MGASA-2021-0100

This kernel-linus update is based on upstream 5.10.19 and fixes at least the following security issues:

There is a vulnerability in the linux kernel versions higher than 5.2 (if kernel compiled with config params CONFIGBPFSYSCALL=y, CONFIGBPF=y, CONFIGCGROUPS=y, CONFIGCGROUPBPF=y, CONFIGHARDENEDUSERCOPY not set, and BPF hook to getsockopt is registered). As result of BPF execution, the local user can trigger bug in _cgroupbpfrunfiltergetsockopt() function that can lead to heap overflow (because of non-hardened usercopy). The impact of attack could be deny of service or possibly privileges escalation. NOTE! Mageia kernel configs have HARDENEDUSERCOPY enabled by default, making this an non-issue when using prebuilt kernels (CVE-2021-20194).

An information disclosure vulnerability exists in the ARM SIGPAGE functionality of Linux Kernel. A userland application can read the contents of the sigpage, which can leak kernel memory contents. An attacker can read a process’s memory at a specific offset to trigger this vulnerability (CVE-2021-21781).

An issue was discovered in the Linux kernel 3.11 through 5.10.16, as used by Xen. To service requests to the PV backend, the driver maps grant references provided by the frontend. In this process, errors may be encountered. In one case, an error encountered earlier might be discarded by later processing, resulting in the caller assuming successful mapping, and hence subsequent operations trying to access space that wasn't mapped. In another case, internal state would be insufficiently updated, preventing safe recovery from the error (CVE-2021-26930).

An issue was discovered in the Linux kernel 2.6.39 through 5.10.16, as used in Xen. Block, net, and SCSI backends consider certain errors a plain bug, deliberately causing a kernel crash. For errors potentially being at least under the influence of guests (such as out of memory conditions), it isn't correct to assume a plain bug. Memory allocations potentially causing such crashes occur only when Linux is running in PV mode, though (CVE-2021-26931).

An issue was discovered in the Linux kernel 3.2 through 5.10.16, as used by Xen. Grant mapping operations often occur in batch hypercalls, where a number of operations are done in a single hypercall, the success or failure of each one is reported to the backend driver, and the backend driver then loops over the results, performing follow-up actions based on the success or failure of each operation. Unfortunately, when running in PV mode, the Linux backend drivers mishandle this: Some errors are ignored, effectively implying their success from the success of related batch elements. In other cases, errors resulting from one batch element lead to further batch elements not being inspected, and hence successful ones to not be possible to properly unmap upon error recovery. Only systems with Linux backends running in PV mode are vulnerable. Linux backends run in HVM / PVH modes are not vulnerable (CVE-2021-26932).

It also adds the following fixes: - enable ACPIECDEBUGFS (mga#28415)

For other upstream fixes, see the referenced changelogs.