This series adds support for ECDSA-signed kernel modules.
The first patch in this series attempts to address the issue where a
developer created an ECDSA key for signing modules and then falls back
to compiling an older version of the kernel that does not support
ECDSA keys. In this case this patch would delete that ECDSA key if it is
in certs/signing_key.pem and trigger the creation of an RSA key. However,
for this to work this patch would have to be applied to previous versions
of the kernel but would also only work for the developer if he/she used a
stable version of the kernel to which this patch was applied. So whether
this patch actually achieves the wanted effect is not always guaranteed.
The 2nd patch adds the support for the ECSDA-signed kernel modules.
This series depends on Nayna's series https://lkml.org/lkml/2021/3/30/571
and my ECDSA support series currently queued here:
https://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git/log/?h=ecc
Stefan
Stefan Berger (2):
certs: Trigger recreation of module signing key if it's not an RSA key
certs: Add support for using elliptic curve keys for signing modules
certs/Kconfig | 25 +++++++++++++++++++++++++
certs/Makefile | 14 ++++++++++++++
crypto/asymmetric_keys/pkcs7_parser.c | 4 ++++
3 files changed, 43 insertions(+)
--
2.29.2
Make sure that the kernel module signing key is an RSA key and
remove it otherwise so that it gets recreated.
Prevent module loading failures if a developer chose an ECDSA key for
module signing with a 5.12 kernel and then falls back to compiling an
older kernel. However, this will also only work if falling back to
kernels that have actually been patched with this kernel, such as the
stable trees.
Fixes: cfc411e7fff3 ("Move certificate handling to its own directory")
Signed-off-by: Stefan Berger <[email protected]>
---
certs/Makefile | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/certs/Makefile b/certs/Makefile
index e3185c57fbd8..f64bc89ccbf1 100644
--- a/certs/Makefile
+++ b/certs/Makefile
@@ -59,6 +59,11 @@ silent_redirect_openssl = 2>/dev/null
# external private key, because 'make randconfig' might enable such a
# boolean option and we unfortunately can't make it depend on !RANDCONFIG.
ifeq ($(CONFIG_MODULE_SIG_KEY),"certs/signing_key.pem")
+
+X509TEXT=$(shell openssl x509 -in $(CONFIG_MODULE_SIG_KEY) -text)
+
+$(if $(findstring rsaEncryption,$(X509TEXT)),,$(shell rm -f $(CONFIG_MODULE_SIG_KEY)))
+
$(obj)/signing_key.pem: $(obj)/x509.genkey
@$(kecho) "###"
@$(kecho) "### Now generating an X.509 key pair to be used for signing modules."
--
2.29.2
Add support for using elliptic curve keys for signing modules. It uses
a NIST P384 (secp384r1) key if the user chooses an elliptic curve key
and will have ECDSA support built into the kernel.
Note: A developer choosing an ECDSA key for signing modules should
delete the signing key (rm certs/signing_key.*) when falling back to
building an older version of a kernel that only supports RSA keys
since otherwise ECDSA-signed modules will not be usable when that
older kernel runs and the ECDSA key was still used for signing modules.
Signed-off-by: Stefan Berger <[email protected]>
---
certs/Kconfig | 25 +++++++++++++++++++++++++
certs/Makefile | 9 +++++++++
crypto/asymmetric_keys/pkcs7_parser.c | 4 ++++
3 files changed, 38 insertions(+)
diff --git a/certs/Kconfig b/certs/Kconfig
index 48675ad319db..6f8337874ae0 100644
--- a/certs/Kconfig
+++ b/certs/Kconfig
@@ -15,6 +15,31 @@ config MODULE_SIG_KEY
then the kernel will automatically generate the private key and
certificate as described in Documentation/admin-guide/module-signing.rst
+choice
+ prompt "Type of module signing key to be generated"
+ default MODULE_SIG_KEY_TYPE_RSA
+ help
+ The type of module signing key type to generate. This option
+ does not apply if a #PKCS11 URI is used.
