Skip to content
  1. Nov 12, 2017
  2. Nov 11, 2017
  3. Nov 10, 2017
  4. Nov 09, 2017
  5. Nov 08, 2017
  6. Nov 07, 2017
  7. Nov 06, 2017
  8. Nov 05, 2017
    • Gregory Nutt's avatar
      This comment converts the underlying IPC used by the UserFS from Unix domain... · 7deb2448
      Gregory Nutt authored
      This comment converts the underlying IPC used by the UserFS from Unix domain local sockets to UDP LocalHost loopback sockets.  The problem with the local sockets is that they do require operations on the top level psuedo-file system inode tree.  That tree must be locked during certain traversals such as enumerate mountpoints or enumerating directory entries.
      
      This conversion is unfortunate in the sense that Unix local domain sockets are relatively lightweight.  LocalHost UDP sockets are much heavier weight since they rely on the full UDP stack.  If anyone is up for a complete redesign, then using some shared memory and a POSIX message queue would be lightweight again.
      
      This commit also fixes several bugs that were not testable before the inode tree deadlock.  I cannot say that the logic is 100% stable but it does not have basic functionality.
      
      Squashed commit of the following:
      
          fs/userfs:  Order locking so that access to the shared I/O buffer is also locked.
          fs/userfs:  Converts to use LocalHost UDP loopback for IPC.
      7deb2448