uefi: support PCIe in UEFI by chris-oo · Pull Request #3268 · microsoft/openvmm

chris-oo · 2026-04-13T22:58:39Z

Add PCIe UEFI support, and a petri test.

Copilot

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

Adds initial PCIe NVMe UEFI boot support (WIP) by wiring PCIe NVMe devices into the Petri/OpenVMM configuration path and passing PCIe BAR aperture metadata to the UEFI config blob, plus an end-to-end test.

Changes:

Add a multi-arch UEFI test that boots Alpine from an NVMe device attached via an emulated PCIe root port.
Extend UEFI config blob definitions and loader to provide per-host-bridge PCIe BAR apertures.
Extend Petri VM config/build path to attach NVMe controllers as PCIe devices for BootDeviceType::PcieNvme.

Reviewed changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
vmm_tests/vmm_tests/tests/tests/multiarch/pcie.rs	Adds a UEFI PCIe-NVMe boot test and basic guest visibility validation.
vm/loader/src/uefi/config.rs	Introduces a config-blob structure for PCIe BAR apertures and a new blob type.
petri/src/vm/openvmm/construct.rs	Converts Petri “PCIe NVMe drives” into OpenVMM `PcieDeviceConfig` entries.
petri/src/vm/mod.rs	Adds `pcie_nvme_drives` to `PetriVmConfig` and a new `BootDeviceType::PcieNvme`.
openvmm/openvmm_core/src/worker/vm_loaders/uefi.rs	Emits PCIe BAR aperture entries into the UEFI config blob when host bridges exist.
.gitignore	Removes an extra blank/whitespace-only ignore line.

Copilot

Pull request overview

Copilot reviewed 5 out of 6 changed files in this pull request and generated 3 comments.

Add support for hardware emulated PCIe buses. This allows booting from NVMe exposed on a PCIe. The host is required to specify which bridges it wants UEFI to enumerate by creating a aperture config structure for each bridge, to avoid requiring UEFI to parse the SSDT. ECAM ranges are parsed from the MCFG table. See the corresponding openvmm change: microsoft/openvmm#3268

Add support for booting from NVMe devices attached to PCIe root ports (ePCI) in the petri test framework: - Add pcie_nvme_drives field to PetriVmConfig for PCIe NVMe drive configs - Add PcieNvme variant to BootDeviceType enum - Translate PCIe NVMe drives to openvmm PcieDeviceConfig in construct.rs - Add pcie_nvme_boot VMM integration test that validates the full PciHostBridgeDxe -> PciBusDxe -> NvmExpressDxe -> OS boot chain Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Root cause: UEFI synthvid allocated VRAM at the ECAM base address (0xE4000000) because PlatformPei declared the ECAM range as MMIO but did not reserve it in GCD. The fix reserves ECAM ranges in PciHostBridgeLib so AllocateMemorySpace cannot hand them out. Changes: - pcie_nvme_boot test: disable VPCI boot to reduce noise - Investigation doc: updated with full root cause analysis - MSVM-epci-debug-x64.fd: pre-built firmware with all PCIe fixes Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Updated MSVM-epci-debug-x64.fd with clean build (only ECAM reservation fix, no diagnostic tracing or double ConnectAll). Updated investigation doc: double ConnectAll confirmed unnecessary, NvmExpress DevicePath leak is a pre-existing bug but not the blocker. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Renamed epci-support-plan.md to pcie-support-plan.md and replaced all ePCI references with PCIe throughout planning and investigation docs. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 6 changed files in this pull request and generated 3 comments.

jstarks · 2026-04-17T23:13:35Z

                BootDeviceType::NvmeViaNvme => todo!(),
+                BootDeviceType::PcieNvme => {
+                    self.config.pcie_nvme_drives.push(PcieNvmeDrive {
+                        port_name: "s0rc0rp0".into(),


This is a bit dodgy.

