[PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Jeff Layton]

...and just make it always turned on. In practice almost everyone
enables this option, and as the Kconfig option says, it needs to be
enabled at runtime anyway. I see no need to make this disableable
via Kconfig option as well, and removing the option simplifies the
testing matrix and reduced ifdef clutter in the code.

Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 fs/cifs/Kconfig       |   27 ---------------------------
 fs/cifs/cifs_debug.c  |    2 --
 fs/cifs/cifsencrypt.c |    2 --
 fs/cifs/cifsglob.h    |   17 +----------------
 fs/cifs/cifspdu.h     |    4 ----
 fs/cifs/cifsproto.h   |    2 --
 fs/cifs/cifssmb.c     |   24 ------------------------
 fs/cifs/connect.c     |    8 ++------
 fs/cifs/sess.c        |   10 ----------
 9 files changed, 3 insertions(+), 93 deletions(-)

diff --git a/fs/cifs/Kconfig b/fs/cifs/Kconfig
index 0554b00..ee7ac6b 100644
--- a/fs/cifs/Kconfig
+++ b/fs/cifs/Kconfig
@@ -49,33 +49,6 @@ config CIFS_STATS2
 	  Unless you are a developer or are doing network performance analysis
 	  or tuning, say N.
 
-config CIFS_WEAK_PW_HASH
-	bool "Support legacy servers which use weaker LANMAN security"
-	depends on CIFS
-	help
-	  Modern CIFS servers including Samba and most Windows versions
-	  (since 1997) support stronger NTLM (and even NTLMv2 and Kerberos)
-	  security mechanisms. These hash the password more securely
-	  than the mechanisms used in the older LANMAN version of the
-	  SMB protocol but LANMAN based authentication is needed to
-	  establish sessions with some old SMB servers.
-
-	  Enabling this option allows the cifs module to mount to older
-	  LANMAN based servers such as OS/2 and Windows 95, but such
-	  mounts may be less secure than mounts using NTLM or more recent
-	  security mechanisms if you are on a public network.  Unless you
-	  have a need to access old SMB servers (and are on a private
-	  network) you probably want to say N.  Even if this support
-	  is enabled in the kernel build, LANMAN authentication will not be
-	  used automatically. At runtime LANMAN mounts are disabled but
-	  can be set to required (or optional) either in
-	  /proc/fs/cifs (see fs/cifs/README for more detail) or via an
-	  option on the mount command. This support is disabled by
-	  default in order to reduce the possibility of a downgrade
-	  attack.
-
-	  If unsure, say N.
-
 config CIFS_UPCALL
 	bool "Kerberos/SPNEGO advanced session setup"
 	depends on CIFS && KEYS
diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c
index 24b3dfc..589774f 100644
--- a/fs/cifs/cifs_debug.c
+++ b/fs/cifs/cifs_debug.c
@@ -126,9 +126,7 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
 #ifdef CONFIG_CIFS_FSCACHE
 	seq_printf(m, " fscache");
 #endif
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 	seq_printf(m, " lanman");
-#endif
 #ifdef CONFIG_CIFS_POSIX
 	seq_printf(m, " posix");
 #endif
diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
index 63c460e..96997cc 100644
--- a/fs/cifs/cifsencrypt.c
+++ b/fs/cifs/cifsencrypt.c
@@ -242,7 +242,6 @@ int setup_ntlm_response(struct cifs_ses *ses, const struct nls_table *nls_cp)
 	return rc;
 }
 
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
 			char *lnm_session_key)
 {
@@ -279,7 +278,6 @@ int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
 
 	return rc;
 }
-#endif /* CIFS_WEAK_PW_HASH */
 
 /* Build a proper attribute value/target info pairs blob.
  * Fill in netbios and dns domain name and workstation name
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 76e7d8b..0e56b1e 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -853,13 +853,8 @@ static inline void free_dfs_info_array(struct dfs_info3_param *param,
 #define   CIFSSEC_MAY_NTLM	0x00002
 #define   CIFSSEC_MAY_NTLMV2	0x00004
 #define   CIFSSEC_MAY_KRB5	0x00008
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 #define   CIFSSEC_MAY_LANMAN	0x00010
 #define   CIFSSEC_MAY_PLNTXT	0x00020
-#else
-#define   CIFSSEC_MAY_LANMAN    0
-#define   CIFSSEC_MAY_PLNTXT    0
-#endif /* weak passwords */
 #define   CIFSSEC_MAY_SEAL	0x00040 /* not supported yet */
 #define   CIFSSEC_MAY_NTLMSSP	0x00080 /* raw ntlmssp with ntlmv2 */
 
@@ -870,23 +865,13 @@ require use of the stronger protocol */
 #define   CIFSSEC_MUST_NTLM	0x02002
 #define   CIFSSEC_MUST_NTLMV2	0x04004
 #define   CIFSSEC_MUST_KRB5	0x08008
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 #define   CIFSSEC_MUST_LANMAN	0x10010
 #define   CIFSSEC_MUST_PLNTXT	0x20020
 #ifdef CONFIG_CIFS_UPCALL
 #define   CIFSSEC_MASK          0xBF0BF /* allows weak security but also krb5 */
-#else
+#else /* UPCALL */
 #define   CIFSSEC_MASK          0xB70B7 /* current flags supported if weak */
 #endif /* UPCALL */
-#else /* do not allow weak pw hash */
-#define   CIFSSEC_MUST_LANMAN	0
-#define   CIFSSEC_MUST_PLNTXT	0
-#ifdef CONFIG_CIFS_UPCALL
-#define   CIFSSEC_MASK          0x8F08F /* flags supported if no weak allowed */
-#else
-#define	  CIFSSEC_MASK          0x87087 /* flags supported if no weak allowed */
-#endif /* UPCALL */
-#endif /* WEAK_PW_HASH */
 #define   CIFSSEC_MUST_SEAL	0x40040 /* not supported yet */
 #define   CIFSSEC_MUST_NTLMSSP	0x80080 /* raw ntlmssp with ntlmv2 */
 
diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h
index 3fb03e2..ef28614 100644
--- a/fs/cifs/cifspdu.h
+++ b/fs/cifs/cifspdu.h
@@ -26,13 +26,9 @@
 #include <asm/unaligned.h>
 #include "smbfsctl.h"
 
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 #define LANMAN_PROT 0
 #define LANMAN2_PROT 1
 #define CIFS_PROT   2
-#else
-#define CIFS_PROT   0
-#endif
 #define POSIX_PROT  (CIFS_PROT+1)
 #define BAD_PROT 0xFFFF
 
diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index 6f4e243..a81549d 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -403,10 +403,8 @@ extern int cifs_crypto_shash_allocate(struct TCP_Server_Info *);
 extern void cifs_crypto_shash_release(struct TCP_Server_Info *);
 extern int calc_seckey(struct cifs_ses *);
 
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 extern int calc_lanman_hash(const char *password, const char *cryptkey,
 				bool encrypt, char *lnm_session_key);
-#endif /* CIFS_WEAK_PW_HASH */
 #ifdef CONFIG_CIFS_DNOTIFY_EXPERIMENTAL /* unused temporarily */
 extern int CIFSSMBNotify(const int xid, struct cifs_tcon *tcon,
 			const int notify_subdirs, const __u16 netfid,
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index 8b7794c..cf12c0c 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -49,10 +49,8 @@ static struct {
 	int index;
 	char *name;
 } protocols[] = {
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 	{LANMAN_PROT, "\2LM1.2X002"},
 	{LANMAN2_PROT, "\2LANMAN2.1"},
-#endif /* weak password hashing for legacy clients */
 	{CIFS_PROT, "\2NT LM 0.12"},
 	{POSIX_PROT, "\2POSIX 2"},
 	{BAD_PROT, "\2"}
@@ -62,10 +60,8 @@ static struct {
 	int index;
 	char *name;
 } protocols[] = {
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 	{LANMAN_PROT, "\2LM1.2X002"},
 	{LANMAN2_PROT, "\2LANMAN2.1"},
-#endif /* weak password hashing for legacy clients */
 	{CIFS_PROT, "\2NT LM 0.12"},
 	{BAD_PROT, "\2"}
 };
@@ -73,17 +69,9 @@ static struct {
 
 /* define the number of elements in the cifs dialect array */
 #ifdef CONFIG_CIFS_POSIX
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 #define CIFS_NUM_PROT 4
-#else
-#define CIFS_NUM_PROT 2
-#endif /* CIFS_WEAK_PW_HASH */
 #else /* not posix */
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 #define CIFS_NUM_PROT 3
-#else
-#define CIFS_NUM_PROT 1
-#endif /* CONFIG_CIFS_WEAK_PW_HASH */
 #endif /* CIFS_POSIX */
 
 /* Forward declarations */
@@ -441,7 +429,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
 		could not negotiate a common dialect */
 		rc = -EOPNOTSUPP;
 		goto neg_err_exit;
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 	} else if ((pSMBr->hdr.WordCount == 13)
 			&& ((server->dialect == LANMAN_PROT)
 				|| (server->dialect == LANMAN2_PROT))) {
@@ -519,13 +506,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
 		/* we will not end up setting signing flags - as no signing
 		was in LANMAN and server did not return the flags on */
 		goto signing_check;
-#else /* weak security disabled */
-	} else if (pSMBr->hdr.WordCount == 13) {
-		cERROR(1, "mount failed, cifs module not built "
-			  "with CIFS_WEAK_PW_HASH support");
-		rc = -EOPNOTSUPP;
-#endif /* WEAK_PW_HASH */
-		goto neg_err_exit;
 	} else if (pSMBr->hdr.WordCount != 17) {
 		/* unknown wct */
 		rc = -EOPNOTSUPP;
@@ -537,9 +517,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
 		cFYI(1, "share mode security");
 
 	if ((server->sec_mode & SECMODE_PW_ENCRYPT) == 0)
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 		if ((secFlags & CIFSSEC_MAY_PLNTXT) == 0)
-#endif /* CIFS_WEAK_PW_HASH */
 			cERROR(1, "Server requests plain text password"
 				  " but client support disabled");
 
@@ -627,9 +605,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
 	} else
 		server->capabilities &= ~CAP_EXTENDED_SECURITY;
 
