arista CVE Vulnerabilities & Metrics

Arista NG Firewall ReportEntry SQL Injection Arbitrary File Read and Write Vulnerability. This vulnerability allows remote attackers to create arbitrary files and disclose sensitive information on affected installations of Arista NG Firewall. Authentication is required to exploit this vulnerability. The specific flaw exists within the ReportEntry class. The issue results from the lack of proper validation of a user-supplied string before using it to construct SQL queries. An attacker can leverage this in conjunction with other vulnerabilities to execute arbitrary code in the context of the www-data user. Was ZDI-CAN-24325.

Arista NG Firewall uvm_login Incorrect Authorization Privilege Escalation Vulnerability. This vulnerability allows local attackers to escalate privileges on affected installations of Arista NG Firewall. An attacker must first obtain the ability to execute low-privileged code on the target system in order to exploit this vulnerability. The specific flaw exists within the uvm_login module. The issue results from incorrect authorization. An attacker can leverage this to escalate privileges to resources normally protected from the user. Was ZDI-CAN-24324.

Arista NG Firewall custom_handler Directory Traversal Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Arista NG Firewall. Authentication is not required to exploit this vulnerability. The specific flaw exists within the implementation of the custom_handler method. The issue results from the lack of proper validation of a user-supplied path prior to using it in file operations. An attacker can leverage this vulnerability to execute code in the context of the www-data user. Was ZDI-CAN-24019.

Arista NG Firewall ExecManagerImpl Command Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Arista NG Firewall. Authentication is required to exploit this vulnerability. The specific flaw exists within the ExecManagerImpl class. The issue results from the lack of proper validation of a user-supplied string before using it to execute a system call. An attacker can leverage this vulnerability to execute code in the context of root. Was ZDI-CAN-24015.

On affected platforms running Arista MOS, the configuration of a BGP password will cause the password to be logged in clear text that can be revealed in local logs or remote logging servers by authenticated users, as well as appear in clear text in the device’s running config.

On affected platforms running Arista EOS with mirroring to multiple destinations configured, an internal system error may trigger a kernel panic and cause system reload.

On affected platforms running Arista EOS with VXLAN configured, malformed or truncated packets received over a VXLAN tunnel and forwarded in hardware can cause egress ports to be unable to forward packets. The device will continue to be susceptible to the issue until remediation is in place.

On affected versions of the CloudVision Portal improper access controls on the connection from devices to CloudVision could enable a malicious actor with network access to CloudVision to get broader access to telemetry and configuration data within the system than intended. This advisory impacts the Arista CloudVision Portal product when run on-premise. It does not impact CloudVision as-a-Service.

On the affected platforms running EOS, a malformed DHCP packet might cause the DHCP relay agent to restart.

On affected platforms running Arista EOS, an authorized attacker with permissions to perform gNMI requests could craft a request allowing it to update arbitrary configurations in the switch. This situation occurs only when the Streaming Telemetry Agent (referred to as the TerminAttr agent) is enabled and gNMI access is configured on the agent. Note: This gNMI over the Streaming Telemetry Agent scenario is mostly commonly used when streaming to a 3rd party system and is not used by default when streaming to CloudVision

On affected modular platforms running Arista EOS equipped with both redundant supervisor modules and having the redundancy protocol configured with RPR or SSO, an existing unprivileged user can login to the standby supervisor as a root user, leading to a privilege escalation. Valid user credentials are required in order to exploit this vulnerability.

On affected platforms running Arista EOS with SNMP configured, a specially crafted packet can cause a memory leak in the snmpd process. This may result in the snmpd processing being terminated (causing SNMP requests to time out until snmpd is automatically restarted) and potential memory resource exhaustion for other processes on the switch. The vulnerability does not have any confidentiality or integrity impacts to the system.

On affected platforms running Arista CloudEOS an issue in the Software Forwarding Engine (Sfe) can lead to a potential denial of service attack by sending malformed packets to the switch. This causes a leak of packet buffers and if enough malformed packets are received, the switch may eventually stop forwarding traffic.

For certain systems running EOS, a Precision Time Protocol (PTP) packet of a management/signaling message with an invalid Type-Length-Value (TLV) causes the PTP agent to restart. Repeated restarts of the service will make the service unavailable.

This advisory documents an internally found vulnerability in the on premises deployment model of Arista CloudVision Portal (CVP) where under a certain set of conditions, user passwords can be leaked in the Audit and System logs. The impact of this vulnerability is that the CVP user login passwords might be leaked to other authenticated users.

