase().to_string(), emitter.file)); - println!( - "{:15}: {} bytes, {} codepoints in {} ranges ({} - {}) using {}", + table_file.push_str(&format!( + "// {:16}: {:5} bytes, {:6} codepoints in {:3} ranges (U+{:06X} - U+{:06X}) using {}\n", property, emitter.bytes_used, datapoints, @@ -261,15 +266,10 @@ fn main() { ranges.first().unwrap().start, ranges.last().unwrap().end, emitter.desc, - ); + )); total_bytes += emitter.bytes_used; } - - let mut table_file = String::new(); - - table_file.push_str( - "///! This file is generated by `./x run src/tools/unicode-table-generator`; do not edit manually!\n", - ); + table_file.push_str(&format!("// {:16}: {:5} bytes\n", "Total", total_bytes)); // Include the range search function table_file.push('\n'); @@ -296,8 +296,6 @@ fn main() { } std::fs::write(&write_location, format!("{}\n", table_file.trim_end())).unwrap(); - - println!("Total table sizes: {total_bytes} bytes"); } fn version() -> String { -- cgit 1.4.1-3-g733a5 From 69e1974bb0bfbcc679d29950b1e4540cd0b9b3ee Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 00:18:04 +0100 Subject: refactor: Include size of case conversion tables Include the sizes of the `to_lowercase` and `to_uppercase` tables in the total size calculations. --- library/core/src/unicode/unicode_data.rs | 12 ++++--- .../unicode-table-generator/src/case_mapping.rs | 41 +++++++++++++++------- src/tools/unicode-table-generator/src/main.rs | 7 +++- 3 files changed, 42 insertions(+), 18 deletions(-) (limited to 'src') diff --git a/library/core/src/unicode/unicode_data.rs b/library/core/src/unicode/unicode_data.rs index 6059f7d6450..787efcc0914 100644 --- a/library/core/src/unicode/unicode_data.rs +++ b/library/core/src/unicode/unicode_data.rs @@ -8,7 +8,9 @@ // N : 457 bytes, 1911 codepoints in 144 ranges (U+000030 - U+01FBFA) using skiplist // Uppercase : 799 bytes, 1978 codepoints in 656 ranges (U+000041 - U+01F18A) using bitset // White_Space : 256 bytes, 25 codepoints in 10 ranges (U+000009 - U+003001) using cascading -// Total : 6530 bytes +// to_lower : 11484 bytes +// to_upper : 13432 bytes +// Total : 31446 bytes #[inline(always)] const fn bitset_search< @@ -782,7 +784,7 @@ pub mod conversions { } } - static LOWERCASE_TABLE: &[(char, u32)] = &[ + static LOWERCASE_TABLE: &[(char, u32); 1434] = &[ ('\u{c0}', 224), ('\u{c1}', 225), ('\u{c2}', 226), ('\u{c3}', 227), ('\u{c4}', 228), ('\u{c5}', 229), ('\u{c6}', 230), ('\u{c7}', 231), ('\u{c8}', 232), ('\u{c9}', 233), ('\u{ca}', 234), ('\u{cb}', 235), ('\u{cc}', 236), ('\u{cd}', 237), ('\u{ce}', 238), @@ -1132,11 +1134,11 @@ pub mod conversions { ('\u{1e921}', 125251), ]; - static LOWERCASE_TABLE_MULTI: &[[char; 3]] = &[ + static LOWERCASE_TABLE_MULTI: &[[char; 3]; 1] = &[ ['i', '\u{307}', '\u{0}'], ]; - static UPPERCASE_TABLE: &[(char, u32)] = &[ + static UPPERCASE_TABLE: &[(char, u32); 1526] = &[ ('\u{b5}', 924), ('\u{df}', 4194304), ('\u{e0}', 192), ('\u{e1}', 193), ('\u{e2}', 194), ('\u{e3}', 195), ('\u{e4}', 196), ('\u{e5}', 197), ('\u{e6}', 198), ('\u{e7}', 199), ('\u{e8}', 200), ('\u{e9}', 201), ('\u{ea}', 202), ('\u{eb}', 203), ('\u{ec}', 204), @@ -1509,7 +1511,7 @@ pub mod conversions { ('\u{1e941}', 125215), ('\u{1e942}', 125216), ('\u{1e943}', 125217), ]; - static UPPERCASE_TABLE_MULTI: &[[char; 3]] = &[ + static UPPERCASE_TABLE_MULTI: &[[char; 3]; 102] = &[ ['S', 'S', '\u{0}'], ['\u{2bc}', 'N', '\u{0}'], ['J', '\u{30c}', '\u{0}'], ['\u{399}', '\u{308}', '\u{301}'], ['\u{3a5}', '\u{308}', '\u{301}'], ['\u{535}', '\u{552}', '\u{0}'], ['H', '\u{331}', '\u{0}'], ['T', '\u{308}', '\u{0}'], diff --git a/src/tools/unicode-table-generator/src/case_mapping.rs b/src/tools/unicode-table-generator/src/case_mapping.rs index 9c6454492e7..a8527ea9a42 100644 --- a/src/tools/unicode-table-generator/src/case_mapping.rs +++ b/src/tools/unicode-table-generator/src/case_mapping.rs @@ -6,20 +6,22 @@ use crate::{UnicodeData, fmt_list}; const INDEX_MASK: u32 = 1 << 22; -pub(crate) fn generate_case_mapping(data: &UnicodeData) -> String { +pub(crate) fn generate_case_mapping(data: &UnicodeData) -> (String, [usize; 2]) { let mut file = String::new(); write!(file, "const INDEX_MASK: u32 = 0x{INDEX_MASK:x};").unwrap(); file.push_str("\n\n"); file.push_str(HEADER.trim_start()); file.push('\n'); - file.push_str(&generate_tables("LOWER", &data.to_lower)); + let (lower_tables, lower_size) = generate_tables("LOWER", &data.to_lower); + file.push_str(&lower_tables); file.push_str("\n\n"); - file.push_str(&generate_tables("UPPER", &data.to_upper)); - file + let (upper_tables, upper_size) = generate_tables("UPPER", &data.to_upper); + file.push_str(&upper_tables); + (file, [lower_size, upper_size]) } -fn generate_tables(case: &str, data: &BTreeMap) -> String { +fn generate_tables(case: &str, data: &BTreeMap) -> (String, usize) { let mut mappings = Vec::with_capacity(data.len()); let mut multis = Vec::new(); @@ -46,16 +48,31 @@ fn generate_tables(case: &str, data: &BTreeMap) -> String } let mut tables = String::new(); - - write!(tables, "static {}CASE_TABLE: &[(char, u32)] = &[{}];", case, fmt_list(mappings)) - .unwrap(); + let mut size = 0; + + size += size_of_val(mappings.as_slice()); + write!( + tables, + "static {}CASE_TABLE: &[(char, u32); {}] = &[{}];", + case, + mappings.len(), + fmt_list(mappings), + ) + .unwrap(); tables.push_str("\n\n"); - write!(tables, "static {}CASE_TABLE_MULTI: &[[char; 3]] = &[{}];", case, fmt_list(multis)) - .unwrap(); - - tables + size += size_of_val(multis.as_slice()); + write!( + tables, + "static {}CASE_TABLE_MULTI: &[[char; 3]; {}] = &[{}];", + case, + multis.len(), + fmt_list(multis), + ) + .unwrap(); + + (tables, size) } struct CharEscape(char); diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index c1017142097..f755ad048e4 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -269,6 +269,11 @@ fn main() { )); total_bytes += emitter.bytes_used; } + let (conversions, sizes) = case_mapping::generate_case_mapping(&unicode_data); + for (name, size) in ["to_lower", "to_upper"].iter().zip(sizes) { + table_file.push_str(&format!("// {:16}: {:5} bytes\n", name, size)); + total_bytes += size; + } table_file.push_str(&format!("// {:16}: {:5} bytes\n", "Total", total_bytes)); // Include the range search function @@ -280,7 +285,7 @@ fn main() { table_file.push('\n'); - modules.push((String::from("conversions"), case_mapping::generate_case_mapping(&unicode_data))); + modules.push((String::from("conversions"), conversions)); for (name, contents) in modules { table_file.push_str("#[rustfmt::skip]\n"); -- cgit 1.4.1-3-g733a5 From 5d54ac5276eade9d9424fc4fbfa6f77bcd5d4940 Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 01:10:15 +0100 Subject: refactor: rewrite `ranges_from_set` The `merge_ranges` function was very complicated and hard to understand. Forunately, we can use `slice::chunk_by` to achieve the same thing. --- src/tools/unicode-table-generator/src/main.rs | 83 ++++++--------------------- 1 file changed, 17 insertions(+), 66 deletions(-) (limited to 'src') diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index f755ad048e4..bf0511a2c77 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -187,33 +187,19 @@ fn load_data() -> UnicodeData { } } - let mut properties: HashMap<&'static str, Vec>> = properties + let mut properties: Vec<(&'static str, Vec>)> = properties .into_iter() - .map(|(k, v)| { - ( - k, - v.into_iter() - .flat_map(|codepoints| match codepoints { - Codepoints::Single(c) => c - .scalar() - .map(|ch| ch as u32..ch as u32 + 1) - .into_iter() - .collect::>(), - Codepoints::Range(c) => c - .into_iter() - .flat_map(|c| c.scalar().map(|ch| ch as u32..ch as u32 + 1)) - .collect::>(), - }) - .collect::>>(), - ) + .map(|(prop, codepoints)| { + let codepoints = codepoints + .into_iter() + .flatten() + .flat_map(|cp| cp.scalar()) + .map(u32::from) + .collect::>(); + (prop, ranges_from_set(&codepoints)) }) .collect(); - for ranges in properties.values_mut() { - merge_ranges(ranges); - } - - let mut properties = properties.into_iter().collect::>(); properties.sort_by_key(|p| p.0); UnicodeData { ranges: properties, to_lower, to_upper } } @@ -402,48 +388,13 @@ fn generate_asserts(s: &mut String, property: &str, points: &[u32], truthy: bool } } +/// Group the elements of `set` into contigous ranges fn ranges_from_set(set: &[u32]) -> Vec> { - let mut ranges = set.iter().map(|e| (*e)..(*e + 1)).collect::>>(); - merge_ranges(&mut ranges); - ranges -} - -fn merge_ranges(ranges: &mut Vec>) { - loop { - let mut new_ranges = Vec::new(); - let mut idx_iter = 0..(ranges.len() - 1); - let mut should_insert_last = true; - while let Some(idx) = idx_iter.next() { - let cur = ranges[idx].clone(); - let next = ranges[idx + 1].clone(); - if cur.end == next.start { - if idx_iter.next().is_none() { - // We're merging the last element - should_insert_last = false; - } - new_ranges.push(cur.start..next.end); - } else { - // We're *not* merging the last element - should_insert_last = true; - new_ranges.push(cur); - } - } - if should_insert_last { - new_ranges.push(ranges.last().unwrap().clone()); - } - if new_ranges.len() == ranges.len() { - *ranges = new_ranges; - break; - } else { - *ranges = new_ranges; - } - } - - let mut last_end = None; - for range in ranges { - if let Some(last) = last_end { - assert!