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 signing_check:
-#endif
 	if ((secFlags & CIFSSEC_MAY_SIGN) == 0) {
 		/* MUST_SIGN already includes the MAY_SIGN FLAG
 		   so if this is zero it means that signing is disabled */
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index aa687c8..5aeca5e 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -1153,10 +1153,8 @@ cifs_parse_mount_options(const char *mountdata, const char *devname,
 			} else if (strnicmp(value, "nontlm", 6) == 0) {
 				/* BB is there a better way to do this? */
 				vol->secFlg |= CIFSSEC_MAY_NTLMV2;
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 			} else if (strnicmp(value, "lanman", 6) == 0) {
 				vol->secFlg |= CIFSSEC_MAY_LANMAN;
-#endif
 			} else if (strnicmp(value, "none", 4) == 0) {
 				vol->nullauth = 1;
 			} else {
@@ -3611,7 +3609,6 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
 		   weaker LANMAN (which we do not send by default) is accepted
 		   by Samba (not sure whether other servers allow
 		   NTLMv2 password here) */
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 		if ((global_secflags & CIFSSEC_MAY_LANMAN) &&
 		    (ses->server->secType == LANMAN))
 			calc_lanman_hash(tcon->password, ses->server->cryptkey,
@@ -3619,9 +3616,8 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
 					    SECMODE_PW_ENCRYPT ? true : false,
 					 bcc_ptr);
 		else
-#endif /* CIFS_WEAK_PW_HASH */
-		rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
-					bcc_ptr, nls_codepage);
+			rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
+						bcc_ptr, nls_codepage);
 
 		bcc_ptr += CIFS_AUTH_RESP_SIZE;
 		if (ses->capabilities & CAP_UNICODE) {
diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c
index d85efad..1cce63b 100644
--- a/fs/cifs/sess.c
+++ b/fs/cifs/sess.c
@@ -592,14 +592,6 @@ ssetup_ntlmssp_authenticate:
 		phase = NtLmAuthenticate; /* if ntlmssp, now final phase */
 
 	if (type == LANMAN) {
-#ifndef CONFIG_CIFS_WEAK_PW_HASH
-		/* LANMAN and plaintext are less secure and off by default.
-		So we make this explicitly be turned on in kconfig (in the
-		build) and turned on at runtime (changed from the default)
-		in proc/fs/cifs or via mount parm.  Unfortunately this is
-		needed for old Win (e.g. Win95), some obscure NAS and OS/2 */
-		return -EOPNOTSUPP;
-#endif
 		wct = 10; /* lanman 2 style sessionsetup */
 	} else if ((type == NTLM) || (type == NTLMv2)) {
 		/* For NTLMv2 failures eventually may need to retry NTLM */
@@ -643,7 +635,6 @@ ssetup_ntlmssp_authenticate:
 	iov[1].iov_len = 0;
 
 	if (type == LANMAN) {
-#ifdef CONFIG_CIFS_WEAK_PW_HASH
 		char lnm_session_key[CIFS_AUTH_RESP_SIZE];
 
 		pSMB->req.hdr.Flags2 &= ~SMBFLG2_UNICODE;
@@ -674,7 +665,6 @@ ssetup_ntlmssp_authenticate:
 		cFYI(1, "Negotiating LANMAN setting up strings");
 		/* Unicode not allowed for LANMAN dialects */
 		ascii_ssetup_strings(&bcc_ptr, ses, nls_cp);
-#endif
 	} else if (type == NTLM) {
 		pSMB->req_no_secext.Capabilities = cpu_to_le32(capabilities);
 		pSMB->req_no_secext.CaseInsensitivePasswordLength =
-- 
1.7.7.5

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Steve French]

Would like some feedback from security experts on this one.
The LANMAN hash is easy to break and enabling it in
a hostile environment seems wrong - obviously there
are home environments and behind NATs where
it is ok and where old LANMAN servers might exist.  Although it
can be controlled at runtime, allowing it to be
not built in gives administrators more control
so that users do not change this workstation
runtime config values in potentially dangerous ways.