This advisory documents the impact of an internally found vulnerability in Arista EOS for security ACL bypass. The impact of this vulnerability is that the security ACL drop rule might be bypassed if a NAT ACL rule filter with permit action matches the packet flow. This could allow a host with an IP address in a range that matches the range allowed by a NAT ACL and a range denied by a Security ACL to be forwarded incorrectly as it should have been denied by the Security ACL. This can enable an ACL bypass.

This advisory documents the impact of an internally found vulnerability in Arista EOS state streaming telemetry agent TerminAttr and OpenConfig transport protocols. The impact of this vulnerability is that, in certain conditions, TerminAttr might leak MACsec sensitive data in clear text in CVP to other authorized users, which could cause MACsec traffic to be decrypted or modified by other authorized users on the device.

This advisory documents the impact of an internally found vulnerability in Arista EOS state streaming telemetry agent TerminAttr and OpenConfig transport protocols. The impact of this vulnerability is that, in certain conditions, TerminAttr might leak IPsec sensitive data in clear text in CVP to other authorized users, which could cause IPsec traffic to be decrypted or modified by other authorized users on the device.

On affected Arista EOS platforms, if a VXLAN match rule exists in an IPv4 access-list that is applied to the ingress of an L2 or an L3 port/SVI, the VXLAN rule and subsequent ACL rules in that access list will ignore the specified IP protocol.

On Arista Strata family products which have “TCAM profile” feature enabled when Port IPv4 access-list has a rule which matches on “vxlan” as protocol then that rule and subsequent rules ( rules declared after it in ACL ) do not match on IP protocol field as expected.

The impact of this vulnerability is that Arista's EOS eAPI may skip re-evaluating user credentials when certificate based authentication is used, which allows remote attackers to access the device via eAPI.

An issue has recently been discovered in Arista EOS where, under certain conditions, the service ACL configured for OpenConfig gNOI and OpenConfig RESTCONF might be bypassed, which results in the denied requests being forwarded to the agent.

An issue has recently been discovered in Arista EOS where certain gNOI APIs incorrectly skip authorization and authentication which could potentially allow a factory reset of the device.

An issue has recently been discovered in Arista EOS where the incorrect use of EOS's AAA API’s by the OpenConfig and TerminAttr agents could result in unrestricted access to the device for local users with nopassword configuration.

An issue has recently been discovered in Arista EOS where the incorrect use of EOS's AAA API’s by the OpenConfig and TerminAttr agents could result in unrestricted access to the device for local users with nopassword configuration.

On systems running Arista EOS and CloudEOS with the affected release version, when using shared secret profiles the password configured for use by BiDirectional Forwarding Detection (BFD) will be leaked when displaying output over eAPI or other JSON outputs to other authenticated users on the device. The affected EOS Versions are: all releases in 4.22.x train, 4.23.9 and below releases in the 4.23.x train, 4.24.7 and below releases in the 4.24.x train, 4.25.4 and below releases in the 4.25.x train, 4.26.1 and below releases in the 4.26.x train

An issue was discovered in the Linux kernel 5.8.9. The WEP, WPA, WPA2, and WPA3 implementations reassemble fragments even though some of them were sent in plaintext. This vulnerability can be abused to inject packets and/or exfiltrate selected fragments when another device sends fragmented frames and the WEP, CCMP, or GCMP data-confidentiality protocol is used.

An issue was discovered on Samsung Galaxy S3 i9305 4.4.4 devices. The WPA, WPA2, and WPA3 implementations reassemble fragments with non-consecutive packet numbers. An adversary can abuse this to exfiltrate selected fragments. This vulnerability is exploitable when another device sends fragmented frames and the WEP, CCMP, or GCMP data-confidentiality protocol is used. Note that WEP is vulnerable to this attack by design.

An issue was discovered on Samsung Galaxy S3 i9305 4.4.4 devices. The WEP, WPA, WPA2, and WPA3 implementations accept plaintext A-MSDU frames as long as the first 8 bytes correspond to a valid RFC1042 (i.e., LLC/SNAP) header for EAPOL. An adversary can abuse this to inject arbitrary network packets independent of the network configuration.

An issue was discovered in the ALFA Windows 10 driver 1030.36.604 for AWUS036ACH. The WEP, WPA, WPA2, and WPA3 implementations accept fragmented plaintext frames in a protected Wi-Fi network. An adversary can abuse this to inject arbitrary data frames independent of the network configuration.

An issue was discovered in the ALFA Windows 10 driver 6.1316.1209 for AWUS036H. The WEP, WPA, WPA2, and WPA3 implementations accept plaintext frames in a protected Wi-Fi network. An adversary can abuse this to inject arbitrary data frames independent of the network configuration.

