long running loops in grant table handling In order to properly monitor resource use, Xen maintains information on the grant mappings a domain may create to map grants offered by other domains. In the process of carrying out certain actions, Xen would iterate over all such entries, including ones which aren't in use anymore and some which may have been created but never used. If the number of entries for a given domain is large enough, this iterating of the entire table may tie up a CPU for too long, starving other domains or causing issues in the hypervisor itself. Note that a domain may map its own grants, i.e. there is no need for multiple domains to be involved here. A pair of "cooperating" guests may, however, cause the effects to be more severe.

An issue was discovered in Xen through 4.10.x allowing x86 HVM guest OS users to cause a denial of service (host OS infinite loop) in situations where a QEMU device model attempts to make invalid transitions between states of a request.

The ohci_service_ed_list function in hw/usb/hcd-ohci.c in QEMU (aka Quick Emulator) before 2.9.0 allows local guest OS users to cause a denial of service (infinite loop) via vectors involving the number of link endpoint list descriptors, a different vulnerability than CVE-2017-9330.

net/ipv4/inet_diag.c in the Linux kernel before 2.6.37-rc2 does not properly audit INET_DIAG bytecode, which allows local users to cause a denial of service (kernel infinite loop) via crafted INET_DIAG_REQ_BYTECODE instructions in a netlink message that contains multiple attribute elements, as demonstrated by INET_DIAG_BC_JMP instructions.

An issue was discovered in Xen through 4.12.x allowing Arm domU attackers to cause a denial of service (infinite loop) involving a LoadExcl or StoreExcl operation.

The pcnet_rdra_addr function in hw/net/pcnet.c in QEMU (aka Quick Emulator) allows local guest OS administrators to cause a denial of service (infinite loop and QEMU process crash) by setting the (1) receive or (2) transmit descriptor ring length to 0.

An infinite loop flaw was found in the e1000 NIC emulator of the QEMU. This issue occurs while processing transmits (tx) descriptors in process_tx_desc if various descriptor fields are initialized with invalid values. This flaw allows a guest to consume CPU cycles on the host, resulting in a denial of service. The highest threat from this vulnerability is to system availability.

A flaw was found in the Linux kernel's implementation of biovecs in versions before 5.9-rc7. A zero-length biovec request issued by the block subsystem could cause the kernel to enter an infinite loop, causing a denial of service. This flaw allows a local attacker with basic privileges to issue requests to a block device, resulting in a denial of service. The highest threat from this vulnerability is to system availability.

The vmsvga_fifo_run function in hw/display/vmware_vga.c in QEMU allows local guest OS administrators to cause a denial of service (infinite loop and QEMU process crash) via a VGA command.

A flaw was found in the Linux kernel in the function hid_debug_events_read() in drivers/hid/hid-debug.c file which may enter an infinite loop with certain parameters passed from a userspace. A local privileged user ("root") can cause a system lock up and a denial of service. Versions from v4.18 and newer are vulnerable.

An issue was discovered in Xen through 4.12.x allowing Arm domU attackers to cause a denial of service (infinite loop) involving a compare-and-exchange operation.

The ehci_process_itd function in hw/usb/hcd-ehci.c in QEMU allows local guest OS administrators to cause a denial of service (infinite loop and CPU consumption) via a circular isochronous transfer descriptor (iTD) list.

The __rds_conn_create function in net/rds/connection.c in the Linux kernel through 4.2.3 allows local users to cause a denial of service (NULL pointer dereference and system crash) or possibly have unspecified other impact by using a socket that was not properly bound.

arch/x86/kvm/x86.c in the Linux kernel before 4.4 does not reset the PIT counter values during state restoration, which allows guest OS users to cause a denial of service (divide-by-zero error and host OS crash) via a zero value, related to the kvm_vm_ioctl_set_pit and kvm_vm_ioctl_set_pit2 functions.

The KVM subsystem in the Linux kernel through 4.2.6, and Xen 4.3.x through 4.6.x, allows guest OS users to cause a denial of service (host OS panic or hang) by triggering many #AC (aka Alignment Check) exceptions, related to svm.c and vmx.c.

fs/splice.c in the splice subsystem in the Linux kernel before 2.6.22.2 does not properly handle a failure of the add_to_page_cache_lru function, and subsequently attempts to unlock a page that was not locked, which allows local users to cause a denial of service (kernel BUG and system crash), as demonstrated by the fio I/O tool.

