Fix cp --sparse=auto not creating sparse-files.

src/uu/cp/src/platform/linux.rs

@@ -3,10 +3,11 @@
 // For the full copyright and license information, please view the LICENSE
 // file that was distributed with this source code.
 // spell-checker:ignore ficlone reflink ftruncate pwrite fiemap
-use std::fs::{File, OpenOptions};
+use std::fs::{self, File, OpenOptions};
 use std::io::Read;
-use std::os::unix::fs::OpenOptionsExt;
+use std::os::unix::fs::{FileTypeExt, OpenOptionsExt};
 use std::os::unix::io::AsRawFd;
+use std::os::unix::prelude::MetadataExt;
 use std::path::Path;
 
 use quick_error::ResultExt;
@@ -32,6 +33,9 @@ enum CloneFallback {
 
     /// Use [`std::fs::copy`].
     FSCopy,
+
+    /// Use sparse_copy
+    SparseCopy,
 }
 
 /// Use the Linux `ioctl_ficlone` API to do a copy-on-write clone.
@@ -53,6 +57,7 @@ where
     match fallback {
         CloneFallback::Error => Err(std::io::Error::last_os_error()),
         CloneFallback::FSCopy => std::fs::copy(source, dest).map(|_| ()),
+        CloneFallback::SparseCopy => sparse_copy(source, dest),
     }
 }
 
@@ -62,8 +67,6 @@ fn sparse_copy<P>(source: P, dest: P) -> std::io::Result<()>
 where
     P: AsRef<Path>,
 {
-    use std::os::unix::prelude::MetadataExt;
-
     let mut src_file = File::open(source)?;
     let dst_file = File::create(dest)?;
     let dst_fd = dst_file.as_raw_fd();
@@ -132,6 +135,48 @@ where
     Ok(num_bytes_copied)
 }
 
+#[cfg(any(target_os = "linux", target_os = "android"))]
+fn check_dest_is_fifo(dest: &Path) -> bool {
+    // If our destination file exists and it is a FIFO, we do a standard copy .
+    let file_type = fs::metadata(dest);
+    match file_type {
+        Ok(f) => f.file_type().is_fifo(),
+
+        _ => false,
+    }
+}
+
+#[cfg(any(target_os = "linux", target_os = "android"))]
+fn check_sparse_detection(source: &Path) -> Result<bool, std::io::Error> {
+    let mut src_file = File::open(source)?;
+
+    let metadata = src_file.metadata()?;
+    let size = metadata.size();
+    let blocks = metadata.blocks();
+    let blksize = metadata.blksize();
+    // cp uses a crude heuristic to detect sparse_files, an estimated formula which seems to accurately
+    // replicate that, is used. Modifications might be required if we observe any edges cases but this
+    // should work for majority of the files.
+    // Reference:https://doc.rust-lang.org/std/os/unix/fs/trait.MetadataExt.html#tymethod.blocks
+
+    let mut buf = vec![0; blksize as usize];
+
+    // Reading the file to account for virtual files which have 0 blocks allocated in disk
+    // and it is unknown if it contains data or not, virtual files need to be standard copied and
+    // categorized as not sparse.
+
+    if blocks == 0 {
+        let _ = src_file.read(&mut buf)?;
+
+        if buf.iter().any(|&x| x != 0x0) {
+            return Ok(false);
+        }
+    } else if blocks < size / 512 {
+        return Ok(true);
+    }
+    Ok(false)
+}
+
 /// Copies `source` to `dest` using copy-on-write if possible.
 ///
 /// The `source_is_fifo` flag must be set to `true` if and only if
@@ -159,19 +204,42 @@ pub(crate) fn copy_on_write(
             copy_debug.reflink = OffloadReflinkDebug::No;
             sparse_copy(source, dest)
         }
-        (ReflinkMode::Never, _) => {
+        (ReflinkMode::Never, SparseMode::Never) => {
             copy_debug.sparse_detection = SparseDebug::No;
             copy_debug.reflink = OffloadReflinkDebug::No;
             std::fs::copy(source, dest).map(|_| ())
         }
+        (ReflinkMode::Never, SparseMode::Auto) => {
+            copy_debug.sparse_detection = SparseDebug::No;
+            copy_debug.reflink = OffloadReflinkDebug::No;
+            match check_sparse_detection(source) {
+                Ok(true) => sparse_copy(source, dest),
+                Ok(false) => std::fs::copy(source, dest).map(|_| ()),
+                Err(e) => Err(e),
+            }
+        }
         (ReflinkMode::Auto, SparseMode::Always) => {
             copy_debug.offload = OffloadReflinkDebug::Avoided;
             copy_debug.sparse_detection = SparseDebug::Zeros;
             copy_debug.reflink = OffloadReflinkDebug::Unsupported;
             sparse_copy(source, dest)
         }
 
