IRQ2 ignored under KVM when using IOAPIC

[roblabla]

When using KVM, and an OS that supports the IOAPIC, interrupts mapped on IRQ2 (for instance, routing an HPET timer on interrupt 2) will cause the interrupts to never be delivered. This is because QEmu, when setting up the KVM interrupt routes, will not set one up for IRQ2[0]. When running without KVM, IRQ2 is identity-mapped to GSI2.

My understanding is that IRQs should be identity mapped to their equivalent GSI unless a redirection entry is present in the MADT. This is supported by ACPI 6.2 spec[1], 5.2.12.5 Interrupt Source Override Structure, which claims: "It is assumed that the ISA interrupts will be identity-mapped into the first I/O APIC sources.".

I stumbled across this while working on my own custom OS, got very confused why the HPET wasn't triggering any interruption - and even more confused why the behavior only happened in KVM and not in non-KVM.

EDIT: Interestingly, the HPET only supports IRQ2 when using the default PIIX chipset, which, combined with this bug, makes it completely unusable.

[0]: https://github.com/qemu/qemu/blob/37560c259d7a0d6aceb96e9d6903ee002f4e5e0c/hw/i386/kvm/ioapic.c#L40

[1]: https://uefi.org/sites/default/files/resources/ACPI_6_2.pdf

[Thomas Huth (th-huth)]

The QEMU project is currently considering to move its bug tracking to
another system. For this we need to know which bugs are still valid
and which could be closed already. Thus we are setting older bugs to
"Incomplete" now.

If you still think this bug report here is valid, then please switch
the state back to "New" within the next 60 days, otherwise this report
will be marked as "Expired". Or please mark it as "Fix Released" if
the problem has been solved with a newer version of QEMU already.

Thank you and sorry for the inconvenience.

[Launchpad Janitor (janitor)]

[Expired for QEMU because there has been no activity for 60 days.]