linux-hardening - [PATCH v2][next] dlm: Replace one-element array with flexible-array member

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 PHC
Open Source and information security mailing list archives

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

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

Date:   Sun, 9 Oct 2022 12:17:37 +1300
From:   Paulo Miguel Almeida <paulo.miguel.almeida.rodenas@...il.com>
To:     Christine Caulfield <ccaulfie@...hat.com>,
        David Teigland <teigland@...hat.com>, cluster-devel@...hat.com
Cc:     linux-kernel@...r.kernel.org, linux-hardening@...r.kernel.org,
        paulo.miguel.almeida.rodenas@...il.com
Subject: [PATCH v2][next] dlm: Replace one-element array with flexible-array
 member

One-element arrays are deprecated, and we are replacing them with
flexible array members instead. So, replace one-element array with
flexible-array member in struct dlm_ls, and refactor the rest of the
code, accordingly.

We strive to make changes that don't produce any before/after binary
output differeces as that makes it easier for the reviewer to accept the
patch. In this particular instance, it wasn't possible to achieve this
due to the fact that the ls_name[1] size, which is used as the
NUL-terminator space, was hidden within the struct's padding as shown
below.

$ diff <(objdump -M intel -j .text -D dlm.old) <(objdump -M intel -j
.text -D dlm.new)

13778c13778
<     c693:	49 8d bc 24 c0 08 00 	lea    rdi,[r12+0x8c0]
---
>     c693:	49 8d bc 24 c1 08 00 	lea    rdi,[r12+0x8c1]

$ pahole dlm.old -C dlm_ls
...
	int                        ls_namelen;           /*  2232     4 */
	char                       ls_name[1];           /*  2236     1 */

	/* size: 2240, cachelines: 35, members: 90 */
	/* sum members: 2166, holes: 17, sum holes: 71 */
	/* padding: 3 */
	/* paddings: 3, sum paddings: 17 */
	/* forced alignments: 1 */
} __attribute__((__aligned__(8)));

$ pahole dlm.new -C dlm_ls
...
	int                        ls_namelen;           /*  2232     4 */
	char                       ls_name[];            /*  2236     0 */

	/* size: 2240, cachelines: 35, members: 90 */
	/* sum members: 2165, holes: 17, sum holes: 71 */
	/* padding: 4 */
	/* paddings: 3, sum paddings: 17 */
	/* forced alignments: 1 */
} __attribute__((__aligned__(8)));


This helps with the ongoing efforts to tighten the FORTIFY_SOURCE
routines on memcpy() and help us make progress towards globally
enabling -fstrict-flex-arrays=3 [1].

Link: https://github.com/KSPP/linux/issues/79
Link: https://github.com/KSPP/linux/issues/228
Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101836 [1]

Signed-off-by: Paulo Miguel Almeida <paulo.miguel.almeida.rodenas@...il.com>
---
My apologies for v2, there was an accidental <Cr> I added on
the CC line which messed up the body of my previus email. 

This patch sets it right.

---
 fs/dlm/dlm_internal.h | 2 +-
 fs/dlm/lockspace.c    | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h
index e34c3d2639a5..ce2e32ed2ede 100644
--- a/fs/dlm/dlm_internal.h
+++ b/fs/dlm/dlm_internal.h
@@ -670,7 +670,7 @@ struct dlm_ls {
 	void			*ls_ops_arg;
 
 	int			ls_namelen;
-	char			ls_name[1];
+	char			ls_name[];
 };
 
 /*
diff --git a/fs/dlm/lockspace.c b/fs/dlm/lockspace.c
index bae050df7abf..c3a36f3197da 100644
--- a/fs/dlm/lockspace.c
+++ b/fs/dlm/lockspace.c
@@ -473,7 +473,7 @@ static int new_lockspace(const char *name, const char *cluster,
 
 	error = -ENOMEM;
 
-	ls = kzalloc(sizeof(struct dlm_ls) + namelen, GFP_NOFS);
+	ls = kzalloc(sizeof(struct dlm_ls) + namelen + 1, GFP_NOFS);
 	if (!ls)
 		goto out;
 	memcpy(ls->ls_name, name, namelen);
-- 
2.37.3