An issue was discovered in the kernel in NetBSD 7.1. An Access Point (AP) forwards EAPOL frames to other clients even though the sender has not yet successfully authenticated to the AP. This might be abused in projected Wi-Fi networks to launch denial-of-service attacks against connected clients and makes it easier to exploit other vulnerabilities in connected clients.

The 802.11 standard that underpins Wi-Fi Protected Access (WPA, WPA2, and WPA3) and Wired Equivalent Privacy (WEP) doesn't require that the A-MSDU flag in the plaintext QoS header field is authenticated. Against devices that support receiving non-SSP A-MSDU frames (which is mandatory as part of 802.11n), an adversary can abuse this to inject arbitrary network packets.

The 802.11 standard that underpins Wi-Fi Protected Access (WPA, WPA2, and WPA3) and Wired Equivalent Privacy (WEP) doesn't require that all fragments of a frame are encrypted under the same key. An adversary can abuse this to decrypt selected fragments when another device sends fragmented frames and the WEP, CCMP, or GCMP encryption key is periodically renewed.

The 802.11 standard that underpins Wi-Fi Protected Access (WPA, WPA2, and WPA3) and Wired Equivalent Privacy (WEP) doesn't require that received fragments be cleared from memory after (re)connecting to a network. Under the right circumstances, when another device sends fragmented frames encrypted using WEP, CCMP, or GCMP, this can be abused to inject arbitrary network packets and/or exfiltrate user data.

A flaw was found in dnsmasq before version 2.83. When receiving a query, dnsmasq does not check for an existing pending request for the same name and forwards a new request. By default, a maximum of 150 pending queries can be sent to upstream servers, so there can be at most 150 queries for the same name. This flaw allows an off-path attacker on the network to substantially reduce the number of attempts that it would have to perform to forge a reply and have it accepted by dnsmasq. This issue is mentioned in the "Birthday Attacks" section of RFC5452. If chained with CVE-2020-25684, the attack complexity of a successful attack is reduced. The highest threat from this vulnerability is to data integrity.

A flaw was found in dnsmasq before version 2.83. When getting a reply from a forwarded query, dnsmasq checks in forward.c:reply_query(), which is the forwarded query that matches the reply, by only using a weak hash of the query name. Due to the weak hash (CRC32 when dnsmasq is compiled without DNSSEC, SHA-1 when it is) this flaw allows an off-path attacker to find several different domains all having the same hash, substantially reducing the number of attempts they would have to perform to forge a reply and get it accepted by dnsmasq. This is in contrast with RFC5452, which specifies that the query name is one of the attributes of a query that must be used to match a reply. This flaw could be abused to perform a DNS Cache Poisoning attack. If chained with CVE-2020-25684 the attack complexity of a successful attack is reduced. The highest threat from this vulnerability is to data integrity.

A flaw was found in dnsmasq before version 2.83. When getting a reply from a forwarded query, dnsmasq checks in the forward.c:reply_query() if the reply destination address/port is used by the pending forwarded queries. However, it does not use the address/port to retrieve the exact forwarded query, substantially reducing the number of attempts an attacker on the network would have to perform to forge a reply and get it accepted by dnsmasq. This issue contrasts with RFC5452, which specifies a query's attributes that all must be used to match a reply. This flaw allows an attacker to perform a DNS Cache Poisoning attack. If chained with CVE-2020-25685 or CVE-2020-25686, the attack complexity of a successful attack is reduced. The highest threat from this vulnerability is to data integrity.

An issue with ARP packets in Arista’s EOS affecting the 7800R3, 7500R3, and 7280R3 series of products may result in issues that cause a kernel crash, followed by a device reload. The affected Arista EOS versions are: 4.24.2.4F and below releases in the 4.24.x train; 4.23.4M and below releases in the 4.23.x train; 4.22.6M and below releases in the 4.22.x train.

In Arista EOS malformed packets can be incorrectly forwarded across VLAN boundaries in one direction. This vulnerability is only susceptible to exploitation by unidirectional traffic (ex. UDP) and not bidirectional traffic (ex. TCP). This affects: EOS 7170 platforms version 4.21.4.1F and below releases in the 4.21.x train; EOS X-Series versions 4.21.11M and below releases in the 4.21.x train; 4.22.6M and below releases in the 4.22.x train; 4.23.4M and below releases in the 4.23.x train; 4.24.2.1F and below releases in the 4.24.x train.

In EVPN VxLAN setups in Arista EOS, specific malformed packets can lead to incorrect MAC to IP bindings and as a result packets can be incorrectly forwarded across VLAN boundaries. This can result in traffic being discarded on the receiving VLAN. This affects versions: 4.21.12M and below releases in the 4.21.x train; 4.22.7M and below releases in the 4.22.x train; 4.23.5M and below releases in the 4.23.x train; 4.24.2F and below releases in the 4.24.x train.