Off-by-one error in the iov_iter_advance function in mm/filemap.c in the Linux kernel before 2.6.27-rc2 allows local users to cause a denial of service (system crash) via a certain sequence of file I/O operations with readv and writev, as demonstrated by testcases/kernel/fs/ftest/ftest03 from the Linux Test Project.

The (1) real_lookup and (2) __lookup_hash functions in fs/namei.c in the vfs implementation in the Linux kernel before 2.6.25.15 do not prevent creation of a child dentry for a deleted (aka S_DEAD) directory, which allows local users to cause a denial of service ("overflow" of the UBIFS orphan area) via a series of attempted file creations within deleted directories.

Integer overflow in the sctp_getsockopt_local_addrs_old function in net/sctp/socket.c in the Stream Control Transmission Protocol (sctp) functionality in the Linux kernel before 2.6.25.9 allows local users to cause a denial of service (resource consumption and system outage) via vectors involving a large addr_num field in an sctp_getaddrs_old data structure.

The ping_unhash function in net/ipv4/ping.c in the Linux kernel before 4.0.3 does not initialize a certain list data structure during an unhash operation, which allows local users to gain privileges or cause a denial of service (use-after-free and system crash) by leveraging the ability to make a SOCK_DGRAM socket system call for the IPPROTO_ICMP or IPPROTO_ICMPV6 protocol, and then making a connect system call after a disconnect.

An issue was discovered in the Linux kernel before 5.11.11. The netfilter subsystem allows attackers to cause a denial of service (panic) because net/netfilter/x_tables.c and include/linux/netfilter/x_tables.h lack a full memory barrier upon the assignment of a new table value, aka CID-175e476b8cdf.

The ahci_commit_buf function in ide/ahci.c in QEMU allows attackers to cause a denial of service (NULL dereference) when the command header 'ad->cur_cmd' is null.

An issue was discovered in the Linux kernel before 5.1.17. There is a NULL pointer dereference caused by a malicious USB device in the sound/usb/line6/pcm.c driver.

fs/direct-io.c in the dio subsystem in the Linux kernel before 2.6.23 does not properly zero out the dio struct, which allows local users to cause a denial of service (OOPS), as demonstrated by a certain fio test.

Memory leak in hw/9pfs/9p.c in QEMU (aka Quick Emulator) allows local privileged guest OS users to cause a denial of service (host memory consumption and possibly QEMU process crash) by leveraging a missing cleanup operation in FileOperations.

Memory leak in hw/9pfs/9p-handle.c in QEMU (aka Quick Emulator) allows local privileged guest OS users to cause a denial of service (host memory consumption and possibly QEMU process crash) by leveraging a missing cleanup operation in the handle backend.

Memory leak in hw/9pfs/9p-proxy.c in QEMU (aka Quick Emulator) allows local privileged guest OS users to cause a denial of service (host memory consumption and possibly QEMU process crash) by leveraging a missing cleanup operation in the proxy backend.

The fix for XSA-365 includes initialization of pointers such that subsequent cleanup code wouldn't use uninitialized or stale values. This initialization went too far and may under certain conditions also overwrite pointers which are in need of cleaning up. The lack of cleanup would result in leaking persistent grants. The leak in turn would prevent fully cleaning up after a respective guest has died, leaving around zombie domains. All Linux versions having the fix for XSA-365 applied are vulnerable. XSA-365 was classified to affect versions back to at least 3.11.

Memory leak in the hwsim_new_radio_nl function in drivers/net/wireless/mac80211_hwsim.c in the Linux kernel through 4.15.9 allows local users to cause a denial of service (memory consumption) by triggering an out-of-array error case.

The resv_map_release function in mm/hugetlb.c in the Linux kernel through 4.15.7 allows local users to cause a denial of service (BUG) via a crafted application that makes mmap system calls and has a large pgoff argument to the remap_file_pages system call.

An issue was discovered in Xen through 4.10.x allowing x86 PV guest OS users to cause a denial of service (host OS CPU hang) via non-preemptable L3/L4 pagetable freeing.

Memory leak in the irda_bind function in net/irda/af_irda.c and later in drivers/staging/irda/net/af_irda.c in the Linux kernel before 4.17 allows local users to cause a denial of service (memory consumption) by repeatedly binding an AF_IRDA socket.

