From 8a931f801340c2be10552c7b5622d5f4852f3a36 Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Wed, 1 Apr 2020 21:07:07 -0700 Subject: [PATCH] mm: memcontrol: recursive memory.low protection MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Right now, the effective protection of any given cgroup is capped by its own explicit memory.low setting, regardless of what the parent says. The reasons for this are mostly historical and ease of implementation: to make delegation of memory.low safe, effective protection is the min() of all memory.low up the tree. Unfortunately, this limitation makes it impossible to protect an entire subtree from another without forcing the user to make explicit protection allocations all the way to the leaf cgroups - something that is highly undesirable in real life scenarios. Consider memory in a data center host. At the cgroup top level, we have a distinction between system management software and the actual workload the system is executing. Both branches are further subdivided into individual services, job components etc. We want to protect the workload as a whole from the system management software, but that doesn't mean we want to protect and prioritize individual workload wrt each other. Their memory demand can vary over time, and we'd want the VM to simply cache the hottest data within the workload subtree. Yet, the current memory.low limitations force us to allocate a fixed amount of protection to each workload component in order to get protection from system management software in general. This results in very inefficient resource distribution. Another concern with mandating downward allocation is that, as the complexity of the cgroup tree grows, it gets harder for the lower levels to be informed about decisions made at the host-level. Consider a container inside a namespace that in turn creates its own nested tree of cgroups to run multiple workloads. It'd be extremely difficult to configure memory.low parameters in those leaf cgroups that on one hand balance pressure among siblings as the container desires, while also reflecting the host-level protection from e.g. rpm upgrades, that lie beyond one or more delegation and namespacing points in the tree. It's highly unusual from a cgroup interface POV that nested levels have to be aware of and reflect decisions made at higher levels for them to be effective. To enable such use cases and scale configurability for complex trees, this patch implements a resource inheritance model for memory that is similar to how the CPU and the IO controller implement work-conserving resource allocations: a share of a resource allocated to a subree always applies to the entire subtree recursively, while allowing, but not mandating, children to further specify distribution rules. That means that if protection is explicitly allocated among siblings, those configured shares are being followed during page reclaim just like they are now. However, if the memory.low set at a higher level is not fully claimed by the children in that subtree, the "floating" remainder is applied to each cgroup in the tree in proportion to its size. Since reclaim pressure is applied in proportion to size as well, each child in that tree gets the same boost, and the effect is neutral among siblings - with respect to each other, they behave as if no memory control was enabled at all, and the VM simply balances the memory demands optimally within the subtree. But collectively those cgroups enjoy a boost over the cgroups in neighboring trees. E.g. a leaf cgroup with a memory.low setting of 0 no longer means that it's not getting a share of the hierarchically assigned resource, just that it doesn't claim a fixed amount of it to protect from its siblings. This allows us to recursively protect one subtree (workload) from another (system management), while letting subgroups compete freely among each other - without having to assign fixed shares to each leaf, and without nested