Arista EOS before 4.21.12M, 4.22.x before 4.22.7M, 4.23.x before 4.23.5M, and 4.24.x before 4.24.2F allows remote attackers to cause traffic loss or incorrect forwarding of traffic via a malformed link-state PDU to the IS-IS router.

Arista’s CloudVision eXchange (CVX) server before 4.21.12M, 4.22.x before 4.22.7M, 4.23.x before 4.23.5M, and 4.24.x before 4.24.2F allows remote attackers to cause a denial of service (crash and restart) in the ControllerOob agent via a malformed control-plane packet.

Arista EOS before 4.21.12M, 4.22.x before 4.22.7M, 4.23.x before 4.23.5M, and 4.24.x before 4.24.2F allows remote attackers to cause a denial of service (restart of agents) by crafting a malformed DHCP packet which leads to an incorrect route being installed.

A vulnerability in Arista’s CloudVision Portal (CVP) prior to 2020.2 allows users with “read-only” or greater access rights to the Configlet Management module to download files not intended for access, located on the CVP server, by accessing a specific API.

A vulnerability exists in Arista’s Cloud EOS VM / vEOS 4.23.2M and below releases in the 4.23.x train, 4.22.4M and below releases in the 4.22.x train, 4.21.3M to 4.21.9M releases in the 4.21.x train, 4.21.3FX-7368.*, 4.21.4-FCRFX.*, 4.21.4.1, 4.21.7.1, 4.22.2.0.1, 4.22.2.2.1, 4.22.3.1, and 4.23.2.1 Router code in a scenario where TCP MSS options are configured.

In support.c in pam_tacplus 1.3.8 through 1.5.1, the TACACS+ shared secret gets logged via syslog if the DEBUG loglevel and journald are used.

An issue was found in Arista EOS. Specific malformed ARP packets can impact the software forwarding of VxLAN packets. This issue is found in Arista’s EOS VxLAN code, which can allow attackers to crash the VxlanSwFwd agent. This affects EOS 4.21.8M and below releases in the 4.21.x train, 4.22.3M and below releases in the 4.22.x train, 4.23.1F and below releases in the 4.23.x train, and all releases in 4.15, 4.16, 4.17, 4.18, 4.19, 4.20 code train.

utility.c in telnetd in netkit telnet through 0.17 allows remote attackers to execute arbitrary code via short writes or urgent data, because of a buffer overflow involving the netclear and nextitem functions.

Arista DCS-7050QX-32S-R 4.20.9M, DCS-7050CX3-32S-R 4.20.11M, and DCS-7280SRAM-48C6-R 4.22.0.1F devices (and possibly other products) allow attackers to bypass intended TACACS+ shell restrictions via a | character. NOTE: the vendor reports that this is a configuration issue relating to an overly permissive regular expression in the TACACS+ server permitted commands

The process_tx_desc function in hw/net/e1000.c in QEMU before 2.4.0.1 does not properly process transmit descriptor data when sending a network packet, which allows attackers to cause a denial of service (infinite loop and guest crash) via unspecified vectors.

Buffer overflow in the send_control_msg function in hw/char/virtio-serial-bus.c in QEMU before 2.4.0 allows guest users to cause a denial of service (QEMU process crash) via a crafted virtio control message.

The ne2000_receive function in hw/net/ne2000.c in QEMU before 2.4.0.1 allows attackers to cause a denial of service (infinite loop and instance crash) or possibly execute arbitrary code via vectors related to receiving packets.

Integer overflow in the VNC display driver in QEMU before 2.1.0 allows attachers to cause a denial of service (process crash) via a CLIENT_CUT_TEXT message, which triggers an infinite loop.

In CloudVision Portal all releases in the 2018.1 and 2018.2 Code train allows users with read-only permissions to bypass permissions for restricted functionality via CVP API calls through the Configlet Builder modules. This vulnerability can potentially enable authenticated users with read-only access to take actions that are otherwise restricted in the GUI.

In CloudVision Portal (CVP) for all releases in the 2018.2 Train, under certain conditions, the application logs user passwords in plain text for certain API calls, potentially leading to user password exposure. This only affects CVP environments where: 1. Devices have enable mode passwords which are different from the user's login password, OR 2. There are configlet builders that use the Device class and specify username and password explicitly Application logs are not accessible or visible from the CVP GUI. Application logs can only be read by authorized users with privileged access to the VM hosting the CVP application.

