FileSystem.Unix: improve CopyFile.

src/libraries/Common/src/Interop/Unix/System.Native/Interop.CopyFile.cs

@@ -10,6 +10,6 @@ internal static partial class Interop
     internal static partial class Sys
     {
         [GeneratedDllImport(Libraries.SystemNative, EntryPoint = "SystemNative_CopyFile", SetLastError = true)]
-        internal static partial int CopyFile(SafeFileHandle source, SafeFileHandle destination);
+        internal static partial int CopyFile(SafeFileHandle source, SafeFileHandle destination, long sourceLength);
     }
 }

src/libraries/Native/Unix/System.Native/pal_io.c

@@ -61,6 +61,14 @@ extern int     getpeereid(int, uid_t *__restrict__, gid_t *__restrict__);
 #endif
 #endif
 
+// The portable build is performed on RHEL7 which doesn't define FICLONE yet.
+// Ensure FICLONE is defined for all Linux builds.
+#ifdef __linux__
+#ifndef FICLONE
+#define FICLONE _IOW(0x94, 9, int)
+#endif
+#endif
+
 #if HAVE_STAT64
 #define stat_ stat64
 #define fstat_ fstat64
@@ -1138,8 +1146,10 @@ static int32_t CopyFile_ReadWrite(int inFd, int outFd)
 }
 #endif // !HAVE_FCOPYFILE
 
-int32_t SystemNative_CopyFile(intptr_t sourceFd, intptr_t destinationFd)
+int32_t SystemNative_CopyFile(intptr_t sourceFd, intptr_t destinationFd, int64_t sourceLength)
 {
+    (void)sourceLength; // unused on some platforms.
+
     int inFd = ToFileDescriptor(sourceFd);
     int outFd = ToFileDescriptor(destinationFd);
 
@@ -1151,28 +1161,27 @@ int32_t SystemNative_CopyFile(intptr_t sourceFd, intptr_t destinationFd)
 #else
     // Get the stats on the source file.
     int ret;
-    struct stat_ sourceStat;
     bool copied = false;
-#if HAVE_SENDFILE_4
-    // If sendfile is available (Linux), try to use it, as the whole copy
-    // can be performed in the kernel, without lots of unnecessary copying.
-    while ((ret = fstat_(inFd, &sourceStat)) < 0 && errno == EINTR);
-    if (ret != 0)
+
+    // Certain files (e.g. procfs) may return a size of 0 even though reading them will
+    // produce data. We use plain read/write for those.
+#ifdef FICLONE
+    // Try copying data using a copy-on-write clone. This shares storage between the files.
+    if (sourceLength != 0)
     {
-        return -1;
+        while ((ret = ioctl(outFd, FICLONE, inFd)) < 0 && errno == EINTR);
+        copied = ret == 0;
     }
-
-    // On 32-bit, if you use 64-bit offsets, the last argument of `sendfile' will be a
-    // `size_t' a 32-bit integer while the `st_size' field of the stat structure will be off64_t.
-    // So `size' will have to be `uint64_t'. In all other cases, it will be `size_t'.
-    uint64_t size = (uint64_t)sourceStat.st_size;
-    if (size != 0)
+#endif
+#if HAVE_SENDFILE_4
+    // Try copying the data using sendfile.
+    if (!copied && sourceLength != 0)
     {
         // Note that per man page for large files, you have to iterate until the
         // whole file is copied (Linux has a limit of 0x7ffff000 bytes copied).
-        while (size > 0)
+        do
         {
-            ssize_t sent = sendfile(outFd, inFd, NULL, (size >= SSIZE_MAX ? SSIZE_MAX : (size_t)size));
+            ssize_t sent = sendfile(outFd, inFd, NULL, (sourceLength >= SSIZE_MAX ? SSIZE_MAX : (size_t)sourceLength));
             if (sent < 0)
             {
                 if (errno != EINVAL && errno != ENOSYS)
@@ -1184,36 +1193,31 @@ int32_t SystemNative_CopyFile(intptr_t sourceFd, intptr_t destinationFd)
                     break;
                 }
             }
+            else if (sent == 0)
+            {
+                // The file was truncated (or maybe some other condition occurred).
+                // Perform the remaining copying using read/write.
+                break;
+            }
             else
             {
-                assert((size_t)sent <= size);
-                size -= (size_t)sent;
+                assert(sent <= sourceLength);
+                sourceLength -= sent;
             }
-        }
+        } while (sourceLength > 0);
 
