readykernel-patch-96.21-108.0-1.vl7

CVE-2020-14305 was later assigned to this issue. See also: https://bugzilla.redhat.com/show_bug.cgi?id=CVE-2020-14305.

It was found that a race was possible between uncharging memory from a cgroup and making that cgroup offline. This could lead to premature destruction of the cgroup and could cause a kernel crash.

A flaw was found in the implementation of GRO, which allows an attacker with local access to trigger a use-after-free read in napi_gro_frags() and, potentially, crash the system.

When a new netns is created, high-order page allocations can happen in nf_ct_alloc_hashtable(). If memory is fragmented, such allocations can become very slow due to memory reclaim, etc. This, in turn, could result in significant slowdowns on the node.

A race condition was discovered in ploop, which could lead to corruption of the index during discard operations in certain cases. Note. It is not recommended to manually unload the ReadyKernel patch with the fix for this issue while any discard operations for ploop images are in progress: the ploop images could be corrupted as a result. Upgrading the patch is OK, only manual unloads and downgrades can be problematic.

It was discovered that xfs_setattr_nonsize() would not unlock 'ILOCK' lock if the user or group were out of their disk quota. As a result, any subsequent operation, which needed to take 'ILOCK', would get stuck, leading to a denial of service.

It was discovered that the implementation of process_vm_readv() system call could try to access memory outside of the structures it was processing in certain cases. A local unprivileged user could use this vulnerability to crash the system.

An issue was found in the implementation of the coalesced MMIO write operation in KVM. The indices used to access an MMIO ring buffer could be supplied by a user-space process in the host system. An attacker with access to /dev/kvm could use this flaw to trigger out-of-bounds memory access and crash the host kernel or, potentially, escalate their privileges.

It was discovered that a local unprivileged attacker could use a specially crafted sequence of system calls to trigger either a kernel crash in __tcp_retransmit_skb() or use-after-free conditions, which could result in privilege escalation.

It was found that a use-after-free condition could be triggered in the block device subsystem while the outstanding command queue was drained. A patient local attacker can use this flaw to crash the system or, potentially, to escalate their privileges.

It was discovered that a race condition was possible between kmem uncharging and mem_cgroup_reparent_charges(). A kernel warning would be triggered as a result.

A buffer overflow vulnerability was found in the networking virtualization functionality (vhost-net) that could be abused during live migration of virtual machines. A privileged guest user may pass descriptors with invalid length to the host when live migration is underway to crash the host kernel or, potentially, escalate their privileges on the host.

It was found that ucma_leave_multicast() function from 'rdma_ucm' module could try to access a certain data structure after the structure had been freed. This allows an attacker to induce kernel memory corruption, leading to a system crash or other unspecified impact.

It was discovered that a local attacker could exploit mincore() system call to obtain information about memory pages of the running applications from the page cache even if the contents of these memory pages were not available to the attacker.

An attacker, who is able to mount an exported NFS filesystem, is able to trigger a null pointer dereference by using an invalid NFS sequence. This can panic the machine and deny access to the NFS server. Any outstanding disk writes to the NFS server will be lost.

It was discovered that if a file located on Virtuozzo Storage cluster was reopened, the kernel could reuse its mappings without notifying MDS. As a result, if errors were then detected on a CS, for example, caused by power failures, the userspace components would be unable to handle them properly.