+
+config MODULE_SIG_KEY_TYPE_RSA
+ bool "RSA"
+ depends on MODULE_SIG || IMA_APPRAISE_MODSIG
+ help
+ Use an RSA key for module signing.
+
+config MODULE_SIG_KEY_TYPE_ECDSA
+ bool "ECDSA"
+ select CRYPTO_ECDSA
+ depends on MODULE_SIG || IMA_APPRAISE_MODSIG
+ help
+ Use an elliptic curve key (NIST P384) for module signing.
+
+ Note: Remove all ECDSA signing keys, e.g. certs/signing_key.pem,
+ when falling back to building Linux 5.11 and older kernels.
+
+endchoice
+
config SYSTEM_TRUSTED_KEYRING
bool "Provide system-wide ring of trusted keys"
depends on KEYS
diff --git a/certs/Makefile b/certs/Makefile
index f64bc89ccbf1..c07736ec5a9a 100644
--- a/certs/Makefile
+++ b/certs/Makefile
@@ -62,7 +62,15 @@ ifeq ($(CONFIG_MODULE_SIG_KEY),"certs/signing_key.pem")
X509TEXT=$(shell openssl x509 -in $(CONFIG_MODULE_SIG_KEY) -text)
+# Support user changing key type
+ifdef CONFIG_MODULE_SIG_KEY_TYPE_ECDSA
+keytype_openssl = -newkey ec -pkeyopt ec_paramgen_curve:secp384r1
+$(if $(findstring ecdsa-with-,$(X509TEXT)),,$(shell rm -f $(CONFIG_MODULE_SIG_KEY)))
+endif
+
+ifdef CONFIG_MODULE_SIG_KEY_TYPE_RSA
$(if $(findstring rsaEncryption,$(X509TEXT)),,$(shell rm -f $(CONFIG_MODULE_SIG_KEY)))
+endif
$(obj)/signing_key.pem: $(obj)/x509.genkey
@$(kecho) "###"
@@ -77,6 +85,7 @@ $(obj)/signing_key.pem: $(obj)/x509.genkey
-batch -x509 -config $(obj)/x509.genkey \
-outform PEM -out $(obj)/signing_key.pem \
-keyout $(obj)/signing_key.pem \
+ $(keytype_openssl) \
$($(quiet)redirect_openssl)
@$(kecho) "###"
@$(kecho) "### Key pair generated."
diff --git a/crypto/asymmetric_keys/pkcs7_parser.c b/crypto/asymmetric_keys/pkcs7_parser.c
index 967329e0a07b..2546ec6a0505 100644
--- a/crypto/asymmetric_keys/pkcs7_parser.c
+++ b/crypto/asymmetric_keys/pkcs7_parser.c
@@ -269,6 +269,10 @@ int pkcs7_sig_note_pkey_algo(void *context, size_t hdrlen,
ctx->sinfo->sig->pkey_algo = "rsa";
ctx->sinfo->sig->encoding = "pkcs1";
break;
+ case OID_id_ecdsa_with_sha256:
+ ctx->sinfo->sig->pkey_algo = "ecdsa";
+ ctx->sinfo->sig->encoding = "x962";
+ break;
default:
printk("Unsupported pkey algo: %u\n", ctx->last_oid);
return -ENOPKG;
--
2.29.2
On Tue, Apr 06, 2021 at 02:53:38PM -0400, Stefan Berger wrote:
> This series adds support for ECDSA-signed kernel modules.
>
> The first patch in this series attempts to address the issue where a
> developer created an ECDSA key for signing modules and then falls back
> to compiling an older version of the kernel that does not support
> ECDSA keys. In this case this patch would delete that ECDSA key if it is
> in certs/signing_key.pem and trigger the creation of an RSA key. However,
> for this to work this patch would have to be applied to previous versions
> of the kernel but would also only work for the developer if he/she used a
> stable version of the kernel to which this patch was applied. So whether
> this patch actually achieves the wanted effect is not always guaranteed.
Just wondering why the key needs to be removed in the fallback.