On Fri, Jan 20, 2012 at 2:25 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> ...and just make it always turned on. In practice almost everyone
> enables this option, and as the Kconfig option says, it needs to be
> enabled at runtime anyway. I see no need to make this disableable
> via Kconfig option as well, and removing the option simplifies the
> testing matrix and reduced ifdef clutter in the code.
>
> Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> ---
>  fs/cifs/Kconfig       |   27 ---------------------------
>  fs/cifs/cifs_debug.c  |    2 --
>  fs/cifs/cifsencrypt.c |    2 --
>  fs/cifs/cifsglob.h    |   17 +----------------
>  fs/cifs/cifspdu.h     |    4 ----
>  fs/cifs/cifsproto.h   |    2 --
>  fs/cifs/cifssmb.c     |   24 ------------------------
>  fs/cifs/connect.c     |    8 ++------
>  fs/cifs/sess.c        |   10 ----------
>  9 files changed, 3 insertions(+), 93 deletions(-)
>
> diff --git a/fs/cifs/Kconfig b/fs/cifs/Kconfig
> index 0554b00..ee7ac6b 100644
> --- a/fs/cifs/Kconfig
> +++ b/fs/cifs/Kconfig
> @@ -49,33 +49,6 @@ config CIFS_STATS2
>          Unless you are a developer or are doing network performance analysis
>          or tuning, say N.
>
> -config CIFS_WEAK_PW_HASH
> -       bool "Support legacy servers which use weaker LANMAN security"
> -       depends on CIFS
> -       help
> -         Modern CIFS servers including Samba and most Windows versions
> -         (since 1997) support stronger NTLM (and even NTLMv2 and Kerberos)
> -         security mechanisms. These hash the password more securely
> -         than the mechanisms used in the older LANMAN version of the
> -         SMB protocol but LANMAN based authentication is needed to
> -         establish sessions with some old SMB servers.
> -
> -         Enabling this option allows the cifs module to mount to older
> -         LANMAN based servers such as OS/2 and Windows 95, but such
> -         mounts may be less secure than mounts using NTLM or more recent
> -         security mechanisms if you are on a public network.  Unless you
> -         have a need to access old SMB servers (and are on a private
> -         network) you probably want to say N.  Even if this support
> -         is enabled in the kernel build, LANMAN authentication will not be
> -         used automatically. At runtime LANMAN mounts are disabled but
> -         can be set to required (or optional) either in
> -         /proc/fs/cifs (see fs/cifs/README for more detail) or via an
> -         option on the mount command. This support is disabled by
> -         default in order to reduce the possibility of a downgrade
> -         attack.
> -
> -         If unsure, say N.
> -
>  config CIFS_UPCALL
>        bool "Kerberos/SPNEGO advanced session setup"
>        depends on CIFS && KEYS
> diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c
> index 24b3dfc..589774f 100644
> --- a/fs/cifs/cifs_debug.c
> +++ b/fs/cifs/cifs_debug.c
> @@ -126,9 +126,7 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
>  #ifdef CONFIG_CIFS_FSCACHE
>        seq_printf(m, " fscache");
>  #endif
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>        seq_printf(m, " lanman");
> -#endif
>  #ifdef CONFIG_CIFS_POSIX
>        seq_printf(m, " posix");
>  #endif
> diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
> index 63c460e..96997cc 100644
> --- a/fs/cifs/cifsencrypt.c
> +++ b/fs/cifs/cifsencrypt.c
> @@ -242,7 +242,6 @@ int setup_ntlm_response(struct cifs_ses *ses, const struct nls_table *nls_cp)
>        return rc;
>  }
>
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
>                        char *lnm_session_key)
>  {
> @@ -279,7 +278,6 @@ int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
>
>        return rc;
>  }
> -#endif /* CIFS_WEAK_PW_HASH */
>
>  /* Build a proper attribute value/target info pairs blob.
>  * Fill in netbios and dns domain name and workstation name
> diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
> index 76e7d8b..0e56b1e 100644
> --- a/fs/cifs/cifsglob.h
> +++ b/fs/cifs/cifsglob.h
> @@ -853,13 +853,8 @@ static inline void free_dfs_info_array(struct dfs_info3_param *param,
>  #define   CIFSSEC_MAY_NTLM     0x00002
>  #define   CIFSSEC_MAY_NTLMV2   0x00004
>  #define   CIFSSEC_MAY_KRB5     0x00008
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  #define   CIFSSEC_MAY_LANMAN   0x00010
>  #define   CIFSSEC_MAY_PLNTXT   0x00020
> -#else
> -#define   CIFSSEC_MAY_LANMAN    0
> -#define   CIFSSEC_MAY_PLNTXT    0
> -#endif /* weak passwords */
>  #define   CIFSSEC_MAY_SEAL     0x00040 /* not supported yet */
>  #define   CIFSSEC_MAY_NTLMSSP  0x00080 /* raw ntlmssp with ntlmv2 */
>
> @@ -870,23 +865,13 @@ require use of the stronger protocol */
>  #define   CIFSSEC_MUST_NTLM    0x02002
>  #define   CIFSSEC_MUST_NTLMV2  0x04004
>  #define   CIFSSEC_MUST_KRB5    0x08008
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  #define   CIFSSEC_MUST_LANMAN  0x10010
>  #define   CIFSSEC_MUST_PLNTXT  0x20020
>  #ifdef CONFIG_CIFS_UPCALL
>  #define   CIFSSEC_MASK          0xBF0BF /* allows weak security but also krb5 */
> -#else
> +#else /* UPCALL */
>  #define   CIFSSEC_MASK          0xB70B7 /* current flags supported if weak */
>  #endif /* UPCALL */
> -#else /* do not allow weak pw hash */
> -#define   CIFSSEC_MUST_LANMAN  0
> -#define   CIFSSEC_MUST_PLNTXT  0
> -#ifdef CONFIG_CIFS_UPCALL
> -#define   CIFSSEC_MASK          0x8F08F /* flags supported if no weak allowed */
> -#else
> -#define          CIFSSEC_MASK          0x87087 /* flags supported if no weak allowed */
> -#endif /* UPCALL */
> -#endif /* WEAK_PW_HASH */
>  #define   CIFSSEC_MUST_SEAL    0x40040 /* not supported yet */
>  #define   CIFSSEC_MUST_NTLMSSP 0x80080 /* raw ntlmssp with ntlmv2 */
>
> diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h
> index 3fb03e2..ef28614 100644
> --- a/fs/cifs/cifspdu.h
> +++ b/fs/cifs/cifspdu.h
> @@ -26,13 +26,9 @@
>  #include <asm/unaligned.h>
>  #include "smbfsctl.h"
>
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  #define LANMAN_PROT 0
>  #define LANMAN2_PROT 1
>  #define CIFS_PROT   2
> -#else
> -#define CIFS_PROT   0
> -#endif
>  #define POSIX_PROT  (CIFS_PROT+1)
>  #define BAD_PROT 0xFFFF
>
> diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
> index 6f4e243..a81549d 100644
> --- a/fs/cifs/cifsproto.h
> +++ b/fs/cifs/cifsproto.h
> @@ -403,10 +403,8 @@ extern int cifs_crypto_shash_allocate(struct TCP_Server_Info *);
>  extern void cifs_crypto_shash_release(struct TCP_Server_Info *);
>  extern int calc_seckey(struct cifs_ses *);
>
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  extern int calc_lanman_hash(const char *password, const char *cryptkey,
>                                bool encrypt, char *lnm_session_key);
> -#endif /* CIFS_WEAK_PW_HASH */
>  #ifdef CONFIG_CIFS_DNOTIFY_EXPERIMENTAL /* unused temporarily */
>  extern int CIFSSMBNotify(const int xid, struct cifs_tcon *tcon,
>                        const int notify_subdirs, const __u16 netfid,
> diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
> index 8b7794c..cf12c0c 100644
> --- a/fs/cifs/cifssmb.c
> +++ b/fs/cifs/cifssmb.c
> @@ -49,10 +49,8 @@ static struct {
>        int index;
>        char *name;
>  } protocols[] = {
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>        {LANMAN_PROT, "\2LM1.2X002"},
>        {LANMAN2_PROT, "\2LANMAN2.1"},
> -#endif /* weak password hashing for legacy clients */
>        {CIFS_PROT, "\2NT LM 0.12"},
>        {POSIX_PROT, "\2POSIX 2"},
>        {BAD_PROT, "\2"}
> @@ -62,10 +60,8 @@ static struct {
>        int index;
>        char *name;
>  } protocols[] = {
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>        {LANMAN_PROT, "\2LM1.2X002"},
>        {LANMAN2_PROT, "\2LANMAN2.1"},
> -#endif /* weak password hashing for legacy clients */
>        {CIFS_PROT, "\2NT LM 0.12"},
>        {BAD_PROT, "\2"}
>  };
> @@ -73,17 +69,9 @@ static struct {
>
>  /* define the number of elements in the cifs dialect array */
>  #ifdef CONFIG_CIFS_POSIX
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  #define CIFS_NUM_PROT 4
> -#else
> -#define CIFS_NUM_PROT 2
> -#endif /* CIFS_WEAK_PW_HASH */
>  #else /* not posix */
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  #define CIFS_NUM_PROT 3
> -#else
> -#define CIFS_NUM_PROT 1
> -#endif /* CONFIG_CIFS_WEAK_PW_HASH */
>  #endif /* CIFS_POSIX */
>
>  /* Forward declarations */
> @@ -441,7 +429,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>                could not negotiate a common dialect */
>                rc = -EOPNOTSUPP;
>                goto neg_err_exit;
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>        } else if ((pSMBr->hdr.WordCount == 13)
>                        && ((server->dialect == LANMAN_PROT)
>                                || (server->dialect == LANMAN2_PROT))) {
> @@ -519,13 +506,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>                /* we will not end up setting signing flags - as no signing
>                was in LANMAN and server did not return the flags on */
>                goto signing_check;
> -#else /* weak security disabled */
> -       } else if (pSMBr->hdr.WordCount == 13) {
> -               cERROR(1, "mount failed, cifs module not built "
> -                         "with CIFS_WEAK_PW_HASH support");
> -               rc = -EOPNOTSUPP;
> -#endif /* WEAK_PW_HASH */
> -               goto neg_err_exit;
>        } else if (pSMBr->hdr.WordCount != 17) {
>                /* unknown wct */
>                rc = -EOPNOTSUPP;
> @@ -537,9 +517,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>                cFYI(1, "share mode security");
>
>        if ((server->sec_mode & SECMODE_PW_ENCRYPT) == 0)
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>                if ((secFlags & CIFSSEC_MAY_PLNTXT) == 0)
> -#endif /* CIFS_WEAK_PW_HASH */
>                        cERROR(1, "Server requests plain text password"
>                                  " but client support disabled");
>
> @@ -627,9 +605,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>        } else
>                server->capabilities &= ~CAP_EXTENDED_SECURITY;
>
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>  signing_check:
> -#endif
>        if ((secFlags & CIFSSEC_MAY_SIGN) == 0) {
>                /* MUST_SIGN already includes the MAY_SIGN FLAG
>                   so if this is zero it means that signing is disabled */
> diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
> index aa687c8..5aeca5e 100644
> --- a/fs/cifs/connect.c
> +++ b/fs/cifs/connect.c
> @@ -1153,10 +1153,8 @@ cifs_parse_mount_options(const char *mountdata, const char *devname,
>                        } else if (strnicmp(value, "nontlm", 6) == 0) {
>                                /* BB is there a better way to do this? */
>                                vol->secFlg |= CIFSSEC_MAY_NTLMV2;
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>                        } else if (strnicmp(value, "lanman", 6) == 0) {
>                                vol->secFlg |= CIFSSEC_MAY_LANMAN;
> -#endif
>                        } else if (strnicmp(value, "none", 4) == 0) {
>                                vol->nullauth = 1;
>                        } else {
> @@ -3611,7 +3609,6 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
>                   weaker LANMAN (which we do not send by default) is accepted
>                   by Samba (not sure whether other servers allow
>                   NTLMv2 password here) */
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>                if ((global_secflags & CIFSSEC_MAY_LANMAN) &&
>                    (ses->server->secType == LANMAN))
>                        calc_lanman_hash(tcon->password, ses->server->cryptkey,
> @@ -3619,9 +3616,8 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
>                                            SECMODE_PW_ENCRYPT ? true : false,
>                                         bcc_ptr);
>                else
> -#endif /* CIFS_WEAK_PW_HASH */
> -               rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
> -                                       bcc_ptr, nls_codepage);
> +                       rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
> +                                               bcc_ptr, nls_codepage);
>
>                bcc_ptr += CIFS_AUTH_RESP_SIZE;
>                if (ses->capabilities & CAP_UNICODE) {
> diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c
> index d85efad..1cce63b 100644
> --- a/fs/cifs/sess.c
> +++ b/fs/cifs/sess.c
> @@ -592,14 +592,6 @@ ssetup_ntlmssp_authenticate:
>                phase = NtLmAuthenticate; /* if ntlmssp, now final phase */
>
>        if (type == LANMAN) {
> -#ifndef CONFIG_CIFS_WEAK_PW_HASH
> -               /* LANMAN and plaintext are less secure and off by default.
> -               So we make this explicitly be turned on in kconfig (in the
> -               build) and turned on at runtime (changed from the default)
> -               in proc/fs/cifs or via mount parm.  Unfortunately this is
> -               needed for old Win (e.g. Win95), some obscure NAS and OS/2 */
> -               return -EOPNOTSUPP;
> -#endif
>                wct = 10; /* lanman 2 style sessionsetup */
>        } else if ((type == NTLM) || (type == NTLMv2)) {
>                /* For NTLMv2 failures eventually may need to retry NTLM */
> @@ -643,7 +635,6 @@ ssetup_ntlmssp_authenticate:
>        iov[1].iov_len = 0;
>
>        if (type == LANMAN) {
> -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>                char lnm_session_key[CIFS_AUTH_RESP_SIZE];
>
>                pSMB->req.hdr.Flags2 &= ~SMBFLG2_UNICODE;
> @@ -674,7 +665,6 @@ ssetup_ntlmssp_authenticate:
>                cFYI(1, "Negotiating LANMAN setting up strings");
>                /* Unicode not allowed for LANMAN dialects */
>                ascii_ssetup_strings(&bcc_ptr, ses, nls_cp);
> -#endif
>        } else if (type == NTLM) {
>                pSMB->req_no_secext.Capabilities = cpu_to_le32(capabilities);
>                pSMB->req_no_secext.CaseInsensitivePasswordLength =
> --
> 1.7.7.5
>



-- 
Thanks,

Steve

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Jeff Layton]

On Fri, 20 Jan 2012 14:34:05 -0600
Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> Would like some feedback from security experts on this one.
> The LANMAN hash is easy to break and enabling it in
> a hostile environment seems wrong - obviously there
> are home environments and behind NATs where
> it is ok and where old LANMAN servers might exist.  Although it
> can be controlled at runtime, allowing it to be
> not built in gives administrators more control
> so that users do not change this workstation
> runtime config values in potentially dangerous ways.
> 

That's fine -- this is intended for 3.4 so you should have plenty of
time. Can you briefly outline the potential scenario where you see this
as useful?

I don't see a lot of benefit from this Kconfig option. You have to be
root to enable this at runtime, and at that point all bets are off as
far as what someone can do. They can plug in their own kmod at that
point if you don't compile it in.


