VMX: VMentry failure with debug exceptions and blocked states

ISSUE DESCRIPTION Please see XSA-260 for background on the MovSS shadow: http://xenbits.xen.org/xsa/advisory-260.html Please see XSA-156 for background on the need for DB interception: http://xenbits.xen.org/xsa/advisory-156.html The VMX VMEntry checks does not like the exact combination of state...

missing descriptor table limit checking in x86 PV emulation

ISSUE DESCRIPTION When emulating certain PV guest operations, descriptor table accesses are performed by the emulating code. Such accesses should respect the guest specified limits, unless otherwise guaranteed to fail in such a case. Without this, emulation of 32-bit guest user mode calls through...

Guest triggerable qemu MSI-X pass-through error messages

ISSUE DESCRIPTION Device model code dealing with guest PCI MSI-X interrupt management activities logs messages on certain supposedly invalid guest operations. IMPACT A buggy or malicious guest repeatedly invoking such operations may result in the host disk to fill up, possibly leading to a Denial...

Certain domctl operations may be abused to lock up the host

ISSUE DESCRIPTION XSA-77 put the majority of the domctl operations on a list excepting them from having security advisories issued for them if any effects their use might have could hamper security. Subsequently some of them got declared disaggregation safe, but for a small subset this was not...

qemu-nbd format-guessing due to missing format specification

ISSUE DESCRIPTION The qemu-nbd tool shipped in the Xen hypervisor tools distribution as qemu-nbd-xen autodetects the image format. If a particular disk image is intended to be raw, a guest operating system administrator could write a header to the image, describing another format than original on...

Multiple speculative security issues

ISSUE DESCRIPTION Note: Multiple issues are contained in this XSA due to their interactions. 1 Researchers at VU Amsterdam have discovered Spectre-BHB, pertaining to the use of Branch History between privilege levels. ARM have assigned CVE-2022-23960. Intel have assigned CVE-2022-0001 Branch...

VMX: intercept issue with INVLPG on non-canonical address

ISSUE DESCRIPTION While INVLPG does not cause a General Protection Fault when used on a non-canonical address, INVVPID in its "individual address" variant, which is used to back the intercepted INVLPG in certain cases, fails in such cases. Failure of INVVPID results in a hypervisor bug check...

Hypervisor crash due to missing exception recovery on XRSTOR

ISSUE DESCRIPTION Processors do certain validity checks on the data passed to XRSTOR. While the hypervisor controls the placement of that memory block, it doesn't restrict the contents in any way. Thus the hypervisor exposes itself to a fault occurring on XRSTOR. Other than for FXRSTOR, which...

information leak in legacy x86 FPU/XMM initialization

ISSUE DESCRIPTION When XSAVE/XRSTOR are not in use by Xen to manage guest extended register state, the initial values in the FPU stack and XMM registers seen by the guest upon first use are those left there by the previous user of those registers. IMPACT A malicious domain may be able to leverage...

HVMOP_get_mem_access crash / HVMOP_set_mem_access information leak

ISSUE DESCRIPTION The HVMOPsetmemaccess operation handler uses an input as an array index before range checking it. IMPACT A malicious guest administrator can cause Xen to crash. If the out of array bounds access does not crash, the arbitrary value read will be used if the caller reads back the...

PV superpage functionality missing sanity checks

ISSUE DESCRIPTION The PV superpage functionality lacks certain validity checks on data being passed to the hypervisor by guests. This is the case for the page identifier MFN passed to MMUEXTMARKSUPER and MMUEXTUNMARKSUPER sub-ops of the HYPERVISORmmuextop hypercall as well as for various forms of...

misplaced free in ocaml xc_vcpu_getaffinity stub

ISSUE DESCRIPTION The ocaml binding for the xcvcpugetaffinity function incorrectly frees a pointer before using it and subsequently freeing it again afterwards. The code therefore contains a use-after-free and double-free flaws. IMPACT An attacker may be able to cause a multithreaded toolstack...

