linux-kernel - [PATCH v4 7/6] iommu/iova: Make cached

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <0c9865fdc53820a81f81adec5a7bb4239aba7bc1.1505904195.git.robin.murphy@arm.com>
Date:   Wed, 20 Sep 2017 12:02:48 +0100
From:   Robin Murphy <robin.murphy@....com>
To:     joro@...tes.org
Cc:     iommu@...ts.linux-foundation.org, thunder.leizhen@...wei.com,
        nwatters@...eaurora.org, tomasz.nowicki@...iumnetworks.com,
        linux-kernel@...r.kernel.org
Subject: [PATCH v4 7/6] iommu/iova: Make cached_node always valid

With the anchor node at the top of the rbtree, there is always a valid
node for rb_next() to return, such that cached_node is only ever NULL
until the first allocation. Initialising it to point at the anchor node
gets rid of that window and makes the NULL checking entirely redundant.

Signed-off-by: Robin Murphy <robin.murphy@....com>
---

Oops, spotted this one slightly too late. This could be squashed into
patch #5 (which I'll do myself if I there's any cause to resend the
whole series again).

Robin.

 drivers/iommu/iova.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/drivers/iommu/iova.c b/drivers/iommu/iova.c
index a7af8273fa98..ec443c0a8319 100644
--- a/drivers/iommu/iova.c
+++ b/drivers/iommu/iova.c
@@ -51,7 +51,7 @@ init_iova_domain(struct iova_domain *iovad, unsigned long granule,
 
 	spin_lock_init(&iovad->iova_rbtree_lock);
 	iovad->rbroot = RB_ROOT;
-	iovad->cached_node = NULL;
+	iovad->cached_node = &iovad->anchor.node;
 	iovad->cached32_node = NULL;
 	iovad->granule = granule;
 	iovad->start_pfn = start_pfn;
@@ -120,10 +120,7 @@ __get_cached_rbnode(struct iova_domain *iovad, unsigned long limit_pfn)
 	if (limit_pfn <= iovad->dma_32bit_pfn && iovad->cached32_node)
 		return iovad->cached32_node;
 
-	if (iovad->cached_node)
-		return iovad->cached_node;
-
-	return &iovad->anchor.node;
+	return iovad->cached_node;
 }
 
 static void
@@ -141,14 +138,11 @@ __cached_rbnode_delete_update(struct iova_domain *iovad, struct iova *free)
 	struct iova *cached_iova;
 	struct rb_node **curr;
 
-	if (free->pfn_hi < iovad->dma_32bit_pfn)
+	if (free->pfn_hi < iovad->dma_32bit_pfn && iovad->cached32_node)
 		curr = &iovad->cached32_node;
 	else
 		curr = &iovad->cached_node;
 
-	if (!*curr)
-		return;
-
 	cached_iova = rb_entry(*curr, struct iova, node);
 	if (free->pfn_lo >= cached_iova->pfn_lo)
 		*curr = rb_next(&free->node);
-- 
2.13.4.dirty