On Thu, Dec 31, 2009 at 12:10:50AM -0500, Ben Blum wrote:
> [This is a revision of http://lkml.org/lkml/2009/12/21/211 ]
>
> This patch series implements support for building, loading, and
> unloading subsystems as modules, both within and outside the kernel
> source tree. It provides an interface cgroup_load_subsys() and
> cgroup_unload_subsys() which modular subsystems can use to register and
> depart during runtime. The net_cls classifier subsystem serves as the
> example for a subsystem which can be converted into a module using these
> changes.
>
> Patch #1 sets up the subsys[] array so its contents can be dynamic as
> modules appear and (eventually) disappear. Iterations over the array are
> modified to handle when subsystems are absent, and the dynamic section
> of the array is protected by cgroup_mutex.
>
> Patch #2 implements an interface for modules to load subsystems, called
> cgroup_load_subsys, similar to cgroup_init_subsys, and adds a module
> pointer in struct cgroup_subsys.
>
> Patch #3 adds a mechanism for unloading modular subsystems, which
> includes a more advanced rework of the rudimentary reference counting
> introduced in patch 2.
>
> Patch #4 modifies the net_cls subsystem, which already had some module
> declarations, to be configurable as a module, which also serves as a
> simple proof-of-concept.
>
> Part of implementing patches 2 and 4 involved updating css pointers in
> each css_set when the module appears or leaves. In doing this, it was
> discovered that css_sets always remain linked to the dummy cgroup,
> regardless of whether or not any subsystems are actually bound to it
> (i.e., not mounted on an actual hierarchy). The subsystem loading and
> unloading code therefore should keep in mind the special cases where the
> added subsystem is the only one in the dummy cgroup (and therefore all
> css_sets need to be linked back into it) and where the removed subsys
> was the only one in the dummy cgroup (and therefore all css_sets should
> be unlinked from it) - however, as all css_sets always stay attached to
> the dummy cgroup anyway, these cases are ignored. Any fix that addresses
> this issue should also make sure these cases are addressed in the
> subsystem loading and unloading code.
>
> -- bblum
>
> ---
> Documentation/cgroups/cgroups.txt | 9
> include/linux/cgroup.h | 18 +
> kernel/cgroup.c | 388
> +++++++++++++++++++++++++++++++++-----
> net/sched/Kconfig | 5
> net/sched/cls_cgroup.c | 36 ++-
> 5 files changed, 400 insertions(+), 56 deletions(-)
>
>
Add interface between cgroups subsystem management and module loading
From: Ben Blum <[email protected]>
This patch implements rudimentary module-loading support for cgroups - namely,
a cgroup_load_subsys (similar to cgroup_init_subsys) for use as a module
initcall, and a struct module pointer in struct cgroup_subsys.
Several functions that might be wanted by modules have had EXPORT_SYMBOL added
to them, but it's unclear exactly which functions want it and which won't.
Signed-off-by: Ben Blum <[email protected]>
Acked-by: Li Zefan <[email protected]>
---
Documentation/cgroups/cgroups.txt | 4 +
include/linux/cgroup.h | 4 +
kernel/cgroup.c | 128 +++++++++++++++++++++++++++++++++++++
3 files changed, 136 insertions(+), 0 deletions(-)
diff --git a/Documentation/cgroups/cgroups.txt
b/Documentation/cgroups/cgroups.txt
index 0b33bfe..6ffcf81 100644
--- a/Documentation/cgroups/cgroups.txt
+++ b/Documentation/cgroups/cgroups.txt
@@ -488,6 +488,10 @@ Each subsystem should:
- add an entry in linux/cgroup_subsys.h
- define a cgroup_subsys object called <name>_subsys
+If a subsystem can be compiled as a module, it should also have in its
+module initcall a call to cgroup_load_subsys(&its_subsys_struct). It
+should also set its_subsys.module = THIS_MODULE in its .c file.
+
Each subsystem may export the following methods. The only mandatory
methods are create/destroy. Any others that are null are presumed to
be successful no-ops.
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 83da43d..9461aed 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -36,6 +36,7 @@ extern void cgroup_post_fork(struct task_struct *p);
extern void cgroup_exit(struct task_struct *p, int run_callbacks);
extern int cgroupstats_build(struct cgroupstats *stats,
struct dentry *dentry);
+extern int cgroup_load_subsys(struct cgroup_subsys *ss);
extern const struct file_operations proc_cgroup_operations;
@@ -477,6 +478,9 @@ struct cgroup_subsys {
/* used when use_id == true */
struct idr idr;
spinlock_t id_lock;
+
+ /* should be defined only by modular subsystems */
+ struct module *module;
};
#define SUBSYS(_x) extern struct cgroup_subsys _x ## _subsys;
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 402e828..d7ca4cf 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -43,6 +43,7 @@
#include <linux/string.h>
#include <linux/sort.h>
#include <linux/kmod.h>
+#include <linux/module.h>
#include <linux/delayacct.h>
#include <linux/cgroupstats.h>
#include <linux/hash.h>
@@ -2084,6 +2085,7 @@ int cgroup_add_file(struct cgroup *cgrp,
error = PTR_ERR(dentry);
return error;
}
+EXPORT_SYMBOL_GPL(cgroup_add_file);
int cgroup_add_files(struct cgroup *cgrp,
struct cgroup_subsys *subsys,
@@ -2098,6 +2100,7 @@ int cgroup_add_files(struct cgroup *cgrp,
}
return 0;
}
+EXPORT_SYMBOL_GPL(cgroup_add_files);
/**
* cgroup_task_count - count the number of tasks in a cgroup.