Linear pagetable use / entry miscounts

ISSUE DESCRIPTION "Linear pagetables" is a technique which involves either pointing a pagetable at itself, or to another pagetable of the same or higher level. Xen has limited support for linear pagetables: A page may either point to itself, or point to another pagetable of the same level i.e., L...

add-to-physmap can be abused to DoS Arm hosts

ISSUE DESCRIPTION p2m-maxmappedgfn is used by the functions p2mresolvetranslationfault and p2mgetentry to sanity check guest physical frame. The rest of the code in the two functions will assume that there is a valid root table and check that with BUGON. The function p2mgetrootpointer will ignore...

ARM: Interrupts are unconditionally unmasked in exception handlers

ISSUE DESCRIPTION When an exception occurs on an ARM system which is handled without changing processor level, some interrupts are unconditionally enabled during exception entry. So exceptions which occur when interrupts are masked will effectively unmask the interrupts. IMPACT A malicious guest...

Linux kernel hits general protection if %ds is corrupt for 32-bit PVOPS.

ISSUE DESCRIPTION Linux kernel when returning from an iret assumes that %ds segment is safe and uses it to reference various per-cpu related fields. Unfortunately the user can modify the LDT and provide a NULL one. Whenever an iret is called we end up in xeniret and try to use the %ds segment and...

Linux: special config may crash when trying to map foreign pages

ISSUE DESCRIPTION With CONFIGXENBALLOONMEMORYHOTPLUG disabled and CONFIGXENUNPOPULATEDALLOC enabled the Linux kernel will use guest physical addresses allocated via the ZONEDEVICE functionality for mapping foreign guest's pages. This will result in problems, as the p2m list will only cover the...

Linux: grant mapping error handling issues

ISSUE DESCRIPTION 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 reported to the backend driver, and the backend driver then loops over the results, performing follow-up actions based on the...

Missing memory barriers in read-write unlock paths

ISSUE DESCRIPTION The read-write unlock paths don't contain a memory barrier. On Arm, this means a processor is allowed to re-order the memory access with the preceding ones. In other words, the unlock may be seen by another processor before all the memory accesses within the "critical" section. ...

Broken error handling in guest_physmap_mark_populate_on_demand()

ISSUE DESCRIPTION guestphysmapmarkpopulateondemand, before carrying out its actual operation, checks that the subject GFNs are not in use. If that check fails, the code prints a message and bypasses the gfnunlock matching the gfnlock carried out before entering the loop. Further, the function is...

missed flush in XSA-321 backport

ISSUE DESCRIPTION An oversight was made when backporting XSA-321, leading entries in the IOMMU not being properly updated under certain circumstances. IMPACT A malicious guest may be able to retain read/write DMA access to frames returned to Xen's free pool, and later reused for another purpose...

Issues with restartable PV type change operations

ISSUE DESCRIPTION To avoid using shadow pagetables for PV guests, Xen exposes the actual hardware pagetables to the guest. In order to prevent the guest from modifying these page tables directly, Xen keeps track of how pages are used using a type system; pages must be "promoted" before being used...

Missing privilege level checks in x86 HLT, LGDT, LIDT, and LMSW emulation

ISSUE DESCRIPTION The emulation of the instructions HLT, LGDT, LIDT, and LMSW fails to perform supervisor mode permission checks. However these instructions are not usually handled by the emulator. Exceptions to this are - when the instruction's memory operand if any lives in emulated or passed...

HVMOP_set_mem_access is not preemptible

ISSUE DESCRIPTION Processing of the HVMOPsetmemaccess HVM control operations does not check the size of its input and can tie up a physical CPU for extended periods of time. IMPACT In a configuration where device models run with limited privilege for example, stubdom device models, a guest attack...

Linux netback DoS via malicious guest ring.

