linux-kernel - [RFC PATCH v2 1/1] mm/vmalloc.c: correct a few logic error for __insert_vmap

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>] [day] [month] [year] [list]

Date:   Tue, 27 Sep 2016 21:41:11 +0800
From:   zijun_hu <zijun_hu@...o.com>
To:     Andrew Morton <akpm@...ux-foundation.org>,
        Nicholas Piggin <npiggin@...il.com>
Cc:     linux-mm@...ck.org, linux-kernel@...r.kernel.org, zijun_hu@....com,
        torvalds@...ux-foundation.org, tj@...nel.org, mingo@...nel.org,
        rientjes@...gle.com, iamjoonsoo.kim@....com,
        mgorman@...hsingularity.net, npiggin@...il.com, mhocko@...nel.org
Subject: [RFC PATCH v2 1/1] mm/vmalloc.c: correct a few logic error for
 __insert_vmap_area()

From: zijun_hu <zijun_hu@....com>

__insert_vmap_area() has a few obvious logic errors as shown by comments
within below code segments
static void __insert_vmap_area(struct vmap_area *va)
{
as a internal function parameter, we assume vmap_area @va has nonzero size
...
		if (va->va_start < tmp->va_end)
			p = &(*p)->rb_left;
		else if (va->va_end > tmp->va_start)
			p = &(*p)->rb_right;
this else if condition is always true and meaningless due to
va->va_end > va->va_start >= tmp_va->va_end > tmp_va->va_start normally
		else
			BUG();
this BUG() is meaningless too due to never be touched normally
...
}

the function don't implement the below desire behavior based on context
if the vmap_area @va to be inserted is lower than that on the rbtree then
we walk around the left branch of the given rbtree node; else if higher
then right branch; else the former vmap_area has overlay with the latter
then the existing BUG() is triggered

it is fixed by correcting vmap_area rbtree walk manner as mentioned above
BTW, we consider (va->va_end == tmp_va->va_start) as legal case since it
indicate vmap_area @va neighbors with @tmp_va tightly

Fixes: db64fe02258f ("mm: rewrite vmap layer")
Signed-off-by: zijun_hu <zijun_hu@....com>
---
 Hi npiggin,
 could you offer some comments for this patch since __insert_vmap_area()
 was introduced by you?
 thanks a lot

 Changes in v2:
  - more detailed commit message is provided

 mm/vmalloc.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 91f44e78c516..cc6ecd60cc0e 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -321,10 +321,10 @@ static void __insert_vmap_area(struct vmap_area *va)
 
 		parent = *p;
 		tmp_va = rb_entry(parent, struct vmap_area, rb_node);
-		if (va->va_start < tmp_va->va_end)
-			p = &(*p)->rb_left;
-		else if (va->va_end > tmp_va->va_start)
-			p = &(*p)->rb_right;
+		if (va->va_end <= tmp_va->va_start)
+			p = &parent->rb_left;
+		else if (va->va_start >= tmp_va->va_end)
+			p = &parent->rb_right;
 		else
 			BUG();
 	}
-- 
1.9.1