@@ -3249,7 +3252,132 @@ static void __init cgroup_init_subsys(struct
cgroup_subsys *ss)
mutex_init(&ss->hierarchy_mutex);
lockdep_set_class(&ss->hierarchy_mutex, &ss->subsys_key);
ss->active = 1;
+
+ /* this function shouldn't be used with modular subsystems, since they
+ * need to register a subsys_id, among other things */
+ BUG_ON(ss->module);
+}
+
+/**
+ * cgroup_load_subsys: load and register a modular subsystem at runtime
+ * @ss: the subsystem to load
+ *
+ * This function should be called in a modular subsystem's initcall. If the
+ * subsytem is built as a module, it will be assigned a new subsys_id and set
+ * up for use. If the subsystem is built-in anyway, work is delegated to the
+ * simpler cgroup_init_subsys.
+ */
+int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss)
+{
+ int i;
+ struct cgroup_subsys_state *css;
+
+ /* check name and function validity */
+ if (ss->name == NULL || strlen(ss->name) > MAX_CGROUP_TYPE_NAMELEN ||
+ ss->create == NULL || ss->destroy == NULL)
+ return -EINVAL;
+
+ /*
+ * we don't support callbacks in modular subsystems. this check is
+ * before the ss->module check for consistency; a subsystem that could
+ * be a module should still have no callbacks even if the user isn't
+ * compiling it as one.
+ */
+ if (ss->fork || ss->exit)
+ return -EINVAL;
+
+ /*
+ * an optionally modular subsystem is built-in: we want to do nothing,
+ * since cgroup_init_subsys will have already taken care of it.
+ */
+ if (ss->module == NULL) {
+ /* a few sanity checks */
+ BUG_ON(ss->subsys_id >= CGROUP_BUILTIN_SUBSYS_COUNT);
+ BUG_ON(subsys[ss->subsys_id] != ss);
+ return 0;
+ }
+
+ /*
+ * need to register a subsys id before anything else - for example,
+ * init_cgroup_css needs it.
+ */
+ mutex_lock(&cgroup_mutex);
+ /* find the first empty slot in the array */
+ for (i = CGROUP_BUILTIN_SUBSYS_COUNT; i < CGROUP_SUBSYS_COUNT; i++) {
+ if (subsys[i] == NULL)
+ break;
+ }
+ if (i == CGROUP_SUBSYS_COUNT) {
+ /* maximum number of subsystems already registered! */
+ mutex_unlock(&cgroup_mutex);
+ return -EBUSY;
+ }
+ /* assign ourselves the subsys_id */
+ ss->subsys_id = i;
+ subsys[i] = ss;
+
+ /*
+ * no ss->create seems to need anything important in the ss struct, so
+ * this can happen first (i.e. before the rootnode attachment).
+ */
+ css = ss->create(ss, dummytop);
+ if (IS_ERR(css)) {
+ /* failure case - need to deassign the subsys[] slot. */
+ subsys[i] = NULL;
+ mutex_unlock(&cgroup_mutex);
+ return PTR_ERR(css);
+ }
+
+ list_add(&ss->sibling, &rootnode.subsys_list);
+ ss->root = &rootnode;
+
+ /* our new subsystem will be attached to the dummy hierarchy. */
+ init_cgroup_css(css, ss, dummytop);
+ /*
+ * Now we need to entangle the css into the existing css_sets. unlike
+ * in cgroup_init_subsys, there are now multiple css_sets, so each one
+ * will need a new pointer to it; done by iterating the css_set_table.
+ * furthermore, modifying the existing css_sets will corrupt the hash
+ * table state, so each changed css_set will need its hash recomputed.
+ * this is all done under the css_set_lock.
+ */
+ write_lock(&css_set_lock);
+ for (i = 0; i < CSS_SET_TABLE_SIZE; i++) {
+ struct css_set *cg;
+ struct hlist_node *node, *tmp;
+ struct hlist_head *bucket = &css_set_table[i], *new_bucket;
+
+ hlist_for_each_entry_safe(cg, node, tmp, bucket, hlist) {
+ /* skip entries that we already rehashed */
+ if (cg->subsys[ss->subsys_id])
+ continue;
+ /* remove existing entry */
+ hlist_del(&cg->hlist);
+ /* set new value */
+ cg->subsys[ss->subsys_id] = css;
+ /* recompute hash and restore entry */
+ new_bucket = css_set_hash(cg->subsys);
+ hlist_add_head(&cg->hlist, new_bucket);
+ }
+ }
+ write_unlock(&css_set_lock);
+
+ mutex_init(&ss->hierarchy_mutex);
+ lockdep_set_class(&ss->hierarchy_mutex, &ss->subsys_key);
+ ss->active = 1;
+
+ /*
+ * pin the subsystem's module so it doesn't go away. this shouldn't
+ * fail, since the module's initcall calls us.
+ * TODO: with module unloading, move this elsewhere
+ */
+ BUG_ON(!try_module_get(ss->module));
+
+ /* success! */
+ mutex_unlock(&cgroup_mutex);
+ return 0;
}
+EXPORT_SYMBOL_GPL(cgroup_load_subsys);
/**
* cgroup_init_early - cgroup initialization at system boot
_______________________________________________
Containers mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/containers
_______________________________________________
Devel mailing list
[email protected]
https://openvz.org/mailman/listinfo/devel