ISSUE DESCRIPTION The Xen netback implementation contains a couple of flaws which can allow a guest to cause a DoS in the backend domain, potentially affecting other domains in the system. CVE-2013-0216 is a failure to sanity check the ring producer/consumer pointers which can allow a guest to...

x86: Register File Data Sampling

ISSUE DESCRIPTION Intel have disclosed RFDS, Register File Data Sampling, affecting some Atom cores. This came from internal validation work. There is no information provided about how an attacker might go about inferring data from the register files. For more details, see:...

VMX: guest user mode may crash guest with non-canonical RIP

ISSUE DESCRIPTION VMX refuses attempts to enter a guest with an instruction pointer which doesn't satisfy certain requirements. In particular, the instruction pointer needs to be canonical when entering a guest currently in 64-bit mode. This is the case even if the VM entry information specifies ...

xen crash due to use after free on hvm guest teardown

ISSUE DESCRIPTION Certain data accessible via hypercalls by the domain controlling the execution of a HVM domain is being freed prematurely, leading to the respective memory regions to possibly be read from and written to in ways unexpected by their new owners. IMPACT Malicious or buggy stub doma...

Insufficient bounding of "REP MOVS" to MMIO emulated inside the hypervisor

ISSUE DESCRIPTION Acceleration support for the "REP MOVS" instruction, when the first iteration accesses memory mapped I/O emulated internally in the hypervisor, incorrectly assumes that the whole range accessed is handled by the same hypervisor sub-component. IMPACT A buggy or malicious HVM gues...

hugetlbfs use may crash PV Linux guests

ISSUE DESCRIPTION Huge 2Mb pages are generally unavailable to PV guests. Since x86 Linux pvops-based kernels are generally multi purpose, they would normally be built with hugetlbfs support enabled. Use of that functionality by an application in a PV guest would cause an infinite page fault loop,...

Linux stack corruption in xen_failsafe_callback for 32bit PVOPS guests.

ISSUE DESCRIPTION xenfailsafecallback incorrectly sets up its stack if an iret fault is injected by the hypervisor. IMPACT Malicious or buggy unprivileged userspace can cause the guest kernel to crash, or operate erroneously. VULNERABLE SYSTEMS All 32bit PVOPS versions of Linux are affected, sinc...

x86 segment base write emulation lacking canonical address checks

ISSUE DESCRIPTION Both writes to the FS and GS register base MSRs as well as the WRFSBASE and WRGSBASE instructions require their input values to be canonical, or a GP fault will be raised. When the use of those instructions by the hypervisor was enabled, the previous guard against GP faults havi...

x86/AMD: Divide speculative information leak

ISSUE DESCRIPTION In the Zen1 microarchitecure, there is one divider in the pipeline which services uops from both threads. In the case of DE, the latched result from the previous DIV to execute will be forwarded speculatively. This is a covert channel that allows two threads to communicate witho...

Linux: blkback driver may leak persistent grants

ISSUE DESCRIPTION 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...

information leak via gnttab_setup_table on ARM

ISSUE DESCRIPTION When initialising an internal data structure on ARM platform Xen was not correctly initialising the memory containing the list of a domain's grant table pages. This list is returned by the GNTTABOPsetuptable subhypercall, leading to an information leak. IMPACT Malicious guest...

HVMOP_set_mem_type allows invalid P2M entries to be created

ISSUE DESCRIPTION The implementation in Xen of the HVMOPsetmemtype HVM control operations attempts to exclude transitioning a page from an inappropriate memory type. However, only an inadequate subset of memory types is excluded. There are certain other types that don't correspond to a particular...