An issue was discovered in Xen 4.8.x through 4.10.x allowing x86 PVH guest OS users to cause a denial of service (NULL pointer dereference and hypervisor crash) by leveraging the mishandling of configurations that lack a Local APIC.

In the Linux Kernel before version 4.15.8, 4.14.25, 4.9.87, 4.4.121, 4.1.51, and 3.2.102, an error in the "_sctp_make_chunk()" function (net/sctp/sm_make_chunk.c) when handling SCTP packets length can be exploited to cause a kernel crash.

In the Linux kernel through 4.14.13, the rds_cmsg_atomic function in net/rds/rdma.c mishandles cases where page pinning fails or an invalid address is supplied, leading to an rds_atomic_free_op NULL pointer dereference.

Guest can force Linux netback driver to hog large amounts of kernel memory T[his CNA information record relates to multiple CVEs; the text explains which aspects/vulnerabilities correspond to which CVE.] Incoming data packets for a guest in the Linux kernel's netback driver are buffered until the guest is ready to process them. There are some measures taken for avoiding to pile up too much data, but those can be bypassed by the guest: There is a timeout how long the client side of an interface can stop consuming new packets before it is assumed to have stalled, but this timeout is rather long (60 seconds by default). Using a UDP connection on a fast interface can easily accumulate gigabytes of data in that time. (CVE-2021-28715) The timeout could even never trigger if the guest manages to have only one free slot in its RX queue ring page and the next package would require more than one free slot, which may be the case when using GSO, XDP, or software hashing. (CVE-2021-28714)

Rogue backends can cause DoS of guests via high frequency events T[his CNA information record relates to multiple CVEs; the text explains which aspects/vulnerabilities correspond to which CVE.] Xen offers the ability to run PV backends in regular unprivileged guests, typically referred to as "driver domains". Running PV backends in driver domains has one primary security advantage: if a driver domain gets compromised, it doesn't have the privileges to take over the system. However, a malicious driver domain could try to attack other guests via sending events at a high frequency leading to a Denial of Service in the guest due to trying to service interrupts for elongated amounts of time. There are three affected backends: * blkfront patch 1, CVE-2021-28711 * netfront patch 2, CVE-2021-28712 * hvc_xen (console) patch 3, CVE-2021-28713

An issue was discovered in the Linux kernel through 5.11.3, as used with Xen PV. A certain part of the netback driver lacks necessary treatment of errors such as failed memory allocations (as a result of changes to the handling of grant mapping errors). A host OS denial of service may occur during misbehavior of a networking frontend driver. NOTE: this issue exists because of an incomplete fix for CVE-2021-26931.

In intel_pmu_drain_pebs_nhm in arch/x86/events/intel/ds.c in the Linux kernel through 5.11.8 on some Haswell CPUs, userspace applications (such as perf-fuzzer) can cause a system crash because the PEBS status in a PEBS record is mishandled, aka CID-d88d05a9e0b6.

An issue was discovered in Xen through 4.11.x on Intel x86 platforms allowing guest OS users to cause a denial of service (host OS hang) because Xen does not work around Intel's mishandling of certain HLE transactions associated with the KACQUIRE instruction prefix.

A certain backport in the TCP Fast Open implementation for the Linux kernel before 3.18 does not properly maintain a count value, which allow local users to cause a denial of service (system crash) via the Fast Open feature, as demonstrated by visiting the chrome://flags/#enable-tcp-fast-open URL when using certain 3.10.x through 3.16.x kernel builds, including longterm-maintenance releases and ckt (aka Canonical Kernel Team) builds.

In the Linux kernel before 4.17, a local attacker able to set attributes on an xfs filesystem could make this filesystem non-operational until the next mount by triggering an unchecked error condition during an xfs attribute change, because xfs_attr_shortform_addname in fs/xfs/libxfs/xfs_attr.c mishandles ATTR_REPLACE operations with conversion of an attr from short to long form.

