[PATCH 2.6.33-rc6] idr: revert misallocation bug fix

[Tejun Heo <tj@kernel.org>]

Commit 859ddf09743a8cc680af33f7259ccd0fd36bfe9d tried to fix
misallocation bug but broke full bit marking by not clearing
pa[idp->layers] and also is causing X failures due to lookup failure
in drm code.  The cause of the latter hasn't been found yet.  Revert
the fix for now.

Signed-off-by: Tejun Heo <tj@kernel.org>
---
Sorry this wasn't found before hitting mainline.  It looks like this
will have to stay broken for a bit more.  I'll make an userland
test rig and find out what's going on.

Thanks.

 lib/idr.c |    7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/lib/idr.c b/lib/idr.c
index ba7d37c..1cac726 100644
--- a/lib/idr.c
+++ b/lib/idr.c
@@ -140,7 +140,8 @@ static int sub_alloc(struct idr *idp, int *starting_id, struct idr_layer **pa)
 	id = *starting_id;
  restart:
 	p = idp->top;
-	l = p->layer;
+	l = idp->layers;
+	pa[l--] = NULL;
 	while (1) {
 		/*
 		 * We run around this while until we reach the leaf node...
@@ -154,8 +155,8 @@ static int sub_alloc(struct idr *idp, int *starting_id, struct idr_layer **pa)
 			oid = id;
 			id = (id | ((1 << (IDR_BITS * l)) - 1)) + 1;
 
-			/* did id go over the limit? */
-			if (id >= (1 << (idp->layers * IDR_BITS))) {
+			/* if already at the top layer, we need to grow */
+			if (!(p = pa[l])) {
 				*starting_id = id;
 				return IDR_NEED_TO_GROW;
 			}