Re: [PATCH memcg v6] net: set proper memcg for net_init hooks allocations
From: Qian Cai <hidden>
Date: 2022-06-06 13:49:40
Also in:
linux-mm, lkml
On Fri, Jun 03, 2022 at 07:19:43AM +0300, Vasily Averin wrote:
__register_pernet_operations() executes init hook of registered pernet_operation structure in all existing net namespaces. Typically, these hooks are called by a process associated with the specified net namespace, and all __GFP_ACCOUNT marked allocation are accounted for corresponding container/memcg. However __register_pernet_operations() calls the hooks in the same context, and as a result all marked allocations are accounted to one memcg for all processed net namespaces. This patch adjusts active memcg for each net namespace and helps to account memory allocated inside ops_init() into the proper memcg. Signed-off-by: Vasily Averin <redacted> Acked-by: Roman Gushchin <redacted> Acked-by: Shakeel Butt <redacted> ---
...
quoted hunk ↗ jump to hunk
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 9ecead1042b9..dad16b484cd5 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h@@ -1755,6 +1755,42 @@ static inline void count_objcg_event(struct obj_cgroup *objcg, rcu_read_unlock(); } +/** + * get_mem_cgroup_from_obj - get a memcg associated with passed kernel object. + * @p: pointer to object from which memcg should be extracted. It can be NULL. + * + * Retrieves the memory group into which the memory of the pointed kernel + * object is accounted. If memcg is found, its reference is taken. + * If a passed kernel object is uncharged, or if proper memcg cannot be found, + * as well as if mem_cgroup is disabled, NULL is returned. + * + * Return: valid memcg pointer with taken reference or NULL. + */ +static inline struct mem_cgroup *get_mem_cgroup_from_obj(void *p) +{ + struct mem_cgroup *memcg; + + rcu_read_lock(); + do { + memcg = mem_cgroup_from_obj(p); + } while (memcg && !css_tryget(&memcg->css)); + rcu_read_unlock(); + return memcg; +}
...
quoted hunk ↗ jump to hunk
diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c index 0ec2f5906a27..6b9f19122ec1 100644 --- a/net/core/net_namespace.c +++ b/net/core/net_namespace.c@@ -18,6 +18,7 @@ #include <linux/user_namespace.h> #include <linux/net_namespace.h> #include <linux/sched/task.h> +#include <linux/sched/mm.h> #include <linux/uidgid.h> #include <linux/cookie.h>@@ -1143,7 +1144,13 @@ static int __register_pernet_operations(struct list_head *list, * setup_net() and cleanup_net() are not possible. */ for_each_net(net) { + struct mem_cgroup *old, *memcg; + + memcg = mem_cgroup_or_root(get_mem_cgroup_from_obj(net)); + old = set_active_memcg(memcg); error = ops_init(ops, net); + set_active_memcg(old); + mem_cgroup_put(memcg); if (error) goto out_undo; list_add_tail(&net->exit_list, &net_exit_list);-- 2.36.1
This triggers a few boot warnings like those. virt_to_phys used for non-linear address: ffffd8efe2d2fe00 (init_net) WARNING: CPU: 87 PID: 3170 at arch/arm64/mm/physaddr.c:12 __virt_to_phys CPU: 87 PID: 3170 Comm: modprobe Tainted: G B W 5.19.0-rc1-next-20220606 #138 pstate: 60400009 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) pc : __virt_to_phys lr : __virt_to_phys sp : ffff800051cc76b0 x29: ffff800051cc76b0 x28: ffffd8efb5ba6ab8 x27: ffffd8efb5ba6b2c x26: ffffd8efb1bccb20 x25: ffffd8efbaaf8200 x24: ffff800051cc77f0 x23: ffffd8efb744a000 x22: ffffd8efbb1bc000 x21: 0000600000000000 x20: 0000d8efe2d2fe00 x19: ffffd8efe2d2fe00 x18: 0000000000000443 x17: 0000000000000000 x16: 0000000000000002 x15: ffffd8efb9db2000 x14: 0000000000000001 x13: 0000000000000000 x12: ffff6806c88f8986 x11: 1fffe806c88f8985 x10: ffff6806c88f8985 x9 : dfff800000000000 x8 : ffff4036447c4c2b x7 : 0000000000000001 x6 : ffff6806c88f8985 x5 : ffff4036447c4c28 x4 : ffff6806c88f8986 x3 : ffffd8efb34b3850 x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffff400335f99a80 Call trace: __virt_to_phys mem_cgroup_from_obj __register_pernet_operations register_pernet_operations register_pernet_subsys nfnetlink_init [nfnetlink] load_module __do_sys_finit_module __arm64_sys_finit_module invoke_syscall el0_svc_common.constprop.0 do_el0_svc el0_svc el0t_64_sync_handler el0t_64_sync irq event stamp: 0 hardirqs last enabled at (0): 0x0 hardirqs last disabled at (0): copy_process softirqs last enabled at (0): copy_process softirqs last disabled at (0): 0x0 virt_to_phys used for non-linear address: ffffd8efe2d2fe00 (init_net) WARNING: CPU: 156 PID: 3176 at arch/arm64/mm/physaddr.c:12 __virt_to_phys CPU: 156 PID: 3176 Comm: modprobe Tainted: G B W 5.19.0-rc1-next-20220606 #138 pstate: 60400009 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) pc : __virt_to_phys lr : __virt_to_phys sp : ffff800051b376e0 x29: ffff800051b376e0 x28: ffffd8efb5ba6ab8 x27: ffffd8efb5ba6b2c x26: ffffd8efb286e910 x25: ffffd8efbaaf8200 x24: ffff800051b37820 x23: ffffd8efb744a000 x22: ffffd8efbb1bc000 x21: 0000600000000000 x20: 0000d8efe2d2fe00 x19: ffffd8efe2d2fe00 x18: 00000000000001cb x17: 0000000000000000 x16: 0000000000000002 x15: ffffd8efb9db2000 x14: 0000000000000001 x13: 0000000000000000 x12: ffff6806c8a03f86 x8 : ffff40364501fc2b x7 : 0000000000000001 x6 : ffff6806c8a03f85 x5 : ffff40364501fc28 x4 : ffff6806c8a03f86 x3 : ffffd8efb34b3850 x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffff40033376b4c0 Call trace: __virt_to_phys mem_cgroup_from_obj __register_pernet_operations register_pernet_operations register_pernet_subsys nf_tables_module_init [nf_tables] do_one_initcall do_init_module load_module __do_sys_finit_module __arm64_sys_finit_module invoke_syscall el0_svc_common.constprop.0 do_el0_svc el0_svc el0t_64_sync_handler el0t_64_sync irq event stamp: 0 hardirqs last enabled at (0): 0x0 hardirqs last disabled at (0): copy_process softirqs last enabled at (0): copy_process softirqs last disabled at (0): 0x0