Bug 1975836
Summary: | add ripemd160 hash back? (how to enable legacy provider) | ||
---|---|---|---|
Product: | Red Hat Enterprise Linux 9 | Reporter: | Yatin Karel <ykarel> |
Component: | openssl | Assignee: | Sahana Prasad <sahana> |
Status: | CLOSED CURRENTRELEASE | QA Contact: | Alicja Kario <hkario> |
Severity: | low | Docs Contact: | Mirek Jahoda <mjahoda> |
Priority: | low | ||
Version: | 9.0 | CC: | agk, apevec, bstinson, dbelyavs, eharney, geguileo, hkario, jbrassow, jwboyer, ltoscano, okozina, prajnoha, rdhasman, sahana, storage-qe, szidek |
Target Milestone: | beta | Keywords: | Triaged |
Target Release: | --- | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Whiteboard: | |||
Fixed In Version: | openssl-3.0.0-0.beta2.5.el9 | Doc Type: | Deprecated Functionality |
Doc Text: |
.OpenSSL deprecates MD2, MD4, MDC2, Whirlpool, Blowfish, CAST, DES, IDEA, RC2, RC4, RC5, SEED, and PBKDF1
The OpenSSL project has deprecated a set of cryptographic algorithms because they are insecure, uncommonly used, or both. Red Hat also discourages the use of those algorithms, and RHEL 9 provides them for migrating encrypted data to use new algorithms. Users must not depend on those algorithms for the security of their systems.
The implementations of the following algorithms have been moved to the legacy provider in OpenSSL: MD2, MD4, MDC2, Whirlpool, Blowfish, CAST, DES, IDEA, RC2, RC4, RC5, SEED, and PBKDF1.
See the `/etc/pki/tls/openssl.cnf` configuration file for instructions on how to load the legacy provider and enable support for the deprecated algorithms.
|
Story Points: | --- |
Clone Of: | 1975799 | Environment: | |
Last Closed: | 2021-12-07 21:24:13 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Bug Depends On: | |||
Bug Blocks: | 1975799 |
Description
Yatin Karel
2021-06-24 14:26:03 UTC
quote from original bug 1975799 > Openssl perhaps will not add RIPEMD hash bbck to suported algorithms, there is a reason for removing it from defaults. Please confirm that ripemd160 will not be added back so we can discuss in OpenStack removal of plain mode crytsetup! The situation is a bit more complex. We are shipping an implementation of ripemd160, but it's in the legacy provider: https://github.com/openssl/openssl/blob/master/doc/man7/OSSL_PROVIDER-legacy.pod The user needs to provide an openssl config file that loads the legacy provider for the ripemd160 to be available to applications. So I think cryptsetup doesn't need to remove plain mode, but it should move it to "unsupported", with documentation stating how to migrate off of it (if changing the default hash is not possible). Removing plain mode in cryptsetup was not even a question, it must remain there. It was about removing plain mode use in OpenStack. Also cryptsetup will not move anything to unsupported, it will simply load legacy provider here. Actually I already tested that and there are some minor issues, but it is easily doable. Cryptsetup plain mode is a simple wrapper around dm-crypt, it must support even old modes (ripemd hashing is here just because of historical and backward compatibility reasons - and here we perhaps switch default, despite it will cause breakage; but older configs must remain supported otherwise we will not be able to decrtypt older data for many people). Nobody should use cryptsetup plain mode until there is very specific use case (like simulating some other priprietary system) - EVERYTHING related to storage should use LUKS. The main reason is that we should no use encryption key that is directly derived from passphrase (as it is in plain mode if a passphrase is used; randomly generated keyfile behaves differently here). TL;DR: I think OpenSSL does not need to do anything here, we should fix OpenSSL backend init in cryptsetup for OpenSSL3 to allow legacy provider use (even without openssl config file). *** Bug 1990877 has been marked as a duplicate of this bug. *** |