> On Fri, Jan 20, 2012 at 2:25 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > ...and just make it always turned on. In practice almost everyone
> > enables this option, and as the Kconfig option says, it needs to be
> > enabled at runtime anyway. I see no need to make this disableable
> > via Kconfig option as well, and removing the option simplifies the
> > testing matrix and reduced ifdef clutter in the code.
> >
> > Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> > ---
> >  fs/cifs/Kconfig       |   27 ---------------------------
> >  fs/cifs/cifs_debug.c  |    2 --
> >  fs/cifs/cifsencrypt.c |    2 --
> >  fs/cifs/cifsglob.h    |   17 +----------------
> >  fs/cifs/cifspdu.h     |    4 ----
> >  fs/cifs/cifsproto.h   |    2 --
> >  fs/cifs/cifssmb.c     |   24 ------------------------
> >  fs/cifs/connect.c     |    8 ++------
> >  fs/cifs/sess.c        |   10 ----------
> >  9 files changed, 3 insertions(+), 93 deletions(-)
> >
> > diff --git a/fs/cifs/Kconfig b/fs/cifs/Kconfig
> > index 0554b00..ee7ac6b 100644
> > --- a/fs/cifs/Kconfig
> > +++ b/fs/cifs/Kconfig
> > @@ -49,33 +49,6 @@ config CIFS_STATS2
> >          Unless you are a developer or are doing network performance analysis
> >          or tuning, say N.
> >
> > -config CIFS_WEAK_PW_HASH
> > -       bool "Support legacy servers which use weaker LANMAN security"
> > -       depends on CIFS
> > -       help
> > -         Modern CIFS servers including Samba and most Windows versions
> > -         (since 1997) support stronger NTLM (and even NTLMv2 and Kerberos)
> > -         security mechanisms. These hash the password more securely
> > -         than the mechanisms used in the older LANMAN version of the
> > -         SMB protocol but LANMAN based authentication is needed to
> > -         establish sessions with some old SMB servers.
> > -
> > -         Enabling this option allows the cifs module to mount to older
> > -         LANMAN based servers such as OS/2 and Windows 95, but such
> > -         mounts may be less secure than mounts using NTLM or more recent
> > -         security mechanisms if you are on a public network.  Unless you
> > -         have a need to access old SMB servers (and are on a private
> > -         network) you probably want to say N.  Even if this support
> > -         is enabled in the kernel build, LANMAN authentication will not be
> > -         used automatically. At runtime LANMAN mounts are disabled but
> > -         can be set to required (or optional) either in
> > -         /proc/fs/cifs (see fs/cifs/README for more detail) or via an
> > -         option on the mount command. This support is disabled by
> > -         default in order to reduce the possibility of a downgrade
> > -         attack.
> > -
> > -         If unsure, say N.
> > -
> >  config CIFS_UPCALL
> >        bool "Kerberos/SPNEGO advanced session setup"
> >        depends on CIFS && KEYS
> > diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c
> > index 24b3dfc..589774f 100644
> > --- a/fs/cifs/cifs_debug.c
> > +++ b/fs/cifs/cifs_debug.c
> > @@ -126,9 +126,7 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
> >  #ifdef CONFIG_CIFS_FSCACHE
> >        seq_printf(m, " fscache");
> >  #endif
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >        seq_printf(m, " lanman");
> > -#endif
> >  #ifdef CONFIG_CIFS_POSIX
> >        seq_printf(m, " posix");
> >  #endif
> > diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
> > index 63c460e..96997cc 100644
> > --- a/fs/cifs/cifsencrypt.c
> > +++ b/fs/cifs/cifsencrypt.c
> > @@ -242,7 +242,6 @@ int setup_ntlm_response(struct cifs_ses *ses, const struct nls_table *nls_cp)
> >        return rc;
> >  }
> >
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
> >                        char *lnm_session_key)
> >  {
> > @@ -279,7 +278,6 @@ int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
> >
> >        return rc;
> >  }
> > -#endif /* CIFS_WEAK_PW_HASH */
> >
> >  /* Build a proper attribute value/target info pairs blob.
> >  * Fill in netbios and dns domain name and workstation name
> > diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
> > index 76e7d8b..0e56b1e 100644
> > --- a/fs/cifs/cifsglob.h
> > +++ b/fs/cifs/cifsglob.h
> > @@ -853,13 +853,8 @@ static inline void free_dfs_info_array(struct dfs_info3_param *param,
> >  #define   CIFSSEC_MAY_NTLM     0x00002
> >  #define   CIFSSEC_MAY_NTLMV2   0x00004
> >  #define   CIFSSEC_MAY_KRB5     0x00008
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  #define   CIFSSEC_MAY_LANMAN   0x00010
> >  #define   CIFSSEC_MAY_PLNTXT   0x00020
> > -#else
> > -#define   CIFSSEC_MAY_LANMAN    0
> > -#define   CIFSSEC_MAY_PLNTXT    0
> > -#endif /* weak passwords */
> >  #define   CIFSSEC_MAY_SEAL     0x00040 /* not supported yet */
> >  #define   CIFSSEC_MAY_NTLMSSP  0x00080 /* raw ntlmssp with ntlmv2 */
> >
> > @@ -870,23 +865,13 @@ require use of the stronger protocol */
> >  #define   CIFSSEC_MUST_NTLM    0x02002
> >  #define   CIFSSEC_MUST_NTLMV2  0x04004
> >  #define   CIFSSEC_MUST_KRB5    0x08008
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  #define   CIFSSEC_MUST_LANMAN  0x10010
> >  #define   CIFSSEC_MUST_PLNTXT  0x20020
> >  #ifdef CONFIG_CIFS_UPCALL
> >  #define   CIFSSEC_MASK          0xBF0BF /* allows weak security but also krb5 */
> > -#else
> > +#else /* UPCALL */
> >  #define   CIFSSEC_MASK          0xB70B7 /* current flags supported if weak */
> >  #endif /* UPCALL */
> > -#else /* do not allow weak pw hash */
> > -#define   CIFSSEC_MUST_LANMAN  0
> > -#define   CIFSSEC_MUST_PLNTXT  0
> > -#ifdef CONFIG_CIFS_UPCALL
> > -#define   CIFSSEC_MASK          0x8F08F /* flags supported if no weak allowed */
> > -#else
> > -#define          CIFSSEC_MASK          0x87087 /* flags supported if no weak allowed */
> > -#endif /* UPCALL */
> > -#endif /* WEAK_PW_HASH */
> >  #define   CIFSSEC_MUST_SEAL    0x40040 /* not supported yet */
> >  #define   CIFSSEC_MUST_NTLMSSP 0x80080 /* raw ntlmssp with ntlmv2 */
> >
> > diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h
> > index 3fb03e2..ef28614 100644
> > --- a/fs/cifs/cifspdu.h
> > +++ b/fs/cifs/cifspdu.h
> > @@ -26,13 +26,9 @@
> >  #include <asm/unaligned.h>
> >  #include "smbfsctl.h"
> >
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  #define LANMAN_PROT 0
> >  #define LANMAN2_PROT 1
> >  #define CIFS_PROT   2
> > -#else
> > -#define CIFS_PROT   0
> > -#endif
> >  #define POSIX_PROT  (CIFS_PROT+1)
> >  #define BAD_PROT 0xFFFF
> >
> > diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
> > index 6f4e243..a81549d 100644
> > --- a/fs/cifs/cifsproto.h
> > +++ b/fs/cifs/cifsproto.h
> > @@ -403,10 +403,8 @@ extern int cifs_crypto_shash_allocate(struct TCP_Server_Info *);
> >  extern void cifs_crypto_shash_release(struct TCP_Server_Info *);
> >  extern int calc_seckey(struct cifs_ses *);
> >
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  extern int calc_lanman_hash(const char *password, const char *cryptkey,
> >                                bool encrypt, char *lnm_session_key);
> > -#endif /* CIFS_WEAK_PW_HASH */
> >  #ifdef CONFIG_CIFS_DNOTIFY_EXPERIMENTAL /* unused temporarily */
> >  extern int CIFSSMBNotify(const int xid, struct cifs_tcon *tcon,
> >                        const int notify_subdirs, const __u16 netfid,
> > diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
> > index 8b7794c..cf12c0c 100644
> > --- a/fs/cifs/cifssmb.c
> > +++ b/fs/cifs/cifssmb.c
> > @@ -49,10 +49,8 @@ static struct {
> >        int index;
> >        char *name;
> >  } protocols[] = {
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >        {LANMAN_PROT, "\2LM1.2X002"},
> >        {LANMAN2_PROT, "\2LANMAN2.1"},
> > -#endif /* weak password hashing for legacy clients */
> >        {CIFS_PROT, "\2NT LM 0.12"},
> >        {POSIX_PROT, "\2POSIX 2"},
> >        {BAD_PROT, "\2"}
> > @@ -62,10 +60,8 @@ static struct {
> >        int index;
> >        char *name;
> >  } protocols[] = {
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >        {LANMAN_PROT, "\2LM1.2X002"},
> >        {LANMAN2_PROT, "\2LANMAN2.1"},
> > -#endif /* weak password hashing for legacy clients */
> >        {CIFS_PROT, "\2NT LM 0.12"},
> >        {BAD_PROT, "\2"}
> >  };
> > @@ -73,17 +69,9 @@ static struct {
> >
> >  /* define the number of elements in the cifs dialect array */
> >  #ifdef CONFIG_CIFS_POSIX
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  #define CIFS_NUM_PROT 4
> > -#else
> > -#define CIFS_NUM_PROT 2
> > -#endif /* CIFS_WEAK_PW_HASH */
> >  #else /* not posix */
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  #define CIFS_NUM_PROT 3
> > -#else
> > -#define CIFS_NUM_PROT 1
> > -#endif /* CONFIG_CIFS_WEAK_PW_HASH */
> >  #endif /* CIFS_POSIX */
> >
> >  /* Forward declarations */
> > @@ -441,7 +429,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
> >                could not negotiate a common dialect */
> >                rc = -EOPNOTSUPP;
> >                goto neg_err_exit;
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >        } else if ((pSMBr->hdr.WordCount == 13)
> >                        && ((server->dialect == LANMAN_PROT)
> >                                || (server->dialect == LANMAN2_PROT))) {
> > @@ -519,13 +506,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
> >                /* we will not end up setting signing flags - as no signing
> >                was in LANMAN and server did not return the flags on */
> >                goto signing_check;
> > -#else /* weak security disabled */
> > -       } else if (pSMBr->hdr.WordCount == 13) {
> > -               cERROR(1, "mount failed, cifs module not built "
> > -                         "with CIFS_WEAK_PW_HASH support");
> > -               rc = -EOPNOTSUPP;
> > -#endif /* WEAK_PW_HASH */
> > -               goto neg_err_exit;
> >        } else if (pSMBr->hdr.WordCount != 17) {
> >                /* unknown wct */
> >                rc = -EOPNOTSUPP;
> > @@ -537,9 +517,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
> >                cFYI(1, "share mode security");
> >
> >        if ((server->sec_mode & SECMODE_PW_ENCRYPT) == 0)
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >                if ((secFlags & CIFSSEC_MAY_PLNTXT) == 0)
> > -#endif /* CIFS_WEAK_PW_HASH */
> >                        cERROR(1, "Server requests plain text password"
> >                                  " but client support disabled");
> >
> > @@ -627,9 +605,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
> >        } else
> >                server->capabilities &= ~CAP_EXTENDED_SECURITY;
> >
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >  signing_check:
> > -#endif
> >        if ((secFlags & CIFSSEC_MAY_SIGN) == 0) {
> >                /* MUST_SIGN already includes the MAY_SIGN FLAG
> >                   so if this is zero it means that signing is disabled */
> > diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
> > index aa687c8..5aeca5e 100644
> > --- a/fs/cifs/connect.c
> > +++ b/fs/cifs/connect.c
> > @@ -1153,10 +1153,8 @@ cifs_parse_mount_options(const char *mountdata, const char *devname,
> >                        } else if (strnicmp(value, "nontlm", 6) == 0) {
> >                                /* BB is there a better way to do this? */
> >                                vol->secFlg |= CIFSSEC_MAY_NTLMV2;
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >                        } else if (strnicmp(value, "lanman", 6) == 0) {
> >                                vol->secFlg |= CIFSSEC_MAY_LANMAN;
> > -#endif
> >                        } else if (strnicmp(value, "none", 4) == 0) {
> >                                vol->nullauth = 1;
> >                        } else {
> > @@ -3611,7 +3609,6 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
> >                   weaker LANMAN (which we do not send by default) is accepted
> >                   by Samba (not sure whether other servers allow
> >                   NTLMv2 password here) */
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >                if ((global_secflags & CIFSSEC_MAY_LANMAN) &&
> >                    (ses->server->secType == LANMAN))
> >                        calc_lanman_hash(tcon->password, ses->server->cryptkey,
> > @@ -3619,9 +3616,8 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
> >                                            SECMODE_PW_ENCRYPT ? true : false,
> >                                         bcc_ptr);
> >                else
> > -#endif /* CIFS_WEAK_PW_HASH */
> > -               rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
> > -                                       bcc_ptr, nls_codepage);
> > +                       rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
> > +                                               bcc_ptr, nls_codepage);
> >
> >                bcc_ptr += CIFS_AUTH_RESP_SIZE;
> >                if (ses->capabilities & CAP_UNICODE) {
> > diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c
> > index d85efad..1cce63b 100644
> > --- a/fs/cifs/sess.c
> > +++ b/fs/cifs/sess.c
> > @@ -592,14 +592,6 @@ ssetup_ntlmssp_authenticate:
> >                phase = NtLmAuthenticate; /* if ntlmssp, now final phase */
> >
> >        if (type == LANMAN) {
> > -#ifndef CONFIG_CIFS_WEAK_PW_HASH
> > -               /* LANMAN and plaintext are less secure and off by default.
> > -               So we make this explicitly be turned on in kconfig (in the
> > -               build) and turned on at runtime (changed from the default)
> > -               in proc/fs/cifs or via mount parm.  Unfortunately this is
> > -               needed for old Win (e.g. Win95), some obscure NAS and OS/2 */
> > -               return -EOPNOTSUPP;
> > -#endif
> >                wct = 10; /* lanman 2 style sessionsetup */
> >        } else if ((type == NTLM) || (type == NTLMv2)) {
> >                /* For NTLMv2 failures eventually may need to retry NTLM */
> > @@ -643,7 +635,6 @@ ssetup_ntlmssp_authenticate:
> >        iov[1].iov_len = 0;
> >
> >        if (type == LANMAN) {
> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >                char lnm_session_key[CIFS_AUTH_RESP_SIZE];
> >
> >                pSMB->req.hdr.Flags2 &= ~SMBFLG2_UNICODE;
> > @@ -674,7 +665,6 @@ ssetup_ntlmssp_authenticate:
> >                cFYI(1, "Negotiating LANMAN setting up strings");
> >                /* Unicode not allowed for LANMAN dialects */
> >                ascii_ssetup_strings(&bcc_ptr, ses, nls_cp);
> > -#endif
> >        } else if (type == NTLM) {
> >                pSMB->req_no_secext.Capabilities = cpu_to_le32(capabilities);
> >                pSMB->req_no_secext.CaseInsensitivePasswordLength =
> > --
> > 1.7.7.5
> >
> 
> 
> 


