]> git.proxmox.com Git - mirror_qemu.git/commit
xics_kvm: cache already enabled vCPU ids
authorGreg Kurz <groug@kaod.org>
Wed, 17 May 2017 14:38:20 +0000 (16:38 +0200)
committerDavid Gibson <david@gibson.dropbear.id.au>
Wed, 24 May 2017 01:39:52 +0000 (11:39 +1000)
commitde86eccc0c836adfa8dbb94848096720177f5ccb
tree34e74c8f060a4f30f7c2584fe28ca83eb8e70261
parent06ec79e865a4a496e762a83126d00d0ed39205f5
xics_kvm: cache already enabled vCPU ids

Since commit a45863bda90d ("xics_kvm: Don't enable KVM_CAP_IRQ_XICS if
already enabled"), we were able to re-hotplug a vCPU that had been hot-
unplugged ealier, thanks to a boolean flag in ICPState that we set when
enabling KVM_CAP_IRQ_XICS.

This could work because the lifecycle of all ICPState objects was the
same as the machine. Commit 5bc8d26de20c ("spapr: allocate the ICPState
object from under sPAPRCPUCore") broke this assumption and now we always
pass a freshly allocated ICPState object (ie, with the flag unset) to
icp_kvm_cpu_setup().

This cause re-hotplug to fail with:

Unable to connect CPU8 to kernel XICS: Device or resource busy

Let's fix this by caching all the vCPU ids for which KVM_CAP_IRQ_XICS was
enabled. This also drops the now useless boolean flag from ICPState.

Reported-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Greg Kurz <groug@kaod.org>
Tested-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
hw/intc/xics_kvm.c
include/hw/ppc/xics.h