rsa-pkcs1pad uses a value returned from a RSA implementation max_size
callback as a size of an input buffer passed to the RSA implementation for
encrypt and sign operations.
CCP RSA implementation uses a hardware input buffer which size depends only
on the current RSA key length, so it should return this key length in
the max_size callback, too.
This also matches what the kernel software RSA implementation does.
Previously, the value returned from this callback was always the maximum
RSA key size the CCP hardware supports.
This resulted in this huge buffer being passed by rsa-pkcs1pad to CCP even
for smaller key sizes and then in a buffer overflow when ccp_run_rsa_cmd()
tried to copy this large input buffer into a RSA key length-sized hardware
input buffer.
Signed-off-by: Maciej S. Szmigiero <[email protected]>
Fixes: ceeec0afd684 ("crypto: ccp - Add support for RSA on the CCP")
Cc: [email protected]
---
drivers/crypto/ccp/ccp-crypto-rsa.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/crypto/ccp/ccp-crypto-rsa.c b/drivers/crypto/ccp/ccp-crypto-rsa.c
index e6db8672d89c..05850dfd7940 100644
--- a/drivers/crypto/ccp/ccp-crypto-rsa.c
+++ b/drivers/crypto/ccp/ccp-crypto-rsa.c
@@ -60,10 +60,9 @@ static int ccp_rsa_complete(struct crypto_async_request *async_req, int ret)
static unsigned int ccp_rsa_maxsize(struct crypto_akcipher *tfm)
{
- if (ccp_version() > CCP_VERSION(3, 0))
- return CCP5_RSA_MAXMOD;
- else
- return CCP_RSA_MAXMOD;
+ struct ccp_ctx *ctx = akcipher_tfm_ctx(tfm);
+
+ return ctx->u.rsa.n_len;
}
static int ccp_rsa_crypt(struct akcipher_request *req, bool encrypt)
On 02/24/2018 10:03 AM, Maciej S. Szmigiero wrote:
> rsa-pkcs1pad uses a value returned from a RSA implementation max_size
> callback as a size of an input buffer passed to the RSA implementation for
> encrypt and sign operations.
>
> CCP RSA implementation uses a hardware input buffer which size depends only
> on the current RSA key length, so it should return this key length in
> the max_size callback, too.
> This also matches what the kernel software RSA implementation does.
>
> Previously, the value returned from this callback was always the maximum
> RSA key size the CCP hardware supports.
> This resulted in this huge buffer being passed by rsa-pkcs1pad to CCP even
> for smaller key sizes and then in a buffer overflow when ccp_run_rsa_cmd()
> tried to copy this large input buffer into a RSA key length-sized hardware
> input buffer.
>
> Signed-off-by: Maciej S. Szmigiero <[email protected]>
Acked-by: Gary R Hook <[email protected]>
> Fixes: ceeec0afd684 ("crypto: ccp - Add support for RSA on the CCP")
> Cc: [email protected]
> ---
> drivers/crypto/ccp/ccp-crypto-rsa.c | 7 +++----
> 1 file changed, 3 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/crypto/ccp/ccp-crypto-rsa.c b/drivers/crypto/ccp/ccp-crypto-rsa.c
> index e6db8672d89c..05850dfd7940 100644
> --- a/drivers/crypto/ccp/ccp-crypto-rsa.c
> +++ b/drivers/crypto/ccp/ccp-crypto-rsa.c
> @@ -60,10 +60,9 @@ static int ccp_rsa_complete(struct crypto_async_request *async_req, int ret)
>
> static unsigned int ccp_rsa_maxsize(struct crypto_akcipher *tfm)
> {
> - if (ccp_version() > CCP_VERSION(3, 0))
> - return CCP5_RSA_MAXMOD;
> - else
> - return CCP_RSA_MAXMOD;
> + struct ccp_ctx *ctx = akcipher_tfm_ctx(tfm);
> +
> + return ctx->u.rsa.n_len;
> }
>
> static int ccp_rsa_crypt(struct akcipher_request *req, bool encrypt)
>
On Sat, Feb 24, 2018 at 05:03:21PM +0100, Maciej S. Szmigiero wrote:
> rsa-pkcs1pad uses a value returned from a RSA implementation max_size
> callback as a size of an input buffer passed to the RSA implementation for
> encrypt and sign operations.