-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Steve French]

My general thinking on this is as follows:

If the kernel is distributed to all the workstations in an organization
with this Kconfig option disabled, it makes it harder for individual users
to make the mistake of enabling lanman (sec=lanman, or the Kconfig
option) on a public network and thus send weak password hashes
which could be discovered simply.   Most distros make the choice
of enabling broader compatibility with old pre-1997 servers but
it is a very small set of servers who would require lanman support,
and a large number of potential attackers who could benefit if
users enable lanman on a public network.  I suspect that there
are environments where removing code (via Kconfig) is preferred
to trusting all owners of all workstations running that organizations
standard linux to never enable lanman at runtime.

But ... the opinion of security specialists on this would be welcome.

On Fri, Jan 20, 2012 at 2:38 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> On Fri, 20 Jan 2012 14:34:05 -0600
> Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
>> Would like some feedback from security experts on this one.
>> The LANMAN hash is easy to break and enabling it in
>> a hostile environment seems wrong - obviously there
>> are home environments and behind NATs where
>> it is ok and where old LANMAN servers might exist.  Although it
>> can be controlled at runtime, allowing it to be
>> not built in gives administrators more control
>> so that users do not change this workstation
>> runtime config values in potentially dangerous ways.
>>
>
> That's fine -- this is intended for 3.4 so you should have plenty of
> time. Can you briefly outline the potential scenario where you see this
> as useful?
>
> I don't see a lot of benefit from this Kconfig option. You have to be
> root to enable this at runtime, and at that point all bets are off as
> far as what someone can do. They can plug in their own kmod at that
> point if you don't compile it in.
>
>
>> On Fri, Jan 20, 2012 at 2:25 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
>> > ...and just make it always turned on. In practice almost everyone
>> > enables this option, and as the Kconfig option says, it needs to be
>> > enabled at runtime anyway. I see no need to make this disableable
>> > via Kconfig option as well, and removing the option simplifies the
>> > testing matrix and reduced ifdef clutter in the code.
>> >
>> > Signed-off-by: Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
>> > ---
>> >  fs/cifs/Kconfig       |   27 ---------------------------
>> >  fs/cifs/cifs_debug.c  |    2 --
>> >  fs/cifs/cifsencrypt.c |    2 --
>> >  fs/cifs/cifsglob.h    |   17 +----------------
>> >  fs/cifs/cifspdu.h     |    4 ----
>> >  fs/cifs/cifsproto.h   |    2 --
>> >  fs/cifs/cifssmb.c     |   24 ------------------------
>> >  fs/cifs/connect.c     |    8 ++------
>> >  fs/cifs/sess.c        |   10 ----------
>> >  9 files changed, 3 insertions(+), 93 deletions(-)
>> >
>> > diff --git a/fs/cifs/Kconfig b/fs/cifs/Kconfig
>> > index 0554b00..ee7ac6b 100644
>> > --- a/fs/cifs/Kconfig
>> > +++ b/fs/cifs/Kconfig
>> > @@ -49,33 +49,6 @@ config CIFS_STATS2
>> >          Unless you are a developer or are doing network performance analysis
>> >          or tuning, say N.
>> >
>> > -config CIFS_WEAK_PW_HASH
>> > -       bool "Support legacy servers which use weaker LANMAN security"
>> > -       depends on CIFS
>> > -       help
>> > -         Modern CIFS servers including Samba and most Windows versions
>> > -         (since 1997) support stronger NTLM (and even NTLMv2 and Kerberos)
>> > -         security mechanisms. These hash the password more securely
>> > -         than the mechanisms used in the older LANMAN version of the
>> > -         SMB protocol but LANMAN based authentication is needed to
>> > -         establish sessions with some old SMB servers.
>> > -
>> > -         Enabling this option allows the cifs module to mount to older
>> > -         LANMAN based servers such as OS/2 and Windows 95, but such
>> > -         mounts may be less secure than mounts using NTLM or more recent
>> > -         security mechanisms if you are on a public network.  Unless you
>> > -         have a need to access old SMB servers (and are on a private
>> > -         network) you probably want to say N.  Even if this support
>> > -         is enabled in the kernel build, LANMAN authentication will not be
>> > -         used automatically. At runtime LANMAN mounts are disabled but
>> > -         can be set to required (or optional) either in
>> > -         /proc/fs/cifs (see fs/cifs/README for more detail) or via an
>> > -         option on the mount command. This support is disabled by
>> > -         default in order to reduce the possibility of a downgrade
>> > -         attack.
>> > -
>> > -         If unsure, say N.
>> > -
>> >  config CIFS_UPCALL
>> >        bool "Kerberos/SPNEGO advanced session setup"
>> >        depends on CIFS && KEYS
>> > diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c
>> > index 24b3dfc..589774f 100644
>> > --- a/fs/cifs/cifs_debug.c
>> > +++ b/fs/cifs/cifs_debug.c
>> > @@ -126,9 +126,7 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
>> >  #ifdef CONFIG_CIFS_FSCACHE
>> >        seq_printf(m, " fscache");
>> >  #endif
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >        seq_printf(m, " lanman");
>> > -#endif
>> >  #ifdef CONFIG_CIFS_POSIX
>> >        seq_printf(m, " posix");
>> >  #endif
>> > diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
>> > index 63c460e..96997cc 100644
>> > --- a/fs/cifs/cifsencrypt.c
>> > +++ b/fs/cifs/cifsencrypt.c
>> > @@ -242,7 +242,6 @@ int setup_ntlm_response(struct cifs_ses *ses, const struct nls_table *nls_cp)
>> >        return rc;
>> >  }
>> >
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
>> >                        char *lnm_session_key)
>> >  {
>> > @@ -279,7 +278,6 @@ int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt,
>> >
>> >        return rc;
>> >  }
>> > -#endif /* CIFS_WEAK_PW_HASH */
>> >
>> >  /* Build a proper attribute value/target info pairs blob.
>> >  * Fill in netbios and dns domain name and workstation name
>> > diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
>> > index 76e7d8b..0e56b1e 100644
>> > --- a/fs/cifs/cifsglob.h
>> > +++ b/fs/cifs/cifsglob.h
>> > @@ -853,13 +853,8 @@ static inline void free_dfs_info_array(struct dfs_info3_param *param,
>> >  #define   CIFSSEC_MAY_NTLM     0x00002
>> >  #define   CIFSSEC_MAY_NTLMV2   0x00004
>> >  #define   CIFSSEC_MAY_KRB5     0x00008
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  #define   CIFSSEC_MAY_LANMAN   0x00010
>> >  #define   CIFSSEC_MAY_PLNTXT   0x00020
>> > -#else
>> > -#define   CIFSSEC_MAY_LANMAN    0
>> > -#define   CIFSSEC_MAY_PLNTXT    0
>> > -#endif /* weak passwords */
>> >  #define   CIFSSEC_MAY_SEAL     0x00040 /* not supported yet */
>> >  #define   CIFSSEC_MAY_NTLMSSP  0x00080 /* raw ntlmssp with ntlmv2 */
>> >
>> > @@ -870,23 +865,13 @@ require use of the stronger protocol */
>> >  #define   CIFSSEC_MUST_NTLM    0x02002
>> >  #define   CIFSSEC_MUST_NTLMV2  0x04004
>> >  #define   CIFSSEC_MUST_KRB5    0x08008
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  #define   CIFSSEC_MUST_LANMAN  0x10010
>> >  #define   CIFSSEC_MUST_PLNTXT  0x20020
>> >  #ifdef CONFIG_CIFS_UPCALL
>> >  #define   CIFSSEC_MASK          0xBF0BF /* allows weak security but also krb5 */
>> > -#else
>> > +#else /* UPCALL */
>> >  #define   CIFSSEC_MASK          0xB70B7 /* current flags supported if weak */
>> >  #endif /* UPCALL */
>> > -#else /* do not allow weak pw hash */
>> > -#define   CIFSSEC_MUST_LANMAN  0
>> > -#define   CIFSSEC_MUST_PLNTXT  0
>> > -#ifdef CONFIG_CIFS_UPCALL
>> > -#define   CIFSSEC_MASK          0x8F08F /* flags supported if no weak allowed */
>> > -#else
>> > -#define          CIFSSEC_MASK          0x87087 /* flags supported if no weak allowed */
>> > -#endif /* UPCALL */
>> > -#endif /* WEAK_PW_HASH */
>> >  #define   CIFSSEC_MUST_SEAL    0x40040 /* not supported yet */
>> >  #define   CIFSSEC_MUST_NTLMSSP 0x80080 /* raw ntlmssp with ntlmv2 */
>> >
>> > diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h
>> > index 3fb03e2..ef28614 100644
>> > --- a/fs/cifs/cifspdu.h
>> > +++ b/fs/cifs/cifspdu.h
>> > @@ -26,13 +26,9 @@
>> >  #include <asm/unaligned.h>
>> >  #include "smbfsctl.h"
>> >
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  #define LANMAN_PROT 0
>> >  #define LANMAN2_PROT 1
>> >  #define CIFS_PROT   2
