Advisories ยป MGASA-2022-0278

Updated kernel packages fix security vulnerabilities

Publication date: 06 Aug 2022
Modification date: 06 Aug 2022
Type: security
Affected Mageia releases : 8
CVE: CVE-2022-21505 , CVE-2022-23825 , CVE-2022-29900 , CVE-2022-29901 , CVE-2022-36123 , CVE-2022-36879 , CVE-2022-36946

Description

This kernel update is based on upstream 5.15.58 and fixes at least the
following security issues:

Kernel lockdown bypass when UEFI secure boot is disabled / unavailable
and IMA appraisal is enabled (CVE-2022-21505).

Aliases in the branch predictor may cause some AMD processors to predict
the wrong branch type potentially leading to information disclosure
(CVE-2022-23825).

Mis-trained branch predictions for return instructions may allow arbitrary
speculative code execution under certain microarchitecture-dependent
conditions (CVE-2022-29900, RetBleed).

Intel microprocessor generations 6 to 8 are affected by a new Spectre
variant that is able to bypass their retpoline mitigation in the kernel
to leak arbitrary data. An attacker with unprivileged user access can
hijack return instructions to achieve arbitrary speculative code execution
under certain microarchitecture-dependent conditions (CVE-2022-29901).

The Linux kernel before 5.18.13 lacks a certain clear operation for the
block starting symbol (.bss). This allows Xen PV guest OS users to cause
a denial of service or gain privileges (CVE-2022-36123).

An issue was discovered in the Linux kernel through 5.18.14.
xfrm_expand_policies in net/xfrm/xfrm_policy.c can cause a refcount to be
dropped twice (CVE-2022-36879).

nfqnl_mangle in net/netfilter/nfnetlink_queue.c in the Linux kernel
through 5.18.14 allows remote attackers to cause a denial of service
(panic) because, in the case of an nf_queue verdict with a one-byte
nfta_payload attribute, an skb_pull can encounter a negative
skb->len (CVE-2022-36946).

Other fixes in this update:
- fs: sendfile handles O_NONBLOCK of out_fd
- hugetlb: fix memoryleak in hugetlb_mcopy_atomic_pte
- mm: fix page leak with multiple threads mapping the same page
- x86/bugs: Do not enable IBPB at firmware entry when IBPB is not available
- x86/speculation: Make all RETbleed mitigations 64-bit only

For other upstream fixes, see the referenced changelogs.

References

SRPMS

8/core