Message ID | 1316051175-17780-3-git-send-email-glommer@parallels.com |
---|---|
State | RFC, archived |
Delegated to: | David Miller |
Headers | show |
On Wed, Sep 14, 2011 at 10:46:10PM -0300, Glauber Costa wrote: > We aim to control the amount of kernel memory pinned at any > time by tcp sockets. To lay the foundations for this work, > this patch adds a pointer to the kmem_cgroup to the socket > structure. > > Signed-off-by: Glauber Costa <glommer@parallels.com> > CC: David S. Miller <davem@davemloft.net> > CC: Hiroyouki Kamezawa <kamezawa.hiroyu@jp.fujitsu.com> > CC: Eric W. Biederman <ebiederm@xmission.com> > --- > include/linux/memcontrol.h | 38 ++++++++++++++++++++++++++++++++++++++ > include/net/sock.h | 2 ++ > net/core/sock.c | 3 +++ > 3 files changed, 43 insertions(+), 0 deletions(-) > > diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h > index 3b535db..be457ce 100644 > --- a/include/linux/memcontrol.h > +++ b/include/linux/memcontrol.h > @@ -395,5 +395,43 @@ mem_cgroup_print_bad_page(struct page *page) > } > #endif > > +#ifdef CONFIG_INET > +#ifdef CONFIG_CGROUP_MEM_RES_CTLR_KMEM > +#include <net/sock.h> > +static inline void sock_update_memcg(struct sock *sk) > +{ > + /* right now a socket spends its whole life in the same cgroup */ > + BUG_ON(sk->sk_cgrp); > + > + rcu_read_lock(); > + sk->sk_cgrp = mem_cgroup_from_task(current); > + > + /* > + * We don't need to protect against anything task-related, because > + * we are basically stuck with the sock pointer that won't change, > + * even if the task that originated the socket changes cgroups. > + * > + * What we do have to guarantee, is that the chain leading us to > + * the top level won't change under our noses. Incrementing the > + * reference count via cgroup_exclude_rmdir guarantees that. > + */ > + cgroup_exclude_rmdir(mem_cgroup_css(sk->sk_cgrp)); > + rcu_read_unlock(); > +} > + > +static inline void sock_release_memcg(struct sock *sk) > +{ > + cgroup_release_and_wakeup_rmdir(mem_cgroup_css(sk->sk_cgrp)); > +} Do we really need to have these functions in the header?
On 09/17/2011 02:52 PM, Kirill A. Shutemov wrote: > On Wed, Sep 14, 2011 at 10:46:10PM -0300, Glauber Costa wrote: >> We aim to control the amount of kernel memory pinned at any >> time by tcp sockets. To lay the foundations for this work, >> this patch adds a pointer to the kmem_cgroup to the socket >> structure. >> >> Signed-off-by: Glauber Costa<glommer@parallels.com> >> CC: David S. Miller<davem@davemloft.net> >> CC: Hiroyouki Kamezawa<kamezawa.hiroyu@jp.fujitsu.com> >> CC: Eric W. Biederman<ebiederm@xmission.com> >> --- >> include/linux/memcontrol.h | 38 ++++++++++++++++++++++++++++++++++++++ >> include/net/sock.h | 2 ++ >> net/core/sock.c | 3 +++ >> 3 files changed, 43 insertions(+), 0 deletions(-) >> >> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h >> index 3b535db..be457ce 100644 >> --- a/include/linux/memcontrol.h >> +++ b/include/linux/memcontrol.h >> @@ -395,5 +395,43 @@ mem_cgroup_print_bad_page(struct page *page) >> } >> #endif >> >> +#ifdef CONFIG_INET >> +#ifdef CONFIG_CGROUP_MEM_RES_CTLR_KMEM >> +#include<net/sock.h> >> +static inline void sock_update_memcg(struct sock *sk) >> +{ >> + /* right now a socket spends its whole life in the same cgroup */ >> + BUG_ON(sk->sk_cgrp); >> + >> + rcu_read_lock(); >> + sk->sk_cgrp = mem_cgroup_from_task(current); >> + >> + /* >> + * We don't need to protect against anything task-related, because >> + * we are basically stuck with the sock pointer that won't change, >> + * even if the task that originated the socket changes cgroups. >> + * >> + * What we do have to guarantee, is that the chain leading us to >> + * the top level won't change under our noses. Incrementing the >> + * reference count via cgroup_exclude_rmdir guarantees that. >> + */ >> + cgroup_exclude_rmdir(mem_cgroup_css(sk->sk_cgrp)); >> + rcu_read_unlock(); >> +} >> + >> +static inline void sock_release_memcg(struct sock *sk) >> +{ >> + cgroup_release_and_wakeup_rmdir(mem_cgroup_css(sk->sk_cgrp)); >> +} > > Do we really need to have these functions in the header? > No, I can move it to memcontrol.c -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sun, Sep 18, 2011 at 12:32:27AM -0300, Glauber Costa wrote: > On 09/17/2011 02:52 PM, Kirill A. Shutemov wrote: > > Do we really need to have these functions in the header? > > > No, I can move it to memcontrol.c > Yes, please.
