erts: Fix UNC path handling on Windows

This is unlikely to be the last of the path problems seen after
OTP 21, and I'm starting to regret my decision to unconditionally
use long paths. The idea to hit all long-path problems all the time
was good in theory as it makes such bugs far more visible, but
there just aren't enough people who test pre-release versions on
Windows, making this the world's slowest game of whack-a-mole.

1 files changed, 102 insertions, 53 deletions

diff --git a/erts/emulator/nifs/win32/win_prim_file.c b/erts/emulator/nifs/win32/win_prim_file.c
index 0c51a21ee1..602a282dd1 100644
--- a/erts/emulator/nifs/win32/win_prim_file.c
+++ b/erts/emulator/nifs/win32/win_prim_file.c
@@ -33,16 +33,32 @@
 
 #define FILE_SHARE_FLAGS (FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE)
 
-#define LP_PREFIX L"\\\\?\\"
-#define LP_PREFIX_SIZE (sizeof(LP_PREFIX) - sizeof(WCHAR))
+/* Long paths can either be in the file (?) or the device (.) namespace. UNC
+ * paths are always in the file namespace. */
+#define LP_FILE_PREFIX L"\\\\?\\"
+#define LP_DEV_PREFIX L"\\\\.\\"
+#define LP_UNC_PREFIX (LP_FILE_PREFIX L"UNC\\")
+
+#define LP_PREFIX_SIZE (sizeof(LP_FILE_PREFIX) - sizeof(WCHAR))
 #define LP_PREFIX_LENGTH (LP_PREFIX_SIZE / sizeof(WCHAR))
 
+#define LP_UNC_PREFIX_SIZE (sizeof(LP_UNC_PREFIX) - sizeof(WCHAR))
+#define LP_UNC_PREFIX_LENGTH (LP_UNC_PREFIX_SIZE / sizeof(WCHAR))
+
+#define IS_LONG_PATH(length, data) \
+    ((length) >= LP_PREFIX_LENGTH && \
+         (!sys_memcmp((data), LP_FILE_PREFIX, LP_PREFIX_SIZE) || \
+          !sys_memcmp((data), LP_DEV_PREFIX, LP_PREFIX_SIZE)))
+
+#define IS_LONG_UNC_PATH(length, data) \
+    ((length) >= LP_UNC_PREFIX_LENGTH && \
+         !sys_memcmp((data), LP_UNC_PREFIX, LP_UNC_PREFIX_SIZE))
+
 #define PATH_LENGTH(path) (path->size / sizeof(WCHAR) - 1)
 
 #define ASSERT_PATH_FORMAT(path) \
     do { \
-        ASSERT(PATH_LENGTH(path) >= 4 && \
-            !memcmp(path->data, LP_PREFIX, LP_PREFIX_SIZE)); \
+        ASSERT(IS_LONG_PATH(PATH_LENGTH(path), (path)->data)); \
         ASSERT(PATH_LENGTH(path) == wcslen((WCHAR*)path->data)); \
     } while(0)
 
@@ -106,7 +122,7 @@ static posix_errno_t get_full_path(ErlNifEnv *env, WCHAR *input, efile_path_t *r
         return ENOENT;
     }
 
-    maximum_length += LP_PREFIX_LENGTH;
+    maximum_length += MAX(LP_PREFIX_LENGTH, LP_UNC_PREFIX_LENGTH);
 
     if(!enif_alloc_binary(maximum_length * sizeof(WCHAR), result)) {
         return ENOMEM;
@@ -115,18 +131,28 @@ static posix_errno_t get_full_path(ErlNifEnv *env, WCHAR *input, efile_path_t *r
     actual_length = GetFullPathNameW(input, maximum_length, (WCHAR*)result->data, NULL);
 
     if(actual_length < maximum_length) {
-        int has_long_path_prefix;
+        int is_long_path, maybe_unc_path;
         WCHAR *path_start;
 
