xfrm/crypto: unaligned access fixes
Subsystems (5)
cover xfrm/crypto: unaligned access fixes
A two-part patchset that fixes some "unaligned access" warnings that showed up my sparc test machines with ipsec set up.
Sowmini Varadhan (2):
crypto/x509: Fix unaligned access in x509_get_sig_params() Fix unaligned access in xfrm_notify_sa() for DELSA
crypto/asymmetric_keys/x509_public_key.c | 5 +++-- net/xfrm/xfrm_user.c | 2 +- 2 files changed, 4 insertions(+), 3 deletions(-)
PATCH 1/2 crypto/x509: Fix unaligned access in x509_get_sig_params()
x509_get_sig_params() has the same code pattern as the one in pkcs7_verify() that is fixed by commit 62f57d05e287 ("crypto: pkcs7 - Fix unaligned access in pkcs7_verify()") so apply a similar fix here: make sure that desc is pointing at an algined value past the digest_size, and take alignment values into consideration when doing kzalloc()
--- crypto/asymmetric_keys/x509_public_key.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-)
crypto/asymmetric_keys/x509_public_key.c +3 -2
@@ -194,14 +194,15 @@ int x509_get_sig_params(struct x509_certificate *cert)* digest storage space. */ ret = -ENOMEM ; - digest =kzalloc (digest_size + desc_size,GFP_KERNEL ); + digest = kzalloc(ALIGN(digest_size, __alignof__(*desc)) + desc_size, +GFP_KERNEL );if (!digest) goto error; cert->sig.digest = digest; cert->sig.digest_size = digest_size; - desc = digest + digest_size; + desc = PTR_ALIGN(digest + digest_size, __alignof__(*desc)); desc->tfm = tfm; desc->flags =CRYPTO_TFM_REQ_MAY_SLEEP ;
Signature · 1 line
1.7.1
Re: [PATCH 1/2] crypto/x509: Fix unaligned access in x509_get_sig_params()
Quoted 7 lines · Sowmini Varadhan
x509_get_sig_params()has the same code pattern as the one inpkcs7_verify()that is fixed by commit62f57d05e287("crypto: pkcs7 - Fix unaligned access inpkcs7_verify()") so apply a similar fix here: make sure that desc is pointing at an algined value past thedigest_size, and take alignment values into consideration when doingkzalloc()Signed-off-by Sowmini Varadhan <[email protected]>
Herbert Xu
maintainerRe: [PATCH 1/2] crypto/x509: Fix unaligned access in x509_get_sig_params()
In reply to Sowmini VaradhanQuoted 7 lines · Sowmini Varadhan
x509_get_sig_params()has the same code pattern as the one inpkcs7_verify()that is fixed by commit62f57d05e287("crypto: pkcs7 - Fix unaligned access inpkcs7_verify()") so apply a similar fix here: make sure that desc is pointing at an algined value past thedigest_size, and take alignment values into consideration when doingkzalloc()Signed-off-by Sowmini Varadhan <[email protected]>
Patch applied. Thanks.
Signature · 3 lines
Email: Herbert Xu <[email protected]> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
PATCH 2/2 xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
In reply to Sowmini VaradhanOn sparc, deleting established SAs (e.g., by restarting ipsec at the peer) results in unaligned access messages via xfrm_del_sa -> km_state_notify -> xfrm_send_state_notify(). Use an aligned pointer to xfrm_usersa_info for this case.
--- net/xfrm/xfrm_user.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
net/xfrm/xfrm_user.c +1 -1
@@ -2659,7 +2659,7 @@ static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c)if (attr ==NULL ) goto out_free_skb; - p =nla_data (attr); + p = PTR_ALIGN(nla_data(attr), __alignof__(*p)); } err =copy_to_user_state_extra (x, p, skb);if (err)
Signature · 1 line
1.7.1
Steffen Klassert
maintainerRe: [PATCH net-next 2/2] xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
Quoted 15 lines · Sowmini Varadhan
On sparc, deleting established SAs (e.g., by restarting ipsec at the peer) results in unaligned access messages via
xfrm_del_sa->km_state_notify->xfrm_send_state_notify(). Use an aligned pointer toxfrm_usersa_infofor this case.Signed-off-by Sowmini Varadhan <[email protected]>--- net/xfrm/xfrm_user.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)net/xfrm/xfrm_user.c +0
-0
net/xfrm/xfrm_user.c static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c) · line 2662
Quoted 5 lines
net/xfrm/xfrm_user.c +1
-1@@ -2659,7 +2659,7 @@ static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c)if (attr ==NULL ) goto out_free_skb; - p =nla_data (attr); + p = PTR_ALIGN(nla_data(attr), __alignof__(*p));
Hm, this breaks userspace notifications on 64-bit systems. Userspace expects this to be aligned to 4, with your patch it is aligned to 8 on 64-bit.