Rogue backends can cause DoS of guests via high frequency events T[his CNA information record relates to multiple CVEs; the text explains which aspects/vulnerabilities correspond to which CVE.] Xen offers the ability to run PV backends in regular unprivileged guests, typically referred to as "driver domains". Running PV backends in driver domains has one primary security advantage: if a driver domain gets compromised, it doesn't have the privileges to take over the system. However, a malicious driver domain could try to attack other guests via sending events at a high frequency leading to a Denial of Service in the guest due to trying to service interrupts for elongated amounts of time. There are three affected backends: * blkfront patch 1, CVE-2021-28711 * netfront patch 2, CVE-2021-28712 * hvc_xen (console) patch 3, CVE-2021-28713

Rogue backends can cause DoS of guests via high frequency events T[his CNA information record relates to multiple CVEs; the text explains which aspects/vulnerabilities correspond to which CVE.] Xen offers the ability to run PV backends in regular unprivileged guests, typically referred to as "driver domains". Running PV backends in driver domains has one primary security advantage: if a driver domain gets compromised, it doesn't have the privileges to take over the system. However, a malicious driver domain could try to attack other guests via sending events at a high frequency leading to a Denial of Service in the guest due to trying to service interrupts for elongated amounts of time. There are three affected backends: * blkfront patch 1, CVE-2021-28711 * netfront patch 2, CVE-2021-28712 * hvc_xen (console) patch 3, CVE-2021-28713

QEMU, as used in Xen 3.3.x through 4.5.x, does not properly restrict access to PCI command registers, which might allow local HVM guest users to cause a denial of service (non-maskable interrupt and host crash) by disabling the (1) memory or (2) I/O decoding for a PCI Express device and then accessing the device, which triggers an Unsupported Request (UR) response.

An issue was discovered in Xen through 4.10.x. Certain PV MMU operations may take a long time to process. For that reason Xen explicitly checks for the need to preempt the current vCPU at certain points. A few rarely taken code paths did bypass such checks. By suitably enforcing the conditions through its own page table contents, a malicious guest may cause such bypasses to be used for an unbounded number of iterations. A malicious or buggy PV guest may cause a Denial of Service (DoS) affecting the entire host. Specifically, it may prevent use of a physical CPU for an indeterminate period of time. All Xen versions from 3.4 onwards are vulnerable. Xen versions 3.3 and earlier are vulnerable to an even wider class of attacks, due to them lacking preemption checks altogether in the affected code paths. Only x86 systems are affected. ARM systems are not affected. Only multi-vCPU x86 PV guests can leverage the vulnerability. x86 HVM or PVH guests as well as x86 single-vCPU PV ones cannot leverage the vulnerability.

inadequate grant-v2 status frames array bounds check The v2 grant table interface separates grant attributes from grant status. That is, when operating in this mode, a guest has two tables. As a result, guests also need to be able to retrieve the addresses that the new status tracking table can be accessed through. For 32-bit guests on x86, translation of requests has to occur because the interface structure layouts commonly differ between 32- and 64-bit. The translation of the request to obtain the frame numbers of the grant status table involves translating the resulting array of frame numbers. Since the space used to carry out the translation is limited, the translation layer tells the core function the capacity of the array within translation space. Unfortunately the core function then only enforces array bounds to be below 8 times the specified value, and would write past the available space if enough frame numbers needed storing.

The (1) BMDMA and (2) AHCI HBA interfaces in the IDE functionality in QEMU 1.0 through 2.1.3 have multiple interpretations of a function's return value, which allows guest OS users to cause a host OS denial of service (memory consumption or infinite loop, and system crash) via a PRDT with zero complete sectors, related to the bmdma_prepare_buf and ahci_dma_prepare_buf functions.

Guest can force Linux netback driver to hog large amounts of kernel memory T[his CNA information record relates to multiple CVEs; the text explains which aspects/vulnerabilities correspond to which CVE.] Incoming data packets for a guest in the Linux kernel's netback driver are buffered until the guest is ready to process them. There are some measures taken for avoiding to pile up too much data, but those can be bypassed by the guest: There is a timeout how long the client side of an interface can stop consuming new packets before it is assumed to have stalled, but this timeout is rather long (60 seconds by default). Using a UDP connection on a fast interface can easily accumulate gigabytes of data in that time. (CVE-2021-28715) The timeout could even never trigger if the guest manages to have only one free slot in its RX queue ring page and the next package would require more than one free slot, which may be the case when using GSO, XDP, or software hashing. (CVE-2021-28714)

The acceleration support for the "REP MOVS" instruction in Xen 4.4.x, 3.2.x, and earlier lacks properly bounds checking for memory mapped I/O (MMIO) emulated in the hypervisor, which allows local HVM guests to cause a denial of service (host crash) via unspecified vectors.