>> > -#else
>> > -#define CIFS_PROT   0
>> > -#endif
>> >  #define POSIX_PROT  (CIFS_PROT+1)
>> >  #define BAD_PROT 0xFFFF
>> >
>> > diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
>> > index 6f4e243..a81549d 100644
>> > --- a/fs/cifs/cifsproto.h
>> > +++ b/fs/cifs/cifsproto.h
>> > @@ -403,10 +403,8 @@ extern int cifs_crypto_shash_allocate(struct TCP_Server_Info *);
>> >  extern void cifs_crypto_shash_release(struct TCP_Server_Info *);
>> >  extern int calc_seckey(struct cifs_ses *);
>> >
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  extern int calc_lanman_hash(const char *password, const char *cryptkey,
>> >                                bool encrypt, char *lnm_session_key);
>> > -#endif /* CIFS_WEAK_PW_HASH */
>> >  #ifdef CONFIG_CIFS_DNOTIFY_EXPERIMENTAL /* unused temporarily */
>> >  extern int CIFSSMBNotify(const int xid, struct cifs_tcon *tcon,
>> >                        const int notify_subdirs, const __u16 netfid,
>> > diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
>> > index 8b7794c..cf12c0c 100644
>> > --- a/fs/cifs/cifssmb.c
>> > +++ b/fs/cifs/cifssmb.c
>> > @@ -49,10 +49,8 @@ static struct {
>> >        int index;
>> >        char *name;
>> >  } protocols[] = {
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >        {LANMAN_PROT, "\2LM1.2X002"},
>> >        {LANMAN2_PROT, "\2LANMAN2.1"},
>> > -#endif /* weak password hashing for legacy clients */
>> >        {CIFS_PROT, "\2NT LM 0.12"},
>> >        {POSIX_PROT, "\2POSIX 2"},
>> >        {BAD_PROT, "\2"}
>> > @@ -62,10 +60,8 @@ static struct {
>> >        int index;
>> >        char *name;
>> >  } protocols[] = {
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >        {LANMAN_PROT, "\2LM1.2X002"},
>> >        {LANMAN2_PROT, "\2LANMAN2.1"},
>> > -#endif /* weak password hashing for legacy clients */
>> >        {CIFS_PROT, "\2NT LM 0.12"},
>> >        {BAD_PROT, "\2"}
>> >  };
>> > @@ -73,17 +69,9 @@ static struct {
>> >
>> >  /* define the number of elements in the cifs dialect array */
>> >  #ifdef CONFIG_CIFS_POSIX
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  #define CIFS_NUM_PROT 4
>> > -#else
>> > -#define CIFS_NUM_PROT 2
>> > -#endif /* CIFS_WEAK_PW_HASH */
>> >  #else /* not posix */
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  #define CIFS_NUM_PROT 3
>> > -#else
>> > -#define CIFS_NUM_PROT 1
>> > -#endif /* CONFIG_CIFS_WEAK_PW_HASH */
>> >  #endif /* CIFS_POSIX */
>> >
>> >  /* Forward declarations */
>> > @@ -441,7 +429,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>> >                could not negotiate a common dialect */
>> >                rc = -EOPNOTSUPP;
>> >                goto neg_err_exit;
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >        } else if ((pSMBr->hdr.WordCount == 13)
>> >                        && ((server->dialect == LANMAN_PROT)
>> >                                || (server->dialect == LANMAN2_PROT))) {
>> > @@ -519,13 +506,6 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>> >                /* we will not end up setting signing flags - as no signing
>> >                was in LANMAN and server did not return the flags on */
>> >                goto signing_check;
>> > -#else /* weak security disabled */
>> > -       } else if (pSMBr->hdr.WordCount == 13) {
>> > -               cERROR(1, "mount failed, cifs module not built "
>> > -                         "with CIFS_WEAK_PW_HASH support");
>> > -               rc = -EOPNOTSUPP;
>> > -#endif /* WEAK_PW_HASH */
>> > -               goto neg_err_exit;
>> >        } else if (pSMBr->hdr.WordCount != 17) {
>> >                /* unknown wct */
>> >                rc = -EOPNOTSUPP;
>> > @@ -537,9 +517,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>> >                cFYI(1, "share mode security");
>> >
>> >        if ((server->sec_mode & SECMODE_PW_ENCRYPT) == 0)
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >                if ((secFlags & CIFSSEC_MAY_PLNTXT) == 0)
>> > -#endif /* CIFS_WEAK_PW_HASH */
>> >                        cERROR(1, "Server requests plain text password"
>> >                                  " but client support disabled");
>> >
>> > @@ -627,9 +605,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses)
>> >        } else
>> >                server->capabilities &= ~CAP_EXTENDED_SECURITY;
>> >
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >  signing_check:
>> > -#endif
>> >        if ((secFlags & CIFSSEC_MAY_SIGN) == 0) {
>> >                /* MUST_SIGN already includes the MAY_SIGN FLAG
>> >                   so if this is zero it means that signing is disabled */
>> > diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
>> > index aa687c8..5aeca5e 100644
>> > --- a/fs/cifs/connect.c
>> > +++ b/fs/cifs/connect.c
>> > @@ -1153,10 +1153,8 @@ cifs_parse_mount_options(const char *mountdata, const char *devname,
>> >                        } else if (strnicmp(value, "nontlm", 6) == 0) {
>> >                                /* BB is there a better way to do this? */
>> >                                vol->secFlg |= CIFSSEC_MAY_NTLMV2;
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >                        } else if (strnicmp(value, "lanman", 6) == 0) {
>> >                                vol->secFlg |= CIFSSEC_MAY_LANMAN;
>> > -#endif
>> >                        } else if (strnicmp(value, "none", 4) == 0) {
>> >                                vol->nullauth = 1;
>> >                        } else {
>> > @@ -3611,7 +3609,6 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
>> >                   weaker LANMAN (which we do not send by default) is accepted
>> >                   by Samba (not sure whether other servers allow
>> >                   NTLMv2 password here) */
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >                if ((global_secflags & CIFSSEC_MAY_LANMAN) &&
>> >                    (ses->server->secType == LANMAN))
>> >                        calc_lanman_hash(tcon->password, ses->server->cryptkey,
>> > @@ -3619,9 +3616,8 @@ CIFSTCon(unsigned int xid, struct cifs_ses *ses,
>> >                                            SECMODE_PW_ENCRYPT ? true : false,
>> >                                         bcc_ptr);
>> >                else
>> > -#endif /* CIFS_WEAK_PW_HASH */
>> > -               rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
>> > -                                       bcc_ptr, nls_codepage);
>> > +                       rc = SMBNTencrypt(tcon->password, ses->server->cryptkey,
>> > +                                               bcc_ptr, nls_codepage);
>> >
>> >                bcc_ptr += CIFS_AUTH_RESP_SIZE;
>> >                if (ses->capabilities & CAP_UNICODE) {
>> > diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c
>> > index d85efad..1cce63b 100644
>> > --- a/fs/cifs/sess.c
>> > +++ b/fs/cifs/sess.c
>> > @@ -592,14 +592,6 @@ ssetup_ntlmssp_authenticate:
>> >                phase = NtLmAuthenticate; /* if ntlmssp, now final phase */
>> >
>> >        if (type == LANMAN) {
>> > -#ifndef CONFIG_CIFS_WEAK_PW_HASH
>> > -               /* LANMAN and plaintext are less secure and off by default.
>> > -               So we make this explicitly be turned on in kconfig (in the
>> > -               build) and turned on at runtime (changed from the default)
>> > -               in proc/fs/cifs or via mount parm.  Unfortunately this is
>> > -               needed for old Win (e.g. Win95), some obscure NAS and OS/2 */
>> > -               return -EOPNOTSUPP;
>> > -#endif
>> >                wct = 10; /* lanman 2 style sessionsetup */
>> >        } else if ((type == NTLM) || (type == NTLMv2)) {
>> >                /* For NTLMv2 failures eventually may need to retry NTLM */
>> > @@ -643,7 +635,6 @@ ssetup_ntlmssp_authenticate:
>> >        iov[1].iov_len = 0;
>> >
>> >        if (type == LANMAN) {
>> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
>> >                char lnm_session_key[CIFS_AUTH_RESP_SIZE];
>> >
>> >                pSMB->req.hdr.Flags2 &= ~SMBFLG2_UNICODE;
>> > @@ -674,7 +665,6 @@ ssetup_ntlmssp_authenticate:
>> >                cFYI(1, "Negotiating LANMAN setting up strings");
>> >                /* Unicode not allowed for LANMAN dialects */
>> >                ascii_ssetup_strings(&bcc_ptr, ses, nls_cp);
>> > -#endif
>> >        } else if (type == NTLM) {
>> >                pSMB->req_no_secext.Capabilities = cpu_to_le32(capabilities);
>> >                pSMB->req_no_secext.CaseInsensitivePasswordLength =
>> > --
>> > 1.7.7.5
>> >
>>
>>
>>
>
>
> --
> Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>