Without your patch I get the correct notification when deleting a SA:
ip x m
Deleted src 172.16.0.2 dst 172.16.0.1
proto esp spi 0x00000002 reqid 2 mode tunnel
replay-window 32
auth-trunc hmac(sha1) 0x31323334353637383930 96
enc cbc(aes) 0x31323334353637383930313233343536
sel src 10.0.0.0/24 dst 192.168.0.0/24With your patch I get for the same SA:
ip x m
Deleted src 50.0.0.0 dst 0.0.0.0
proto 0 reqid 0 mode transport
replay-window 0 flag decap-dscp
auth-trunc hmac(sha1) 0x31323334353637383930 96
enc cbc(aes) 0x31323334353637383930313233343536
sel src 0.0.0.0/0 dst 0.234.255.255/0 proto igmpRe: [PATCH net-next 2/2] xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
Quoted 6 lines · Steffen Klassert
Quoted 6 lines
net/xfrm/xfrm_user.c +0
-0@@ -2659,7 +2659,7 @@ static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c)if (attr ==NULL ) goto out_free_skb;
Quoted 9 lines
net/xfrm/xfrm_user.c static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c) · line 2662
Quoted 2 lines
net/xfrm/xfrm_user.c +1
-1- p =nla_data (attr); + p = PTR_ALIGN(nla_data(attr), __alignof__(*p));Hm, this breaks userspace notifications on 64-bit systems. Userspace expects this to be aligned to 4, with your patch it is aligned to 8 on 64-bit.
Without your patch I get the correct notification when deleting a SA:
But __alignof__(*p) is 8 on sparc, and without the patch I get all types of unaligned access. So what do you suggest as the fix?
(and openswan/pluto dont flag any errors with the patch, which is a separate bug).
--Sowmini
Re: [PATCH net-next 2/2] xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
Quoted 2 lines · Sowmini Varadhan
But
__alignof__(*p)is 8 on sparc, and without the patch I get all types of unaligned access. So what do you suggest as the fix?
Even though the alignment is, in fact, 8 (and that comes from struct xfrm_lifetime_cfg), if uspace is firmly attached to the 4 byte alignment, I think we can retain that behavior and still avoid unaligned access in the kernel with the following (admittedly ugly hack). Can you please take a look? I tested it with 'ip x m' and a transport mode tunnel on my sparc.
net/xfrm/xfrm_user.c +10 -5
@@ -2620,7 +2620,7 @@ static inline size_t xfrm_sa_len(struct xfrm_state *x)static int xfrm_notify_sa (struct xfrm_state *x,const struct km_event *c) {struct net *net =xs_net (x); -struct xfrm_usersa_info *p; +struct xfrm_usersa_info *p, tmp;struct xfrm_usersa_id *id;struct nlmsghdr *nlh;struct sk_buff *skb;
@@ -2659,11 +2659,16 @@ static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c)if (attr ==NULL ) goto out_free_skb; - p = PTR_ALIGN(nla_data(attr), __alignof__(*p)); + p =nla_data (attr); + err =copy_to_user_state_extra (x, &tmp, skb); +if (err) + goto out_free_skb; +memcpy ((u8 *)p, &tmp,sizeof (tmp)); + }else { + err =copy_to_user_state_extra (x, p, skb); +if (err) + goto out_free_skb; } - err =copy_to_user_state_extra (x, p, skb); -if (err) - goto out_free_skb;nlmsg_end (skb, nlh);
David Miller
maintainerRe: [PATCH net-next 2/2] xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
From: Sowmini Varadhan <[email protected]>
Date: Wed, 21 Oct 2015 08:36:28 -0400
Quoted 1 line
+
memcpy((u8 *)p, &tmp, sizeof(tmp));
memcpy() _never_ works for avoiding unaligned accessed.
