Re: CVE-2021-30130 php-phpseclib and phpseclib

2021-05-27 Thread Abhijith PA
Hi Ola,

On 26/05/21 01:45 PM, Ola Lundqvist wrote:
>Hi fellow LTS contributors
> 
>I have checked this CVE and my conclusions are as follows.
>The CVE actually cover five different problems. I guess CVEs should not
>do that, but it did anyway.
> 
>Quote from upstream:
> 
>Two were vulnerabilities in v3.0 involving the new
>RSA::SIGNATURE_RELAXED_PKCS1 mode (which doesn't exist in 2.0)
> 
>Two were bugs in v3.0 involving the new RSA::SIGNATURE_RELAXED_PKCS1
>mode (which again, doesn't exist in 2.0)
> 
>One was a bug in v1.0, v2.0 and v3.0.
> 
>The bug refers to "We have also found incompatibility issue in
>phpseclib v1, v2, v3 (strict mode)'s RSA PKCS#1 v1.5 signature
>verification suffering from rejecting valid signatures whose encoded
>message uses implicit hash algorithm's NULL parameter."
> 
>My conclusion is that one bug can be fixed. But I do not think it is a
>security problem. The problem is that some signatures fail valid
>signatures, if they are encoded in a special way.
> 
>What I have done is to mark the CVE as not-affected with a note about
>this.
> 
>Let me know if you think my analysis is correct.

I've gone through those comments and fixes. Since valid signature 
failing bug in v1 and v2 is not a security issue. I think marking 
CVE-2021-30130 as not-affected is the way to go. Sorry for holding the 
package.

--abhijith 

signature.asc
Description: PGP signature


CVE-2021-30130 php-phpseclib and phpseclib

2021-05-26 Thread Ola Lundqvist
Hi fellow LTS contributors

I have checked this CVE and my conclusions are as follows.
The CVE actually cover five different problems. I guess CVEs should not do
that, but it did anyway.

Quote from upstream:

Two were vulnerabilities in v3.0 involving the new
RSA::SIGNATURE_RELAXED_PKCS1 mode (which doesn't exist in 2.0)
Two were bugs in v3.0 involving the new RSA::SIGNATURE_RELAXED_PKCS1 mode
(which again, doesn't exist in 2.0)
One was a bug in v1.0, v2.0 and v3.0.


The bug refers to "We have also found incompatibility issue in phpseclib
v1, v2, v3 (strict mode)'s RSA PKCS#1 v1.5 signature verification suffering
from rejecting valid signatures whose encoded message uses implicit hash
algorithm's NULL parameter."

My conclusion is that one bug can be fixed. But I do not think it is a
security problem. The problem is that some signatures fail valid
signatures, if they are encoded in a special way.

What I have done is to mark the CVE as not-affected with a note about this.

Let me know if you think my analysis is correct.

I'm sending this to you all since I'm not 100% sure how to treat it.

Best regards

// Ola


-- 
 --- Inguza Technology AB --- MSc in Information Technology 
|  o...@inguza.como...@debian.org|
|  http://inguza.com/Mobile: +46 (0)70-332 1551 |
 ---