-- 
Thanks,

Steve

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Jeff Layton]

On Fri, 20 Jan 2012 14:45:48 -0600
Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> My general thinking on this is as follows:
> 
> If the kernel is distributed to all the workstations in an organization
> with this Kconfig option disabled, it makes it harder for individual users
> to make the mistake of enabling lanman (sec=lanman, or the Kconfig
> option) on a public network and thus send weak password hashes
> which could be discovered simply.   Most distros make the choice
> of enabling broader compatibility with old pre-1997 servers but
> it is a very small set of servers who would require lanman support,
> and a large number of potential attackers who could benefit if
> users enable lanman on a public network.  I suspect that there
> are environments where removing code (via Kconfig) is preferred
> to trusting all owners of all workstations running that organizations
> standard linux to never enable lanman at runtime.
> 
> But ... the opinion of security specialists on this would be welcome.
> 

There are myriads of ways for someone to screw themselves if you give
them root access. This one is pretty low on the list.

Code quality is important for security too, and the writhing mass of
ifdefs below is impossible to parse by eye and therefore debug. At some
point we have to weigh what requires a Kconfig option, and what doesn't.

Given that this can be administratively prohibited at runtime, I see
absolutely no need to keep this Kconfig option around. It just adds a
maintenance burden for no benefit.

While it's well and good to say you want to hear from "security
specialists", I'm not sure what that means and it sounds awfully open
ended. Is a lack of objection sufficient for you to take this patch or
do you need something else?

[...]

> >> > diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
> >> > index 76e7d8b..0e56b1e 100644
> >> > --- a/fs/cifs/cifsglob.h
> >> > +++ b/fs/cifs/cifsglob.h
> >> > @@ -853,13 +853,8 @@ static inline void free_dfs_info_array(struct dfs_info3_param *param,
> >> >  #define   CIFSSEC_MAY_NTLM     0x00002
> >> >  #define   CIFSSEC_MAY_NTLMV2   0x00004
> >> >  #define   CIFSSEC_MAY_KRB5     0x00008
> >> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >> >  #define   CIFSSEC_MAY_LANMAN   0x00010
> >> >  #define   CIFSSEC_MAY_PLNTXT   0x00020
> >> > -#else
> >> > -#define   CIFSSEC_MAY_LANMAN    0
> >> > -#define   CIFSSEC_MAY_PLNTXT    0
> >> > -#endif /* weak passwords */
> >> >  #define   CIFSSEC_MAY_SEAL     0x00040 /* not supported yet */
> >> >  #define   CIFSSEC_MAY_NTLMSSP  0x00080 /* raw ntlmssp with ntlmv2 */
> >> >
> >> > @@ -870,23 +865,13 @@ require use of the stronger protocol */
> >> >  #define   CIFSSEC_MUST_NTLM    0x02002
> >> >  #define   CIFSSEC_MUST_NTLMV2  0x04004
> >> >  #define   CIFSSEC_MUST_KRB5    0x08008
> >> > -#ifdef CONFIG_CIFS_WEAK_PW_HASH
> >> >  #define   CIFSSEC_MUST_LANMAN  0x10010
> >> >  #define   CIFSSEC_MUST_PLNTXT  0x20020
> >> >  #ifdef CONFIG_CIFS_UPCALL
> >> >  #define   CIFSSEC_MASK          0xBF0BF /* allows weak security but also krb5 */
> >> > -#else
> >> > +#else /* UPCALL */
> >> >  #define   CIFSSEC_MASK          0xB70B7 /* current flags supported if weak */
> >> >  #endif /* UPCALL */
> >> > -#else /* do not allow weak pw hash */
> >> > -#define   CIFSSEC_MUST_LANMAN  0
> >> > -#define   CIFSSEC_MUST_PLNTXT  0
> >> > -#ifdef CONFIG_CIFS_UPCALL
> >> > -#define   CIFSSEC_MASK          0x8F08F /* flags supported if no weak allowed */
> >> > -#else
> >> > -#define          CIFSSEC_MASK          0x87087 /* flags supported if no weak allowed */
> >> > -#endif /* UPCALL */
> >> > -#endif /* WEAK_PW_HASH */
> >> >  #define   CIFSSEC_MUST_SEAL    0x40040 /* not supported yet */
> >> >  #define   CIFSSEC_MUST_NTLMSSP 0x80080 /* raw ntlmssp with ntlmv2 */
> >> >
-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Steve French]

On Fri, Jan 20, 2012 at 3:01 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> On Fri, 20 Jan 2012 14:45:48 -0600
> Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
>> My general thinking on this is as follows:
>>
>> If the kernel is distributed to all the workstations in an organization
>> with this Kconfig option disabled, it makes it harder for individual users
>> to make the mistake of enabling lanman (sec=lanman, or the Kconfig
>> option) on a public network and thus send weak password hashes
>> which could be discovered simply.   Most distros make the choice
>> of enabling broader compatibility with old pre-1997 servers but
>> it is a very small set of servers who would require lanman support,
>> and a large number of potential attackers who could benefit if
>> users enable lanman on a public network.  I suspect that there
>> are environments where removing code (via Kconfig) is preferred
>> to trusting all owners of all workstations running that organizations
>> standard linux to never enable lanman at runtime.
>>
>> But ... the opinion of security specialists on this would be welcome.
>>
>
> There are myriads of ways for someone to screw themselves if you give
> them root access. This one is pretty low on the list.
>
> Code quality is important for security too, and the writhing mass of
> ifdefs below is impossible to parse by eye and therefore debug. At some
> point we have to weigh what requires a Kconfig option, and what doesn't.
>
> Given that this can be administratively prohibited at runtime, I see
> absolutely no need to keep this Kconfig option around. It just adds a
> maintenance burden for no benefit.
>
> While it's well and good to say you want to hear from "security
> specialists", I'm not sure what that means and it sounds awfully open
> ended. Is a lack of objection sufficient for you to take this patch or
> do you need something else?

We can give some time for users to respond or perhaps paraphrase the
general question and put on lkml and see if anyone comments.

I agree that removing the ifdef would simplify the code but would be
helpful to see if there are those who turn on the Kconfig for reasons
similar to what I described

-- 
Thanks,

Steve

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Andrew Bartlett]

On Fri, 2012-01-20 at 14:45 -0600, Steve French wrote:
> My general thinking on this is as follows:
> 
> If the kernel is distributed to all the workstations in an organization
> with this Kconfig option disabled, it makes it harder for individual users
> to make the mistake of enabling lanman (sec=lanman, or the Kconfig
> option) on a public network and thus send weak password hashes
> which could be discovered simply.   Most distros make the choice
> of enabling broader compatibility with old pre-1997 servers but
> it is a very small set of servers who would require lanman support,
> and a large number of potential attackers who could benefit if
> users enable lanman on a public network.  I suspect that there
> are environments where removing code (via Kconfig) is preferred
> to trusting all owners of all workstations running that organizations
> standard linux to never enable lanman at runtime.
> 
> But ... the opinion of security specialists on this would be welcome.

We have been though some of this with the kerberos libs, which now allow
(default?) to not even compile with weak crypto.  If the weak crypto is
not compiled in, it can therefore be asserted that the weak crypto
cannot be used, and this makes it easier to comply with security
audits/certification etc.

I don't want to make your code more complex than it needs to be, but LM
encryption really, really needs to go away.  If it is not a major
bother, I would like to make it easier for that to happen if possible.

Andrew Bartlett

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Jeff Layton]

On Sat, 21 Jan 2012 15:03:31 +1100
Andrew Bartlett <abartlet-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> wrote:

> On Fri, 2012-01-20 at 14:45 -0600, Steve French wrote:
> > My general thinking on this is as follows:
> > 
> > If the kernel is distributed to all the workstations in an organization
> > with this Kconfig option disabled, it makes it harder for individual users
> > to make the mistake of enabling lanman (sec=lanman, or the Kconfig
> > option) on a public network and thus send weak password hashes
> > which could be discovered simply.   Most distros make the choice
> > of enabling broader compatibility with old pre-1997 servers but
> > it is a very small set of servers who would require lanman support,
> > and a large number of potential attackers who could benefit if
> > users enable lanman on a public network.  I suspect that there
> > are environments where removing code (via Kconfig) is preferred
> > to trusting all owners of all workstations running that organizations
> > standard linux to never enable lanman at runtime.
> > 
> > But ... the opinion of security specialists on this would be welcome.
> 
> We have been though some of this with the kerberos libs, which now allow
> (default?) to not even compile with weak crypto.  If the weak crypto is
> not compiled in, it can therefore be asserted that the weak crypto
> cannot be used, and this makes it easier to comply with security
> audits/certification etc.
> 
> I don't want to make your code more complex than it needs to be, but LM
> encryption really, really needs to go away.  If it is not a major
> bother, I would like to make it easier for that to happen if possible.
> 

The only way for it to go away completely is for all servers that
support only that encryption to go away completely. Unfortunately,
that's a tall order -- there are still at least some in the field and
people need to get at data on them.