-        /* Make sure we have a long-path prefix; GetFullPathNameW only adds one
-         * if the path is relative. */
-        has_long_path_prefix = actual_length >= LP_PREFIX_LENGTH &&
-            !sys_memcmp(result->data, LP_PREFIX, LP_PREFIX_SIZE);
-
-        if(!has_long_path_prefix) {
+        /* The APIs we use have varying path length limits and sometimes
+         * behave differently when given a long-path prefix, so it's simplest
+         * to always use long paths. */
+
+        is_long_path = IS_LONG_PATH(actual_length, result->data);
+        maybe_unc_path = !sys_memcmp(result->data, L"\\\\", sizeof(WCHAR) * 2);
+
+        if(maybe_unc_path && !is_long_path) {
+            /* \\localhost\c$\gurka -> \\?\UNC\localhost\c$\gurka */
+            sys_memmove(result->data + LP_UNC_PREFIX_SIZE,
+                &((WCHAR*)result->data)[2],
+                (actual_length - 1) * sizeof(WCHAR));
+            sys_memcpy(result->data, LP_UNC_PREFIX, LP_UNC_PREFIX_SIZE);
+            actual_length += LP_UNC_PREFIX_LENGTH;
+        } else if(!is_long_path) {
+            /* C:\gurka -> \\?\C:\gurka */
             sys_memmove(result->data + LP_PREFIX_SIZE, result->data,
                 (actual_length + 1) * sizeof(WCHAR));
-            sys_memcpy(result->data, LP_PREFIX, LP_PREFIX_SIZE);
+            sys_memcpy(result->data, LP_FILE_PREFIX, LP_PREFIX_SIZE);
             actual_length += LP_PREFIX_LENGTH;
         }
 
@@ -200,13 +226,19 @@ static int normalize_path_result(ErlNifBinary *path) {
     ASSERT(length < path->size / sizeof(WCHAR));
 
     /* Get rid of the long-path prefix, if present. */
-    if(length >= LP_PREFIX_LENGTH) {
-        if(!sys_memcmp(path_start, LP_PREFIX, LP_PREFIX_SIZE)) {
-            length -= LP_PREFIX_LENGTH;
 
-            sys_memmove(path_start, &path_start[LP_PREFIX_LENGTH],
-                length * sizeof(WCHAR));
-        }
+    if(IS_LONG_UNC_PATH(length, path_start)) {
+        /* The first two characters (\\) are the same for both long and short
+         * UNC paths. */
+        sys_memmove(&path_start[2], &path_start[LP_UNC_PREFIX_LENGTH],
+            (length - LP_UNC_PREFIX_LENGTH) * sizeof(WCHAR));
+
+        length -= LP_UNC_PREFIX_LENGTH - 2;
+    } else if(IS_LONG_PATH(length, path_start)) {
+        length -= LP_PREFIX_LENGTH;
+
+        sys_memmove(path_start, &path_start[LP_PREFIX_LENGTH],
+            length * sizeof(WCHAR));
     }
 
     path_end = &path_start[length];
@@ -318,49 +350,55 @@ static int has_same_mount_point(const efile_path_t *path_a, const efile_path_t *
 /* Mirrors the PathIsRootW function of the shell API, but doesn't choke on
  * paths longer than MAX_PATH. */
 static int is_path_root(const efile_path_t *path) {
-    const WCHAR *path_start, *path_end;
+    const WCHAR *path_start, *path_end, *path_iterator;
     int length;
 
     ASSERT_PATH_FORMAT(path);
 
-    path_start = (WCHAR*)path->data + LP_PREFIX_LENGTH;
-    length = PATH_LENGTH(path) - LP_PREFIX_LENGTH;
+    if(!IS_LONG_UNC_PATH(PATH_LENGTH(path), path->data)) {
+        path_start = (WCHAR*)path->data + LP_PREFIX_LENGTH;
+        length = PATH_LENGTH(path) - LP_PREFIX_LENGTH;
 
