Skip to contentSkip to conversation

fs/cifs: suppress a string overflow warning

PATCH Applied Stephen Rothwell 2 participants 4 messages Sep 3, 2018 – Sep 4, 2018

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 fs/cifs: suppress a string overflow warning

A powerpc build of cifs with gcc v8.2.0 produces this warning:

fs/cifs/cifssmb.c: In function ‘CIFSSMBNegotiate’:
fs/cifs/cifssmb.c:605:3: warning: ‘strncpy’ writing 16 bytes into a region of size 1 overflows the destination [-Wstringop-overflow=]
strncpy(pSMB->DialectsArray+count, protocols[i].name, 16); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Since we are already doing a strlen() on the source, change the strncpy to a memcpy().

---
 fs/cifs/cifssmb.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/cifssmb.c +7 -3
index dc2f4cf08fe9..dcf939cb9d2f 100644
@@ -601,10 +601,14 @@ CIFSSMBNegotiate(const unsigned int xid, struct cifs_ses *ses)
 	}
 
 	count = 0;
+	/*
+	 * We know that all the name entries in the protocols array
+	 * are short (< 16 bytes anyway) and are NUL terminated.
+	 */
 	for (i = 0; i < CIFS_NUM_PROT; i++) {
-		strncpy(pSMB->DialectsArray+count, protocols[i].name, 16);
-		count += strlen(protocols[i].name) + 1;
-		/* null at end of source and target buffers anyway */
+		size_t len = strlen(protocols[i].name) + 1;
+		memcpy(pSMB->DialectsArray+count, protocols[i].name, len);
+		count += len;
 	}
 	inc_rfc1001_len(pSMB, count);
 	pSMB->ByteCount = cpu_to_le16(count);
Signature · 5 lines
2.19.0.rc1

-- 
Cheers,
Stephen Rothwell
Applied
#1 lore raw

Merged into cifs-2.6.git for-next

Is the message annoying enough on powerpc to merit going into 4.19-rc3, presumably should wait for 4.20?

Quoted 44 lines · Stephen Rothwell

A powerpc build of cifs with gcc v8.2.0 produces this warning:

fs/cifs/cifssmb.c: In function ‘CIFSSMBNegotiate’:
fs/cifs/cifssmb.c:605:3: warning: ‘strncpy’ writing 16 bytes into a region of size 1 overflows the destination [-Wstringop-overflow=]

   strncpy(pSMB->DialectsArray+count, protocols[i].name, 16);
   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Since we are already doing a strlen() on the source, change the strncpy to a memcpy().

---
 fs/cifs/cifssmb.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/cifssmb.c +7 -4
index dc2f4cf08fe9..dcf939cb9d2f 100644
@@ -601,10 +601,14 @@ CIFSSMBNegotiate(const unsigned int xid, struct cifs_ses *ses)
        }
 
        count = 0;
+       /*
+        * We know that all the name entries in the protocols array
+        * are short (< 16 bytes anyway) and are NUL terminated.
+        */
        for (i = 0; i < CIFS_NUM_PROT; i++) {
-               strncpy(pSMB->DialectsArray+count, protocols[i].name, 16);
-               count += strlen(protocols[i].name) + 1;
-               /* null at end of source and target buffers anyway */
+               size_t len = strlen(protocols[i].name) + 1;
+               memcpy(pSMB->DialectsArray+count, protocols[i].name, len);
+               count += len;
        }
        inc_rfc1001_len(pSMB, count);
        pSMB->ByteCount = cpu_to_le16(count);
--

2.19.0.rc1

--
Cheers,
Stephen Rothwell

Signature · 3 lines
Thanks,

Steve
#2 lore raw

Hi Steve,

Quoted 5 lines · On Mon, 3 Sep 2018 09:44:17 -0500 Steve French <[email protected]> wrote:

Merged into cifs-2.6.git for-next

Is the message annoying enough on powerpc to merit going into 4.19-rc3, presumably should wait for 4.20?

Up to you, but 4.20 is probably fine.

Signature · 2 lines
Cheers,
Stephen Rothwell
#3 lore raw

ok - it looks like it will be a little simpler if I wait

Quoted 15 lines · Stephen Rothwell

Hi Steve,

Quoted 5 lines · On Mon, 3 Sep 2018 09:44:17 -0500 Steve French <[email protected]> wrote:

Merged into cifs-2.6.git for-next

Is the message annoying enough on powerpc to merit going into 4.19-rc3, presumably should wait for 4.20?

Up to you, but 4.20 is probably fine.

--
Cheers,
Stephen Rothwell

Signature · 3 lines
Thanks,

Steve