On 20 October 2014 19:30, Peter Maydell <peter.maydell@xxxxxxxxxx> wrote: > On 20 October 2014 18:19, Claudio Fontana <claudio.fontana@xxxxxxxxx> wrote: >> maybe I should have added that I am looking at AArch64 only, and that the >> long term plan is to move from device trees to ACPI for the guest as well. >> >> ACPI 5.1 provides as far as I understand extended support for the GIC, >> including the address of the GIC distributor and CPU interface; >> but I wondered however how to reliably query the supported GIC version. >> Maybe there is not an agreed way to do it yet. When you have a target with GICv3, I think ACPI differentiates it by using the proper subtable type. In the case of GICv2, the GICC and GICD subtables are used (0xB and 0xC), on the other hand for GICv3, I believe the GICR (0xE) subtable is used instead. Have a look at the ACPI v5.1 spec, with all the subtables listed in table 5-46. Best regards. _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