/Jarkko
On Wed, 2021-04-07 at 18:53 +0300, Jarkko Sakkinen wrote:
> On Tue, Apr 06, 2021 at 02:53:38PM -0400, Stefan Berger wrote:
> > This series adds support for ECDSA-signed kernel modules.
> >
> > The first patch in this series attempts to address the issue where a
> > developer created an ECDSA key for signing modules and then falls back
> > to compiling an older version of the kernel that does not support
> > ECDSA keys. In this case this patch would delete that ECDSA key if it is
> > in certs/signing_key.pem and trigger the creation of an RSA key. However,
> > for this to work this patch would have to be applied to previous versions
> > of the kernel but would also only work for the developer if he/she used a
> > stable version of the kernel to which this patch was applied. So whether
> > this patch actually achieves the wanted effect is not always guaranteed.
>
> Just wondering why the key needs to be removed in the fallback.
The main concern is with bisecting the kernel. Either elliptic curve
support or the first patch needs to be backported. This patch will
cause the kernel module signing key to be regenerated.
Mimi
On 4/7/21 12:10 PM, Mimi Zohar wrote:
> On Wed, 2021-04-07 at 18:53 +0300, Jarkko Sakkinen wrote:
>> On Tue, Apr 06, 2021 at 02:53:38PM -0400, Stefan Berger wrote:
>>> This series adds support for ECDSA-signed kernel modules.
>>>
>>> The first patch in this series attempts to address the issue where a
>>> developer created an ECDSA key for signing modules and then falls back
>>> to compiling an older version of the kernel that does not support
>>> ECDSA keys. In this case this patch would delete that ECDSA key if it is
>>> in certs/signing_key.pem and trigger the creation of an RSA key. However,
>>> for this to work this patch would have to be applied to previous versions
>>> of the kernel but would also only work for the developer if he/she used a
>>> stable version of the kernel to which this patch was applied. So whether
>>> this patch actually achieves the wanted effect is not always guaranteed.
>> Just wondering why the key needs to be removed in the fallback.
Because if you signed an older kernel's modules with the ECDSA key it
won't be able to load the modules...
> The main concern is with bisecting the kernel. Either elliptic curve
> support or the first patch needs to be backported. This patch will
> cause the kernel module signing key to be regenerated.
This assumes of course that one will bisect via the stable kernels where
this 1st patch has been applied. Not sure whether that's what people
will do.
?? Stefan
>
> Mimi
>
On Wed, Apr 07, 2021 at 01:53:24PM -0400, Stefan Berger wrote:
>
> On 4/7/21 12:10 PM, Mimi Zohar wrote:
> > On Wed, 2021-04-07 at 18:53 +0300, Jarkko Sakkinen wrote:
> > > On Tue, Apr 06, 2021 at 02:53:38PM -0400, Stefan Berger wrote:
> > > > This series adds support for ECDSA-signed kernel modules.
> > > >
> > > > The first patch in this series attempts to address the issue where a
> > > > developer created an ECDSA key for signing modules and then falls back
> > > > to compiling an older version of the kernel that does not support
> > > > ECDSA keys. In this case this patch would delete that ECDSA key if it is
> > > > in certs/signing_key.pem and trigger the creation of an RSA key. However,
> > > > for this to work this patch would have to be applied to previous versions
> > > > of the kernel but would also only work for the developer if he/she used a
> > > > stable version of the kernel to which this patch was applied. So whether
> > > > this patch actually achieves the wanted effect is not always guaranteed.
> > > Just wondering why the key needs to be removed in the fallback.
> Because if you signed an older kernel's modules with the ECDSA key it won't
> be able to load the modules...
> > The main concern is with bisecting the kernel. Either elliptic curve
> > support or the first patch needs to be backported. This patch will
> > cause the kernel module signing key to be regenerated.
>
>
> This assumes of course that one will bisect via the stable kernels where
> this 1st patch has been applied. Not sure whether that's what people will
> do.
In any case, sounds non-trivial issue enough ought to be documented in the
commit message.
/Jarkko