2015-12-18 00:03:07

by David Howells

[permalink] [raw]
Subject: [PATCH] X.509: Fix determination of self-signedness

Fix determination of whether an X.509 certificate is self-signed or not.

It is currently assumed that a cert is self-signed if has no
authorityKeyIdentifier or the authorityKeyIdentifier matches the
subjectKeyIdentifier. However, it is possible to encounter a certificate
that has neither AKID not SKID but is not self-signed.

This symptoms of this show up as an attempt to load a certificate failing
with -ERANGE or -EBADMSG, produced from the RSA module when the result of
calculating "m = s^e mod n" is checked.

To fix this, don't check to see if a certificate is self-signed if the
Issuer and Subject names differ.

Signed-off-by: David Howells <[email protected]>
cc: David Woodhouse <[email protected]>
---

crypto/asymmetric_keys/x509_public_key.c | 11 ++++++++---
1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/crypto/asymmetric_keys/x509_public_key.c b/crypto/asymmetric_keys/x509_public_key.c
index 2a44b3752471..6236e7996f19 100644
--- a/crypto/asymmetric_keys/x509_public_key.c
+++ b/crypto/asymmetric_keys/x509_public_key.c
@@ -313,9 +313,14 @@ static int x509_key_preparse(struct key_preparsed_payload *prep)
cert->pub->id_type = PKEY_ID_X509;

/* Check the signature on the key if it appears to be self-signed */
- if ((!cert->akid_skid && !cert->akid_id) ||
- asymmetric_key_id_same(cert->skid, cert->akid_skid) ||
- asymmetric_key_id_same(cert->id, cert->akid_id)) {
+ if ((!cert->akid_skid && !cert->akid_id)) {
+ if (cert->raw_issuer_size == cert->raw_subject_size &&
+ memcmp(cert->raw_issuer, cert->raw_subject,
+ cert->raw_subject_size) == 0)
+ goto self_signed;
+ } else if (asymmetric_key_id_same(cert->skid, cert->akid_skid) ||
+ asymmetric_key_id_same(cert->id, cert->akid_id)) {
+self_signed:
ret = x509_check_signature(cert->pub, cert); /* self-signed */
if (ret < 0)
goto error_free_cert;


2015-12-18 14:42:56

by Josh Boyer

[permalink] [raw]
Subject: Re: [PATCH] X.509: Fix determination of self-signedness

On Thu, Dec 17, 2015 at 7:03 PM, David Howells <[email protected]> wrote:
> Fix determination of whether an X.509 certificate is self-signed or not.
>
> It is currently assumed that a cert is self-signed if has no
> authorityKeyIdentifier or the authorityKeyIdentifier matches the
> subjectKeyIdentifier. However, it is possible to encounter a certificate
> that has neither AKID not SKID but is not self-signed.
>
> This symptoms of this show up as an attempt to load a certificate failing
> with -ERANGE or -EBADMSG, produced from the RSA module when the result of
> calculating "m = s^e mod n" is checked.
>
> To fix this, don't check to see if a certificate is self-signed if the
> Issuer and Subject names differ.
>
> Signed-off-by: David Howells <[email protected]>
> cc: David Woodhouse <[email protected]>

Should this also be Cc'd to stable?

josh

> ---
>
> crypto/asymmetric_keys/x509_public_key.c | 11 ++++++++---
> 1 file changed, 8 insertions(+), 3 deletions(-)
>
> diff --git a/crypto/asymmetric_keys/x509_public_key.c b/crypto/asymmetric_keys/x509_public_key.c
> index 2a44b3752471..6236e7996f19 100644
> --- a/crypto/asymmetric_keys/x509_public_key.c
> +++ b/crypto/asymmetric_keys/x509_public_key.c
> @@ -313,9 +313,14 @@ static int x509_key_preparse(struct key_preparsed_payload *prep)
> cert->pub->id_type = PKEY_ID_X509;
>
> /* Check the signature on the key if it appears to be self-signed */
> - if ((!cert->akid_skid && !cert->akid_id) ||
> - asymmetric_key_id_same(cert->skid, cert->akid_skid) ||
> - asymmetric_key_id_same(cert->id, cert->akid_id)) {
> + if ((!cert->akid_skid && !cert->akid_id)) {
> + if (cert->raw_issuer_size == cert->raw_subject_size &&
> + memcmp(cert->raw_issuer, cert->raw_subject,
> + cert->raw_subject_size) == 0)
> + goto self_signed;
> + } else if (asymmetric_key_id_same(cert->skid, cert->akid_skid) ||
> + asymmetric_key_id_same(cert->id, cert->akid_id)) {
> +self_signed:
> ret = x509_check_signature(cert->pub, cert); /* self-signed */
> if (ret < 0)
> goto error_free_cert;
>

2015-12-18 14:46:50

by David Howells

[permalink] [raw]
Subject: Re: [PATCH] X.509: Fix determination of self-signedness

Josh Boyer <[email protected]> wrote:

> Should this also be Cc'd to stable?

Argh. Probably.

David