Merge pull request #1657 from Kobzol/artifact-size

Store artifact sizes into the DB

Author: Kobzol

collector/src/bin/collector.rs

@@ -11,7 +11,7 @@ use collector::compile::benchmark::{
     compile_benchmark_dir, get_compile_benchmarks, Benchmark, BenchmarkName,
 };
 use collector::{runtime, utils, CollectorCtx, CollectorStepBuilder};
-use database::{ArtifactId, Commit, CommitType, Connection, Pool};
+use database::{ArtifactId, ArtifactIdNumber, Commit, CommitType, Connection, Pool};
 use rayon::iter::{IndexedParallelIterator, IntoParallelRefIterator, ParallelIterator};
 use std::cmp::Ordering;
 use std::ffi::OsStr;
@@ -921,8 +921,8 @@ fn main_result() -> anyhow::Result<i32> {
             sysroot.preserve(); // don't delete it
 
             // Print the directory containing the toolchain.
-            sysroot.rustc.pop();
-            let s = format!("{:?}", sysroot.rustc);
+            sysroot.components.rustc.pop();
+            let s = format!("{:?}", sysroot.components.rustc);
             println!("{}", &s[1..s.len() - 1]);
 
             Ok(0)
@@ -978,6 +978,12 @@ fn run_benchmarks(
     compile: Option<CompileBenchmarkConfig>,
     runtime: Option<RuntimeBenchmarkConfig>,
 ) -> anyhow::Result<()> {
+    rt.block_on(record_toolchain_sizes(
+        connection.as_mut(),
+        &shared.artifact_id,
+        &shared.toolchain,
+    ));
+
     let collector = rt.block_on(init_collection(
         connection.as_mut(),
         &shared,
@@ -1190,6 +1196,38 @@ fn bench_compile(
     errors
 }
 
+/// Records the sizes of individual components (rustc, libLLVM, etc.) for the given toolchain
+/// and artifact id into the database.
+async fn record_toolchain_sizes(
+    conn: &mut dyn Connection,
+    artifact_id: &ArtifactId,
+    toolchain: &Toolchain,
+) {
+    let aid = conn.artifact_id(artifact_id).await;
+
+    async fn record(
+        conn: &mut dyn Connection,
+        aid: ArtifactIdNumber,
+        component: &str,
+        path: Option<&Path>,
+    ) {
+        if let Some(path) = path {
+            if let Ok(size) = fs::metadata(path).map(|m| m.len()) {
+                conn.record_artifact_size(aid, component, size).await;
+            }
+        }
+    }
+
+    let paths = &toolchain.components;
+    record(conn, aid, "rustc", Some(&paths.rustc)).await;
+    record(conn, aid, "rustdoc", paths.rustdoc.as_deref()).await;
+    record(conn, aid, "cargo", Some(&paths.cargo)).await;
+    record(conn, aid, "librustc_driver", paths.lib_rustc.as_deref()).await;
+    record(conn, aid, "libstd", paths.lib_std.as_deref()).await;
+    record(conn, aid, "libtest", paths.lib_test.as_deref()).await;
+    record(conn, aid, "libLLVM", paths.lib_llvm.as_deref()).await;
+}
+
 fn add_perf_config(directory: &Path, category: Category) {
     let data = serde_json::json!({
         "category": category.to_string()

collector/src/compile/execute/mod.rs

@@ -144,12 +144,12 @@ impl<'a> CargoProcess<'a> {
     }
 
     fn base_command(&self, cwd: &Path, subcommand: &str) -> Command {
-        let mut cmd = Command::new(Path::new(&self.toolchain.cargo));
+        let mut cmd = Command::new(Path::new(&self.toolchain.components.cargo));
         cmd
             // Not all cargo invocations (e.g. `cargo clean`) need all of these
             // env vars set, but it doesn't hurt to have them.
             .env("RUSTC", &*FAKE_RUSTC)
-            .env("RUSTC_REAL", &self.toolchain.rustc)
+            .env("RUSTC_REAL", &self.toolchain.components.rustc)
             // We separately pass -Cincremental to the leaf crate --
             // CARGO_INCREMENTAL is cached separately for both the leaf crate
             // and any in-tree dependencies, and we don't want that; it wastes
@@ -164,7 +164,7 @@ impl<'a> CargoProcess<'a> {
             .arg("--manifest-path")
             .arg(&self.manifest_path);
 