-        (ReflinkMode::Auto, _) => {
+        (ReflinkMode::Auto, SparseMode::Auto) => {
+            copy_debug.sparse_detection = SparseDebug::No;
+            copy_debug.reflink = OffloadReflinkDebug::Unsupported;
+            if source_is_fifo {
+                copy_fifo_contents(source, dest).map(|_| ())
+            } else {
+                match (check_sparse_detection(source), check_dest_is_fifo(dest)) {
+                    (Ok(true), false) => clone(source, dest, CloneFallback::SparseCopy),
+                    (Ok(true), true) => clone(source, dest, CloneFallback::FSCopy),
+                    (Ok(false), _) => clone(source, dest, CloneFallback::FSCopy),
+                    (Err(e), _) => Err(e),
+                }
+            }
+        }
+        (ReflinkMode::Auto, SparseMode::Never) => {
             copy_debug.sparse_detection = SparseDebug::No;
             copy_debug.reflink = OffloadReflinkDebug::Unsupported;
             if source_is_fifo {

tests/by-util/test_cp.rs

@@ -2330,6 +2330,26 @@ fn test_cp_sparse_always_non_empty() {
     assert_eq!(at.metadata("dst_file_sparse").blocks(), touched_block_count);
 }
 
+#[test]
+#[cfg(target_os = "linux")]
+fn test_cp_for_virtual_files() {
+    use std::os::unix::prelude::MetadataExt;
+    let ts = TestScenario::new(util_name!());
+    let at = &ts.fixtures;
+    ts.ucmd()
+        .arg("--sparse=auto")
+        .arg("/sys/kernel/address_bits")
+        .arg("b")
+        .succeeds();
+
+    let dest_size = std::fs::metadata(at.plus("b"))
+        .expect("Metadata of copied file cannot be read")
+        .size();
+    if dest_size == 0 {
+        panic!("Copy unsuccessful");
+    }
+}
+
 #[cfg(any(target_os = "linux", target_os = "android"))]
 #[test]
 fn test_cp_sparse_invalid_option() {
@@ -3218,6 +3238,65 @@ fn test_copy_contents_fifo() {
     assert_eq!(at.read("outfile"), "foo");
 }
 
+#[cfg(any(target_os = "linux", target_os = "android"))]
+#[test]
+fn test_sparse_auto_for_sparse_files() {
+    let ts = TestScenario::new(util_name!());
+    let at = &ts.fixtures;
+    at.touch("a");
+    at.write("a", "hello");
+    at.truncate("a", "-s10000");
+    ts.ucmd().arg("--sparse=auto").arg("a").arg("b").succeeds();
+
+    let src_size = std::fs::metadata(at.plus("a"))
+        .expect("Metadata of source file cannot be read")
+        .size();
+
+    let src_blocks = std::fs::metadata(at.plus("a"))
+        .expect("Metadata of source file cannot be read")
+        .blocks();
+
+    let dest_size = std::fs::metadata(at.plus("b"))
+        .expect("Metadata of copied file cannot be read")
+        .size();
+    let dest_blocks = std::fs::metadata(at.plus("b"))
+        .expect("Metadata of copied file cannot be read")
+        .blocks();
+
+    assert_eq!(src_size, dest_size);
+    assert_eq!(src_blocks, dest_blocks);
+}
+
+#[cfg(any(target_os = "linux", target_os = "android"))]
+#[test]
+fn test_sparse_auto_for_non_sparse_files() {
+    let ts = TestScenario::new(util_name!());
+    let at = &ts.fixtures;
+    let a_bytes = [1 as u8; 10000];
+    at.touch("a");
+    at.write("a", "hello");
+    at.append_bytes("a", &a_bytes);
+    ts.ucmd().arg("--sparse=auto").arg("a").arg("b").succeeds();
+
+    let src_size = std::fs::metadata(at.plus("a"))
+        .expect("Metadata of source file cannot be read")
+        .size();
+
+    let src_blocks = std::fs::metadata(at.plus("a"))
+        .expect("Metadata of source file cannot be read")
+        .blocks();
+
+    let dest_size = std::fs::metadata(at.plus("b"))
+        .expect("Metadata of copied file cannot be read")
+        .size();
+    let dest_blocks = std::fs::metadata(at.plus("b"))
+        .expect("Metadata of copied file cannot be read")
+        .blocks();
+
+    assert_eq!(src_size, dest_size);
+    assert_eq!(src_blocks, dest_blocks);
+}
+
 #[cfg(target_os = "linux")]
 #[test]
 fn test_reflink_never_sparse_always() {