-        if (size == 0)
-        {
-            copied = true;
-        }
+        copied = sourceLength == 0;
     }
-
-    // sendfile couldn't be used; fall back to a manual copy below. This could happen
-    // if we're on an old kernel, for example, where sendfile could only be used
-    // with sockets and not regular files.  Additionally, certain files (e.g. procfs)
-    // may return a size of 0 even though reading from then will produce data.  As such,
-    // we avoid using sendfile with the queried size if the size is reported as 0.
 #endif // HAVE_SENDFILE_4
 
-    // Manually read all data from the source and write it to the destination.
+    // Perform a manual copy.
     if (!copied && CopyFile_ReadWrite(inFd, outFd) != 0)
     {
         return -1;
     }
 
-    // Now that the data from the file has been copied, copy over metadata
-    // from the source file.  First copy the file times.
-    // If futimes nor futimes are available on this platform, file times will
-    // not be copied over.
+    // Copy file times.
+    struct stat_ sourceStat;
     while ((ret = fstat_(inFd, &sourceStat)) < 0 && errno == EINTR);
     if (ret == 0)
     {
@@ -1242,7 +1246,10 @@ int32_t SystemNative_CopyFile(intptr_t sourceFd, intptr_t destinationFd)
     {
         return -1;
     }
-    // Then copy permissions.
+
+    // Copy permissions.
+    // Even though managed code created the file with permissions matching those of the source file,
+    // we need to copy permissions because the open permissions may be filtered by 'umask'.
     while ((ret = fchmod(outFd, sourceStat.st_mode & (S_IRWXU | S_IRWXG | S_IRWXO))) < 0 && errno == EINTR);
     if (ret != 0 && errno != EPERM) // See EPERM comment above
     {

src/libraries/Native/Unix/System.Native/pal_io.h

@@ -675,7 +675,7 @@ PALEXPORT int32_t SystemNative_Write(intptr_t fd, const void* buffer, int32_t bu
  *
  * Returns 0 on success; otherwise, returns -1 and sets errno.
  */
-PALEXPORT int32_t SystemNative_CopyFile(intptr_t sourceFd, intptr_t destinationFd);
+PALEXPORT int32_t SystemNative_CopyFile(intptr_t sourceFd, intptr_t destinationFd, int64_t sourceLength);
 
 /**
 * Initializes a new inotify instance and returns a file

src/libraries/System.Private.CoreLib/src/Microsoft/Win32/SafeHandles/SafeFileHandle.Unix.cs

@@ -57,12 +57,8 @@ internal bool SupportsRandomAccess
 
         internal void EnsureThreadPoolBindingInitialized() { /* nop */ }
 
-        /// <summary>Opens the specified file with the requested flags and mode.</summary>
-        /// <param name="path">The path to the file.</param>
-        /// <param name="flags">The flags with which to open the file.</param>
-        /// <param name="mode">The mode for opening the file.</param>
-        /// <returns>A SafeFileHandle for the opened file.</returns>
-        private static SafeFileHandle Open(string path, Interop.Sys.OpenFlags flags, int mode)
+        private static SafeFileHandle Open(string path, Interop.Sys.OpenFlags flags, int mode,
+                                           Func<Interop.ErrorInfo, Interop.Sys.OpenFlags, string, Exception?>? createOpenException)
         {
             Debug.Assert(path != null);
             SafeFileHandle handle = Interop.Sys.Open(path, flags, mode);
@@ -73,6 +69,11 @@ private static SafeFileHandle Open(string path, Interop.Sys.OpenFlags flags, int
                 Interop.ErrorInfo error = Interop.Sys.GetLastErrorInfo();
                 handle.Dispose();
 
