NVIDIA DOCA App Shield Agent Application Guide

1.0

This guide provides process introspection system implementation on top of NVIDIA® BlueField® DPU.

App Shield Agent monitors a process in the host system using the DOCA App Shield library.

This security capability helps identify corruption of core processes in the system from an independent and trusted DPU. This is a major and innovate intrusion detection system (IDS) ability since it cannot be provided from inside the host.

The DOCA App Shield Library gives the capability to read, analyze, and authenticate the host (bare metal/VM) memory directly from the DPU.

Using the library, this application hashes the un-writeable memory pages (also unloaded pages) of a specific process and its libraries. Then, at regular intervals, the app authenticates the loaded pages.

The app reports pass/fail after every iteration until the first attestation failure. The reports are both printed to the console and exported to the DOCA Telemetry Service (DTS) using inter-process communication (IPC).

This guide describes how to build secure process monitoring using the DOCA App Shield library, which leverages the DPU's advantages such as hardware-based DMA, integrity, and more.

The App Shield agent is designed to run independently on the DPU's Arm without hindering the host.

The host's involvement is limited to configuring monitoring of a new process when there is a need to generate the needed ZIP and JSON files to pass to the DPU. This is done at inception ("time 0") which is when the host is still in a "safe" state.

Generating the needed files can be done by running DOCA App Shield's doca_apsh_config.py tool on the host. See DOCA App Shield for more info.

high-level-diagram-version-1-modificationdate-1707756008163-api-v2.png

The user creates three mandatory files using the DOCA tool doca_apsh_config.py and copies them to the DPU. The application can report attestation results to the:

  • File

  • Terminal

  • DTS

app-shield-arch-version-1-modificationdate-1707756008450-api-v2.png

  1. The files are generated by running doca_apsh_config.py on the host against the process at time zero.

    Note

    The actions 2-5 recur at regular time intervals.

  2. The App Shield agent requests new attestation from DOCA App Shield library.

  3. The DOCA App Shield library creates a new attestation:

    1. Scans and hashes process memory pages (that are currently in use).

    2. Compares the hash to the original hash.

    3. Creates attestation for each lib/exe involved in the process. Each of attestation includes the number of valid pages and the number of pages.

  4. The App Shield agent searches each attestation for inconsistency between number of used pages and number of valid pages.

  5. The App Shield agent reports results with a timestamp and scan count to:

    1. Local telemetry files – a folder and files representing the data a real DTS would have received. These files are used for the purposes of this example only as normally this data is not exported into user-readable files.

    2. DOCA log (without scan count).

    3. DTS IPC interface (even if no DTS is active).

  6. The App Shield agent exits on first attestation failure.

This application leverages the following DOCA libraries:

Refer to their respective programming guide for more information.

Info

Please refer to the NVIDIA DOCA Installation Guide for Linux for details on how to install BlueField-related software.

The installation of DOCA's reference applications contains the sources of the applications, alongside the matching compilation instructions. This allows for compiling the applications "as-is" and provides the ability to modify the sources, then compile a new version of the application.

Tip

For more information about the applications as well as development and compilation tips, refer to the DOCA Applications page.

The sources of the application can be found under the application's directory: /opt/mellanox/doca/applications/app_shield_agent/.

Compiling All Applications

All DOCA applications are defined under a single meson project. So, by default, the compilation includes all of them.

To build all the applications together, run:

Copy
Copied!
            

cd /opt/mellanox/doca/applications/ meson /tmp/build ninja -C /tmp/build

Info

doca_app_shield_agent is created under /tmp/build/app_shield_agent/.


Compiling Only the Current Application

To build only the App Shield Agent application:

Copy
Copied!
            

cd /opt/mellanox/doca/applications/ meson /tmp/build -Denable_all_applications=false -Denable_app_shield_agent=true ninja -C /tmp/build

Info

doca_app_shield_agent is created under /tmp/build/app_shield_agent/.

Alternatively, the user can set the desired flags in the meson_options.txt file instead of providing them in the compilation command line:

  1. Edit the following flags in /opt/mellanox/doca/applications/meson_options.txt:

    • Set enable_all_applications to false

    • Set enable_app_shield_agent to true

  2. Run the following compilation commands :

    Copy
    Copied!
                

    cd /opt/mellanox/doca/applications/ meson /tmp/build ninja -C /tmp/build

    Info

    doca_app_shield_agent is created under /tmp/build/app_shield_agent/.

Troubleshooting

Refer to the NVIDIA DOCA Troubleshooting Guide for any issue encountered with the compilation of the application .