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 3b535db..be457ce 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -395,5 +395,43 @@ mem_cgroup_print_bad_page(struct page *page) } #endif +#ifdef CONFIG_INET +#ifdef CONFIG_CGROUP_MEM_RES_CTLR_KMEM +#include <net/sock.h> +static inline void sock_update_memcg(struct sock *sk) +{ + /* right now a socket spends its whole life in the same cgroup */ + BUG_ON(sk->sk_cgrp); + + rcu_read_lock(); + sk->sk_cgrp = mem_cgroup_from_task(current); + + /* + * We don't need to protect against anything task-related, because + * we are basically stuck with the sock pointer that won't change, + * even if the task that originated the socket changes cgroups. + * + * What we do have to guarantee, is that the chain leading us to + * the top level won't change under our noses. Incrementing the + * reference count via cgroup_exclude_rmdir guarantees that. + */ + cgroup_exclude_rmdir(mem_cgroup_css(sk->sk_cgrp)); + rcu_read_unlock(); +} + +static inline void sock_release_memcg(struct sock *sk) +{ + cgroup_release_and_wakeup_rmdir(mem_cgroup_css(sk->sk_cgrp)); +} +#else +#include <net/sock.h> +static inline void sock_update_memcg(struct sock *sk) +{ +} +static inline void sock_release_memcg(struct sock *sk) +{ +} +#endif /* CONFIG_CGROUP_MEM_RES_CTLR_KMEM */ +#endif /* CONFIG_INET */ #endif /* _LINUX_MEMCONTROL_H */ diff --git a/include/net/sock.h b/include/net/sock.h index 8e4062f..afe1467 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -228,6 +228,7 @@ struct sock_common { * @sk_security: used by security modules * @sk_mark: generic packet mark * @sk_classid: this socket's cgroup classid + * @sk_cgrp: this socket's kernel memory (kmem) cgroup * @sk_write_pending: a write to stream socket waits to start * @sk_state_change: callback to indicate change in the state of the sock * @sk_data_ready: callback to indicate there is data to be processed @@ -339,6 +340,7 @@ struct sock { #endif __u32 sk_mark; u32 sk_classid; + struct mem_cgroup *sk_cgrp; void (*sk_state_change)(struct sock *sk); void (*sk_data_ready)(struct sock *sk, int bytes); void (*sk_write_space)(struct sock *sk); diff --git a/net/core/sock.c b/net/core/sock.c index 3449df8..54ec8ac 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -125,6 +125,7 @@ #include <net/xfrm.h> #include <linux/ipsec.h> #include <net/cls_cgroup.h> +#include <linux/memcontrol.h> #include <linux/filter.h> @@ -1139,6 +1140,7 @@ struct sock *sk_alloc(struct net *net, int family, gfp_t priority, atomic_set(&sk->sk_wmem_alloc, 1); sock_update_classid(sk); + sock_update_memcg(sk); } return sk; @@ -1170,6 +1172,7 @@ static void __sk_free(struct sock *sk) put_cred(sk->sk_peer_cred); put_pid(sk->sk_peer_pid); put_net(sock_net(sk)); + sock_release_memcg(sk); sk_prot_free(sk->sk_prot_creator, sk); }
We aim to control the amount of kernel memory pinned at any time by tcp sockets. To lay the foundations for this work, this patch adds a pointer to the kmem_cgroup to the socket structure. Signed-off-by: Glauber Costa <glommer@parallels.com> CC: David S. Miller <davem@davemloft.net> CC: Hiroyouki Kamezawa <kamezawa.hiroyu@jp.fujitsu.com> CC: Eric W. Biederman <ebiederm@xmission.com> --- include/linux/memcontrol.h | 38 ++++++++++++++++++++++++++++++++++++++ include/net/sock.h | 2 ++ net/core/sock.c | 3 +++ 3 files changed, 43 insertions(+), 0 deletions(-)