-        if let Some(r) = &self.toolchain.rustdoc {
+        if let Some(r) = &self.toolchain.components.rustdoc {
             cmd.env("RUSTDOC", &*FAKE_RUSTDOC).env("RUSTDOC_REAL", r);
         }
         cmd

collector/src/compile/execute/rustc.rs

@@ -95,11 +95,11 @@ fn record(
     .arg("rust.deny-warnings=false")
     .arg("--set")
     .arg(&format!("build.rustc={}", fake_rustc.to_str().unwrap()))
-    .env("RUSTC_PERF_REAL_RUSTC", &toolchain.rustc)
+    .env("RUSTC_PERF_REAL_RUSTC", &toolchain.components.rustc)
     .arg("--set")
     .arg(&format!(
         "build.cargo={}",
-        toolchain.cargo.to_str().unwrap()
+        toolchain.components.cargo.to_str().unwrap()
     ))
     .status()
     .context("configuring")?;
@@ -114,7 +114,7 @@ fn record(
                     .context("x.py script canonicalize")?,
             )
             .current_dir(checkout)
-            .env("RUSTC_PERF_REAL_RUSTC", &toolchain.rustc)
+            .env("RUSTC_PERF_REAL_RUSTC", &toolchain.components.rustc)
             .arg("build")
             .arg("--stage")
             .arg("0")

collector/src/runtime/benchmark.rs

@@ -240,9 +240,9 @@ fn start_cargo_build(
     benchmark_dir: &Path,
     target_dir: Option<&Path>,
 ) -> anyhow::Result<Child> {
-    let mut command = Command::new(&toolchain.cargo);
+    let mut command = Command::new(&toolchain.components.cargo);
     command
-        .env("RUSTC", &toolchain.rustc)
+        .env("RUSTC", &toolchain.components.rustc)
         .arg("build")
         .arg("--release")
         .arg("--message-format")

collector/src/toolchain.rs

@@ -1,6 +1,7 @@
 use crate::compile::benchmark::profile::Profile;
 use anyhow::{anyhow, Context};
 use log::debug;
+use std::ffi::OsStr;
 use std::fs::{self, File};
 use std::io::{BufReader, Read};
 use std::path::{Path, PathBuf};
@@ -12,9 +13,7 @@ use xz2::bufread::XzDecoder;
 /// Sysroot downloaded from CI.
 pub struct Sysroot {
     pub sha: String,
-    pub rustc: PathBuf,
-    pub rustdoc: PathBuf,
-    pub cargo: PathBuf,
+    pub components: ToolchainComponents,
     pub triple: String,
     pub preserve: bool,
 }
@@ -121,10 +120,15 @@ impl SysrootDownload {
             })
         };
 
+        let components = ToolchainComponents::from_binaries_and_libdir(
+            sysroot_bin("rustc")?,
+            Some(sysroot_bin("rustdoc")?),
+            sysroot_bin("cargo")?,
+            &self.directory.join(&self.rust_sha).join("lib"),
+        )?;
+
         Ok(Sysroot {
-            rustc: sysroot_bin("rustc")?,
-            rustdoc: sysroot_bin("rustdoc")?,
-            cargo: sysroot_bin("cargo")?,
+            components,
             sha: self.rust_sha,
             triple: self.triple,
             preserve: false,
@@ -218,25 +222,72 @@ impl SysrootDownload {
 /// Representation of a toolchain that can be used to compile Rust programs.
 #[derive(Debug, Clone)]
 pub struct Toolchain {
-    pub rustc: PathBuf,
-    pub rustdoc: Option<PathBuf>,
-    pub cargo: PathBuf,
+    pub components: ToolchainComponents,
     pub id: String,
     pub triple: String,
 }
 
 impl Toolchain {
     pub fn from_sysroot(sysroot: &Sysroot, id: String) -> Self {
         Self {
-            rustc: sysroot.rustc.clone(),
-            rustdoc: Some(sysroot.rustdoc.clone()),
-            cargo: sysroot.cargo.clone(),
+            components: sysroot.components.clone(),
             id,
             triple: sysroot.triple.clone(),
         }
     }
 }
 
