| Message ID | 1407819721-27833-1-git-send-email-xiaoguangrong@linux.vnet.ibm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Aug 11, 2014 at 10:02 PM, Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> wrote: > @@ -722,9 +719,10 @@ static struct kvm_memslots *install_new_memslots(struct kvm *kvm, > { > struct kvm_memslots *old_memslots = kvm->memslots; > I think you want slots->generation = old_memslots->generation; here. On the KVM_MR_DELETE path, install_new_memslots is called twice so this patch introduces a short window of time where the generation number actually decreases. > - update_memslots(slots, new, kvm->memslots->generation); > + update_memslots(slots, new); > rcu_assign_pointer(kvm->memslots, slots); > synchronize_srcu_expedited(&kvm->srcu); > + slots->generation++; > > kvm_arch_memslots_updated(kvm); > > -- > 1.8.3.1 > -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 08/13/2014 05:18 AM, David Matlack wrote: > On Mon, Aug 11, 2014 at 10:02 PM, Xiao Guangrong > <xiaoguangrong@linux.vnet.ibm.com> wrote: >> @@ -722,9 +719,10 @@ static struct kvm_memslots *install_new_memslots(struct kvm *kvm, >> { >> struct kvm_memslots *old_memslots = kvm->memslots; >> > > I think you want > > slots->generation = old_memslots->generation; > > here. > > On the KVM_MR_DELETE path, install_new_memslots is called twice so this > patch introduces a short window of time where the generation number > actually decreases. Yes, indeed. Thank you for pointing it out, will update. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 33712fb..ca3cdac 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -96,7 +96,7 @@ static void hardware_disable_all(void); static void kvm_io_bus_destroy(struct kvm_io_bus *bus); static void update_memslots(struct kvm_memslots *slots, - struct kvm_memory_slot *new, u64 last_generation); + struct kvm_memory_slot *new); static void kvm_release_pfn_dirty(pfn_t pfn); static void mark_page_dirty_in_slot(struct kvm *kvm, @@ -687,8 +687,7 @@ static void sort_memslots(struct kvm_memslots *slots) } static void update_memslots(struct kvm_memslots *slots, - struct kvm_memory_slot *new, - u64 last_generation) + struct kvm_memory_slot *new) { if (new) { int id = new->id; @@ -699,8 +698,6 @@ static void update_memslots(struct kvm_memslots *slots, if (new->npages != npages) sort_memslots(slots); } - - slots->generation = last_generation + 1; } static int check_memory_region_flags(struct kvm_userspace_memory_region *mem) @@ -722,9 +719,10 @@ static struct kvm_memslots *install_new_memslots(struct kvm *kvm, { struct kvm_memslots *old_memslots = kvm->memslots; - update_memslots(slots, new, kvm->memslots->generation); + update_memslots(slots, new); rcu_assign_pointer(kvm->memslots, slots); synchronize_srcu_expedited(&kvm->srcu); + slots->generation++; kvm_arch_memslots_updated(kvm);
We may cache the current mmio generation number and stale memslot info into spte, like this scenario? CPU 0 CPU 1 page fault: add a new memslot read memslot and detecting its a mmio access update memslots update generation number read generation number cache the gpa and current gen number into spte So, if guest accesses the gpa later, it will generate a incorrect mmio exit This patch fixes it by updating the generation number after synchronize_srcu_expedited() that makes sure the generation number updated only if memslots update is finished Cc: stable@vger.kernel.org Cc: David Matlack <dmatlack@google.com> Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> --- virt/kvm/kvm_main.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)