[PATCH] fs/jffs2: Fix NULL deref in jffs2_scan_dirty_space

[PATCH] fs/jffs2: Fix NULL deref in jffs2_scan_dirty_space

[Artem Chernyshev]

As was mentioned in 2ebf09c2, it is possible to get oops,
when marking space dirty in scan, but no previous node exists

It still can be in jffs2_link_node_ref() via deref jeb->last_node.
Since all jffs2_scan_dirty_space() callers check value of a
function, we can return error code safely.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Signed-off-by: Artem Chernyshev <artem.chernyshev@red-soft.ru>
---
 fs/jffs2/nodelist.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fs/jffs2/nodelist.c b/fs/jffs2/nodelist.c
index b86c78d178c6..6bebf1d64000 100644
--- a/fs/jffs2/nodelist.c
+++ b/fs/jffs2/nodelist.c
@@ -669,8 +669,11 @@ int jffs2_scan_dirty_space(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb
 			size, jeb->free_size, jeb->wasted_size);
 		BUG();
 	}
+	if (!jeb->last_node) {
+		return -EINVAL;
+	}
 	/* REF_EMPTY_NODE is !obsolete, so that works OK */
-	if (jeb->last_node && ref_obsolete(jeb->last_node)) {
+	if (ref_obsolete(jeb->last_node) {
 #ifdef TEST_TOTLEN
 		jeb->last_node->__totlen += size;
 #endif
-- 
2.37.3


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

Re: [PATCH] fs/jffs2: Fix NULL deref in jffs2_scan_dirty_space

[Richard Weinberger]

----- Ursprüngliche Mail -----
> Von: "Artem Chernyshev" <artem.chernyshev@red-soft.ru>
> An: "David Woodhouse" <dwmw2@infradead.org>, "richard" <richard@nod.at>
> CC: "Artem Chernyshev" <artem.chernyshev@red-soft.ru>, "linux-mtd" <linux-mtd@lists.infradead.org>, "linux-kernel"
> <linux-kernel@vger.kernel.org>, lvc-project@linuxtesting.org
> Gesendet: Freitag, 29. März 2024 13:16:29
> Betreff: [PATCH] fs/jffs2: Fix NULL deref in jffs2_scan_dirty_space

> As was mentioned in 2ebf09c2, it is possible to get oops,
> when marking space dirty in scan, but no previous node exists
> 
> It still can be in jffs2_link_node_ref() via deref jeb->last_node.
> Since all jffs2_scan_dirty_space() callers check value of a
> function, we can return error code safely.
> 
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
> 
> Signed-off-by: Artem Chernyshev <artem.chernyshev@red-soft.ru>
> ---
> fs/jffs2/nodelist.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/jffs2/nodelist.c b/fs/jffs2/nodelist.c
> index b86c78d178c6..6bebf1d64000 100644
> --- a/fs/jffs2/nodelist.c
> +++ b/fs/jffs2/nodelist.c
> @@ -669,8 +669,11 @@ int jffs2_scan_dirty_space(struct jffs2_sb_info *c, struct
> jffs2_eraseblock *jeb
> 			size, jeb->free_size, jeb->wasted_size);
> 		BUG();
> 	}
> +	if (!jeb->last_node) {
> +		return -EINVAL;
> +	}

Did you test this change? I'm not so sure whether returning -EINVAL is
correct.

Thanks,
//richard

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/