Hello Tejun,
On 10/04/2018 08:20 PM, Tejun Heo wrote:
> Hello,
>
> On Thu, Oct 04, 2018 at 08:14:55PM +0200, Michael Kerrisk (man-pages) wrote:
>> On 10/02/2018 11:07 PM, Tejun Heo wrote:
>>> Hello, Michael.
>>>
>>> Great catch. Can you please verify whether the following patch fixes
>>> the issue?
>>>
>>> Thanks.
>> Against which kernel version should this apply? I get these build
>> errors on kernel 4.18:
>
> lol, sorry about that. The header change ended up in the wrong patch.
> Updated patch follows.
That fixes the problem for me (built against 4.19-rc6).
When you roll the final patch, please include Amin and Joao in the
Reported-by:
Thanks,
Michael
> ------ 8< ------
> Subject: cgroup: Fix dom_cgrp propagation when enabling threaded mode
>
> A cgroup which is already a threaded domain may be converted into a
> threaded cgroup if the prerequisite conditions are met. When this
> happens, all threaded descendant should also have their ->dom_cgrp
> updated to the new threaded domain cgroup. Unfortunately, this
> propagation was missing leading to the following failure.
>
> # cd /sys/fs/cgroup/unified
> # cat cgroup.subtree_control # show that no controllers are enabled
>
> # mkdir -p mycgrp/a/b/c
> # echo threaded > mycgrp/a/b/cgroup.type
>
> At this point, the hierarchy looks as follows:
>
> mycgrp [d]
> a [dt]
> b [t]
> c [inv]
>
> Now let's make node "a" threaded (and thus "mycgrp" s made "domain threaded"):
>
> # echo threaded > mycgrp/a/cgroup.type
>
> By this point, we now have a hierarchy that looks as follows:
>
> mycgrp [dt]
> a [t]
> b [t]
> c [inv]
>
> But, when we try to convert the node "c" from "domain invalid" to
> "threaded", we get ENOTSUP on the write():
>
> # echo threaded > mycgrp/a/b/c/cgroup.type
> sh: echo: write error: Operation not supported
>
> This patch fixes the problem by
>
> * Moving the opencoded ->dom_cgrp save and restoration in
> cgroup_enable_threaded() into cgroup_{save|restore}_control() so
> that mulitple cgroups can be handled.
>
> * Updating all threaded descendants' ->dom_cgrp to point to the new
> dom_cgrp when enabling threaded mode.
>
> Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
> Reported-by: "Michael Kerrisk (man-pages)" <mtk.manpages@xxxxxxxxx>
> Link: https://lore.kernel.org/r/CAKgNAkhHYCMn74TCNiMJ=ccLd7DcmXSbvw3CbZ1YREeG7iJM5g@xxxxxxxxxxxxxx
> Fixes: 454000adaa2a ("cgroup: introduce cgroup->dom_cgrp and threaded css_set handling")
> Cc: stable@xxxxxxxxxxxxxxx # v4.14+
> ---
> include/linux/cgroup-defs.h | 1 +
> kernel/cgroup/cgroup.c | 25 ++++++++++++++++---------
> 2 files changed, 17 insertions(+), 9 deletions(-)
>
> --- a/include/linux/cgroup-defs.h
> +++ b/include/linux/cgroup-defs.h
> @@ -412,6 +412,7 @@ struct cgroup {
> * specific task are charged to the dom_cgrp.
> */
> struct cgroup *dom_cgrp;
> + struct cgroup *old_dom_cgrp; /* used while enabling threaded */
>
> /* per-cpu recursive resource statistics */
> struct cgroup_rstat_cpu __percpu *rstat_cpu;
> --- a/kernel/cgroup/cgroup.c
> +++ b/kernel/cgroup/cgroup.c
> @@ -2836,11 +2836,12 @@ restart:
> }
>
> /**
> - * cgroup_save_control - save control masks of a subtree
> + * cgroup_save_control - save control masks and dom_cgrp of a subtree
> * @cgrp: root of the target subtree
> *
> - * Save ->subtree_control and ->subtree_ss_mask to the respective old_
> - * prefixed fields for @cgrp's subtree including @cgrp itself.
> + * Save ->subtree_control, ->subtree_ss_mask and ->dom_cgrp to the
> + * respective old_ prefixed fields for @cgrp's subtree including @cgrp
> + * itself.
> */
> static void cgroup_save_control(struct cgroup *cgrp)
> {
> @@ -2850,6 +2851,7 @@ static void cgroup_save_control(struct c
> cgroup_for_each_live_descendant_pre(dsct, d_css, cgrp) {
> dsct->old_subtree_control = dsct->subtree_control;
> dsct->old_subtree_ss_mask = dsct->subtree_ss_mask;
> + dsct->old_dom_cgrp = dsct->dom_cgrp;
> }
> }
>
> @@ -2875,11 +2877,12 @@ static void cgroup_propagate_control(str
> }
>
> /**
> - * cgroup_restore_control - restore control masks of a subtree
> + * cgroup_restore_control - restore control masks and dom_cgrp of a subtree
> * @cgrp: root of the target subtree
> *
> - * Restore ->subtree_control and ->subtree_ss_mask from the respective old_
> - * prefixed fields for @cgrp's subtree including @cgrp itself.
> + * Restore ->subtree_control, ->subtree_ss_mask and ->dom_cgrp from the
> + * respective old_ prefixed fields for @cgrp's subtree including @cgrp
> + * itself.
> */
> static void cgroup_restore_control(struct cgroup *cgrp)
> {
> @@ -2889,6 +2892,7 @@ static void cgroup_restore_control(struc
> cgroup_for_each_live_descendant_post(dsct, d_css, cgrp) {
> dsct->subtree_control = dsct->old_subtree_control;
> dsct->subtree_ss_mask = dsct->old_subtree_ss_mask;
> + dsct->dom_cgrp = dsct->old_dom_cgrp;
> }
> }
>
> @@ -3196,6 +3200,8 @@ static int cgroup_enable_threaded(struct
> {
> struct cgroup *parent = cgroup_parent(cgrp);
> struct cgroup *dom_cgrp = parent->dom_cgrp;
> + struct cgroup *dsct;
> + struct cgroup_subsys_state *d_css;
> int ret;
>
> lockdep_assert_held(&cgroup_mutex);
> @@ -3225,12 +3231,13 @@ static int cgroup_enable_threaded(struct
> */
> cgroup_save_control(cgrp);
>
> - cgrp->dom_cgrp = dom_cgrp;
> + cgroup_for_each_live_descendant_pre(dsct, d_css, cgrp)
> + if (dsct == cgrp || cgroup_is_threaded(dsct))
> + dsct->dom_cgrp = dom_cgrp;
> +
> ret = cgroup_apply_control(cgrp);
> if (!ret)
> parent->nr_threaded_children++;
> - else
> - cgrp->dom_cgrp = cgrp;
>
> cgroup_finalize_control(cgrp, ret);
> return ret;
>
--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/
