KVM: MMU: fix missing locking in alloc_mmu_pages
commit005270f475352a33f852cbfb9f49945d7953ea61
authorMarcelo Tosatti <mtosatti@redhat.com>
Fri, 18 Sep 2009 23:08:03 +0000 (18 20:08 -0300)
committerGreg Kroah-Hartman <gregkh@suse.de>
Mon, 5 Oct 2009 15:27:55 +0000 (5 08:27 -0700)
tree65cbd24b70a456862737d21e3dec80cff7245c7d
parent3f1538a4dd00709b1c7f6a4fda31cf5daf4b3e46
KVM: MMU: fix missing locking in alloc_mmu_pages

(cherry picked from commit 6a1ac77110ee3e8d8dfdef8442f3b30b3d83e6a2)

n_requested_mmu_pages/n_free_mmu_pages are used by
kvm_mmu_change_mmu_pages to calculate the number of pages to zap.

alloc_mmu_pages, called from the vcpu initialization path, modifies this
variables without proper locking, which can result in a negative value
in kvm_mmu_change_mmu_pages (say, with cpu hotplug).

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
arch/x86/kvm/mmu.c