vfio/pci: Disable auto-enable of exclusive INTx IRQ
Linux
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
Linux
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
≥ 89e1f7d4c66d85f42c3d52ea3866eb10cadf6153 · < 3b3491ad0f80d913e7d255941d4470f4a4d9bfda| No fix yet |
| Linux Linux |
|
|
In the Linux kernel, the following vulnerability has been resolved:
vfio/pci: Disable auto-enable of exclusive INTx IRQ
Currently for devices requiring masking at the irqchip for INTx, ie. devices without DisINTx support, the IRQ is enabled in request_irq() and subsequently disabled as necessary to align with the masked status flag. This presents a window where the interrupt could fire between these events, resulting in the IRQ incrementing the disable depth twice. This would be unrecoverable for a user since the masked flag prevents nested enables through vfio.
Instead, invert the logic using IRQF_NO_AUTOEN such that exclusive INTx is never auto-enabled, then unmask as required.
From the CVE record · Linux
The Linux kernel’s vfio/pci code can mishandle exclusive INTx interrupts on devices that require masking at the irqchip. If an interrupt fires during a brief enable-and-disable window, the IRQ can become unrecoverable for the user. The input does not specify an attack path, authentication requirements, or user interaction.
Engineers running Linux systems that use vfio/pci with devices lacking DisINTx support should check their kernel; affected versions are not specified.