We use RCU's for rarely updated lists like iommus, rmrr, atsr units.
I'm not sure why domain_remove_dev_info() in domain_exit() was surrounded
by rcu_read_lock. Lock was present before refactoring in 
d160aca527,
but it was related to rcu list, not domain_remove_dev_info function.
dmar_remove_one_dev_info() doesn't touch any of those lists, so it doesn't
require a lock. In fact it is called 6 times without it anyway.
Fixes: d160aca5276d ("iommu/vt-d: Unify domain->iommu attach/detachment")
Signed-off-by: Lukasz Odzioba <lukasz.odzioba@intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
        struct page *freelist;
 
        /* Remove associated devices and clear attached or cached domains */
-       rcu_read_lock();
        domain_remove_dev_info(domain);
-       rcu_read_unlock();
 
        /* destroy iovas */
        put_iova_domain(&domain->iovad);
 
                old_domain = find_domain(dev);
                if (old_domain) {
-                       rcu_read_lock();
                        dmar_remove_one_dev_info(dev);
-                       rcu_read_unlock();
 
                        if (!domain_type_is_vm_or_si(old_domain) &&
                            list_empty(&old_domain->devices))