On 13.04.2018 14:54, Michal Hocko wrote: > On Fri 13-04-18 14:49:32, Kirill Tkhai wrote: >> On 13.04.2018 14:38, Michal Hocko wrote: >>> On Fri 13-04-18 14:29:11, Kirill Tkhai wrote: > [...] >>>> mem_cgroup_id_put_many() unpins css, but this may be not the last >>>> reference to the css. >>>> Thus, we release ID earlier, then all references to css are freed. >>> >>> Right and so what. If we have released the idr then we are not going to >>> do that again in css_free. That is why we have that memcg->id.id > 0 >>> check before idr_remove and memcg->id.id = 0 for the last memcg ref. >>> count. So again, why cannot we do the clean up in mem_cgroup_free and >>> have a less confusing code? Or am I just not getting your point and >>> being dense here? >> >> We can, but mem_cgroup_free() called from mem_cgroup_css_alloc() is unlikely >> case. >> The likely case is mem_cgroup_free() is called from mem_cgroup_css_free(), >> where >> this idr manipulations will be a noop. Noop in likely case looks more >> confusing >> for me. > > Well, I would really prefer to have _free being symmetric to _alloc so > that you can rely that the full state is gone after _free is called. > This confused the hell out of me. Because I _did_ expect that > mem_cgroup_free would do that and so I was looking at completely > different place. > >> Less confusing will be to move >> >> memcg->id.id = idr_alloc(&mem_cgroup_idr, NULL, >> 1, MEM_CGROUP_ID_MAX, >> GFP_KERNEL); >> >> into mem_cgroup_css_alloc(). How are you think about this? > > I would have to double check. Maybe it can be done on top. But for the > actual fix and a stable backport potentially should be as clear as > possible. Your original patch would be just fine but if I would prefer > mem_cgroup_free for the symmetry.
We definitely can move id allocation to mem_cgroup_css_alloc(), but this is really not for an easy fix, which will be backported to stable. Moving idr destroy to mem_cgroup_free() hides IDR trick. My IMHO it's less readable for a reader. The main problem is allocation asymmetric, and we shouldn't handle it on free path... Kirill