+                if (createOpenException?.Invoke(error, flags, path) is Exception ex)
+                {
+                    throw ex;
+                }
+
                 // If we fail to open the file due to a path not existing, we need to know whether to blame
                 // the file itself or its directory.  If we're creating the file, then we blame the directory,
                 // otherwise we blame the file.
@@ -155,30 +156,52 @@ public override bool IsInvalid
             }
         }
 
-        internal static SafeFileHandle Open(string fullPath, FileMode mode, FileAccess access, FileShare share, FileOptions options, long preallocationSize)
-        {
-            // Translate the arguments into arguments for an open call.
-            Interop.Sys.OpenFlags openFlags = PreOpenConfigurationFromOptions(mode, access, share, options);
-
-            // If the file gets created a new, we'll select the permissions for it.  Most Unix utilities by default use 666 (read and
-            // write for all), so we do the same (even though this doesn't match Windows, where by default it's possible to write out
-            // a file and then execute it). No matter what we choose, it'll be subject to the umask applied by the system, such that the
-            // actual permissions will typically be less than what we select here.
-            const Interop.Sys.Permissions OpenPermissions =
+        // If the file gets created a new, we'll select the permissions for it.  Most Unix utilities by default use 666 (read and
+        // write for all), so we do the same (even though this doesn't match Windows, where by default it's possible to write out
+        // a file and then execute it). No matter what we choose, it'll be subject to the umask applied by the system, such that the
+        // actual permissions will typically be less than what we select here.
+        private const Interop.Sys.Permissions DefaultOpenPermissions =
                 Interop.Sys.Permissions.S_IRUSR | Interop.Sys.Permissions.S_IWUSR |
                 Interop.Sys.Permissions.S_IRGRP | Interop.Sys.Permissions.S_IWGRP |
                 Interop.Sys.Permissions.S_IROTH | Interop.Sys.Permissions.S_IWOTH;
 
+        // Specialized Open that returns the file length and permissions of the opened file.
+        // This information is retrieved from the 'stat' syscall that must be performed to ensure the path is not a directory.
+        internal static SafeFileHandle OpenReadOnly(string fullPath, FileOptions options, out long fileLength, out Interop.Sys.Permissions filePermissions)
+        {
+            SafeFileHandle handle = Open(fullPath, FileMode.Open, FileAccess.Read, FileShare.Read, options, preallocationSize: 0, DefaultOpenPermissions, out fileLength, out filePermissions, null);
+            Debug.Assert(fileLength >= 0);
+            return handle;
+        }
+
+        internal static SafeFileHandle Open(string fullPath, FileMode mode, FileAccess access, FileShare share, FileOptions options, long preallocationSize,
+                                            Interop.Sys.Permissions openPermissions = DefaultOpenPermissions,
+                                            Func<Interop.ErrorInfo, Interop.Sys.OpenFlags, string, Exception?>? createOpenException = null)
+        {
+            long fileLength;
+            Interop.Sys.Permissions filePermissions;
+            return Open(fullPath, mode, access, share, options, preallocationSize, openPermissions, out fileLength, out filePermissions, null);
+        }
+
+        private static SafeFileHandle Open(string fullPath, FileMode mode, FileAccess access, FileShare share, FileOptions options, long preallocationSize,
+                                            Interop.Sys.Permissions openPermissions,
+                                            out long fileLength,
+                                            out Interop.Sys.Permissions filePermissions,
+                                            Func<Interop.ErrorInfo, Interop.Sys.OpenFlags, string, Exception?>? createOpenException = null)
+        {
+            // Translate the arguments into arguments for an open call.
+            Interop.Sys.OpenFlags openFlags = PreOpenConfigurationFromOptions(mode, access, share, options);
+
             SafeFileHandle? safeFileHandle = null;
             try
             {
                 while (true)
                 {
-                    safeFileHandle = Open(fullPath, openFlags, (int)OpenPermissions);
+                    safeFileHandle = Open(fullPath, openFlags, (int)openPermissions, createOpenException);
 
