Opened 3 years ago
Closed 3 years ago
Last modified 3 years ago
#4643 closed task (fixed)
|Reported by:||Douglas R. Reno||Owned by:|
New security release
Major changes between OpenSSL 1.1.1f and OpenSSL 1.1.1g [21 Apr 2020] Fixed segmentation fault in SSL_check_chain() (CVE-2020-1967)
OpenSSL Security Advisory [21 April 2020] ========================================= Segmentation fault in SSL_check_chain (CVE-2020-1967) ===================================================== Severity: High Server or client applications that call the SSL_check_chain() function during or after a TLS 1.3 handshake may crash due to a NULL pointer dereference as a result of incorrect handling of the "signature_algorithms_cert" TLS extension. The crash occurs if an invalid or unrecognised signature algorithm is received from the peer. This could be exploited by a malicious peer in a Denial of Service attack. OpenSSL version 1.1.1d, 1.1.1e, and 1.1.1f are affected by this issue. This issue did not affect OpenSSL versions prior to 1.1.1d. Affected OpenSSL 1.1.1 users should upgrade to 1.1.1g This issue was found by Bernd Edlinger and reported to OpenSSL on 7th April 2020. It was found using the new static analysis pass being implemented in GCC, - -fanalyzer. Additional analysis was performed by Matt Caswell and Benjamin Kaduk. Note ===== This issue did not affect OpenSSL 1.0.2 however these versions are out of support and no longer receiving public updates. Extended support is available for premium support customers: https://www.openssl.org/support/contracts.html This issue did not affect OpenSSL 1.1.0 however these versions are out of support and no longer receiving updates. Users of these versions should upgrade to OpenSSL 1.1.1. References ========== URL for this Security Advisory: https://www.openssl.org/news/secadv/20200421.txt Note: the online version of the advisory may be updated with additional details over time.
Change History (6)
comment:1 by , 3 years ago
comment:2 by , 3 years ago
|Status:||new → assigned|
comment:3 by , 3 years ago
On one of my machines, a different test fails 30-test_afalg.t : that same test failed on the same machine with 1.1.1f (no other failures). The 20-test_enc.t test seems to have generally failed in 1.1.1d.
comment:4 by , 3 years ago
from https://mta.openssl.org/pipermail/openssl-users/2017-February/005324.html the af_alg test uses crypto options in the kernel to offload to the kernel. Also see the follow-up https://mta.openssl.org/pipermail/openssl-users/2017-February/005325.html : it sounds as if the async driver is assumed to exist for non-ancient kernels, but that on some kernel versions or kernel configs it is not present.
Unfortunately, the github link in that follow-up post gives a 404, so trying to determine which kernel config options are needed is "left as an exercise by the reader".
comment:5 by , 3 years ago
|Status:||assigned → closed|
comment:6 by , 3 years ago
|Milestone:||9.2 → 10.0|
(edit: remove comment on old sed re memset, it is no-longer in the book and did nothing)
As a reminder for people doing upgrades, find what processes are currently using the old libs and restart them:
On my sysvinit server, things with rc scripts which can be restarted.
On my sysvinit desktop, also anything using qtwebengine (falkon in my case) and my original login on tty1 from where I ran startx. So, for a desktop use killall -HUP on your browsers to get them to save state, then logout.