groups having to echo higher-level settings. The floating protection composes naturally with fixed protection. Consider the following example tree: A A: low = 2G / \ A1: low = 1G A1 A2 A2: low = 0G As outside pressure is applied to this tree, A1 will enjoy a fixed protection from A2 of 1G, but the remaining, unclaimed 1G from A is split evenly among A1 and A2, coming out to 1.5G and 0.5G. There is a slight risk of regressing theoretical setups where the top-level cgroups don't know about the true budgeting and set bogusly high "bypass" values that are meaningfully allocated down the tree. Such setups would rely on unclaimed protection to be discarded, and distributing it would change the intended behavior. Be safe and hide the new behavior behind a mount option, 'memory_recursiveprot'. Signed-off-by: Johannes Weiner Signed-off-by: Andrew Morton Acked-by: Tejun Heo Acked-by: Roman Gushchin Acked-by: Chris Down Cc: Michal Hocko Cc: Michal Koutný Link: http://lkml.kernel.org/r/20200227195606.46212-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds --- Documentation/admin-guide/cgroup-v2.rst | 11 ++++++ include/linux/cgroup-defs.h | 5 +++ kernel/cgroup/cgroup.c | 17 ++++++++- mm/memcontrol.c | 51 +++++++++++++++++++++++-- 4 files changed, 79 insertions(+), 5 deletions(-) diff --git a/Documentation/admin-guide/cgroup-v2.rst b/Documentation/admin-guide/cgroup-v2.rst index fbb111616705..bcc80269bb6a 100644 --- a/Documentation/admin-guide/cgroup-v2.rst +++ b/Documentation/admin-guide/cgroup-v2.rst @@ -188,6 +188,17 @@ cgroup v2 currently supports the following mount options. modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts. + memory_recursiveprot + + Recursively apply memory.min and memory.low protection to + entire subtrees, without requiring explicit downward + propagation into leaf cgroups. This allows protecting entire + subtrees from one another, while retaining free competition + within those subtrees. This should have been the default + behavior but is a mount-option to avoid regressing setups + relying on the original semantics (e.g. specifying bogusly + high 'bypass' protection values at higher tree levels). + Organizing Processes and Threads -------------------------------- diff --git a/include/linux/cgroup-defs.h b/include/linux/cgroup-defs.h index 63097cb243cb..e1fafed22db1 100644 --- a/include/linux/cgroup-defs.h +++ b/include/linux/cgroup-defs.h @@ -94,6 +94,11 @@ enum { * Enable legacy local memory.events. */ CGRP_ROOT_MEMORY_LOCAL_EVENTS = (1 << 5), + + /* + * Enable recursive subtree protection + */ + CGRP_ROOT_MEMORY_RECURSIVE_PROT = (1 << 6), }; /* cftype->flags */ diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index 915dda3f7f19..755c07d845ce 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -1813,12 +1813,14 @@ int cgroup_show_path(struct seq_file *sf, struct kernfs_node *kf_node, enum cgroup2_param { Opt_nsdelegate, Opt_memory_localevents, + Opt_memory_recursiveprot, nr__cgroup2_params }; static const struct fs_parameter_spec cgroup2_fs_parameters[] = { fsparam_flag("nsdelegate", Opt_nsdelegate), fsparam_flag("memory_localevents", Opt_memory_localevents), + fsparam_flag("memory_recursiveprot", Opt_memory_recursiveprot), {} }; @@ -1839,6 +1841,9 @@ static int cgroup2_parse_param(struct fs_context *fc, struct fs_parameter *param case Opt_memory_localevents: ctx->flags |= CGRP_ROOT_MEMORY_LOCAL_EVENTS; return 0; + case Opt_memory_recursiveprot: + ctx->flags |= CGRP_ROOT_MEMORY_RECURSIVE_PROT; + return 0; } return -EINVAL; } @@ -1855,6 +1860,11 @@ static void apply_cgroup_root_flags(unsigned int root_flags) cgrp_dfl_root.flags |= CGRP_ROOT_MEMORY_LOCAL_EVENTS; else cgrp_dfl_root.flags &= ~CGRP_ROOT_MEMORY_LOCAL_EVENTS; + + if (root_flags & CGRP_ROOT_MEMORY_RECURSIVE_PROT) + cgrp_dfl_root.flags |= CGRP_ROOT_MEMORY_RECURSIVE_PROT; + else + cgrp_dfl_root.flags &= ~CGRP_ROOT_MEMORY_RECURSIVE_PROT; } } @@ -1864,6 +1874,8 @@ static int cgroup_show_options(struct seq_file *seq, struct