                     // When Init return false, the path has changed to another file entry, and
                     // we need to re-open the path to reflect that.
-                    if (safeFileHandle.Init(fullPath, mode, access, share, options, preallocationSize))
+                    if (safeFileHandle.Init(fullPath, mode, access, share, options, preallocationSize, out fileLength, out filePermissions))
                     {
                         return safeFileHandle;
                     }
@@ -275,10 +298,13 @@ private static Interop.Sys.OpenFlags PreOpenConfigurationFromOptions(FileMode mo
             return flags;
         }
 
-        private bool Init(string path, FileMode mode, FileAccess access, FileShare share, FileOptions options, long preallocationSize)
+        private bool Init(string path, FileMode mode, FileAccess access, FileShare share, FileOptions options, long preallocationSize,
+                          out long fileLength, out Interop.Sys.Permissions filePermissions)
         {
             Interop.Sys.FileStatus status = default;
             bool statusHasValue = false;
+            fileLength = -1;
+            filePermissions = 0;
 
             // Make sure our handle is not a directory.
             // We can omit the check when write access is requested. open will have failed with EISDIR.
@@ -300,6 +326,9 @@ private bool Init(string path, FileMode mode, FileAccess access, FileShare share
                     _canSeek = NullableBool.True;
                     Debug.Assert(Interop.Sys.LSeek(this, 0, Interop.Sys.SeekWhence.SEEK_CUR) >= 0);
                 }
+
+                fileLength = status.Size;
+                filePermissions = (Interop.Sys.Permissions)(status.Mode & (int)Interop.Sys.Permissions.Mask);
             }
 
             IsAsync = (options & FileOptions.Asynchronous) != 0;

src/libraries/System.Private.CoreLib/src/System/IO/FileSystem.Unix.cs

@@ -11,25 +11,30 @@ namespace System.IO
     /// <summary>Provides an implementation of FileSystem for Unix systems.</summary>
     internal static partial class FileSystem
     {
-        internal const int DefaultBufferSize = 4096;
-
         // On Linux, the maximum number of symbolic links that are followed while resolving a pathname is 40.
         // See: https://man7.org/linux/man-pages/man7/path_resolution.7.html
         private const int MaxFollowedLinks = 40;
 
         public static void CopyFile(string sourceFullPath, string destFullPath, bool overwrite)
         {
-            // If the destination path points to a directory, we throw to match Windows behaviour
-            if (DirectoryExists(destFullPath))
-            {
-                throw new IOException(SR.Format(SR.Arg_FileIsDirectory_Name, destFullPath));
-            }
+            long fileLength;
+            Interop.Sys.Permissions filePermissions;
+            using SafeFileHandle src = SafeFileHandle.OpenReadOnly(sourceFullPath, FileOptions.None, out fileLength, out filePermissions);
+            using SafeFileHandle dst = SafeFileHandle.Open(destFullPath, overwrite ? FileMode.Create : FileMode.CreateNew,
+                                            FileAccess.ReadWrite, FileShare.None, FileOptions.None, preallocationSize: 0, openPermissions: filePermissions,
+                                            (Interop.ErrorInfo error, Interop.Sys.OpenFlags flags, string path) => CreateOpenException(error, flags, path));
+
+            Interop.CheckIo(Interop.Sys.CopyFile(src, dst, fileLength));
 
-            // Copy the contents of the file from the source to the destination, creating the destination in the process
-            using (SafeFileHandle src = File.OpenHandle(sourceFullPath, FileMode.Open, FileAccess.Read, FileShare.Read, FileOptions.None))
-            using (SafeFileHandle dst = File.OpenHandle(destFullPath, overwrite ? FileMode.Create : FileMode.CreateNew, FileAccess.ReadWrite, FileShare.None, FileOptions.None))
+            static Exception? CreateOpenException(Interop.ErrorInfo error, Interop.Sys.OpenFlags flags, string path)
             {
-                Interop.CheckIo(Interop.Sys.CopyFile(src, dst));
+                // If the destination path points to a directory, we throw to match Windows behaviour.
+                if (error.Error == Interop.Error.EEXIST && DirectoryExists(path))
+                {
+                    return new IOException(SR.Format(SR.Arg_FileIsDirectory_Name, path));
+                }
+
+                return null; // Let SafeFileHandle create the exception for this error.
             }
         }