[Ocfs2-commits] mfasheh commits r855 - trunk

[svn-commits at oss.oracle.com]

Author: mfasheh
Date: 2004-04-19 19:22:30 -0500 (Mon, 19 Apr 2004)
New Revision: 855

Modified:
   trunk/TODO
Log:
* update the TODO list


Modified: trunk/TODO
===================================================================
--- trunk/TODO	2004-04-19 23:28:32 UTC (rev 854)
+++ trunk/TODO	2004-04-20 00:22:30 UTC (rev 855)
@@ -1,8 +1,3 @@
-* when we create the root oin we don't stick the inode on it and do the
-  whole atomic_inc of i_count business. we should start doing this.
-
-* disable sigkill and sigterm in ocfs listener thread
-
 * figure out what to do about old style cdsl. Either we error when we see it
   (in the module), or we provide an fsck to find and fix them.
 
@@ -15,9 +10,6 @@
 
 * finish mkfs to really create a filesystem.
 
-* See about just always having an OIN on the inode... Maybe we can start
-  using main_res or i_sem for locking now too, instead of the bh_sem_hash
-
 * System files should be locked with cache lock, and we need inodes
   for them so that we can cache our reads to them. This can also count
   for the main bitmap. Data writes to the bitmap files can be
@@ -42,13 +34,14 @@
 * Make all our 64-bit divides into shifts, this way we won't need 
   divdi3.c anymore.
 
-* Move the last of the used, non redundant fields off the oin and put them
-  on the inode in our private data area. This should have the nice effect of
-  getting rid of oin.
-
 * Combine the ocfs_inode_private and ocfs_inode_num structs, keeping the
   inode hash, but just using a unified struct as the inode private data, and
   the hash element (there are a bunch of redundant fields)
 
 * Add statistics in /proc for total number of allocs, number from local
   alloc, whether local alloc is enabled and number of main bitmap allocs
+
+* move things to use i_sem to lock out inode changes instead of the bh sem
+  hash we have now.
+
+* get rid of as much of lockres as possible.