kernfs_root *kf_root seq_puts(seq, ",nsdelegate"); if (cgrp_dfl_root.flags & CGRP_ROOT_MEMORY_LOCAL_EVENTS) seq_puts(seq, ",memory_localevents"); + if (cgrp_dfl_root.flags & CGRP_ROOT_MEMORY_RECURSIVE_PROT) + seq_puts(seq, ",memory_recursiveprot"); return 0; } @@ -6412,7 +6424,10 @@ static struct kobj_attribute cgroup_delegate_attr = __ATTR_RO(delegate); static ssize_t features_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { - return snprintf(buf, PAGE_SIZE, "nsdelegate\nmemory_localevents\n"); + return snprintf(buf, PAGE_SIZE, + "nsdelegate\n" + "memory_localevents\n" + "memory_recursiveprot\n"); } static struct kobj_attribute cgroup_features_attr = __ATTR_RO(features); diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 0e3a8c11fb3b..07032c088608 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -6264,13 +6264,27 @@ struct cgroup_subsys memory_cgrp_subsys = { * budget is NOT proportional. A cgroup's protection from a sibling * is capped to its own memory.min/low setting. * + * 5. However, to allow protecting recursive subtrees from each other + * without having to declare each individual cgroup's fixed share + * of the ancestor's claim to protection, any unutilized - + * "floating" - protection from up the tree is distributed in + * proportion to each cgroup's *usage*. This makes the protection + * neutral wrt sibling cgroups and lets them compete freely over + * the shared parental protection budget, but it protects the + * subtree as a whole from neighboring subtrees. + * + * Note that 4. and 5. are not in conflict: 4. is about protecting + * against immediate siblings whereas 5. is about protecting against + * neighboring subtrees. */ static unsigned long effective_protection(unsigned long usage, + unsigned long parent_usage, unsigned long setting, unsigned long parent_effective, unsigned long siblings_protected) { unsigned long protected; + unsigned long ep; protected = min(usage, setting); /* @@ -6301,7 +6315,34 @@ static unsigned long effective_protection(unsigned long usage, * protection is always dependent on how memory is actually * consumed among the siblings anyway. */ - return protected; + ep = protected; + + /* + * If the children aren't claiming (all of) the protection + * afforded to them by the parent, distribute the remainder in + * proportion to the (unprotected) memory of each cgroup. That + * way, cgroups that aren't explicitly prioritized wrt each + * other compete freely over the allowance, but they are + * collectively protected from neighboring trees. + * + * We're using unprotected memory for the weight so that if + * some cgroups DO claim explicit protection, we don't protect + * the same bytes twice. + */ + if (!(cgrp_dfl_root.flags & CGRP_ROOT_MEMORY_RECURSIVE_PROT)) + return ep; + + if (parent_effective > siblings_protected && usage > protected) { + unsigned long unclaimed; + + unclaimed = parent_effective - siblings_protected; + unclaimed *= usage - protected; + unclaimed /= parent_usage - siblings_protected; + + ep += unclaimed; + } + + return ep; } /** @@ -6321,8 +6362,8 @@ static unsigned long effective_protection(unsigned long usage, enum mem_cgroup_protection mem_cgroup_protected(struct mem_cgroup *root, struct mem_cgroup *memcg) { + unsigned long usage, parent_usage; struct mem_cgroup *parent; - unsigned long usage; if (mem_cgroup_disabled()) return MEMCG_PROT_NONE; @@ -6347,11 +6388,13 @@ enum mem_cgroup_protection mem_cgroup_protected(struct mem_cgroup *root, goto out; } - memcg->memory.emin = effective_protection(usage, + parent_usage = page_counter_read(&parent->memory); + + memcg->memory.emin = effective_protection(usage, parent_usage, memcg->memory.min, READ_ONCE(parent->memory.emin), atomic_long_read(&parent->memory.children_min_usage)); - memcg->memory.elow = effective_protection(usage, + memcg->memory.elow = effective_protection(usage, parent_usage, memcg->memory.low, READ_ONCE(parent->memory.elow), atomic_long_read(&parent->memory.children_low_usage));