Prerequisites

  1. Configure the BlueField's firmware.

    1. On the BlueField system, configure the PF base address register and NVMe emulation. Run:

      Copy
      Copied!
                  

      dpu> mlxconfig -d /dev/mst/mt41686_pciconf0 s PF_BAR2_SIZE=2 PF_BAR2_ENABLE=1 NVME_EMULATION_ENABLE=1 

    2. Perform a BlueField system reboot for the mlxconfig settings to take effect.

    3. You may verify these configurations using the following command:

      Copy
      Copied!
                  

      dpu> mlxconfig -d /dev/mst/mt41686_pciconf0 q | grep -E "NVME|BAR"

  2. Download target system (host/VM) symbols.

    • For Ubuntu:

      Copy
      Copied!
                  

      host> sudo tee /etc/apt/sources.list.d/ddebs.list << EOF deb http://ddebs.ubuntu.com/ $(lsb_release -cs) main restricted universe multiverse deb http://ddebs.ubuntu.com/ $(lsb_release -cs)-updates main restricted universe multiverse deb http://ddebs.ubuntu.com/ $(lsb_release -cs)-proposed main restricted universe multiverse EOF host> sudo apt install ubuntu-dbgsym-keyring host> sudo apt-get update host> sudo apt-get install linux-image-$(uname -r)-dbgsym

    • For CentOS:

      Copy
      Copied!
                  

      host> yum install --enablerepo=base-debuginfo kernel-devel-$(uname -r) kernel-debuginfo-$(uname -r) kernel-debuginfo-common-$(uname -m)-$(uname -r)

    • No action is needed for Windows

  3. Perform IOMMU passthrough. This stage is only necessary if IOMMU is not enabled by default (e.g., when the host is using an AMD CPU).

    Note

    Skip this step if you are not sure whether it is needed. Return to it only if DMA fails with a message similar to the following in dmesg:

    Copy
    Copied!
                

    host> dmesg [ 3839.822897] mlx5_core 0000:81:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0047 address=0x2a0aff8 flags=0x0000]

    1. Locate your OS's grub file (most likely /boot/grub/grub.conf, /boot/grub2/grub.cfg, or /etc/default/grub) and open it for editing. Run:

      Copy
      Copied!
                  

      host> vim /etc/default/grub

    2. Search for the line defining GRUB_CMDLINE_LINUX_DEFAULT and add the argument iommu=pt. For example:

      Copy
      Copied!
                  

      GRUB_CMDLINE_LINUX_DEFAULT="iommu=pt <intel/amd>_iommu=on"

    3. Run:

      Note

      Prior to performing a power cycle, make sure to do a graceful shutdown.

      • For Ubuntu:

        Copy
        Copied!
                    

        host> sudo update-grub host> ipmitool power cycle

      • For CentOS:

        Copy
        Copied!
                    

        host> grub2-mkconfig -o /boot/grub2/grub.cfg host> ipmitool power cycle

      • For Windows targets, turn off Hyper-V capability.

  4. Prepare target:

    1. Install DOCA on the target system.

    2. Create the ZIP and JSON files. Run:

      Copy
      Copied!
                  

      target-system> cd /opt/mellanox/doca/tools/ target-system> python3 doca_apsh_config.py --pid <pid-of-process-to-monitor> --os <windows/linux> --path <path to dwarf2json executable or pdbparse-to-json.py> target-system> cp /opt/mellanox/doca/tools/*.* <shared-folder-with-baremetal> dpu> scp <shared-folder-with-baremetal>/* <path-to-app-shield-binary>

      If the target system does not have DOCA installed, the script can be copied from the BlueField.

      The required dwaf2json and pdbparse-to-json.py are not provided with DOCA.

      Note

      If the kernel and process .exe have not changed, there no need to redo this step.

Application Execution

  1. The App Shield Agent application is provided in source form, hence a compilation is required before the application can be executed.

    1. Application usage instructions:

      Copy
      Copied!
                  

      Usage: doca_app_shield_agent [DOCA Flags] [Program Flags]   DOCA Flags: -h, --help Print a help synopsis -v, --version Print program version information -l, --log-level Set the (numeric) log level for the program <10=DISABLE, 20=CRITICAL, 30=ERROR, 40=WARNING, 50=INFO, 60=DEBUG, 70=TRACE> --sdk-log-level Set the SDK (numeric) log level for the program <10=DISABLE, 20=CRITICAL, 30=ERROR, 40=WARNING, 50=INFO, 60=DEBUG, 70=TRACE> -j, --json <path> Parse all command flags from an input json file   Program Flags: -p, --pid Process ID of process to be attested -e, --ehm <path> Exec hash map path -m, --memr <path> System memory regions map -f, --vuid VUID of the System device -d, --dma DMA device name -o, --osym <path> System OS symbol map path -s, --osty <windows|linux> System OS type - windows/linux -t, --time <seconds> Scan time interval in seconds

      Info

      This usage printout can be printed to the command line using the -h (or --help) options:

      Copy
      Copied!
                  

      ./doca_app_shield_agent -h

      Info

      For additional information, please refer to section "Command Line Flags".

    2. CLI example for running the application on the BlueField:

      Copy
      Copied!
                  

      ./doca_app_shield_agent -p 13577 -e hash.zip -m mem_regions.json -o symbols.json -f MT2125X03335MLNXS0D0F0VF1 -d mlx5_0 -t 3 -s linux

      Note

      All used identifiers (-f , -p and -d flags) should match the identifier of the desired devices and processes.