(range.start > last, "{range:?}"); - } - last_end = Some(range.end); - } + set.chunk_by(|a, b| a + 1 == *b) + .map(|chunk| { + let start = *chunk.first().unwrap(); + let end = *chunk.last().unwrap(); + start..(end + 1) + }) + .collect() } -- cgit 1.4.1-3-g733a5 From 30d1bc7ba869c0f86bc6d2e1d9ed1ad3b58f7865 Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 01:46:00 +0100 Subject: refactor: `generate_tests` Rewrite `generate_tests` to be more idiomatic. --- src/tools/unicode-table-generator/src/main.rs | 97 +++++++++++++-------------- 1 file changed, 45 insertions(+), 52 deletions(-) (limited to 'src') diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index bf0511a2c77..c9530fec48a 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -72,6 +72,8 @@ //! or not. use std::collections::{BTreeMap, HashMap}; +use std::fmt; +use std::fmt::Write; use std::ops::Range; use ucd_parse::Codepoints; @@ -222,7 +224,7 @@ fn main() { let ranges_by_property = &unicode_data.ranges; if let Some(path) = test_path { - std::fs::write(&path, generate_tests(&write_location, ranges_by_property)).unwrap(); + std::fs::write(&path, generate_tests(ranges_by_property).unwrap()).unwrap(); } let mut table_file = String::new(); @@ -326,66 +328,57 @@ fn fmt_list(values: impl IntoIterator) -> String { out } -fn generate_tests(data_path: &str, ranges: &[(&str, Vec>)]) -> String { +fn generate_tests(ranges: &[(&str, Vec>)]) -> Result { let mut s = String::new(); - s.push_str("#![allow(incomplete_features, unused)]\n"); - s.push_str("#![feature(const_generics)]\n\n"); - s.push_str("\n#[allow(unused)]\nuse std::hint;\n"); - s.push_str(&format!("#[path = \"{data_path}\"]\n")); - s.push_str("mod unicode_data;\n\n"); - - s.push_str("\nfn main() {\n"); - + writeln!(s, "#![feature(core_intrinsics)]")?; + writeln!(s, "#![allow(internal_features, dead_code)]")?; + writeln!(s, "// ignore-tidy-filelength")?; + writeln!(s, "use std::intrinsics;")?; + writeln!(s, "mod unicode_data;")?; + writeln!(s, "fn main() {{")?; for (property, ranges) in ranges { - s.push_str(&format!(r#" println!("Testing {property}");"#)); - s.push('\n'); - s.push_str(&format!(" {}_true();\n", property.to_lowercase())); - s.push_str(&format!(" {}_false();\n", property.to_lowercase())); - let mut is_true = Vec::new(); - let mut is_false = Vec::new(); - for ch_num in 0..(std::char::MAX as u32) { - if std::char::from_u32(ch_num).is_none() { - continue; - } - if ranges.iter().any(|r| r.contains(&ch_num)) { - is_true.push(ch_num); - } else { - is_false.push(ch_num); - } - } - - s.push_str(&format!(" fn {}_true() {{\n", property.to_lowercase())); - generate_asserts(&mut s, property, &is_true, true); - s.push_str(" }\n\n"); - s.push_str(&format!(" fn {}_false() {{\n", property.to_lowercase())); - generate_asserts(&mut s, property, &is_false, false); - s.push_str(" }\n\n"); + let prop = property.to_lowercase(); + writeln!(s, r#" println!("Testing {prop}");"#)?; + writeln!(s, " {prop}_true();")?; + writeln!(s, " {prop}_false();")?; + let (is_true, is_false): (Vec<_>, Vec<_>) = (char::MIN..=char::MAX) + .filter(|c| !c.is_ascii()) + .map(u32::from) + .partition(|c| ranges.iter().any(|r| r.contains(c))); + + writeln!(s, " fn {prop}_true() {{")?; + generate_asserts(&mut s, &prop, &is_true, true)?; + writeln!(s, " }}")?; + + writeln!(s, " fn {prop}_false() {{")?; + generate_asserts(&mut s, &prop, &is_false, false)?; + writeln!(s, " }}")?; } - s.push('}'); - s + writeln!(s, "}}")?; + Ok(s) } -fn generate_asserts(s: &mut String, property: &str, points: &[u32], truthy: bool) { +fn generate_asserts( + s: &mut String, + prop: &str, + points: &[u32], + truthy: bool, +) -> Result<(), fmt::Error> { + let truthy = if truthy { "" } else { "!" }; for range in ranges_from_set(points) { - if range.end == range.start + 1 { - s.push_str(&format!( - " assert!({}unicode_data::{}::lookup({:?}), \"{}\");\n", - if truthy { "" } else { "!" }, - property.to_lowercase(), - std::char::from_u32(range.start).unwrap(), - range.start, - )); - } else { - s.push_str(&format!(" for chn in {range:?}u32 {{\n")); - s.push_str(&format!( - " assert!({}unicode_data::{}::lookup(std::char::from_u32(chn).unwrap()), \"{{:?}}\", chn);\n", - if truthy { "" } else { "!" }, - property.to_lowercase(), - )); - s.push_str(" }\n"); + let start = char::from_u32(range.start).unwrap(); + let end = char::from_u32(range.end - 1).unwrap(); + match range.len() { + 1 => writeln!(s, " assert!({truthy}unicode_data::{prop}::lookup({start:?}));")?, + _ => { + writeln!(s, " for c in {start:?}..={end:?} {{")?; + writeln!(s, " assert!({truthy}unicode_data::{prop}::lookup(c));")?; + writeln!(s, " }}")?; + } } } + Ok(()) } /// Group the elements of `set` into contigous ranges -- cgit 1.4.1-3-g733a5 From c3ce0796544152460554d8b8db4e56528fe362db Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 02:13:03 +0100 Subject: refactor: Add tests for case conversions --- .../unicode-table-generator/src/case_mapping.rs | 4 +- src/tools/unicode-table-generator/src/main.rs | 48 ++++++++++++++++++---- 2 files changed, 41 insertions(+), 11 deletions(-) (limited to 'src') diff --git a/src/tools/unicode-table-generator/src/case_mapping.rs b/src/tools/unicode-table-generator/src/case_mapping.rs index a8527ea9a42..49aef3ec33e 100644 --- a/src/tools/unicode-table-generator/src/case_mapping.rs +++ b/src/tools/unicode-table-generator/src/case_mapping.rs @@ -21,11 +21,11 @@ pub(crate) fn generate_case_mapping(data: &UnicodeData) -> (String, [usize; 2]) (file, [lower_size, upper_size]) } -fn generate_tables(case: &str, data: &BTreeMap) -> (String, usize) { +fn generate_tables(case: &str, data: &BTreeMap) -> (String, usize) { let mut mappings = Vec::with_capacity(data.len()); let mut multis = Vec::new(); - for (&key, &(a, b, c)) in data.iter() { + for (&key, &[a, b, c]) in data.iter() { let key = char::from_u32(key).unwrap(); if key.is_ascii() { diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index c9530fec48a..1d70eebdbc6 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -100,11 +100,11 @@ static PROPERTIES: &[&str] = &[ struct UnicodeData { ranges: Vec<(&'static str, Vec>)>, - to_upper: BTreeMap, - to_lower: BTreeMap, + to_upper: BTreeMap, + to_lower: BTreeMap, } -fn to_mapping(origin: u32, codepoints: Vec) -> Option<(u32, u32, u32)> { +fn to_mapping(origin: u32, codepoints: Vec) -> Option<[u32; 3]> { let mut a = None; let mut b = None; let mut c = None; @@ -125,7 +125,7 @@ fn to_mapping(origin: u32, codepoints: Vec) -> Option<(u32 } } - Some((a.unwrap(), b.unwrap_or(0), c.unwrap_or(0))) + Some([a.unwrap(), b.unwrap_or(0), c.unwrap_or(0)]) } static UNICODE_DIRECTORY: &str = "unicode-downloads"; @@ -165,12 +165,12 @@ fn load_data() -> UnicodeData { if let Some(mapped) = row.simple_lowercase_mapping && mapped != row.codepoint { - to_lower.insert(row.codepoint.value(), (mapped.value(), 0, 0)); + to_lower.insert(row.codepoint.value(), [mapped.value(), 0, 0]); } if let Some(mapped) = row.simple_uppercase_mapping && mapped != row.codepoint { - to_upper.insert(row.codepoint.value(), (mapped.value(), 0, 0)); + to_upper.insert(row.codepoint.value(), [mapped.value(), 0, 0]); } } @@ -224,7 +224,7 @@ fn main() { let ranges_by_property = &unicode_data.ranges; if let Some(path) = test_path { - std::fs::write(&path, generate_tests(ranges_by_property).unwrap()).unwrap(); + std::fs::write(&path, generate_tests(&unicode_data).unwrap()).unwrap(); } let mut table_file = String::new(); @@ -328,7 +328,7 @@ fn fmt_list(values: impl IntoIterator) -> String { out } -fn generate_tests(ranges: &[(&str, Vec>)]) -> Result { +fn generate_tests(data: &UnicodeData) -> Result { let mut s = String::new(); writeln!(s, "#![feature(core_intrinsics)]")?; writeln!(s, "#![allow(internal_features, dead_code)]")?; @@ -336,7 +336,7 @@ fn generate_tests(ranges: &[(&str, Vec>)]) -> Result>)]) -> Result = (char::MIN..=char::MAX) + .filter(|c| !c.is_ascii()) + .map(u32::from) + .filter(|c| !conversion.contains_key(c)) + .collect(); + let unmapped_ranges = ranges_from_set(&unmapped); + for range in unmapped_ranges { + let start = char::from_u32(range.start).unwrap(); + let end = char::from_u32(range.end - 1).unwrap(); + writeln!(s, " for c in {start:?}..={end:?} {{")?; + writeln!( + s, + r#" assert_eq!(unicode_data::conversions::{name}(c), [c, '\0', '\0']);"# + )?; + + writeln!(s, " }}")?; + } + } + writeln!