Go before 1.12.11 and 1.3.x before 1.13.2 can panic upon an attempt to process network traffic containing an invalid DSA public key. There are several attack scenarios, such as traffic from a client to a server that verifies client certificates.

A vulnerability has been found in the implementation of the Label Distribution Protocol (LDP) protocol in EOS. Under race conditions, the LDP agent can establish an LDP session with a malicious peer potentially allowing the possibility of a Denial of Service (DoS) attack on route updates and in turn potentially leading to an Out of Memory (OOM) condition that is disruptive to traffic forwarding. Affected EOS versions include: 4.22 release train: 4.22.1F and earlier releases 4.21 release train: 4.21.0F - 4.21.2.3F, 4.21.3F - 4.21.7.1M 4.20 release train: 4.20.14M and earlier releases 4.19 release train: 4.19.12M and earlier releases End of support release trains (4.18 and 4.17)

Arista EOS through 4.21.0F allows a crash because 802.1x authentication is mishandled.

Arista CloudVision Portal through 2018.1.1 has Incorrect Permissions.

Arista EOS before 4.20.2F allows remote BGP peers to cause a denial of service (Rib agent restart) via a malformed path attribute in an UPDATE message.

The Mlag agent in Arista EOS 4.19 before 4.19.4M and 4.20 before 4.20.2F allows remote attackers to cause a denial of service (agent restart) via crafted UDP packets.

The tcpmss_mangle_packet function in net/netfilter/xt_TCPMSS.c in the Linux kernel before 4.11, and 4.9.x before 4.9.36, allows remote attackers to cause a denial of service (use-after-free and memory corruption) or possibly have unspecified other impact by leveraging the presence of xt_TCPMSS in an iptables action.

Heap-based buffer overflow in dnsmasq before 2.78 allows remote attackers to cause a denial of service (crash) or execute arbitrary code via a crafted DNS response.

CloudVision Portal (CVP) before 2016.1.2.1 allows remote authenticated users to gain access to the internal configuration mechanisms via the management plane, related to a request to /web/system/console/bundle.

Arista EOS 4.15 before 4.15.8M, 4.16 before 4.16.7M, and 4.17 before 4.17.0F on DCS-7050 series devices allow remote attackers to cause a denial of service (device reboot) by sending crafted packets to the control plane.

Arista EOS before 4.11.12, 4.12 before 4.12.11, 4.13 before 4.13.14M, 4.14 before 4.14.5FX.5, and 4.15 before 4.15.0FX1.1 allows remote attackers to execute arbitrary code as root by leveraging management-plane access, aka Bug 138716.

hw/ide/core.c in QEMU does not properly restrict the commands accepted by an ATAPI device, which allows guest users to cause a denial of service or possibly have unspecified other impact via certain IDE commands, as demonstrated by a WIN_READ_NATIVE_MAX command to an empty drive, which triggers a divide-by-zero error and instance crash.

The pit_ioport_read in i8254.c in the Linux kernel before 2.6.33 and QEMU before 2.3.1 does not distinguish between read lengths and write lengths, which might allow guest OS users to execute arbitrary code on the host OS by triggering use of an invalid index.

The C+ mode offload emulation in the RTL8139 network card device model in QEMU, as used in Xen 4.5.x and earlier, allows remote attackers to read process heap memory via unspecified vectors.

Heap-based buffer overflow in the PCNET controller in QEMU allows remote attackers to execute arbitrary code by sending a packet with TXSTATUS_STARTPACKET set and then a crafted packet with TXSTATUS_DEVICEOWNS set.

GNU Bash through 4.3 bash43-025 processes trailing strings after certain malformed function definitions in the values of environment variables, which allows remote attackers to write to files or possibly have unknown other impact via a crafted environment, as demonstrated by vectors involving the ForceCommand feature in OpenSSH sshd, the mod_cgi and mod_cgid modules in the Apache HTTP Server, scripts executed by unspecified DHCP clients, and other situations in which setting the environment occurs across a privilege boundary from Bash execution. NOTE: this vulnerability exists because of an incomplete fix for CVE-2014-6271.

GNU Bash through 4.3 processes trailing strings after function definitions in the values of environment variables, which allows remote attackers to execute arbitrary code via a crafted environment, as demonstrated by vectors involving the ForceCommand feature in OpenSSH sshd, the mod_cgi and mod_cgid modules in the Apache HTTP Server, scripts executed by unspecified DHCP clients, and other situations in which setting the environment occurs across a privilege boundary from Bash execution, aka "ShellShock." NOTE: the original fix for this issue was incorrect; CVE-2014-7169 has been assigned to cover the vulnerability that is still present after the incorrect fix.