>
> CCP RSA implementation uses a hardware input buffer which size depends only
> on the current RSA key length, so it should return this key length in
> the max_size callback, too.
> This also matches what the kernel software RSA implementation does.
>
> Previously, the value returned from this callback was always the maximum
> RSA key size the CCP hardware supports.
> This resulted in this huge buffer being passed by rsa-pkcs1pad to CCP even
> for smaller key sizes and then in a buffer overflow when ccp_run_rsa_cmd()
> tried to copy this large input buffer into a RSA key length-sized hardware
> input buffer.
>
> Signed-off-by: Maciej S. Szmigiero <[email protected]>
> Fixes: ceeec0afd684 ("crypto: ccp - Add support for RSA on the CCP")
> Cc: [email protected]
Patch applied. Thanks.
--
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
On 02.03.2018 17:44, Herbert Xu wrote:
> On Sat, Feb 24, 2018 at 05:03:21PM +0100, Maciej S. Szmigiero wrote:
>> rsa-pkcs1pad uses a value returned from a RSA implementation max_size
>> callback as a size of an input buffer passed to the RSA implementation for
>> encrypt and sign operations.
>>
>> CCP RSA implementation uses a hardware input buffer which size depends only
>> on the current RSA key length, so it should return this key length in
>> the max_size callback, too.
>> This also matches what the kernel software RSA implementation does.
>>
>> Previously, the value returned from this callback was always the maximum
>> RSA key size the CCP hardware supports.
>> This resulted in this huge buffer being passed by rsa-pkcs1pad to CCP even
>> for smaller key sizes and then in a buffer overflow when ccp_run_rsa_cmd()
>> tried to copy this large input buffer into a RSA key length-sized hardware
>> input buffer.
>>
>> Signed-off-by: Maciej S. Szmigiero <[email protected]>
>> Fixes: ceeec0afd684 ("crypto: ccp - Add support for RSA on the CCP")
>> Cc: [email protected]
>
> Patch applied. Thanks.
Thanks.
However, what about the first patch from this series?
Without it, while it no longer should cause a buffer overflow, in-kernel
X.509 certificate verification will still fail with CCP driver loaded
(since CCP RSA implementation has a higher priority than the software
RSA implementation).
Maciej
On 3/2/2018 5:15 PM, Maciej S. Szmigiero wrote:
> On 02.03.2018 17:44, Herbert Xu wrote:
>> On Sat, Feb 24, 2018 at 05:03:21PM +0100, Maciej S. Szmigiero wrote:
>>> rsa-pkcs1pad uses a value returned from a RSA implementation max_size
>>> callback as a size of an input buffer passed to the RSA implementation for
>>> encrypt and sign operations.
>>>
>>> CCP RSA implementation uses a hardware input buffer which size depends only
>>> on the current RSA key length, so it should return this key length in
>>> the max_size callback, too.
>>> This also matches what the kernel software RSA implementation does.
>>>
>>> Previously, the value returned from this callback was always the maximum
>>> RSA key size the CCP hardware supports.
>>> This resulted in this huge buffer being passed by rsa-pkcs1pad to CCP even
>>> for smaller key sizes and then in a buffer overflow when ccp_run_rsa_cmd()
>>> tried to copy this large input buffer into a RSA key length-sized hardware
>>> input buffer.
>>>
>>> Signed-off-by: Maciej S. Szmigiero <[email protected]>
>>> Fixes: ceeec0afd684 ("crypto: ccp - Add support for RSA on the CCP")
>>> Cc: [email protected]
>>
>> Patch applied. Thanks.