(s, "}}")?; Ok(s) } -- cgit 1.4.1-3-g733a5 From aa27cca79e7888deff3bafed84ea16a5a89cb7c7 Mon Sep 17 00:00:00 2001 From: Alex Crichton Date: Tue, 26 Aug 2025 11:34:12 -0700 Subject: Add a sanity check to bootstrap for wasm-component-ld For `wasm32-wasip2`-and-beyond this tool is required, so in case it's disabled in `config.toml` add a sanity-check that it's present in the environment. --- src/bootstrap/src/core/sanity.rs | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'src') diff --git a/src/bootstrap/src/core/sanity.rs b/src/bootstrap/src/core/sanity.rs index 099ec488397..04cf63f1c6d 100644 --- a/src/bootstrap/src/core/sanity.rs +++ b/src/bootstrap/src/core/sanity.rs @@ -397,6 +397,16 @@ $ pacman -R cmake && pacman -S mingw-w64-x86_64-cmake ); } } + + // For testing `wasm32-wasip2`-and-beyond it's required to have + // `wasm-component-ld`. This is enabled by default via `tool_enabled` + // but if it's disabled then double-check it's present on the system. + if target.contains("wasip") + && !target.contains("wasip1") + && !build.tool_enabled("wasm-component-ld") + { + cmd_finder.must_have("wasm-component-ld"); + } } if let Some(ref s) = build.config.ccache { -- cgit 1.4.1-3-g733a5 From 5d81f033a5ab2fce07898a22ad8c607380812cf7 Mon Sep 17 00:00:00 2001 From: Alex Crichton Date: Tue, 26 Aug 2025 13:53:11 -0700 Subject: std: Start supporting WASIp2 natively This commit is the start of an effort to support WASIp2 natively in the standard library. Before this commit the `wasm32-wasip2` target behaved exactly like `wasm32-wasip1` target by importing APIs from the core wasm module `wasi_snapshot_preview1`. These APIs are satisfied by the `wasm-component-ld` target by using an [adapter] which implements WASIp1 in terms of WASIp2. This adapter comes at a cost, however, in terms of runtime indirection and instantiation cost, so ideally the adapter would be removed entirely. The purpose of this adapter was to provide a smoother on-ramp from WASIp1 to WASIp2 when it was originally created. The `wasm32-wasip2` target has been around for long enough now that it's much more established. Additionally the only thing historically blocking using WASIp2 directly was implementation effort. Work is now underway to migrate wasi-libc itself to using WASIp2 directly and now seems as good a time as any to migrate the Rust standard library too. Implementation-wise the milestones here are: * The `wasm32-wasip2` target now also depends on the `wasi` crate at version 0.14.* in addition to the preexisting dependency of 0.11.*. The 0.14.* release series binds WASIp2 APIs instead of WASIp1 APIs. * Some preexisting naming around `mod wasi` or `wasi.rs` was renamed to `wasip1` where appropriate. For example `std::sys::pal::wasi` is now called `std::sys::pal::wasip1`. * More platform-specific WASI modules are now split between WASIp1 and WASIp2. For example getting the current time, randomness, and process arguments now use WASIp2 APIs directly instead of using WASIp1 APIs that require an adapter. It's worth pointing out that this PR does not migrate the entire standard library away from using WASIp1 APIs on the `wasm32-wasip2` target. Everything related to file descriptors and filesystem APIs is still using WASIp1. Migrating that is left for a future PR. In the meantime the goal of this change is to lay the groundwork necessary for migrating in the future. Eventually the goal is to drop the `wasi` 0.11.* dependency on the `wasm32-wasip2` target (the `wasm32-wasip1` target will continue to retain this dependency). [adapter]: https://github.com/bytecodealliance/wasmtime/blob/main/crates/wasi-preview1-component-adapter/README.md --- library/Cargo.lock | 24 +++- library/std/Cargo.toml | 5 + library/std/src/sys/args/mod.rs | 10 +- library/std/src/sys/args/wasi.rs | 26 ---- library/std/src/sys/args/wasip1.rs | 26 ++++ library/std/src/sys/args/wasip2.rs | 6 + library/std/src/sys/pal/mod.rs | 6 +- library/std/src/sys/pal/wasi/helpers.rs | 114 ---------------- library/std/src/sys/pal/wasi/mod.rs | 39 ------ library/std/src/sys/pal/wasi/os.rs | 151 --------------------- library/std/src/sys/pal/wasi/thread.rs | 214 ------------------------------ library/std/src/sys/pal/wasi/time.rs | 76 ----------- library/std/src/sys/pal/wasip1/helpers.rs | 114 ++++++++++++++++ library/std/src/sys/pal/wasip1/mod.rs | 39 ++++++ library/std/src/sys/pal/wasip1/os.rs | 151 +++++++++++++++++++++ library/std/src/sys/pal/wasip1/thread.rs | 214 ++++++++++++++++++++++++++++++ library/std/src/sys/pal/wasip1/time.rs | 76 +++++++++++ library/std/src/sys/pal/wasip2/mod.rs | 6 +- library/std/src/sys/pal/wasip2/thread.rs | 73 ++++++++++ library/std/src/sys/pal/wasip2/time.rs | 69 ++++++++++ library/std/src/sys/random/mod.rs | 11 +- library/std/src/sys/random/wasi.rs | 5 - library/std/src/sys/random/wasip1.rs | 5 + library/std/src/sys/random/wasip2.rs | 9 ++ src/tools/tidy/src/deps.rs | 6 +- triagebot.toml | 3 +- 26 files changed, 837 insertions(+), 641 deletions(-) delete mode 100644 library/std/src/sys/args/wasi.rs create mode 100644 library/std/src/sys/args/wasip1.rs create mode 100644 library/std/src/sys/args/wasip2.rs delete mode 100644 library/std/src/sys/pal/wasi/helpers.rs delete mode 100644 library/std/src/sys/pal/wasi/mod.rs delete mode 100644 library/std/src/sys/pal/wasi/os.rs delete mode 100644 library/std/src/sys/pal/wasi/thread.rs delete mode 100644 library/std/src/sys/pal/wasi/time.rs create mode 100644 library/std/src/sys/pal/wasip1/helpers.rs create mode 100644 library/std/src/sys/pal/wasip1/mod.rs create mode 100644 library/std/src/sys/pal/wasip1/os.rs create mode 100644 library/std/src/sys/pal/wasip1/thread.rs create mode 100644 library/std/src/sys/pal/wasip1/time.rs create mode 100644 library/std/src/sys/pal/wasip2/thread.rs create mode 100644 library/std/src/sys/pal/wasip2/time.rs delete mode 100644 library/std/src/sys/random/wasi.rs create mode 100644 library/std/src/sys/random/wasip1.rs create mode 100644 library/std/src/sys/random/wasip2.rs (limited to 'src') diff --git a/library/Cargo.lock b/library/Cargo.lock index 8b860f69492..c16bb3eafbd 100644 --- a/library/Cargo.lock +++ b/library/Cargo.lock @@ -327,7 +327,8 @@ dependencies = [ "rustc-demangle", "std_detect", "unwind", - "wasi", + "wasi 0.11.1+wasi-snapshot-preview1", + "wasi 0.14.3+wasi-0.2.4", "windows-targets 0.0.0", ] @@ -399,6 +400,17 @@ dependencies = [ "rustc-std-workspace-core", ] +[[package]] +name = "wasi" +version = "0.14.3+wasi-0.2.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6a51ae83037bdd272a9e28ce236db8c07016dd0d50c27038b3f407533c030c95" +dependencies = [ + "rustc-std-workspace-alloc", + "rustc-std-workspace-core", + "wit-bindgen", +] + [[package]] name = "windows-sys" version = "0.59.0" @@ -475,3 +487,13 @@ name = "windows_x86_64_msvc" version = "0.52.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec" + +[[package]] +name = "wit-bindgen" +version = "0.45.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "052283831dbae3d879dc7f51f3d92703a316ca49f91540417d38591826127814" +dependencies = [ + "rustc-std-workspace-alloc", + "rustc-std-workspace-core", +] diff --git a/library/std/Cargo.toml b/library/std/Cargo.toml index eff2562f031..8b78fe53a39 100644 --- a/library/std/Cargo.toml +++ b/library/std/Cargo.toml @@ -80,6 +80,11 @@ wasi = { version = "0.11.0", features = [ 'rustc-dep-of-std', ], default-features = false } +[target.'cfg(all(target_os = "wasi", target_env = "p2"))'.dependencies] +wasip2 = { version = '0.14.3', features = [ + 'rustc-dep-of-std', +], default-features = false, package = 'wasi' } + [target.'cfg(target_os = "uefi")'.dependencies] r-efi = { version = "5.2.0", features = ['rustc-dep-of-std'] } r-efi-alloc = { version = "2.0.0", features = ['rustc-dep-of-std'] } diff --git a/library/std/src/sys/args/mod.rs b/library/std/src/sys/args/mod.rs index c9627322276..e11e8e5430f 100644 --- a/library/std/src/sys/args/mod.rs +++ b/library/std/src/sys/args/mod.rs @@ -32,9 +32,13 @@ cfg_select! { mod uefi; pub use uefi::*; } - target_os = "wasi" => { - mod wasi; - pub use wasi::*; + all(target_os = "wasi", target_env = "p1") => { + mod wasip1; + pub use wasip1::*; + } + all(target_os = "wasi", target_env = "p2") => { + mod wasip2; + pub use wasip2::*; } target_os = "xous" => { mod xous; diff --git a/library/std/src/sys/args/wasi.rs b/library/std/src/sys/args/wasi.rs deleted file mode 100644 index 72063a87dc9..00000000000 --- a/library/std/src/sys/args/wasi.rs +++ /dev/null @@ -1,26 +0,0 @@ -#![forbid(unsafe_op_in_unsafe_fn)] - -pub use super::common::Args; -use crate::ffi::{CStr, OsStr, OsString}; -use crate::os::wasi::ffi::OsStrExt; - -/// Returns the command line arguments -pub fn args() -> Args { - Args::new(maybe_args().unwrap_or(Vec::new())) -} - -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) - } -} diff --git a/library/std/src/sys/args/wasip1.rs b/library/std/src/sys/args/wasip1.rs new file mode 100644 index 00000000000..72063a87dc9 --- /dev/null +++ b/library/std/src/sys/args/wasip1.rs @@ -0,0 +1,26 @@ +#![forbid(unsafe_op_in_unsafe_fn)] + +pub use super::common::Args; +use crate::ffi::{CStr, OsStr, OsString}; +use crate::os::wasi::ffi::OsStrExt; + +/// Returns the command line arguments +pub fn args() -> Args { + Args::new(maybe_args().unwrap_or(Vec::new())) +} + +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) + } +} diff --git a/library/std/src/sys/args/wasip2.rs b/library/std/src/sys/args/wasip2.rs new file mode 100644 index 00000000000..a57e4b97786 --- /dev/null +++ b/library/std/src/sys/args/wasip2.rs @@ -0,0 +1,6 @@ +pub use super::common::Args; + +/// Returns the command line arguments +pub fn args() -> Args { + Args::new(wasip2::cli::environment::get_arguments().into_iter().map(|arg| arg.into()).collect()) +} diff --git a/library/std/src/sys/pal/mod.rs b/library/std/src/sys/pal/mod.rs index 9376f5249f1..513121c6d30 100644 --- a/library/std/src/sys/pal/mod.rs +++ b/library/std/src/sys/pal/mod.rs @@ -49,9 +49,9 @@ cfg_select! { mod wasip2; pub use self::wasip2::*; } - target_os = "wasi" => { - mod wasi; - pub use self::wasi::*; + all(target_os = "wasi", target_env = "p1") => { + mod wasip1; + pub use self::wasip1::*; } target_family = "wasm" => { mod wasm; diff --git a/library/std/src/sys/pal/wasi/helpers.rs b/library/std/src/sys/pal/wasi/helpers.rs deleted file mode 100644 index 404747f0dc7..00000000000 --- a/library/std/src/sys/pal/wasi/helpers.rs +++ /dev/null @@ -1,114 +0,0 @@ -#![forbid(unsafe_op_in_unsafe_fn)] - -use crate::io as std_io; - -#[inline] -pub fn is_interrupted(errno: i32) -> bool { - errno == wasi::ERRNO_INTR.raw().into() -} - -pub fn decode_error_kind(errno: i32) -> std_io::ErrorKind { - use std_io::ErrorKind; - - let Ok(errno) = u16::try_from(errno) else { - return ErrorKind::Uncategorized; - }; - - macro_rules! match_errno { - ($($($errno:ident)|+ => $errkind:ident),*, _ => $wildcard:ident $(,)?) => { - match errno { - $(e if $(e == ::wasi::$errno.raw())||+ => ErrorKind::$errkind),*, - _ => ErrorKind::$wildcard, - } - }; - } - - match_errno! { - ERRNO_2BIG => ArgumentListTooLong, - ERRNO_ACCES => PermissionDenied, - ERRNO_ADDRINUSE => AddrInUse, - ERRNO_ADDRNOTAVAIL => AddrNotAvailable, - ERRNO_AFNOSUPPORT => Unsupported, - ERRNO_AGAIN => WouldBlock, - // ALREADY => "connection already in progress", - // BADF => "bad file descriptor", - // BADMSG => "bad message", - ERRNO_BUSY => ResourceBusy, - // CANCELED => "operation canceled", - // CHILD => "no child processes", - ERRNO_CONNABORTED => ConnectionAborted, - ERRNO_CONNREFUSED => ConnectionRefused, - ERRNO_CONNRESET => ConnectionReset, - ERRNO_DEADLK => Deadlock, - // DESTADDRREQ => "destination address required", - ERRNO_DOM => InvalidInput, - // DQUOT => /* reserved */, - ERRNO_EXIST => AlreadyExists, - // FAULT => "bad address", - ERRNO_FBIG => FileTooLarge, - ERRNO_HOSTUNREACH => HostUnreachable, - // IDRM => "identifier removed", - // ILSEQ => "illegal byte sequence", - // INPROGRESS => "operation in progress", - ERRNO_INTR => Interrupted, - ERRNO_INVAL => InvalidInput, - ERRNO_IO => Uncategorized, - // ISCONN => "socket is connected", - ERRNO_ISDIR => IsADirectory, - ERRNO_LOOP => FilesystemLoop, - // MFILE => "file descriptor value too large", - ERRNO_MLINK => TooManyLinks, - // MSGSIZE => "message too large", - // MULTIHOP => /* reserved */, - ERRNO_NAMETOOLONG => InvalidFilename, - ERRNO_NETDOWN => NetworkDown, - // NETRESET => "connection aborted by network", - ERRNO_NETUNREACH => NetworkUnreachable, - // NFILE => "too many files open in system", - // NOBUFS => "no buffer space available", - ERRNO_NODEV => NotFound, - ERRNO_NOENT => NotFound, - // NOEXEC => "executable file format error", - // NOLCK => "no locks available", - // NOLINK => /* reserved */, - ERRNO_NOMEM => OutOfMemory, - // NOMSG => "no message of the desired type", - // NOPROTOOPT => "protocol not available", - ERRNO_NOSPC => StorageFull, - ERRNO_NOSYS => Unsupported, - ERRNO_NOTCONN => NotConnected, - ERRNO_NOTDIR => NotADirectory, - ERRNO_NOTEMPTY => DirectoryNotEmpty, - // NOTRECOVERABLE => "state not recoverable", - // NOTSOCK => "not a socket", - ERRNO_NOTSUP => Unsupported, - // NOTTY => "inappropriate I/O control operation", - ERRNO_NXIO => NotFound, - // OVERFLOW => "value too large to be stored in data type", - // OWNERDEAD => "previous owner died", - ERRNO_PERM => PermissionDenied, - ERRNO_PIPE => BrokenPipe, - // PROTO => "protocol error", - ERRNO_PROTONOSUPPORT => Unsupported, - // PROTOTYPE => "protocol wrong type for socket", - // RANGE => "result too large", - ERRNO_ROFS => ReadOnlyFilesystem, - ERRNO_SPIPE => NotSeekable, - ERRNO_SRCH => NotFound, - // STALE => /* reserved */, - ERRNO_TIMEDOUT => TimedOut, - ERRNO_TXTBSY => ResourceBusy, - ERRNO_XDEV => CrossesDevices, - ERRNO_NOTCAPABLE => PermissionDenied, - _ => Uncategorized, - } -} - -pub fn abort_internal() -> ! { - unsafe { libc::abort() } -} - -#[inline] -pub(crate) fn err2io(err: wasi::Errno) -> std_io::Error { - std_io::Error::from_raw_os_error(err.raw().into()) -} diff --git a/library/std/src/sys/pal/wasi/mod.rs b/library/std/src/sys/pal/wasi/mod.rs deleted file mode 100644 index 61dd1c3f98b..00000000000 --- a/library/std/src/sys/pal/wasi/mod.rs +++ /dev/null @@ -1,39 +0,0 @@ -//! System bindings for the wasm/web platform -//! -//! This module contains the facade (aka platform-specific) implementations of -//! OS level functionality for wasm. -//! -//! This is all super highly experimental and not actually intended for -//! wide/production use yet, it's still all in the experimental category. This -//! will likely change over time. -//! -//! Currently all functions here are basically stubs that immediately return -//! errors. The hope is that with a portability lint we can turn actually just -//! remove all this and just omit parts of the standard library if we're -//! compiling for wasm. That way it's a compile time error for something that's -//! guaranteed to be a runtime error! - -#[allow(unused)] -#[path = "../wasm/atomics/futex.rs"] -pub mod futex; - -pub mod os; -#[path = "../unsupported/pipe.rs"] -pub mod pipe; -pub mod thread; -pub mod time; - -#[path = "../unsupported/common.rs"] -#[deny(unsafe_op_in_unsafe_fn)] -#[allow(unused)] -mod common; - -pub use common::*; - -mod helpers; - -// The following exports are listed individually to work around Rust's glob -// import conflict rules. If we glob export `helpers` and `common` together, -// then the compiler complains about conflicts. - -pub(crate) use helpers::{abort_internal, decode_error_kind, err2io, is_interrupted}; diff --git a/library/std/src/sys/pal/wasi/os.rs b/library/std/src/sys/pal/wasi/os.rs deleted file mode 100644 index 151ba254ec4..00000000000 --- a/library/std/src/sys/pal/wasi/os.rs +++ /dev/null @@ -1,151 +0,0 @@ -#![forbid(unsafe_op_in_unsafe_fn)] - -use crate::ffi::{CStr, OsStr, OsString}; -use crate::marker::PhantomData; -use crate::os::wasi::prelude::*; -use crate::path::{self, PathBuf}; -use crate::sys::common::small_c_string::run_path_with_cstr; -use crate::sys::unsupported; -use crate::{fmt, io, str}; - -// Add a few symbols not in upstream `libc` just yet. -pub mod libc { - pub use libc::*; - - unsafe extern "C" { - pub fn getcwd(buf: *mut c_char, size: size_t) -> *mut c_char; - pub fn chdir(dir: *const c_char) -> c_int; - pub fn __wasilibc_get_environ() -> *mut *mut c_char; - } -} - -pub fn errno() -> i32 { - unsafe extern "C" { - #[thread_local] - static errno: libc::c_int; - } - - unsafe { errno as i32 } -} - -pub fn error_string(errno: i32) -> String { - let mut buf = [0 as libc::c_char; 1024]; - - let p = buf.as_mut_ptr(); - unsafe { - if libc::strerror_r(errno as libc::c_int, p, buf.len()) < 0 { - panic!("strerror_r failure"); - } - str::from_utf8(CStr::from_ptr(p).to_bytes()).unwrap().to_owned() - } -} - -pub fn getcwd() -> io::Result { - let mut buf = Vec::with_capacity(512); - loop { - unsafe { - let ptr = buf.as_mut_ptr() as *mut libc::c_char; - if !libc::getcwd(ptr, buf.capacity()).is_null() { - let len = CStr::from_ptr(buf.as_ptr() as *const libc::c_char).to_bytes().len(); - buf.set_len(len); - buf.shrink_to_fit(); - return Ok(PathBuf::from(OsString::from_vec(buf))); - } else { - let error = io::Error::last_os_error(); - if error.raw_os_error() != Some(libc::ERANGE) { - return Err(error); - } - } - - // Trigger the internal buffer resizing logic of `Vec` by requiring - // more space than the current capacity. - let cap = buf.capacity(); - buf.set_len(cap); - buf.reserve(1); - } - } -} - -pub fn chdir(p: &path::Path) -> io::Result<()> { - let result = run_path_with_cstr(p, &|p| unsafe { Ok(libc::chdir(p.as_ptr())) })?; - match result == (0 as libc::c_int) { - true => Ok(()), - false => Err(io::Error::last_os_error()), - } -} - -pub struct SplitPaths<'a>(!, PhantomData<&'a ()>); - -pub fn split_paths(_unparsed: &OsStr) -> SplitPaths<'_> { - panic!("unsupported") -} - -impl<'a> Iterator for SplitPaths<'a> { - type Item = PathBuf; - fn next(&mut self) -> Option { - self.0 - } -} - -#[derive(Debug)] -pub struct JoinPathsError; - -pub fn join_paths(_paths: I) -> Result -where - I: Iterator, - T: AsRef, -{ - Err(JoinPathsError) -} - -impl fmt::Display for JoinPathsError { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - "not supported on wasm yet".fmt(f) - } -} - -impl crate::error::Error for JoinPathsError {} - -pub fn current_exe() -> io::Result { - unsupported() -} - -#[allow(dead_code)] -pub fn page_size() -> usize { - unsafe { libc::sysconf(libc::_SC_PAGESIZE) as usize } -} - -pub fn temp_dir() -> PathBuf { - panic!("no filesystem on wasm") -} - -pub fn home_dir() -> Option { - None -} - -pub fn exit(code: i32) -> ! { - unsafe { libc::exit(code) } -} - -pub fn getpid() -> u32 { - panic!("unsupported"); -} - -#[doc(hidden)] -pub trait IsMinusOne { - fn is_minus_one(&self) -> bool; -} - -macro_rules! impl_is_minus_one { - ($($t:ident)*) => ($(impl IsMinusOne for $t { - fn is_minus_one(&self) -> bool { - *self == -1 - } - })*) -} - -impl_is_minus_one! { i8 i16 i32 i64 isize } - -pub fn cvt(t: T) -> io::Result { - if t.is_minus_one() { Err(io::Error::last_os_error()) } else { Ok(t) } -} diff --git a/library/std/src/sys/pal/wasi/thread.rs b/library/std/src/sys/pal/wasi/thread.rs deleted file mode 100644 index e062b49bd7a..00000000000 --- a/library/std/src/sys/pal/wasi/thread.rs +++ /dev/null @@ -1,214 +0,0 @@ -#![forbid(unsafe_op_in_unsafe_fn)] - -use crate::ffi::CStr; -use crate::num::NonZero; -use crate::time::{Duration, Instant}; -use crate::{io, mem}; - -cfg_select! { - target_feature = "atomics" => { - use crate::cmp; - use crate::ptr; - use crate::sys::os; - // Add a few symbols not in upstream `libc` just yet. - mod libc { - pub use crate::ffi; - pub use libc::*; - - // defined in wasi-libc - // https://github.com/WebAssembly/wasi-libc/blob/a6f871343313220b76009827ed0153586361c0d5/libc-top-half/musl/include/alltypes.h.in#L108 - #[repr(C)] - union pthread_attr_union { - __i: [ffi::c_int; if size_of::() == 8 { 14 } else { 9 }], - __vi: [ffi::c_int; if size_of::() == 8 { 14 } else { 9 }], - __s: [ffi::c_ulong; if size_of::() == 8 { 7 } else { 9 }], - } - - #[repr(C)] - pub struct pthread_attr_t { - __u: pthread_attr_union, - } - - #[allow(non_camel_case_types)] - pub type pthread_t = *mut ffi::c_void; - - pub const _SC_NPROCESSORS_ONLN: ffi::c_int = 84; - - unsafe extern "C" { - pub fn pthread_create( - native: *mut pthread_t, - attr: *const pthread_attr_t, - f: extern "C" fn(*mut ffi::c_void) -> *mut ffi::c_void, - value: *mut ffi::c_void, - ) -> ffi::c_int; - pub fn pthread_join(native: pthread_t, value: *mut *mut ffi::c_void) -> ffi::c_int; - pub fn pthread_attr_init(attrp: *mut pthread_attr_t) -> ffi::c_int; - pub fn pthread_attr_setstacksize( - attr: *mut pthread_attr_t, - stack_size: libc::size_t, - ) -> ffi::c_int; - pub fn pthread_attr_destroy(attr: *mut pthread_attr_t) -> ffi::c_int; - pub fn pthread_detach(thread: pthread_t) -> ffi::c_int; - } - } - - pub struct Thread { - id: libc::pthread_t, - } - - impl Drop for Thread { - fn drop(&mut self) { - let ret = unsafe { libc::pthread_detach(self.id) }; - debug_assert_eq!(ret, 0); - } - } - } - _ => { - pub struct Thread(!); - } -} - -pub const DEFAULT_MIN_STACK_SIZE: usize = 1024 * 1024; - -impl Thread { - // unsafe: see thread::Builder::spawn_unchecked for safety requirements - cfg_select! { - target_feature = "atomics" => { - pub unsafe fn new(stack: usize, _name: Option<&str>, p: Box) -> io::Result { - let p = Box::into_raw(Box::new(p)); - let mut native: libc::pthread_t = unsafe { mem::zeroed() }; - let mut attr: libc::pthread_attr_t = unsafe { mem::zeroed() }; - assert_eq!(unsafe { libc::pthread_attr_init(&mut attr) }, 0); - - let stack_size = cmp::max(stack, DEFAULT_MIN_STACK_SIZE); - - match unsafe { libc::pthread_attr_setstacksize(&mut attr, stack_size) } { - 0 => {} - n => { - assert_eq!(n, libc::EINVAL); - // EINVAL means |stack_size| is either too small or not a - // multiple of the system page size. Because it's definitely - // >= PTHREAD_STACK_MIN, it must be an alignment issue. - // Round up to the nearest page and try again. - let page_size = os::page_size(); - let stack_size = - (stack_size + page_size - 1) & (-(page_size as isize - 1) as usize - 1); - assert_eq!(unsafe { libc::pthread_attr_setstacksize(&mut attr, stack_size) }, 0); - } - }; - - let ret = unsafe { libc::pthread_create(&mut native, &attr, thread_start, p as *mut _) }; - // Note: if the thread creation fails and this assert fails, then p will - // be leaked. However, an alternative design could cause double-free - // which is clearly worse. - assert_eq!(unsafe {libc::pthread_attr_destroy(&mut attr) }, 0); - - return if ret != 0 { - // The thread failed to start and as a result p was not consumed. Therefore, it is - // safe to reconstruct the box so that it gets deallocated. - unsafe { drop(Box::from_raw(p)); } - Err(io::Error::from_raw_os_error(ret)) - } else { - Ok(Thread { id: native }) - }; - - extern "C" fn thread_start(main: *mut libc::c_void) -> *mut libc::c_void { - unsafe { - // Finally, let's run some code. - Box::from_raw(main as *mut Box)(); - } - ptr::null_mut() - } - } - } - _ => { - pub unsafe fn new(_stack: usize, _name: Option<&str>, _p: Box) -> io::Result { - crate::sys::unsupported() - } - } - } - - pub fn yield_now() { - let ret = unsafe { wasi::sched_yield() }; - debug_assert_eq!(ret, Ok(())); - } - - pub fn set_name(_name: &CStr) { - // nope - } - - pub fn sleep(dur: Duration) { - let mut nanos = dur.as_nanos(); - while nanos > 0 { - const USERDATA: wasi::Userdata = 0x0123_45678; - - let clock = wasi::SubscriptionClock { - id: wasi::CLOCKID_MONOTONIC, - timeout: u64::try_from(nanos).unwrap_or(u64::MAX), - precision: 0, - flags: 0, - }; - nanos -= u128::from(clock.timeout); - - let in_ = wasi::Subscription { - userdata: USERDATA, - u: wasi::SubscriptionU { tag: 0, u: wasi::SubscriptionUU { clock } }, - }; - unsafe { - let mut event: wasi::Event = mem::zeroed(); - let res = wasi::poll_oneoff(&in_, &mut event, 1); - match (res, event) { - ( - Ok(1), - wasi::Event { - userdata: USERDATA, - error: wasi::ERRNO_SUCCESS, - type_: wasi::EVENTTYPE_CLOCK, - .. - }, - ) => {} - _ => panic!("thread::sleep(): unexpected result of poll_oneoff"), - } - } - } - } - - pub fn sleep_until(deadline: Instant) { - let now = Instant::now(); - - if let Some(delay) = deadline.checked_duration_since(now) { - Self::sleep(delay); - } - } - - pub fn join(self) { - cfg_select! { - target_feature = "atomics" => { - let id = mem::ManuallyDrop::new(self).id; - let ret = unsafe { libc::pthread_join(id, ptr::null_mut()) }; - if ret != 0 { - rtabort!("failed to join thread: {}", io::Error::from_raw_os_error(ret)); - } - } - _ => { - self.0 - } - } - } -} - -pub(crate) fn current_os_id() -> Option { - None -} - -pub fn available_parallelism() -> io::Result> { - cfg_select! { - target_feature = "atomics" => { - match unsafe { libc::sysconf(libc::_SC_NPROCESSORS_ONLN) } { - -1 => Err(io::Error::last_os_error()), - cpus => NonZero::new(cpus as usize).ok_or(io::Error::UNKNOWN_THREAD_COUNT), - } - } - _ => crate::sys::unsupported(), - } -} diff --git a/library/std/src/sys/pal/wasi/time.rs b/library/std/src/sys/pal/wasi/time.rs deleted file mode 100644 index 892661b312b..00000000000 --- a/library/std/src/sys/pal/wasi/time.rs +++ /dev/null @@ -1,76 +0,0 @@ -#![forbid(unsafe_op_in_unsafe_fn)] - -use crate::time::Duration; - -#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Debug, Hash)] -pub struct Instant(Duration); - -#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Debug, Hash)] -pub struct SystemTime(Duration); - -pub const UNIX_EPOCH: SystemTime = SystemTime(Duration::from_secs(0)); - -fn current_time(clock: wasi::Clockid) -> Duration { - let ts = unsafe { - wasi::clock_time_get( - clock, 1, // precision... seems ignored though? - ) - .unwrap() - }; - Duration::new((ts / 1_000_000_000) as u64, (ts % 1_000_000_000) as u32) -} - -impl Instant { - pub fn now() -> Instant { - Instant(current_time(wasi::CLOCKID_MONOTONIC)) - } - - pub fn checked_sub_instant(&self, other: &Instant) -> Option { - self.0.checked_sub(other.0) - } - - pub fn checked_add_duration(&self, other: &Duration) -> Option { - Some(Instant(self.0.checked_add(*other)?)) - } - - pub fn checked_sub_duration(&self, other: &Duration) -> Option { - Some(Instant(self.0.checked_sub(*other)?)) - } -} - -impl SystemTime { - pub fn now() -> SystemTime { - SystemTime(current_time(wasi::CLOCKID_REALTIME)) - } - - #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] - pub const fn from_wasi_timestamp(ts: wasi::Timestamp) -> SystemTime { - SystemTime(Duration::from_nanos(ts)) - } - - #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] - pub const fn to_wasi_timestamp(&self) -> Option { - // FIXME: const TryInto - let ns = self.0.as_nanos(); - if ns <= u64::MAX as u128 { Some(ns as u64) } else { None } - } - - #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] - pub const fn sub_time(&self, other: &SystemTime) -> Result { - // FIXME: ok_or_else with const closures - match self.0.checked_sub(other.0) { - Some(duration) => Ok(duration), - None => Err(other.0 - self.0), - } - } - - #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] - pub const fn checked_add_duration(&self, other: &Duration) -> Option { - Some(SystemTime(self.0.checked_add(*other)?)) - } - - #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] - pub const fn checked_sub_duration(&self, other: &Duration) -> Option { - Some(SystemTime(self.0.checked_sub(*other)?)) - } -} diff --git a/library/std/src/sys/pal/wasip1/helpers.rs b/library/std/src/sys/pal/wasip1/helpers.rs new file mode 100644 index 00000000000..404747f0dc7 --- /dev/null +++ b/library/std/src/sys/pal/wasip1/helpers.rs @@ -0,0 +1,114 @@ +#![forbid(unsafe_op_in_unsafe_fn)] + +use crate::io as std_io; + +#[inline] +pub fn is_interrupted(errno: i32) -> bool { + errno == wasi::ERRNO_INTR.raw().into() +} + +pub fn decode_error_kind(errno: i32) -> std_io::ErrorKind { + use std_io::ErrorKind; + + let Ok(errno) = u16::try_from(errno) else { + return ErrorKind::Uncategorized; + }; + + macro_rules! match_errno { + ($($($errno:ident)|+ => $errkind:ident),*, _ => $wildcard:ident $(,)?) => { + match errno { + $(e if $(e == ::wasi::$errno.raw())||+ => ErrorKind::$errkind),*, + _ => ErrorKind::$wildcard, + } + }; + } + + match_errno! { + ERRNO_2BIG => ArgumentListTooLong, + ERRNO_ACCES => PermissionDenied, + ERRNO_ADDRINUSE => AddrInUse, + ERRNO_ADDRNOTAVAIL => AddrNotAvailable, + ERRNO_AFNOSUPPORT => Unsupported, + ERRNO_AGAIN => WouldBlock, + // ALREADY => "connection already in progress", + // BADF => "bad file descriptor", + // BADMSG => "bad message", + ERRNO_BUSY => ResourceBusy, + // CANCELED => "operation canceled", + // CHILD => "no child processes", + ERRNO_CONNABORTED => ConnectionAborted, + ERRNO_CONNREFUSED => ConnectionRefused, + ERRNO_CONNRESET => ConnectionReset, + ERRNO_DEADLK => Deadlock, + // DESTADDRREQ => "destination address required", + ERRNO_DOM => InvalidInput, + // DQUOT => /* reserved */, + ERRNO_EXIST => AlreadyExists, + // FAULT => "bad address", + ERRNO_FBIG => FileTooLarge, + ERRNO_HOSTUNREACH => HostUnreachable, + // IDRM => "identifier removed", + // ILSEQ => "illegal byte sequence", + // INPROGRESS => "operation in progress", + ERRNO_INTR => Interrupted, + ERRNO_INVAL => InvalidInput, + ERRNO_IO => Uncategorized, + // ISCONN => "socket is connected", + ERRNO_ISDIR => IsADirectory, + ERRNO_LOOP => FilesystemLoop, + // MFILE => "file descriptor value too large", + ERRNO_MLINK => TooManyLinks, + // MSGSIZE => "message too large", + // MULTIHOP => /* reserved */, + ERRNO_NAMETOOLONG => InvalidFilename, + ERRNO_NETDOWN => NetworkDown, + // NETRESET => "connection aborted by network", + ERRNO_NETUNREACH => NetworkUnreachable, + // NFILE => "too many files open in system", + // NOBUFS => "no buffer space available", + ERRNO_NODEV => NotFound, + ERRNO_NOENT => NotFound, + // NOEXEC => "executable file format error", + // NOLCK => "no locks available", + // NOLINK => /* reserved */, + ERRNO_NOMEM => OutOfMemory, + // NOMSG => "no message of the desired type", + // NOPROTOOPT => "protocol not available", + ERRNO_NOSPC => StorageFull, + ERRNO_NOSYS => Unsupported, + ERRNO_NOTCONN => NotConnected, + ERRNO_NOTDIR => NotADirectory, + ERRNO_NOTEMPTY => DirectoryNotEmpty, + // NOTRECOVERABLE => "state not recoverable", + // NOTSOCK => "not a socket", + ERRNO_NOTSUP => Unsupported, + // NOTTY => "inappropriate I/O control operation", + ERRNO_NXIO => NotFound, + // OVERFLOW => "value too large to be stored in data type", + // OWNERDEAD => "previous owner died", + ERRNO_PERM => PermissionDenied, + ERRNO_PIPE => BrokenPipe, + // PROTO => "protocol error", + ERRNO_PROTONOSUPPORT => Unsupported, + // PROTOTYPE => "protocol wrong type for socket", + // RANGE => "result too large", + ERRNO_ROFS => ReadOnlyFilesystem, + ERRNO_SPIPE => NotSeekable, + ERRNO_SRCH => NotFound, + // STALE => /* reserved */, + ERRNO_TIMEDOUT => TimedOut, + ERRNO_TXTBSY => ResourceBusy, + ERRNO_XDEV => CrossesDevices, + ERRNO_NOTCAPABLE => PermissionDenied, + _ => Uncategorized, + } +} + +pub fn abort_internal() -> ! { + unsafe { libc::abort() } +} + +#[inline] +pub(crate) fn err2io(err: wasi::Errno) -> std_io::Error { + std_io::Error::from_raw_os_error(err.raw().into()) +} diff --git a/library/std/src/sys/pal/wasip1/mod.rs b/library/std/src/sys/pal/wasip1/mod.rs new file mode 100644 index 00000000000..61dd1c3f98b --- /dev/null +++ b/library/std/src/sys/pal/wasip1/mod.rs @@ -0,0 +1,39 @@ +//! System bindings for the wasm/web platform +//! +//! This module contains the facade (aka platform-specific) implementations of +//! OS level functionality for wasm. +//! +//! This is all super highly experimental and not actually intended for +//! wide/production use yet, it's still all in the experimental category. This +//! will likely change over time. +//! +//! Currently all functions here are basically stubs that immediately return +//! errors. The hope is that with a portability lint we can turn actually just +//! remove all this and just omit parts of the standard library if we're +//! compiling for wasm. That way it's a compile time error for something that's +//! guaranteed to be a runtime error! + +#[allow(unused)] +#[path = "../wasm/atomics/futex.rs"] +pub mod futex; + +pub mod os; +#[path = "../unsupported/pipe.rs"] +pub mod pipe; +pub mod thread; +pub mod time; + +#[path = "../unsupported/common.rs"] +#[deny(unsafe_op_in_unsafe_fn)] +#[allow(unused)] +mod common; + +pub use common::*; + +mod helpers; + +// The following exports are listed individually to work around Rust's glob +// import conflict rules. If we glob export `helpers` and `common` together, +// then the compiler complains about conflicts. + +pub(crate) use helpers::{abort_internal, decode_error_kind, err2io, is_interrupted}; diff --git a/library/std/src/sys/pal/wasip1/os.rs b/library/std/src/sys/pal/wasip1/os.rs new file mode 100644 index 00000000000..151ba254ec4 --- /dev/null +++ b/library/std/src/sys/pal/wasip1/os.rs @@ -0,0 +1,151 @@ +#![forbid(unsafe_op_in_unsafe_fn)] + +use crate::ffi::{CStr, OsStr, OsString}; +use crate::marker::PhantomData; +use crate::os::wasi::prelude::*; +use crate::path::{self, PathBuf}; +use crate::sys::common::small_c_string::run_path_with_cstr; +use crate::sys::unsupported; +use crate::{fmt, io, str}; + +// Add a few symbols not in upstream `libc` just yet. +pub mod libc { + pub use libc::*; + + unsafe extern "C" { + pub fn getcwd(buf: *mut c_char, size: size_t) -> *mut c_char; + pub fn chdir(dir: *const c_char) -> c_int; + pub fn __wasilibc_get_environ() -> *mut *mut c_char; + } +} + +pub fn errno() -> i32 { + unsafe extern "C" { + #[thread_local] + static errno: libc::c_int; + } + + unsafe { errno as i32 } +} + +pub fn error_string(errno: i32) -> String { + let mut buf = [0 as libc::c_char; 1024]; + + let p = buf.as_mut_ptr(); + unsafe { + if libc::strerror_r(errno as libc::c_int, p, buf.len()) < 0 { + panic!("strerror_r failure"); + } + str::from_utf8(CStr::from_ptr(p).to_bytes()).unwrap().to_owned() + } +} + +pub fn getcwd() -> io::Result { + let mut buf = Vec::with_capacity(512); + loop { + unsafe { + let ptr = buf.as_mut_ptr() as *mut libc::c_char; + if !libc::getcwd(ptr, buf.capacity()).is_null() { + let len = CStr::from_ptr(buf.as_ptr() as *const libc::c_char).to_bytes().len(); + buf.set_len(len); + buf.shrink_to_fit(); + return Ok(PathBuf::from(OsString::from_vec(buf))); + } else { + let error = io::Error::last_os_error(); + if error.raw_os_error() != Some(libc::ERANGE) { + return Err(error); + } + } + + // Trigger the internal buffer resizing logic of `Vec` by requiring + // more space than the current capacity. + let cap = buf.capacity(); + buf.set_len(cap); + buf.reserve(1); + } + } +} + +pub fn chdir(p: &path::Path) -> io::Result<()> { + let result = run_path_with_cstr(p, &|p| unsafe { Ok(libc::chdir(p.as_ptr())) })?; + match result == (0 as libc::c_int) { + true => Ok(()), + false => Err(io::Error::last_os_error()), + } +} + +pub struct SplitPaths<'a>(!, PhantomData<&'a ()>); + +pub fn split_paths(_unparsed: &OsStr) -> SplitPaths<'_> { + panic!("unsupported") +} + +impl<'a> Iterator for SplitPaths<'a> { + type Item = PathBuf; + fn next(&mut self) -> Option { + self.0 + } +} + +#[derive(Debug)] +pub struct JoinPathsError; + +pub fn join_paths(_paths: I) -> Result +where + I: Iterator, + T: AsRef, +{ + Err(JoinPathsError) +} + +impl fmt::Display for JoinPathsError { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + "not supported on wasm yet".fmt(f) + } +} + +impl crate::error::Error for JoinPathsError {} + +pub fn current_exe() -> io::Result { + unsupported() +} + +#[allow(dead_code)] +pub fn page_size() -> usize { + unsafe { libc::sysconf(libc::_SC_PAGESIZE) as usize } +} + +pub fn temp_dir() -> PathBuf { + panic!("no filesystem on wasm") +} + +pub fn home_dir() -> Option { + None +} + +pub fn exit(code: i32) -> ! { + unsafe { libc::exit(code) } +} + +pub fn getpid() -> u32 { + panic!