On Thu, 2007-05-31 at 16:25 +0200, Jörn Engel wrote: > --- linux-2.6.21logfs/fs/jfs/jfs_txnmgr.c~I_LOCK 2007-05-07 > 10:28:55.000000000 +0200 > +++ linux-2.6.21logfs/fs/jfs/jfs_txnmgr.c 2007-05-29 > 13:10:32.000000000 +0200 > @@ -1286,7 +1286,14 @@ int txCommit(tid_t tid, /* > transaction > * commit the transaction synchronously, so the last > iput > * will be done by the calling thread (or later) > */ > - if (tblk->u.ip->i_state & I_LOCK) > + /* > + * I believe this code is no longer needed. Splitting > I_LOCK > + * into two bits, I_LOCK and I_SYNC should prevent > this > + * deadlock as well. But since I don't have a JFS > testload > + * to verify this, only a trivial s/I_LOCK/I_SYNC/ was > done. > + * Joern > + */ > + if (tblk->u.ip->i_state & I_SYNC) > tblk->xflag &= ~COMMIT_LAZY; > } I think the code is still needed, and I think this change is correct. The deadlock that this code is avoiding is caused by clear_inode() calling wait_on_inode(). Since clear_inode() now calls inode_sync_wait(inode), we want to avoid the lazily committing this transaction when the I_SYNC flag is set. Unfortunately, recreating the deadlock is hard, and I haven't been able to recreate it with this code commented out. Thanks, Shaggy -- David Kleikamp IBM Linux Technology Center - To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
