Re: [PATCH v2 0/7] Enable Qualcomm WiFi PCIe passthrough to VMs

[Jeff Johnson]

On 10/11/2026 12:45 AM, Jose Ignacio Tornos Martinez wrote:
> This series enables Qualcomm ath11k/ath12k WiFi devices to work in VM
> passthrough scenarios, addressing a long-standing issue where these
> devices fail to initialize when passed through to VMs via VFIO.
> 
> Qualcomm ath11k/ath12k WiFi devices have been broken in VM passthrough
> since they were introduced. The Qualcomm firmware has its own interrupt
> controller with private registers that need to be programmed with MSI
> addr/data. The driver extracts the addr/data from the MSI descriptor
> and programs the firmware's private registers. In a VM, the driver
> only sees virtual (IOVA) addr/data, but the firmware's private writes
> bypass the IOMMU, so they need the physical host values. This causes
> firmware initialization to fail with errors like "BHI offset:
> 0xffffffff is out of range".
> 
> Multiple attempts have been made to solve this over the past 2 years:
> 
> March 2024: Workaround using module parameters [1]
>   - Required manually copying MSI values from host to VM
>   - Rejected by Johannes Berg as too manual, suggested VMM solution
> 
> August 2024: QEMU + kernel solution by Alex Williamson [2]
>   - Kernel disables MSI virtualization, QEMU intercepts config writes
>   - Uses brute force pattern matching to infer MSI data writes
>   - Stuck as RFC - no progress in 8+ months
> 
> No action from Qualcomm: Despite multiple reports, no vendor solution
> 
> The solution in this series provides a clean, kernel-only approach that
> addresses the previous concerns. The PCI reset support for these devices
> has already landed in linux-next (commit f8522a518e59 "PCI: Add device-
> specific reset for Qualcomm WCN6855/WCN7850 WLAN") [3], solving the device
> reset path for VFIO passthrough. This MSI series is the remaining
> piece, with both in place, these Qualcomm WiFi devices would be fully
> operational in VMs for the first time.
> 
> In practice, this specific behavior (firmware requiring physical host
> MSI addresses in private registers, bypassing the IOMMU) has only
> been observed in Qualcomm ath11k/ath12k devices. A device-specific
> approach seems appropriate, and the VFIO variant driver pattern fits
> well for this, following the same established pattern as mlx5-vfio-pci
> and nvgrace-gpu for their own device-specific needs.
> 
> Architecture:
> 1. qcom-vfio-pci variant driver caches physical host MSI values and
>    exposes them via extended config space with magic signature "QMSI".
>    Device-specific hardware quirks belong only in kernel drivers, not
>    userspace, independent of the tools used.
> 2. VM drivers (ath11k/ath12k) automatically discover and use cached values
> 3. PCIe link recovery handles VM timing variations
> 
> Thoroughly tested: 3 VMs across 2 WiFi generations, all devices
> successfully initialized and are ready for use.
> - VM1: ath11k (WCN6855)
> - VM2: ath11k (WCN6855)
> - VM3: ath12k (WCN7850)
> 
> Series structure:
>   Patch 1/7: PCI helper to find free space in extended config space
>     for device-specific data exchange between host and VM drivers
>   Patch 2/7: Shared header defining the MSI cache protocol between
>     the host variant driver and VM drivers
>   Patch 3/7: qcom-vfio-pci variant driver that caches host physical
>     MSI address/data and exposes them to VMs through extended config
>     space
>   Patch 4/7: ath11k PCIe link recovery retry to handle initialization
>     timing variations in VM environments
>   Patch 5/7: ath11k support to read cached host MSI values and
>     configure firmware with correct physical addresses, making the
>     device operational in VMs
>   Patch 6/7: ath12k PCIe link recovery retry to handle initialization
>     timing variations in VM environments
>   Patch 7/7: ath12k support to read cached host MSI values and
>     configure firmware with correct physical addresses, making the
>     device operational in VMs
> 
> [1] https://lore.kernel.org/all/20240322104912.94811-1-jtornosm@xxxxxxxxxx/
> [2] https://lore.kernel.org/kvm/20240812170014.1583783-1-alex.williamson@xxxxxxxxxx/
> [3] https://lore.kernel.org/all/20261005105436.83962-1-jtornosm@xxxxxxxxxx/
> [4] Original problem report:
>     https://lore.kernel.org/all/fc6bd06f-d52b-4dee-ab1b-4bb845cc0b95@xxxxxxxxxxx/
> 
> ---
> v2: Complete cover letter with more information

Thank you for providing this level of detail in your cover letter. I've raised
this issue with the engineering team in the past, but I've once again raised
the issue with your in-depth analysis.

/jeff