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[simo]

On Sat, 2012-01-21 at 07:37 -0500, Jeff Layton wrote: 
> On Sat, 21 Jan 2012 15:03:31 +1100
> Andrew Bartlett <abartlet-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> wrote:
> 
> > On Fri, 2012-01-20 at 14:45 -0600, Steve French wrote:
> > > My general thinking on this is as follows:
> > > 
> > > If the kernel is distributed to all the workstations in an organization
> > > with this Kconfig option disabled, it makes it harder for individual users
> > > to make the mistake of enabling lanman (sec=lanman, or the Kconfig
> > > option) on a public network and thus send weak password hashes
> > > which could be discovered simply.   Most distros make the choice
> > > of enabling broader compatibility with old pre-1997 servers but
> > > it is a very small set of servers who would require lanman support,
> > > and a large number of potential attackers who could benefit if
> > > users enable lanman on a public network.  I suspect that there
> > > are environments where removing code (via Kconfig) is preferred
> > > to trusting all owners of all workstations running that organizations
> > > standard linux to never enable lanman at runtime.
> > > 
> > > But ... the opinion of security specialists on this would be welcome.
> > 
> > We have been though some of this with the kerberos libs, which now allow
> > (default?) to not even compile with weak crypto.  If the weak crypto is
> > not compiled in, it can therefore be asserted that the weak crypto
> > cannot be used, and this makes it easier to comply with security
> > audits/certification etc.
> > 
> > I don't want to make your code more complex than it needs to be, but LM
> > encryption really, really needs to go away.  If it is not a major
> > bother, I would like to make it easier for that to happen if possible.
> > 
> 
> The only way for it to go away completely is for all servers that
> support only that encryption to go away completely. Unfortunately,
> that's a tall order -- there are still at least some in the field and
> people need to get at data on them.

Jeff, can you identify them ?

LM only servers means pre Win 95 machines, I'd be curious to know what
servers are there that really support only LM hashes and not NT hashes.

Simo.

-- 
Simo Sorce
Samba Team GPL Compliance Officer <simo-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org>
Principal Software Engineer at Red Hat, Inc. <simo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Jeff Layton]

On Sat, 21 Jan 2012 12:26:43 -0500
simo <idra-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> wrote:

> On Sat, 2012-01-21 at 07:37 -0500, Jeff Layton wrote: 
> > On Sat, 21 Jan 2012 15:03:31 +1100
> > Andrew Bartlett <abartlet-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> wrote:
> > 
> > > On Fri, 2012-01-20 at 14:45 -0600, Steve French wrote:
> > > > My general thinking on this is as follows:
> > > > 
> > > > If the kernel is distributed to all the workstations in an organization
> > > > with this Kconfig option disabled, it makes it harder for individual users
> > > > to make the mistake of enabling lanman (sec=lanman, or the Kconfig
> > > > option) on a public network and thus send weak password hashes
> > > > which could be discovered simply.   Most distros make the choice
> > > > of enabling broader compatibility with old pre-1997 servers but
> > > > it is a very small set of servers who would require lanman support,
> > > > and a large number of potential attackers who could benefit if
> > > > users enable lanman on a public network.  I suspect that there
> > > > are environments where removing code (via Kconfig) is preferred
> > > > to trusting all owners of all workstations running that organizations
> > > > standard linux to never enable lanman at runtime.
> > > > 
> > > > But ... the opinion of security specialists on this would be welcome.
> > > 
> > > We have been though some of this with the kerberos libs, which now allow
> > > (default?) to not even compile with weak crypto.  If the weak crypto is
> > > not compiled in, it can therefore be asserted that the weak crypto
> > > cannot be used, and this makes it easier to comply with security
> > > audits/certification etc.
> > > 
> > > I don't want to make your code more complex than it needs to be, but LM
> > > encryption really, really needs to go away.  If it is not a major
> > > bother, I would like to make it easier for that to happen if possible.
> > > 
> > 
> > The only way for it to go away completely is for all servers that
> > support only that encryption to go away completely. Unfortunately,
> > that's a tall order -- there are still at least some in the field and
> > people need to get at data on them.
> 
> Jeff, can you identify them ?
> 
> LM only servers means pre Win 95 machines, I'd be curious to know what
> servers are there that really support only LM hashes and not NT hashes.
> 

Sorry for the long delay in responding here. Yes, mostly pre-win95
machines. We occasionally get reports from people using OS/2 and I'm
pretty sure it's LM-only.

Steve, in any case...shall I consider this patch NAK'ed for now? I was
carrying it in my tree for 3.5, but it seems like there's resistance to
removing this option and I'm not particularly religious on the matter.

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Steve French]

On Fri, Apr 27, 2012 at 6:15 AM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> On Sat, 21 Jan 2012 12:26:43 -0500
> simo <idra-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> wrote:
>
>> On Sat, 2012-01-21 at 07:37 -0500, Jeff Layton wrote:
>> > On Sat, 21 Jan 2012 15:03:31 +1100
>> > Andrew Bartlett <abartlet-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> wrote:
>> >
>> > > On Fri, 2012-01-20 at 14:45 -0600, Steve French wrote:
>> > > > My general thinking on this is as follows:
>> > > >
>> > > > If the kernel is distributed to all the workstations in an organization
>> > > > with this Kconfig option disabled, it makes it harder for individual users
>> > > > to make the mistake of enabling lanman (sec=lanman, or the Kconfig
>> > > > option) on a public network and thus send weak password hashes
>> > > > which could be discovered simply.   Most distros make the choice
>> > > > of enabling broader compatibility with old pre-1997 servers but
>> > > > it is a very small set of servers who would require lanman support,
>> > > > and a large number of potential attackers who could benefit if
>> > > > users enable lanman on a public network.  I suspect that there
>> > > > are environments where removing code (via Kconfig) is preferred
>> > > > to trusting all owners of all workstations running that organizations
>> > > > standard linux to never enable lanman at runtime.
>> > > >
>> > > > But ... the opinion of security specialists on this would be welcome.
>> > >
>> > > We have been though some of this with the kerberos libs, which now allow
>> > > (default?) to not even compile with weak crypto.  If the weak crypto is
>> > > not compiled in, it can therefore be asserted that the weak crypto
>> > > cannot be used, and this makes it easier to comply with security
>> > > audits/certification etc.
>> > >
>> > > I don't want to make your code more complex than it needs to be, but LM
>> > > encryption really, really needs to go away.  If it is not a major
>> > > bother, I would like to make it easier for that to happen if possible.
>> > >
>> >
>> > The only way for it to go away completely is for all servers that
>> > support only that encryption to go away completely. Unfortunately,
>> > that's a tall order -- there are still at least some in the field and
>> > people need to get at data on them.
>>
>> Jeff, can you identify them ?
>>
>> LM only servers means pre Win 95 machines, I'd be curious to know what
>> servers are there that really support only LM hashes and not NT hashes.
>>
>
> Sorry for the long delay in responding here. Yes, mostly pre-win95
> machines. We occasionally get reports from people using OS/2 and I'm
> pretty sure it's LM-only.
>
> Steve, in any case...shall I consider this patch NAK'ed for now? I was
> carrying it in my tree for 3.5, but it seems like there's resistance to
> removing this option and I'm not particularly religious on the matter.

I don't think it is worth changing - I prefer to build with
WEAK_PS_HASH disabled.

-- 
Thanks,

Steve

Re: [PATCH] cifs: eliminate CONFIG_CIFS_WEAK_PW_HASH

[Jeff Layton]

On Fri, 27 Apr 2012 08:58:55 -0500
Steve French <smfrench@gmail.com> wrote:

> On Fri, Apr 27, 2012 at 6:15 AM, Jeff Layton <jlayton@redhat.com> wrote:
> > On Sat, 21 Jan 2012 12:26:43 -0500
> > simo <idra@samba.org> wrote:
> >
> >> On Sat, 2012-01-21 at 07:37 -0500, Jeff Layton wrote:
> >> > On Sat, 21 Jan 2012 15:03:31 +1100
> >> > Andrew Bartlett <abartlet@samba.org> wrote:
> >> >
> >> > > On Fri, 2012-01-20 at 14:45 -0600, Steve French wrote:
> >> > > > My general thinking on this is as follows:
> >> > > >
> >> > > > If the kernel is distributed to all the workstations in an organization
> >> > > > with this Kconfig option disabled, it makes it harder for individual users
> >> > > > to make the mistake of enabling lanman (sec=lanman, or the Kconfig
> >> > > > option) on a public network and thus send weak password hashes
> >> > > > which could be discovered simply.   Most distros make the choice
> >> > > > of enabling broader compatibility with old pre-1997 servers but
> >> > > > it is a very small set of servers who would require lanman support,
> >> > > > and a large number of potential attackers who could benefit if
> >> > > > users enable lanman on a public network.  I suspect that there
> >> > > > are environments where removing code (via Kconfig) is preferred
> >> > > > to trusting all owners of all workstations running that organizations
> >> > > > standard linux to never enable lanman at runtime.
> >> > > >
> >> > > > But ... the opinion of security specialists on this would be welcome.
> >> > >
> >> > > We have been though some of this with the kerberos libs, which now allow
> >> > > (default?) to not even compile with weak crypto.  If the weak crypto is
> >> > > not compiled in, it can therefore be asserted that the weak crypto
> >> > > cannot be used, and this makes it easier to comply with security
> >> > > audits/certification etc.
> >> > >
> >> > > I don't want to make your code more complex than it needs to be, but LM
> >> > > encryption really, really needs to go away.  If it is not a major
> >> > > bother, I would like to make it easier for that to happen if possible.
> >> > >
> >> >
> >> > The only way for it to go away completely is for all servers that
> >> > support only that encryption to go away completely. Unfortunately,
> >> > that's a tall order -- there are still at least some in the field and
> >> > people need to get at data on them.
> >>
> >> Jeff, can you identify them ?
> >>
> >> LM only servers means pre Win 95 machines, I'd be curious to know what
> >> servers are there that really support only LM hashes and not NT hashes.
> >>
> >
> > Sorry for the long delay in responding here. Yes, mostly pre-win95
> > machines. We occasionally get reports from people using OS/2 and I'm
> > pretty sure it's LM-only.
> >
> > Steve, in any case...shall I consider this patch NAK'ed for now? I was
> > carrying it in my tree for 3.5, but it seems like there's resistance to
> > removing this option and I'm not particularly religious on the matter.
> 
> I don't think it is worth changing - I prefer to build with
> WEAK_PS_HASH disabled.
> 

Alrighty then, dropped...

-- 
Jeff Layton <jlayton@redhat.com>