Host crash due to guest VMX instruction execution

ISSUE DESCRIPTION Permission checks on the emulation paths intended for guests using nested virtualization for VMLAUNCH and VMRESUME were deferred too much. The hypervisor would try to use internal state which is not set up unless nested virtualization is actually enabled for a guest. IMPACT A...

several HVM operations do not validate the range of their inputs

ISSUE DESCRIPTION Several HVM control operations do not check the size of their inputs and can tie up a physical CPU for extended periods of time. In addition dirty video RAM tracking involves clearing the bitmap provided by the domain controlling the guest e.g. dom0 or a stubdom. If the size of...

network backend may cause Linux netfront to use freed SKBs

ISSUE DESCRIPTION While adding logic to support XDP eXpress Data Path, a code label was moved in a way allowing for SKBs having references pointers retained for further processing to nevertheless be freed. IMPACT A misbehaving or malicious backend may cause a Denial of Service DoS in the guest...

find_next_bit() issues

ISSUE DESCRIPTION In a number of places bitmaps are being used by the hypervisor to track certain state. Iteration over all bits involves functions which may misbehave in certain corner cases: - On 32-bit Arm accesses to bitmaps with bit a count which is a multiple of 32, an out of bounds access...

Device quarantine for alternate pci assignment methods

ISSUE DESCRIPTION XSA-302 relies on the use of libxl's "assignable-add" feature to prepare devices to be assigned to untrusted guests. Unfortunately, this is not considered a strictly required step for device assignment. The PCI passthrough documentation on the wiki describes alternate ways of...

xl command line config handling stack overflow

ISSUE DESCRIPTION The xl command line utility mishandles long configuration values when passed as command line arguments, with a buffer overrun. VULNERABLE SYSTEMS Systems built on top of xl which pass laundered or checked but otherwise untrusted configuration values onto xl's command line, witho...

Hypercalls exposed to privilege rings 1 and 2 of HVM guests

ISSUE DESCRIPTION The privilege check applied to hypercall attempts by a HVM guest only refused access from ring 3; rings 1 and 2 were allowed through. IMPACT Code running in the intermediate privilege rings of HVM guest OSes may be able to elevate its privileges inside the guest by careful...

heap buffer overflow vulnerability in pcnet emulator

ISSUE DESCRIPTION The QEMU security team has predisclosed the following advisory: The AMD PC-Net II emulatorhw/net/pcnet.c, while receiving packets in loopback mode, appends CRC code to the receive buffer. If the data size given is same as the buffer size4096, the appended CRC code overwrites 4...

arm: various unimplemented hypercalls log without rate limiting

ISSUE DESCRIPTION The HYPERVISORphysdevop hypercall and most suboperations of the HYPERVISORhvmop hypercall are not currently implemented by Xen on ARM and when called will log the use to the hypervisor console. However these guest accessible log messages are not rate-limited. IMPACT A malicious...

HVM qemu unexpectedly enabling emulated VGA graphics backends

ISSUE DESCRIPTION When instantiating an emulated VGA device for an x86 HVM guest qemu will by default enable a backend to expose that device, either SDL or VNC depending on the version of qemu and the build time configuration. The libxl toolstack library does not explicitly disable these default...

Potential use of freed memory in event channel operations

ISSUE DESCRIPTION Wrong ordering of operations upon extending the per-domain event channel tracking table can cause a pointer to freed memory to be left in place, when the hypervisor is under memory pressure and XSM Xen Security Module is enabled. IMPACT Malicious guest kernels could inject...

VCPUOP_initialise DoS

ISSUE DESCRIPTION hypercallcreatecontinuation is a variadic function which uses a printf-like format string to interpret its parameters. Error handling for a bad format character was done using BUG, which crashes Xen. One path, via the VCPUOPinitialise hypercall, has a bad format character. The B...

Unsanitised guest input in libxl device handling code

ISSUE DESCRIPTION Various parts of libxl device-handling code inappropriately use information from partially guest controlled areas of xenstore principally the frontend directory /local/domain/GUEST/device/TYPE/DEVID, henceforth referred to as FE. The problems vary by device type: For almost all...

x86: Long latency populate-on-demand operation is not preemptible

ISSUE DESCRIPTION When running an HVM domain in Populate-on-Demand mode, Xen would sometimes search the domain for memory to reclaim, in response to demands for population of other pages in the same domain. This search runs without preemption. The guest can, by suitable arrangement of its memory...