-    path_end = &path_start[length];
-
-    if(length == 1) {
         /* A single \ refers to the root of the current working directory. */
-        return IS_SLASH(path_start[0]);
-    } else if(length == 3 && iswalpha(path_start[0]) && path_start[1] == L':') {
-        /* Drive letter. */
-        return IS_SLASH(path_start[2]);
-    } else if(length >= 4) {
-        /* Check whether we're a UNC root, eg. \\server, \\server\share */
-        const WCHAR *path_iterator;
+        if(length == 1) {
+            return IS_SLASH(path_start[0]);
+        }
 
-        if(!IS_SLASH(path_start[0]) || !IS_SLASH(path_start[1])) {
-            return 0;
+        /* Drive letter. */
+        if(length == 3 && iswalpha(path_start[0]) && path_start[1] == L':') {
+            return IS_SLASH(path_start[2]);
         }
 
-        path_iterator = path_start + 2;
+        return 0;
+    }
 
-        /* Slide to the slash between the server and share names, if present. */
-        while(path_iterator < path_end && !IS_SLASH(*path_iterator)) {
-            path_iterator++;
-        }
+    /* Check whether we're a UNC root, eg. \\server, \\server\share */
 
-        /* Slide past the end of the string, stopping at the first slash we
-         * encounter. */
-        do {
-            path_iterator++;
-        }  while(path_iterator < path_end && !IS_SLASH(*path_iterator));
+    path_start = (WCHAR*)path->data + LP_UNC_PREFIX_LENGTH;
+    length = PATH_LENGTH(path) - LP_UNC_PREFIX_LENGTH;
 
-        /* If we're past the end of the string and it didnt't end with a slash,
-         * then we're a root path. */
-        return path_iterator >= path_end && !IS_SLASH(path_start[length - 1]);
+    path_end = &path_start[length];
+    path_iterator = path_start;
+
+    /* Server name must be at least one character. */
+    if(length <= 1) {
+        return 0;
     }
 
-    return 0;
+    /* Slide to the slash between the server and share names, if present. */
+    while(path_iterator < path_end && !IS_SLASH(*path_iterator)) {
+        path_iterator++;
+    }
+
+    /* Slide past the end of the string, stopping at the first slash we
+     * encounter. */
+    do {
+        path_iterator++;
+    }  while(path_iterator < path_end && !IS_SLASH(*path_iterator));
+
+    /* If we're past the end of the string and it didnt't end with a slash,
+     * then we're a root path. */
+    return path_iterator >= path_end && !IS_SLASH(path_start[length - 1]);
 }
 
 posix_errno_t efile_open(const efile_path_t *path, enum efile_modes_t modes,
@@ -1248,11 +1286,22 @@ posix_errno_t efile_set_cwd(const efile_path_t *path) {
 
     /* We have to use _wchdir since that's the only function that updates the
      * per-drive working directory, but it naively assumes that all paths
-     * starting with \\ are UNC paths, so we have to skip the \\?\-prefix. */
-    path_start = (WCHAR*)path->data + LP_PREFIX_LENGTH;
+     * starting with \\ are UNC paths, so we have to skip the long-path prefix.
+     *
+     * _wchdir doesn't handle long-prefixed UNC paths either so we hand those
+     * to SetCurrentDirectoryW instead. The per-drive working directory is
+     * irrelevant for such paths anyway. */
 
-    if(_wchdir(path_start)) {
-        return windows_to_posix_errno(GetLastError());
+    if(!IS_LONG_UNC_PATH(PATH_LENGTH(path), path->data)) {
+        path_start = (WCHAR*)path->data + LP_PREFIX_LENGTH;
+
+        if(_wchdir(path_start)) {
+            return windows_to_posix_errno(GetLastError());
+        }
+    } else {
+        if(!SetCurrentDirectoryW((WCHAR*)path->data)) {
+            return windows_to_posix_errno(GetLastError());
+        }
     }
 
     return 0;
@@ -1333,7 +1382,7 @@ posix_errno_t efile_altname(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TE
         int name_length;
 
         /* Reject path wildcards. */
-        if(wcspbrk(&((const WCHAR*)path->data)[4], L"?*")) {
+        if(wcspbrk(&((const WCHAR*)path->data)[LP_PREFIX_LENGTH], L"?*")) {
             return ENOENT;
         }