("unsupported"); +} + +#[doc(hidden)] +pub trait IsMinusOne { + fn is_minus_one(&self) -> bool; +} + +macro_rules! impl_is_minus_one { + ($($t:ident)*) => ($(impl IsMinusOne for $t { + fn is_minus_one(&self) -> bool { + *self == -1 + } + })*) +} + +impl_is_minus_one! { i8 i16 i32 i64 isize } + +pub fn cvt(t: T) -> io::Result { + if t.is_minus_one() { Err(io::Error::last_os_error()) } else { Ok(t) } +} diff --git a/library/std/src/sys/pal/wasip1/thread.rs b/library/std/src/sys/pal/wasip1/thread.rs new file mode 100644 index 00000000000..e062b49bd7a --- /dev/null +++ b/library/std/src/sys/pal/wasip1/thread.rs @@ -0,0 +1,214 @@ +#![forbid(unsafe_op_in_unsafe_fn)] + +use crate::ffi::CStr; +use crate::num::NonZero; +use crate::time::{Duration, Instant}; +use crate::{io, mem}; + +cfg_select! { + target_feature = "atomics" => { + use crate::cmp; + use crate::ptr; + use crate::sys::os; + // Add a few symbols not in upstream `libc` just yet. + mod libc { + pub use crate::ffi; + pub use libc::*; + + // defined in wasi-libc + // https://github.com/WebAssembly/wasi-libc/blob/a6f871343313220b76009827ed0153586361c0d5/libc-top-half/musl/include/alltypes.h.in#L108 + #[repr(C)] + union pthread_attr_union { + __i: [ffi::c_int; if size_of::() == 8 { 14 } else { 9 }], + __vi: [ffi::c_int; if size_of::() == 8 { 14 } else { 9 }], + __s: [ffi::c_ulong; if size_of::() == 8 { 7 } else { 9 }], + } + + #[repr(C)] + pub struct pthread_attr_t { + __u: pthread_attr_union, + } + + #[allow(non_camel_case_types)] + pub type pthread_t = *mut ffi::c_void; + + pub const _SC_NPROCESSORS_ONLN: ffi::c_int = 84; + + unsafe extern "C" { + pub fn pthread_create( + native: *mut pthread_t, + attr: *const pthread_attr_t, + f: extern "C" fn(*mut ffi::c_void) -> *mut ffi::c_void, + value: *mut ffi::c_void, + ) -> ffi::c_int; + pub fn pthread_join(native: pthread_t, value: *mut *mut ffi::c_void) -> ffi::c_int; + pub fn pthread_attr_init(attrp: *mut pthread_attr_t) -> ffi::c_int; + pub fn pthread_attr_setstacksize( + attr: *mut pthread_attr_t, + stack_size: libc::size_t, + ) -> ffi::c_int; + pub fn pthread_attr_destroy(attr: *mut pthread_attr_t) -> ffi::c_int; + pub fn pthread_detach(thread: pthread_t) -> ffi::c_int; + } + } + + pub struct Thread { + id: libc::pthread_t, + } + + impl Drop for Thread { + fn drop(&mut self) { + let ret = unsafe { libc::pthread_detach(self.id) }; + debug_assert_eq!(ret, 0); + } + } + } + _ => { + pub struct Thread(!); + } +} + +pub const DEFAULT_MIN_STACK_SIZE: usize = 1024 * 1024; + +impl Thread { + // unsafe: see thread::Builder::spawn_unchecked for safety requirements + cfg_select! { + target_feature = "atomics" => { + pub unsafe fn new(stack: usize, _name: Option<&str>, p: Box) -> io::Result { + let p = Box::into_raw(Box::new(p)); + let mut native: libc::pthread_t = unsafe { mem::zeroed() }; + let mut attr: libc::pthread_attr_t = unsafe { mem::zeroed() }; + assert_eq!(unsafe { libc::pthread_attr_init(&mut attr) }, 0); + + let stack_size = cmp::max(stack, DEFAULT_MIN_STACK_SIZE); + + match unsafe { libc::pthread_attr_setstacksize(&mut attr, stack_size) } { + 0 => {} + n => { + assert_eq!(n, libc::EINVAL); + // EINVAL means |stack_size| is either too small or not a + // multiple of the system page size. Because it's definitely + // >= PTHREAD_STACK_MIN, it must be an alignment issue. + // Round up to the nearest page and try again. + let page_size = os::page_size(); + let stack_size = + (stack_size + page_size - 1) & (-(page_size as isize - 1) as usize - 1); + assert_eq!(unsafe { libc::pthread_attr_setstacksize(&mut attr, stack_size) }, 0); + } + }; + + let ret = unsafe { libc::pthread_create(&mut native, &attr, thread_start, p as *mut _) }; + // Note: if the thread creation fails and this assert fails, then p will + // be leaked. However, an alternative design could cause double-free + // which is clearly worse. + assert_eq!(unsafe {libc::pthread_attr_destroy(&mut attr) }, 0); + + return if ret != 0 { + // The thread failed to start and as a result p was not consumed. Therefore, it is + // safe to reconstruct the box so that it gets deallocated. + unsafe { drop(Box::from_raw(p)); } + Err(io::Error::from_raw_os_error(ret)) + } else { + Ok(Thread { id: native }) + }; + + extern "C" fn thread_start(main: *mut libc::c_void) -> *mut libc::c_void { + unsafe { + // Finally, let's run some code. + Box::from_raw(main as *mut Box)(); + } + ptr::null_mut() + } + } + } + _ => { + pub unsafe fn new(_stack: usize, _name: Option<&str>, _p: Box) -> io::Result { + crate::sys::unsupported() + } + } + } + + pub fn yield_now() { + let ret = unsafe { wasi::sched_yield() }; + debug_assert_eq!(ret, Ok(())); + } + + pub fn set_name(_name: &CStr) { + // nope + } + + pub fn sleep(dur: Duration) { + let mut nanos = dur.as_nanos(); + while nanos > 0 { + const USERDATA: wasi::Userdata = 0x0123_45678; + + let clock = wasi::SubscriptionClock { + id: wasi::CLOCKID_MONOTONIC, + timeout: u64::try_from(nanos).unwrap_or(u64::MAX), + precision: 0, + flags: 0, + }; + nanos -= u128::from(clock.timeout); + + let in_ = wasi::Subscription { + userdata: USERDATA, + u: wasi::SubscriptionU { tag: 0, u: wasi::SubscriptionUU { clock } }, + }; + unsafe { + let mut event: wasi::Event = mem::zeroed(); + let res = wasi::poll_oneoff(&in_, &mut event, 1); + match (res, event) { + ( + Ok(1), + wasi::Event { + userdata: USERDATA, + error: wasi::ERRNO_SUCCESS, + type_: wasi::EVENTTYPE_CLOCK, + .. + }, + ) => {} + _ => panic!("thread::sleep(): unexpected result of poll_oneoff"), + } + } + } + } + + pub fn sleep_until(deadline: Instant) { + let now = Instant::now(); + + if let Some(delay) = deadline.checked_duration_since(now) { + Self::sleep(delay); + } + } + + pub fn join(self) { + cfg_select! { + target_feature = "atomics" => { + let id = mem::ManuallyDrop::new(self).id; + let ret = unsafe { libc::pthread_join(id, ptr::null_mut()) }; + if ret != 0 { + rtabort!("failed to join thread: {}", io::Error::from_raw_os_error(ret)); + } + } + _ => { + self.0 + } + } + } +} + +pub(crate) fn current_os_id() -> Option { + None +} + +pub fn available_parallelism() -> io::Result> { + cfg_select! { + target_feature = "atomics" => { + match unsafe { libc::sysconf(libc::_SC_NPROCESSORS_ONLN) } { + -1 => Err(io::Error::last_os_error()), + cpus => NonZero::new(cpus as usize).ok_or(io::Error::UNKNOWN_THREAD_COUNT), + } + } + _ => crate::sys::unsupported(), + } +} diff --git a/library/std/src/sys/pal/wasip1/time.rs b/library/std/src/sys/pal/wasip1/time.rs new file mode 100644 index 00000000000..892661b312b --- /dev/null +++ b/library/std/src/sys/pal/wasip1/time.rs @@ -0,0 +1,76 @@ +#![forbid(unsafe_op_in_unsafe_fn)] + +use crate::time::Duration; + +#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Debug, Hash)] +pub struct Instant(Duration); + +#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Debug, Hash)] +pub struct SystemTime(Duration); + +pub const UNIX_EPOCH: SystemTime = SystemTime(Duration::from_secs(0)); + +fn current_time(clock: wasi::Clockid) -> Duration { + let ts = unsafe { + wasi::clock_time_get( + clock, 1, // precision... seems ignored though? + ) + .unwrap() + }; + Duration::new((ts / 1_000_000_000) as u64, (ts % 1_000_000_000) as u32) +} + +impl Instant { + pub fn now() -> Instant { + Instant(current_time(wasi::CLOCKID_MONOTONIC)) + } + + pub fn checked_sub_instant(&self, other: &Instant) -> Option { + self.0.checked_sub(other.0) + } + + pub fn checked_add_duration(&self, other: &Duration) -> Option { + Some(Instant(self.0.checked_add(*other)?)) + } + + pub fn checked_sub_duration(&self, other: &Duration) -> Option { + Some(Instant(self.0.checked_sub(*other)?)) + } +} + +impl SystemTime { + pub fn now() -> SystemTime { + SystemTime(current_time(wasi::CLOCKID_REALTIME)) + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn from_wasi_timestamp(ts: wasi::Timestamp) -> SystemTime { + SystemTime(Duration::from_nanos(ts)) + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn to_wasi_timestamp(&self) -> Option { + // FIXME: const TryInto + let ns = self.0.as_nanos(); + if ns <= u64::MAX as u128 { Some(ns as u64) } else { None } + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn sub_time(&self, other: &SystemTime) -> Result { + // FIXME: ok_or_else with const closures + match self.0.checked_sub(other.0) { + Some(duration) => Ok(duration), + None => Err(other.0 - self.0), + } + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn checked_add_duration(&self, other: &Duration) -> Option { + Some(SystemTime(self.0.checked_add(*other)?)) + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn checked_sub_duration(&self, other: &Duration) -> Option { + Some(SystemTime(self.0.checked_sub(*other)?)) + } +} diff --git a/library/std/src/sys/pal/wasip2/mod.rs b/library/std/src/sys/pal/wasip2/mod.rs index 47fe3221c90..5f3fb6d6ddf 100644 --- a/library/std/src/sys/pal/wasip2/mod.rs +++ b/library/std/src/sys/pal/wasip2/mod.rs @@ -10,13 +10,11 @@ #[path = "../wasm/atomics/futex.rs"] pub mod futex; -#[path = "../wasi/os.rs"] +#[path = "../wasip1/os.rs"] pub mod os; #[path = "../unsupported/pipe.rs"] pub mod pipe; -#[path = "../wasi/thread.rs"] pub mod thread; -#[path = "../wasi/time.rs"] pub mod time; #[path = "../unsupported/common.rs"] @@ -26,7 +24,7 @@ mod common; pub use common::*; -#[path = "../wasi/helpers.rs"] +#[path = "../wasip1/helpers.rs"] mod helpers; // The following exports are listed individually to work around Rust's glob diff --git a/library/std/src/sys/pal/wasip2/thread.rs b/library/std/src/sys/pal/wasip2/thread.rs new file mode 100644 index 00000000000..ad52918f15a --- /dev/null +++ b/library/std/src/sys/pal/wasip2/thread.rs @@ -0,0 +1,73 @@ +use crate::ffi::CStr; +use crate::io; +use crate::num::NonZero; +use crate::time::{Duration, Instant}; + +pub struct Thread(!); + +pub const DEFAULT_MIN_STACK_SIZE: usize = 1024 * 1024; + +impl Thread { + pub unsafe fn new( + _stack: usize, + _name: Option<&str>, + _p: Box, + ) -> io::Result { + // Note that unlike WASIp1 even if the wasm `atomics` feature is enabled + // there is no support for threads, not even experimentally, not even in + // wasi-libc. Thus this is unconditionally unsupported. + crate::sys::unsupported() + } + + pub fn yield_now() { + // no API for this in WASIp2, but there's also no threads, so that's + // sort of expected. + } + + pub fn set_name(_name: &CStr) { + // nope + } + + pub fn sleep(dur: Duration) { + // Sleep in increments of `u64::MAX` nanoseconds until the `dur` is + // entirely drained. + let mut remaining = dur.as_nanos(); + while remaining > 0 { + let amt = u64::try_from(remaining).unwrap_or(u64::MAX); + wasip2::clocks::monotonic_clock::subscribe_duration(amt).block(); + remaining -= u128::from(amt); + } + } + + pub fn sleep_until(deadline: Instant) { + match u64::try_from(deadline.into_inner().as_duration().as_nanos()) { + // If the point in time we're sleeping to fits within a 64-bit + // number of nanoseconds then directly use `subscribe_instant`. + Ok(deadline) => { + wasip2::clocks::monotonic_clock::subscribe_instant(deadline).block(); + } + // ... otherwise we're sleeping for 500+ years relative to the + // "start" of what the system is using as a clock so speed/accuracy + // is not so much of a concern. Use `sleep` instead. + Err(_) => { + let now = Instant::now(); + + if let Some(delay) = deadline.checked_duration_since(now) { + Self::sleep(delay); + } + } + } + } + + pub fn join(self) { + self.0 + } +} + +pub(crate) fn current_os_id() -> Option { + None +} + +pub fn available_parallelism() -> io::Result> { + crate::sys::unsupported() +} diff --git a/library/std/src/sys/pal/wasip2/time.rs b/library/std/src/sys/pal/wasip2/time.rs new file mode 100644 index 00000000000..f1f6839774b --- /dev/null +++ b/library/std/src/sys/pal/wasip2/time.rs @@ -0,0 +1,69 @@ +use crate::time::Duration; + +#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Debug, Hash)] +pub struct Instant(Duration); + +#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Debug, Hash)] +pub struct SystemTime(Duration); + +pub const UNIX_EPOCH: SystemTime = SystemTime(Duration::from_secs(0)); + +impl Instant { + pub fn now() -> Instant { + Instant(Duration::from_nanos(wasip2::clocks::monotonic_clock::now())) + } + + pub fn checked_sub_instant(&self, other: &Instant) -> Option { + self.0.checked_sub(other.0) + } + + pub fn checked_add_duration(&self, other: &Duration) -> Option { + Some(Instant(self.0.checked_add(*other)?)) + } + + pub fn checked_sub_duration(&self, other: &Duration) -> Option { + Some(Instant(self.0.checked_sub(*other)?)) + } + + pub(super) fn as_duration(&self) -> &Duration { + &self.0 + } +} + +impl SystemTime { + pub fn now() -> SystemTime { + let now = wasip2::clocks::wall_clock::now(); + SystemTime(Duration::new(now.seconds, now.nanoseconds)) + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn from_wasi_timestamp(ts: wasi::Timestamp) -> SystemTime { + SystemTime(Duration::from_nanos(ts)) + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn to_wasi_timestamp(&self) -> Option { + // FIXME: const TryInto + let ns = self.0.as_nanos(); + if ns <= u64::MAX as u128 { Some(ns as u64) } else { None } + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn sub_time(&self, other: &SystemTime) -> Result { + // FIXME: ok_or_else with const closures + match self.0.checked_sub(other.0) { + Some(duration) => Ok(duration), + None => Err(other.0 - self.0), + } + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn checked_add_duration(&self, other: &Duration) -> Option { + Some(SystemTime(self.0.checked_add(*other)?)) + } + + #[rustc_const_unstable(feature = "const_system_time", issue = "144517")] + pub const fn checked_sub_duration(&self, other: &Duration) -> Option { + Some(SystemTime(self.0.checked_sub(*other)?)) + } +} diff --git a/library/std/src/sys/random/mod.rs b/library/std/src/sys/random/mod.rs index a7fbae86099..1e0eec07b50 100644 --- a/library/std/src/sys/random/mod.rs +++ b/library/std/src/sys/random/mod.rs @@ -86,9 +86,13 @@ cfg_select! { mod vxworks; pub use vxworks::fill_bytes; } - target_os = "wasi" => { - mod wasi; - pub use wasi::fill_bytes; + all(target_os = "wasi", target_env = "p1") => { + mod wasip1; + pub use wasip1::fill_bytes; + } + all(target_os = "wasi", target_env = "p2") => { + mod wasip2; + pub use wasip2::{fill_bytes, hashmap_random_keys}; } target_os = "zkvm" => { mod zkvm; @@ -110,6 +114,7 @@ cfg_select! { target_os = "linux", target_os = "android", all(target_family = "wasm", target_os = "unknown"), + all(target_os = "wasi", target_env = "p2"), target_os = "xous", )))] pub fn hashmap_random_keys() -> (u64, u64) { diff --git a/library/std/src/sys/random/wasi.rs b/library/std/src/sys/random/wasi.rs deleted file mode 100644 index d41da3751fc..00000000000 --- a/library/std/src/sys/random/wasi.rs +++ /dev/null @@ -1,5 +0,0 @@ -pub fn fill_bytes(bytes: &mut [u8]) { - unsafe { - wasi::random_get(bytes.as_mut_ptr(), bytes.len()).expect("failed to generate random data") - } -} diff --git a/library/std/src/sys/random/wasip1.rs b/library/std/src/sys/random/wasip1.rs new file mode 100644 index 00000000000..d41da3751fc --- /dev/null +++ b/library/std/src/sys/random/wasip1.rs @@ -0,0 +1,5 @@ +pub fn fill_bytes(bytes: &mut [u8]) { + unsafe { + wasi::random_get(bytes.as_mut_ptr(), bytes.len()).expect("failed to generate random data") + } +} diff --git a/library/std/src/sys/random/wasip2.rs b/library/std/src/sys/random/wasip2.rs new file mode 100644 index 00000000000..a67c8a6428d --- /dev/null +++ b/library/std/src/sys/random/wasip2.rs @@ -0,0 +1,9 @@ +pub fn fill_bytes(bytes: &mut [u8]) { + bytes.copy_from_slice(&wasip2::random::random::get_random_bytes( + u64::try_from(bytes.len()).unwrap(), + )); +} + +pub fn hashmap_random_keys() -> (u64, u64) { + wasip2::random::insecure_seed::insecure_seed() +} diff --git a/src/tools/tidy/src/deps.rs b/src/tools/tidy/src/deps.rs index 6974ede624a..560f11ecf50 100644 --- a/src/tools/tidy/src/deps.rs +++ b/src/tools/tidy/src/deps.rs @@ -490,6 +490,7 @@ const PERMITTED_STDLIB_DEPENDENCIES: &[&str] = &[ "windows_x86_64_gnu", "windows_x86_64_gnullvm", "windows_x86_64_msvc", + "wit-bindgen", // tidy-alphabetical-end ]; @@ -798,7 +799,10 @@ fn check_runtime_no_duplicate_dependencies(metadata: &Metadata, bad: &mut bool) continue; } - if !seen_pkgs.insert(&*pkg.name) { + // Skip the `wasi` crate here which the standard library explicitly + // depends on two version of (one for the `wasm32-wasip1` target and + // another for the `wasm32-wasip2` target). + if pkg.name.to_string() != "wasi" && !seen_pkgs.insert(&*pkg.name) { tidy_error!( bad, "duplicate package `{}` is not allowed for the standard library", diff --git a/triagebot.toml b/triagebot.toml index 3b7b216c0d2..cf2e95203fa 100644 --- a/triagebot.toml +++ b/triagebot.toml @@ -450,7 +450,8 @@ trigger_files = [ [autolabel."O-wasi"] trigger_files = [ - "library/std/src/sys/pal/wasi", + "library/std/src/sys/pal/wasip1", + "library/std/src/sys/pal/wasip2", "library/std/src/os/wasi" ] -- cgit 1.4.1-3-g733a5 From 03bffa88011c6f59c0b62940aeb98fce212e4e82 Mon Sep 17 00:00:00 2001 From: Miguel Ojeda Date: Wed, 3 Sep 2025 10:46:34 +0200 Subject: CI: rfl: move job forward to Linux v6.17-rc3 plus 2 commits This upgrades the Rust CI from v6.16-rc1 plus a temporary commit for the >= 1.91 target spec [1] to v6.17-rc3 with two commits pending to be merged upstream -- one for the same target spec format change [1] and another for the `file_as_c_str` change [2]. Link: https://github.com/rust-lang/rust/pull/144443 [1] Link: https://github.com/rust-lang/rust/pull/145928 [2] Signed-off-by: Miguel Ojeda --- src/ci/docker/scripts/rfl-build.sh | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'src') diff --git a/src/ci/docker/scripts/rfl-build.sh b/src/ci/docker/scripts/rfl-build.sh index 70ff5d0d2c7..7e0fb5794f4 100755 --- a/src/ci/docker/scripts/rfl-build.sh +++ b/src/ci/docker/scripts/rfl-build.sh @@ -3,7 +3,8 @@ set -euo pipefail # https://github.com/rust-lang/rust/pull/144443 -LINUX_VERSION=7770d51bce622b13195b2d3c85407282fc9c27e5 +# https://github.com/rust-lang/rust/pull/145928 +LINUX_VERSION=8851e27d2cb947ea8bbbe8e812068f7bf5cbd00b # Build rustc, rustdoc, cargo, clippy-driver and rustfmt ../x.py build --stage 2 library rustdoc clippy rustfmt -- cgit 1.4.1-3-g733a5