Command Line Flags

Flag Type

Short Flag

Long Flag

Description

General flags

h

help

Print a help synopsis

v

version

Print program version information

l

log-level

Set the log level for the application:

  • DISABLE=10

  • CRITICAL=20

  • ERROR=30

  • WARNING=40

  • INFO=50

  • DEBUG=60

  • TRACE=70 (requires compilation with TRACE log level support)

N/A

sdk-log-level

Set the log level for the program:

  • DISABLE=10

  • CRITICAL=20

  • ERROR=30

  • WARNING=40

  • INFO=50

  • DEBUG=60

  • TRACE=70

j

json

Parse all command flags from an input JSON file

Program flags

p

pid

PID of the process to be attested

e

ehm

Path to the pre-generated hash.zip file transferred from the host

m

memr

Path to the pre-generated mem_regions.json file transferred from the host

f

pcif

System PCIe function vendor unique identifier (VUID) of the VF/PF exposed to the target system. Used for DMA operations.

To obtain this argument, run:

Copy
Copied!
            

target-system> lspci -vv | grep "\[VU\] Vendor specific:"

Example output:

Copy
Copied!
            

[VU] Vendor specific: MT2125X03335MLNXS0D0F0 [VU] Vendor specific: MT2125X03335MLNXS0D0F1

Two VUIDs are printed for each DPU connected to the target system. The first is of the DPU on pf0 and the second is of the DPU on port pf1.

Note

Running this command on the DPU outputs VUIDs with an additional "EC" string in the middle. You must remove the "EC" to arrive at the correct VUID.

The VUID of a VF allocated on PF0/1 is the VUID of the PF with an additional suffix, VF<vf-number>, where vf-number is the VF index +1.

For example, for the output in the example above:

  • PF0 VUID = MT2125X03335MLNXS0D0F0

  • PF1 VUID = MT2125X03335MLNXS0D0F1

  • VUID of VF0 on PF0 = MT2125X03335MLNXS0D0F0VF1

VUIDs are persistent even on reset.

d

dma

DMA device name to use

o

osym

Path to the pre-generated symbols.json file transferred from the host

s

osty

OS type (windows or linux) of the system where the process is running

t

time

Number of seconds to sleep between scans

Info

Refer to DOCA Arg Parser for more information regarding the supported flags and execution modes.


Troubleshooting

Refer to the NVIDIA DOCA Troubleshooting Guide for any issue encountered with the installation or execution of the DOCA applications.

  1. Parse application argument.

    1. Initialize arg parser resources and register DOCA general parameters.

      Copy
      Copied!
                  

      doca_argp_init();

    2. Register application parameters.

      Copy
      Copied!
                  

      register_apsh_params();

    3. Parse the arguments.

      Copy
      Copied!
                  

      doca_argp_start();

  2. Initialize DOCA App Shield lib context.

    1. Create lib context.

      Copy
      Copied!
                  

      doca_apsh_create();

    2. Set DMA device for lib.

      Copy
      Copied!
                  

      doca_devinfo_list_create(); doca_dev_open(); doca_devinfo_list_destroy(); doca_apsh_dma_dev_set();

    3. Start the context

      Copy
      Copied!
                  

      doca_apsh_start(); apsh_system_init();

  3. Initialize DOCA App Shield lib system context handler.

    1. Get the representor of the remote PCIe function exposed to the system.

      Copy
      Copied!
                  

      doca_devinfo_remote_list_create(); doca_dev_remote_open(); doca_devinfo_remote_list_destroy();

    2. Create and start the system context handler.

      Copy
      Copied!
                  

      doca_apsh_system_create(); doca_apsh_sys_os_symbol_map_set(); doca_apsh_sys_mem_region_set(); doca_apsh_sys_dev_set(); doca_apsh_sys_os_type_set(); doca_apsh_system_start();

  4. Find target process by pid.

    Copy
    Copied!
                

    doca_apsh_processes_get();

  5. Telemetry initialization.

    Copy
    Copied!
                

    telemetry_start();

    1. Initialize a new telemetry schema.

    2. Register attestation type event.

    3. Set up output to file (in addition to default IPC).

    4. Start the telemetry schema.

    5. Initialize and start a new DTS source with the gethostname() name as source ID.

  6. Get initial attestation of the process.

    Copy
    Copied!
                

    doca_apsh_attestation_get();

  7. Loop until attestation validation fail.

    Copy
    Copied!
                

    doca_apsh_attst_refresh(); /* validation logic */ doca_telemetry_source_report(); DOCA_LOG_INFO(); sleep();

  8. DOCA App Shield Agent destroy.

    Copy
    Copied!
                

    doca_apsh_attestation_free(); doca_apsh_processes_free(); doca_apsh_system_destroy(); doca_apsh_destroy(); doca_dev_close(); doca_dev_remote_close();

  9. Telemetry destroy.

    Copy
    Copied!
                

    telemetry_destroy();

  10. Arg parser destroy.

    Copy
    Copied!
                

    doca_argp_destroy();

  • /opt/mellanox/doca/applications/app_shield_agent/

© Copyright 2024, NVIDIA. Last updated on Feb 9, 2024.