On Mon, Oct 23, 2017 at 04:08:24PM +0200, Eric Auger wrote: > Currently the ITS caches are not emptied on reset. > > After a reset, in case we attempt to save the state before > the bound devices have registered their MSIs and after the > 1st level table has been allocated by the ITS driver > (device BASER is valid), the first level entries are still > invalid. If the device cache is not empty (devices registered > before the reset), vgic_its_save_device_tables fails. > > This failure has no consequence as those devices do not > deserve to be saved: they correspond to the state before > the reset. > > However the ITS driver already sent MAPC for collections > and those need to be saved. With the current code, they > will not and the restored guest will not work properly. > > So this patch saves collection tables before device tables. > > Signed-off-by: Eric Auger <eric.auger@xxxxxxxxxx> > > --- > > candidate to be CC'ed stable > --- > virt/kvm/arm/vgic/vgic-its.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/virt/kvm/arm/vgic/vgic-its.c b/virt/kvm/arm/vgic/vgic-its.c > index b6650c2..8472417 100644 > --- a/virt/kvm/arm/vgic/vgic-its.c > +++ b/virt/kvm/arm/vgic/vgic-its.c > @@ -2324,11 +2324,11 @@ static int vgic_its_save_tables_v0(struct vgic_its *its) > return -EBUSY; > } > > - ret = vgic_its_save_device_tables(its); > + ret = vgic_its_save_collection_table(its); > if (ret) > goto out; > > - ret = vgic_its_save_collection_table(its); > + ret = vgic_its_save_device_tables(its); > I don't understand this. It seems to indicate an ordering of device tables vs. collection tables. What is that? I thought the point was that you'd want to save the valid table, and not save the other one. So, aren't we looking for something like this: ret = vgic_its_save_device_tables(its); if (ret < 0) goto out; ret = vgic_its_save_collection_table(its); Thanks, -Christoffer