+#[derive(Debug, Clone, Default)]
+pub struct ToolchainComponents {
+    pub rustc: PathBuf,
+    pub rustdoc: Option<PathBuf>,
+    pub cargo: PathBuf,
+    pub lib_rustc: Option<PathBuf>,
+    pub lib_std: Option<PathBuf>,
+    pub lib_test: Option<PathBuf>,
+    pub lib_llvm: Option<PathBuf>,
+}
+
+impl ToolchainComponents {
+    fn from_binaries_and_libdir(
+        rustc: PathBuf,
+        rustdoc: Option<PathBuf>,
+        cargo: PathBuf,
+        libdir: &Path,
+    ) -> anyhow::Result<Self> {
+        let mut component = ToolchainComponents {
+            rustc,
+            rustdoc,
+            cargo,
+            ..Default::default()
+        };
+        component.fill_libraries(libdir)?;
+        Ok(component)
+    }
+
+    /// Finds known library components in the given `dir` and stores them in `self`.
+    fn fill_libraries(&mut self, dir: &Path) -> anyhow::Result<()> {
+        for entry in fs::read_dir(dir).context("Cannot read lib dir to find components")? {
+            let entry = entry?;
+            let path = entry.path();
+            if path.is_file() && path.extension() == Some(OsStr::new("so")) {
+                if let Some(filename) = path.file_name().and_then(|s| s.to_str()) {
+                    if filename.starts_with("libLLVM") {
+                        self.lib_llvm = Some(path);
+                    } else if filename.starts_with("librustc_driver") {
+                        self.lib_rustc = Some(path);
+                    } else if filename.starts_with("libstd") {
+                        self.lib_std = Some(path);
+                    } else if filename.starts_with("libtest") {
+                        self.lib_test = Some(path);
+                    }
+                }
+            }
+        }
+        Ok(())
+    }
+}
+
 /// Get a toolchain from the input.
 /// - `rustc`: check if the given one is acceptable.
 /// - `rustdoc`: if one is given, check if it is acceptable. Otherwise, if
@@ -374,10 +425,10 @@ pub fn get_local_toolchain(
         cargo
     };
 
+    let lib_dir = get_lib_dir_from_rustc(&rustc).context("Cannot find libdir for rustc")?;
+
     Ok(Toolchain {
-        rustc,
-        rustdoc,
-        cargo,
+        components: ToolchainComponents::from_binaries_and_libdir(rustc, rustdoc, cargo, &lib_dir)?,
         id,
         triple: target_triple,
     })
@@ -420,11 +471,25 @@ pub fn create_toolchain_from_published_version(
     debug!("Found rustdoc: {}", rustdoc.display());
     debug!("Found cargo: {}", cargo.display());
 
+    let lib_dir = get_lib_dir_from_rustc(&rustc)?;
+
+    let components =
+        ToolchainComponents::from_binaries_and_libdir(rustc, Some(rustdoc), cargo, &lib_dir)?;
+
     Ok(Toolchain {
-        rustc,
-        rustdoc: Some(rustdoc),
-        cargo,
+        components,
         id: toolchain.to_string(),
         triple: target_triple.to_string(),
     })
 }
+
+fn get_lib_dir_from_rustc(rustc: &Path) -> anyhow::Result<PathBuf> {
+    let sysroot = Command::new(rustc)
+        .arg("--print")
+        .arg("sysroot")
+        .output()?
+        .stdout;
+    let sysroot_path = String::from_utf8_lossy(&sysroot);
+
+    Ok(Path::new(sysroot_path.as_ref().trim()).join("lib"))
+}

database/schema.md

@@ -61,6 +61,22 @@ id          name        date        type
 1           LOCAL_TEST              release
 ```
 
+### artifact size
+
+Records the size of individual components (like `librustc_driver.so` or `libLLVM.so`) of a single
+artifact.
+
+This description includes:
+* component: normalized name of the component (hashes are removed)
+* size: size of the component in bytes
+
+```
+sqlite> select * from artifact_size limit 1;
+aid         component   size
+----------  ----------  ----------
+1           libLLVM.so  177892352
+```
+
 ### collection
 
 A "collection" of benchmarks tied only differing by the statistic collected.
@@ -267,26 +283,13 @@ bors_sha    pr  parent_sha  complete  requested    include  exclude  runs  commi
 1w0p83...   42  fq24xq...   true      <timestamp>                    3     <timestamp>
 ```
 
-### error_series
-
-Records a compile-time benchmark that caused an error.
-
-This table exists to avoid duplicating benchmarks many times in the `error` table.
-
-```
-sqlite> select * from error_series limit 1;
-id          crate
-----------  -----------
-1           hello-world
-```
-
 ### error
 
-Records a compilation error for an artifact and an entry in `error_series`.
+Records a compilation or runtime error for an artifact and a benchmark.
 
 ```
 sqlite> select * from error limit 1;
-series      aid  error
-----------  ---  -----
-1           42   Failed to compile...
+aid         benchmark   error
+----------  ---         -----
+1           syn-1.0.89  Failed to compile...
 ```

database/src/pool.rs

@@ -86,6 +86,11 @@ pub trait Connection: Send + Sync {
         krate: &str,
         value: Duration,
     );
+
+    /// Records the size of an artifact component (like `librustc_driver.so` or `libLLVM.so`) in
+    /// bytes.
+    async fn record_artifact_size(&self, artifact: ArtifactIdNumber, component: &str, size: u64);
+
     /// Returns vector of bootstrap build times for the given artifacts. The kth
     /// element is the minimum build time for the kth artifact in `aids`, across
     /// all collections for the artifact, or none if there is no bootstrap data