I repeat, no matter what you do, no matter what kinds of casts or fancy typing you use, memcpy() _never_ works for this purpose.
The compiler knows that the pointer you are using is supposed to be at least 8 byte aligned, it can look through the cast and that's completely legitimate for it to do.
So it can legitimately inline emit loads and stores to implement the memcpy() and those will still get the unaligned accesses.
There is one and only one portable way to access unaligned data, and that is with the get_unaligned() and put_unaligned() helpers.
Userland must do something similar to deal with this situation as well.
Re: [PATCH net-next 2/2] xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
Quoted 4 lines · David Miller
memcpy()_never_ works for avoiding unaligned accessed.I repeat, no matter what you do, no matter what kinds of casts or fancy typing you use,
memcpy()_never_ works for this purpose.
:
Quoted 2 lines
There is one and only one portable way to access unaligned data, and that is with the
get_unaligned()andput_unaligned()helpers.
ok. I'll fix it up to use the *_unaligned functions and resend this out later today.
--Sowmini
David Miller
maintainerRe: [PATCH net-next 2/2] xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
In reply to Sowmini VaradhanFrom: Sowmini Varadhan <[email protected]>
Date: Wed, 21 Oct 2015 06:54:42 -0400
Quoted 14 lines · Steffen Klassert
Quoted 13 lines
net/xfrm/xfrm_user.c static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c) · line 2662
Quoted 8 lines
net/xfrm/xfrm_user.c +1
-1@@ -2659,7 +2659,7 @@ static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c)if (attr ==NULL ) goto out_free_skb; - p =nla_data (attr); + p = PTR_ALIGN(nla_data(attr), __alignof__(*p));Hm, this breaks userspace notifications on 64-bit systems. Userspace expects this to be aligned to 4, with your patch it is aligned to 8 on 64-bit.
Quoted 5 lines
Quoted 2 lines
Without your patch I get the correct notification when deleting a SA:
But
__alignof__(*p)is 8 on sparc, and without the patch I get all types of unaligned access. So what do you suggest as the fix?
The accesses have to be done using something like get_unaligned() and put_unaligned().
Sorry, but the protocol is set in stone and this is unfortunately how it is.
David Miller
maintainerRe: [PATCH net-next 2/2] xfrm: Fix unaligned access in xfrm_notify_sa() for DELSA
In reply to Steffen KlassertFrom: Steffen Klassert <[email protected]>
Date: Wed, 21 Oct 2015 08:57:04 +0200
Quoted 19 lines · Sowmini Varadhan
Quoted 18 lines
On sparc, deleting established SAs (e.g., by restarting ipsec at the peer) results in unaligned access messages via
xfrm_del_sa->km_state_notify->xfrm_send_state_notify(). Use an aligned pointer toxfrm_usersa_infofor this case.Signed-off-by Sowmini Varadhan <[email protected]>--- net/xfrm/xfrm_user.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)net/xfrm/xfrm_user.c +0
-0@@ -2659,7 +2659,7 @@ static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c)if (attr ==NULL ) goto out_free_skb;
Quoted 6 lines
net/xfrm/xfrm_user.c static int xfrm_notify_sa(struct xfrm_state *x, const struct km_event *c) · line 2662
Quoted 2 lines
net/xfrm/xfrm_user.c +1
-1- p =nla_data (attr); + p = PTR_ALIGN(nla_data(attr), __alignof__(*p));Hm, this breaks userspace notifications on 64-bit systems. Userspace expects this to be aligned to 4, with your patch it is aligned to 8 on 64-bit.
That's correct, netlink attributes are fundamentally only 4 byte aligned and this cannot be changed. nla_data() is exactly where the attribute must be placed, aligned or not.
The only workaround is, when designing netlink attributes. Various netlink libraries have workarounds for accessing, for example, 64-bit stats which are going to be unaligned in netlink messages.