HGFS: Fix Posix server from blocking on pipes

The HGFS protocol function to get attributes of a file or folder does
a Posix open without the O_NONBLOCK flag. This can cause the open to block
if there isn't anyone on the other end of the pipe.

Fix is to use the same flags we use for the protocol open to ensure the
correct basic flags are used. Also, if the share is not allowed to follow
symlinks we get this flag set correctly too now.

Signed-off-by: Dmitry Torokhov <dtor@vmware.com>

diff --git a/open-vm-tools/lib/hgfsServer/hgfsServerLinux.c b/open-vm-tools/lib/hgfsServer/hgfsServerLinux.c
index 0d55a8e92174b4771eb4c46df598b74b28ac25b3..de4c4a6ae96362f172ff72a30add6b219157c352 100644 (file)
--- a/open-vm-tools/lib/hgfsServer/hgfsServerLinux.c
+++ b/open-vm-tools/lib/hgfsServer/hgfsServerLinux.c
@@ -328,6 +328,7 @@ static int HgfsFStat(int fd,
                      uint64 *creationTime);
 
 static void HgfsGetSequentialOnlyFlagFromName(const char *fileName,
+                                              Bool followSymlinks,
                                               HgfsFileAttrInfo *attr);
 
 static void HgfsGetSequentialOnlyFlagFromFd(int fd,
@@ -1942,16 +1943,40 @@ HgfsFStat(int fd,                 // IN: file descriptor
 
 static void
 HgfsGetSequentialOnlyFlagFromName(const char *fileName,        // IN
+                                  Bool followSymlinks,         // IN: If true then follow symlink
                                   HgfsFileAttrInfo *attr)      // IN/OUT
 {
 #if defined(__linux) || defined(__APPLE__)
    int fd;
+   int openFlags;
 
    if ((NULL == fileName) || (NULL == attr)) {
       return;
    }
 
-   fd = Posix_Open(fileName, O_RDONLY);
+   /*
+    * We're not interested in creating a new file. So let's just get the
+    * flags for a simple open request. This really should always work.
+    */
+   HgfsServerGetOpenFlags(0, &openFlags);
+
+   /*
+    * By default we don't follow symlinks, O_NOFOLLOW is always set.
+    * Unset it if followSymlinks config option is specified.
+    */
+   if (followSymlinks) {
+      openFlags &= ~O_NOFOLLOW;
+   }
+
+   /*
+    * Checking for a FIFO we open in nonblocking mode. In this case, opening for
+    * read only will succeed even if no-one has opened on the write side yet,
+    * opening for write only will fail with ENXIO (no such device or address)
+    * unless the other end has already been opened.
+    * Note, Under Linux, opening a FIFO for read and write will succeed both
+    * in blocking and nonblocking mode. POSIX leaves this behavior undefined.
+    */
+   fd = Posix_Open(fileName, openFlags | O_RDONLY);
    if (fd < 0) {
       LOG(4, ("%s: Couldn't open the file \"%s\"\n", __FUNCTION__, fileName));
       return;
@@ -2060,15 +2085,17 @@ HgfsPlatformGetattrFromName(char *fileName,                 // IN/OUT:  Input fi
    int error;
    char *myTargetName = NULL;
    uint64 creationTime;
+   Bool followSymlinks;
 
    ASSERT(fileName);
    ASSERT(attr);
 
    LOG(4, ("%s: getting attrs for \"%s\"\n", __FUNCTION__, fileName));
+   followSymlinks = HgfsServerPolicy_IsShareOptionSet(configOptions,
+                                                      HGFS_SHARE_FOLLOW_SYMLINKS),
 
    error = HgfsStat(fileName,
-                    HgfsServerPolicy_IsShareOptionSet(configOptions,
-                                                      HGFS_SHARE_FOLLOW_SYMLINKS),
+                    followSymlinks,
                     &stats,
                     &creationTime);
    if (error) {
@@ -2204,7 +2231,7 @@ HgfsPlatformGetattrFromName(char *fileName,                 // IN/OUT:  Input fi
     */
    HgfsGetHiddenAttr(fileName, attr);
 
-   HgfsGetSequentialOnlyFlagFromName(fileName, attr);
+   HgfsGetSequentialOnlyFlagFromName(fileName, followSymlinks, attr);
 
    /* Get effective permissions if we can */
    if (!(S_ISLNK(stats.st_mode))) {