Messages in this thread | | | Date | Fri, 21 Nov 2008 11:01:28 -0800 | Subject | Re: [PATCH] cgroups: don't put struct cgroupfs_root protected by RCU | From | Paul Menage <> |
| |
On Fri, Nov 21, 2008 at 12:49 AM, Lai Jiangshan <laijs@cn.fujitsu.com> wrote: > > we don't access struct cgroupfs_root in fast path, > so we should not put struct cgroupfs_root protected by RCU > > but the comment in struct cgroup_subsys.root confuse us. > > struct cgroup_subsys.root is used in these places: > > 1 find_css_set(): if (ss->root->subsys_list.next == &ss->sibling) > 2 rebind_subsystems(): if (ss->root != &rootnode) > rcu_assign_pointer(ss->root, root); > rcu_assign_pointer(subsys[i]->root, &rootnode); > 3 cgroup_has_css_refs(): if (ss->root != cgrp->root) > 4 cgroup_init_subsys(): ss->root = &rootnode; > 5 proc_cgroupstats_show(): ss->name, ss->root->subsys_bits, > ss->root->number_of_cgroups, !ss->disabled); > 6 cgroup_clone(): root = subsys->root; > if ((root != subsys->root) || > > all these place we have held cgroup_lock() or we don't dereference to > struct cgroupfs_root. It's means wo don't need RCU when use > struct cgroup_subsys.root, and we should not put struct cgroupfs_root > protected by RCU. > > Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Reviewed-by: Paul Menage <menage@google.com>
Thanks, seems reasonable.
> --- > diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h > index 1164963..22901ff 100644 > --- a/include/linux/cgroup.h > +++ b/include/linux/cgroup.h > @@ -343,7 +345,6 @@ struct cgroup_subsys { > #define MAX_CGROUP_TYPE_NAMELEN 32 > const char *name; > > - /* Protected by RCU */ > struct cgroupfs_root *root; > > struct list_head sibling; > diff --git a/kernel/cgroup.c b/kernel/cgroup.c > index 358e775..ddc10ac 100644 > --- a/kernel/cgroup.c > +++ b/kernel/cgroup.c > @@ -717,7 +722,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, > cgrp->subsys[i] = dummytop->subsys[i]; > cgrp->subsys[i]->cgroup = cgrp; > list_add(&ss->sibling, &root->subsys_list); > - rcu_assign_pointer(ss->root, root); > + ss->root = root; > if (ss->bind) > ss->bind(ss, cgrp); > > @@ -729,7 +734,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, > ss->bind(ss, dummytop); > dummytop->subsys[i]->cgroup = dummytop; > cgrp->subsys[i] = NULL; > - rcu_assign_pointer(subsys[i]->root, &rootnode); > + subsys[i]->root = &rootnode; > list_del(&ss->sibling); > } else if (bit & final_bits) { > /* Subsystem state should already exist */ > > >
| |