>
> Thanks.
>
> However, what about the first patch from this series?
> Without it, while it no longer should cause a buffer overflow, in-kernel
> X.509 certificate verification will still fail with CCP driver loaded
> (since CCP RSA implementation has a higher priority than the software
> RSA implementation).
>
> Maciej
>
I commented on that one here:
https://marc.info/?l=linux-crypto-vger&m=151986452422791&w=2
Effectively a NACK. We are a reviewing a proposed patch right now.
On 03.03.2018 00:49, Hook, Gary wrote:
> On 3/2/2018 5:15 PM, Maciej S. Szmigiero wrote:
>> On 02.03.2018 17:44, Herbert Xu wrote:
>>> On Sat, Feb 24, 2018 at 05:03:21PM +0100, Maciej S. Szmigiero wrote:
>>>> rsa-pkcs1pad uses a value returned from a RSA implementation max_size
>>>> callback as a size of an input buffer passed to the RSA implementation for
>>>> encrypt and sign operations.
>>>>
>>>> CCP RSA implementation uses a hardware input buffer which size depends only
>>>> on the current RSA key length, so it should return this key length in
>>>> the max_size callback, too.
>>>> This also matches what the kernel software RSA implementation does.
>>>>
>>>> Previously, the value returned from this callback was always the maximum
>>>> RSA key size the CCP hardware supports.
>>>> This resulted in this huge buffer being passed by rsa-pkcs1pad to CCP even
>>>> for smaller key sizes and then in a buffer overflow when ccp_run_rsa_cmd()
>>>> tried to copy this large input buffer into a RSA key length-sized hardware
>>>> input buffer.
>>>>
>>>> Signed-off-by: Maciej S. Szmigiero <[email protected]>
>>>> Fixes: ceeec0afd684 ("crypto: ccp - Add support for RSA on the CCP")
>>>> Cc: [email protected]
>>>
>>> Patch applied. Thanks.
>>
>> Thanks.
>>
>> However, what about the first patch from this series?
>> Without it, while it no longer should cause a buffer overflow, in-kernel
>> X.509 certificate verification will still fail with CCP driver loaded
>> (since CCP RSA implementation has a higher priority than the software
>> RSA implementation).
>>
>> Maciej
>>
>
>
> I commented on that one here:
> https://marc.info/?l=linux-crypto-vger&m=151986452422791&w=2
>
> Effectively a NACK. We are a reviewing a proposed patch right now.
Your earlier comment referred to the third patch from this series.
My message above was about the first one.
Maciej
On 03/02/2018 05:58 PM, Maciej S. Szmigiero wrote:
> On 03.03.2018 00:49, Hook, Gary wrote:
>> On 3/2/2018 5:15 PM, Maciej S. Szmigiero wrote:
>>>
>>> Thanks.
>>>
>>> However, what about the first patch from this series?
>>> Without it, while it no longer should cause a buffer overflow, in-kernel
>>> X.509 certificate verification will still fail with CCP driver loaded
>>> (since CCP RSA implementation has a higher priority than the software
>>> RSA implementation).
>>>
>>> Maciej
>>>
>>
>>
>> I commented on that one here:
>> https://marc.info/?l=linux-crypto-vger&m=151986452422791&w=2
>>
>> Effectively a NACK. We are a reviewing a proposed patch right now.
>
> Your earlier comment referred to the third patch from this series.
> My message above was about the first one.
Apologies; my mistake.
On Sat, Mar 03, 2018 at 12:15:20AM +0100, Maciej S. Szmigiero wrote:
>
> However, what about the first patch from this series?
> Without it, while it no longer should cause a buffer overflow, in-kernel
> X.509 certificate verification will still fail with CCP driver loaded
> (since CCP RSA implementation has a higher priority than the software
> RSA implementation).
That normally goes through the security tree.
Cheers,
--
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt