QEMU pvpanic guest panic notifier
A small paravirtual device exposed by QEMU/KVM to Linux guests, available as either a PCI device or an ACPI/MMIO region, that lets the guest notify the host hypervisor when it panics or shuts down. It is widely used in modern virtualization fleets so management tools can detect and react to guest crashes.
recommendation
It should stay because pvpanic is the standard way for a Linux guest running under QEMU/KVM to tell the host when it has panicked or shut down, and it is still being actively maintained: a shutdown-notification feature landed in 2024 and a cleanup patch was even backported to stable. QEMU's current documentation continues to describe the device, and no other driver fills the same guest-to-host signalling role.
repository signals
sources
- lore.kernel.org
Mainline still received functional pvpanic work in 2024 (`misc/pvpanic: add support for normal shutdowns`), indicating ongoing maintenance rather than abandonment.
- lore.kernel.org
The 2024 `misc/pvpanic: deduplicate common code` change was backported to stable, showing the driver still matters to supported kernels.
- qemu.org
QEMU current documentation still describes pvpanic as a supported simulated guest device, with PCI and ACPI/MMIO interfaces, for reporting guest panic and shutdown events to the host.
- qemu.org
The current QEMU manual set still ships pvpanic documentation in the active `master` docs, consistent with ongoing use in modern virtualized deployments.
codex reasoning notes (technical)
Local source inspection via shell (`rg`) shows this is a paravirtual guest-notification driver for QEMU with PCI and MMIO frontends, not legacy physical hardware. `lore_activity` on `drivers/misc/pvpanic/pvpanic.c` returned 2023-2024 feature and cleanup patches plus stable backports, and no removal discussion surfaced; a broader `lore_regex` attempt timed out, so confidence is based on file-scoped lore evidence rather than corpus-wide subject scan. QEMU docs were obtained via web search and show pvpanic remains documented in current releases, so deployments are still plausible in KVM/QEMU guest fleets even though no physical hardware is sold. No upstream replacement driver covers the same guest-to-host panic signalling role.