(Maybe it's OK for now, I think we need to think a little harder about the "default" PCIe topology, at least within petri)

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Copilot · 2026-04-17T23:15:37Z

+                BootDeviceType::PcieNvme => {
+                    self.config.pcie_nvme_drives.push(PcieNvmeDrive {
+                        port_name: "s0rc0rp0".into(),
+                        nsid: 1,
+                        drive: boot_drive,
+                    });
+                    self
+                }


BootDeviceType::PcieNvme hard-codes the target port name ("s0rc0rp0") and assumes a PCIe root topology has been created elsewhere. If the caller forgets to add with_pcie_root_topology(...) (or otherwise create that port), OpenVMM will later fail when wiring the PCIe device to a non-existent port, and the resulting error will be far from the boot-device selection site. Consider either (a) making the port name configurable as part of this boot device type, or (b) adding an explicit early validation / clearer error when PcieNvme is selected but no matching port will exist.

Copilot AI review requested due to automatic review settings April 13, 2026 22:58

Copilot started reviewing on behalf of chris-oo April 13, 2026 23:05 View session

Copilot AI reviewed Apr 13, 2026

View reviewed changes

chris-oo mentioned this pull request Apr 13, 2026

mu_msvm: add support for hardware emulated pcie microsoft/mu_msvm#65

Merged

Copilot AI review requested due to automatic review settings April 14, 2026 20:58

chris-oo changed the title ~~[WIP] pcie uefi support~~ uefi: support PCIe in UEFI Apr 14, 2026

Copilot AI reviewed Apr 14, 2026

View reviewed changes

Comment thread vmm_tests/vmm_tests/tests/tests/multiarch/pcie.rs Outdated

Comment thread petri/src/vm/mod.rs Outdated

Comment thread petri/src/vm/openvmm/construct.rs Outdated

Copilot started reviewing on behalf of chris-oo April 14, 2026 23:18 View session

chris-oo force-pushed the pcie-uefi branch from e87e700 to 98ec5b8 Compare April 17, 2026 21:53

chris-oo and others added 14 commits April 17, 2026 14:56

pcie: add planning for PCIe in mu_msvm

82be12f

pcie: plan refinements

3e492e8

pcie: more plan updates

b2cda6d

pcie: add PcieBarApertures config blob entry

cbb95ae

pcie-uefi: investigation of why nvme boot doesn't work

ddd47df

pcie: rename ePCI references to PCIe across docs

a235307

Renamed epci-support-plan.md to pcie-support-plan.md and replaced all ePCI references with PCIe throughout planning and investigation docs. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

petri: simplify nvme boot device construction

577c831

pcie: report ACPI uid for each bar aperture

182db25

pcie: remove intermediate plans/fw

f314610

pcie: comment updates for uefi config

3ca76cb

pcie: cleanup test comment, test windows

bc1ac53

chris-oo force-pushed the pcie-uefi branch from 98ec5b8 to bc1ac53 Compare April 17, 2026 21:56

chris-oo marked this pull request as ready for review April 17, 2026 21:59

chris-oo requested a review from a team as a code owner April 17, 2026 21:59

Copilot AI review requested due to automatic review settings April 17, 2026 21:59

Copilot started reviewing on behalf of chris-oo April 17, 2026 22:00 View session

Copilot AI reviewed Apr 17, 2026

View reviewed changes

Comment thread openvmm/openvmm_core/src/worker/vm_loaders/uefi.rs Outdated

Comment thread vmm_tests/vmm_tests/tests/tests/multiarch/pcie.rs Outdated

Comment thread vmm_tests/vmm_tests/tests/tests/multiarch/pcie.rs Outdated

chris-oo added 2 commits April 17, 2026 16:04

fix review feedback/issues

73081e3

oops

107a494

Copilot AI review requested due to automatic review settings April 17, 2026 23:09

Copilot started reviewing on behalf of chris-oo April 17, 2026 23:10 View session

revert gitignore useless diff

7ad5d7f

jstarks reviewed Apr 17, 2026

View reviewed changes

jstarks previously approved these changes Apr 17, 2026

View reviewed changes

Copilot AI reviewed Apr 17, 2026

View reviewed changes

english is hard

3db94e7

chris-oo dismissed jstarks’s stale review via 3db94e7 April 17, 2026 23:18

jstarks approved these changes Apr 17, 2026

View reviewed changes

chris-oo enabled auto-merge (squash) April 17, 2026 23:19

chris-oo merged commit 96c1737 into microsoft:main Apr 18, 2026
61 checks passed

Conversation

chris-oo commented Apr 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jstarks Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

jstarks Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

chris-oo commented Apr 13, 2026 •

edited

Loading