---
title: "Discussion on \"Below the Kernel: How to Stop Hardware Bootkits with TPM 2.0 and fwupd\" | Hashnode"
description: "Discussion on \"Below the Kernel: How to Stop Hardware Bootkits with TPM 2.0 and fwupd\". Defeat UEFI bootkits with a Hardware Root of Trust. SRE guide to TPM 2.0, fwupd firmware updates, and LUKS auto-unlock with Clevis & Tang."
url: "https://hashnode.com/posts/stop-hardware-bootkits-tpm2-fwupd-bare-metal/6ad1af7bbca9b148c8c6b6ff"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Below the Kernel: How to Stop Hardware Bootkits with TPM 2.0 and fwupd](https://irexta.hashnode.dev/stop-hardware-bootkits-tpm2-fwupd-bare-metal?utm_source=hashnode&utm_medium=feed)

Defeat UEFI bootkits with a Hardware Root of Trust. SRE guide to TPM 2.0, fwupd firmware updates, and LUKS auto-unlock with Clevis & Tang.

Oct 16 · 5 min read · irexta.hashnode.dev  
Read the article

-   [#cybersecurity](https://hashnode.com/tag/cybersecurity)
-   [#linux](https://hashnode.com/tag/linux)
-   [#devops](https://hashnode.com/tag/devops)
-   [#sre](https://hashnode.com/tag/sre)
-   [#bare-metal](https://hashnode.com/tag/bare-metal)

## Written by

[Andrew Wiggey](https://hashnode.com/@irexta)

@irexta

  
[View profile](https://hashnode.com/@irexta)

Technical Team @ iRexta | Decoding High-Performance Bare Metal & GPU Infrastructure for the AI Era.

I am part of the Technical Team at iRexta, bridging the gap between high-end infrastructure and enterprise solutions. At iRexta, we specialize in providing unshared Bare Metal and Dedicated GPU servers to help businesses scale their AI environments, defend against advanced threats, and escape the exorbitant "Cloud Tax". Follow our space for deep dives into tech architecture and server engineering!

## Responses 1

Markdown supported · ⌘↵ to post

[Damiano Bianda](https://hashnode.com/@damianobianda)

[Oct 28](https://hashnode.com/posts/stop-hardware-bootkits-tpm2-fwupd-bare-metal/6ad1af7bbca9b148c8c6b6ff/comment/6ae201cfb0632e98b381d1a7)

Solid playbook, and the PCR 7 binding warning is the detail that saves fleets - bind to PCR 0 and the next BIOS update bricks your auto-unlock. One gap worth adding to Phase 2: fwupd only helps when a signed, patched capsule exists. CVE-2026-97631, the buffer overflow in the EDK II PXE boot driver disclosed last week, is still unpatched upstream, and it executes in DXE, before Secure Boot and before the TPM measures anything. For that class of bug "keep firmware updated" does nothing today; the working mitigations are disabling network boot on machines that do not provision from it and alerting on unexpected PXE traffic. Curious how you handle that unpatched window across a fleet: do you disable PXE by default and enable it only during provisioning runs?
