Re: [RFC PATCH 6/9] livepatch: create per-task consistency model

From: Josh Poimboeuf
Date: Tue Feb 17 2015 - 10:18:08 EST

On Sat, Feb 14, 2015 at 12:40:01PM +0100, Jiri Slaby wrote:
> On 02/09/2015, 06:31 PM, Josh Poimboeuf wrote:
> > Add a basic per-task consistency model. This is the foundation which
> > will eventually enable us to patch those ~10% of security patches which
> > change function prototypes and/or data semantics.
> >
> > When a patch is enabled, livepatch enters into a transition state where
> > tasks are converging from the old universe to the new universe. If a
> > given task isn't using any of the patched functions, it's switched to
> > the new universe. Once all the tasks have been converged to the new
> > universe, patching is complete.
> >
> > The same sequence occurs when a patch is disabled, except the tasks
> > converge from the new universe to the old universe.
> >
> > The /sys/kernel/livepatch/<patch>/transition file shows whether a patch
> > is in transition. Only a single patch (the topmost patch on the stack)
> > can be in transition at a given time. A patch can remain in the
> > transition state indefinitely, if any of the tasks are stuck in the
> > previous universe.
> >
> > A transition can be reversed and effectively canceled by writing the
> > opposite value to the /sys/kernel/livepatch/<patch>/enabled file while
> > the transition is in progress. Then all the tasks will attempt to
> > converge back to the original universe.
> >
> > Signed-off-by: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
> > ---
> > include/linux/livepatch.h | 18 ++-
> > include/linux/sched.h | 3 +
> > kernel/fork.c | 2 +
> > kernel/livepatch/Makefile | 2 +-
> > kernel/livepatch/core.c | 71 ++++++----
> > kernel/livepatch/patch.c | 34 ++++-
> > kernel/livepatch/patch.h | 1 +
> > kernel/livepatch/transition.c | 300 ++++++++++++++++++++++++++++++++++++++++++
> > kernel/livepatch/transition.h | 16 +++
> > kernel/sched/core.c | 2 +
> > 10 files changed, 423 insertions(+), 26 deletions(-)
> > create mode 100644 kernel/livepatch/transition.c
> > create mode 100644 kernel/livepatch/transition.h
> >
> > diff --git a/include/linux/livepatch.h b/include/linux/livepatch.h
> > index 0e65b4d..b8c2f15 100644
> > --- a/include/linux/livepatch.h
> > +++ b/include/linux/livepatch.h
> > @@ -40,6 +40,7 @@
> > * @old_size: size of the old function
> > * @new_size: size of the new function
> > * @patched: the func has been added to the klp_ops list
> > + * @transition: the func is currently being applied or reverted
> > */
> > struct klp_func {
> > /* external */
> > @@ -60,6 +61,7 @@ struct klp_func {
> > struct list_head stack_node;
> > unsigned long old_size, new_size;
> > int patched;
> > + int transition;
> > };
> >
> > /**
> > @@ -128,6 +130,20 @@ extern int klp_unregister_patch(struct klp_patch *);
> > extern int klp_enable_patch(struct klp_patch *);
> > extern int klp_disable_patch(struct klp_patch *);
> >
> > -#endif /* CONFIG_LIVEPATCH */
> > +extern int klp_universe_goal;
> > +
> > +static inline void klp_update_task_universe(struct task_struct *t)
> > +{
> > + /* corresponding smp_wmb() is in klp_set_universe_goal() */
> > + smp_rmb();
> > +
> > + t->klp_universe = klp_universe_goal;
> > +}
> > +
> > +#else /* !CONFIG_LIVEPATCH */
> > +
> > +static inline void klp_update_task_universe(struct task_struct *t) {}
> > +
> > +#endif /* !CONFIG_LIVEPATCH */
> >
> > #endif /* _LINUX_LIVEPATCH_H_ */
> > diff --git a/include/linux/sched.h b/include/linux/sched.h
> > index 8db31ef..a95e59a 100644
> > --- a/include/linux/sched.h
> > +++ b/include/linux/sched.h
> > @@ -1701,6 +1701,9 @@ struct task_struct {
> > #ifdef CONFIG_DEBUG_ATOMIC_SLEEP
> > unsigned long task_state_change;
> > #endif
> > +#ifdef CONFIG_LIVEPATCH
> > + int klp_universe;
> > +#endif
> > };
> >
> > /* Future-safe accessor for struct task_struct's cpus_allowed. */
> > diff --git a/kernel/fork.c b/kernel/fork.c
> > index 4dc2dda..1dcbebe 100644
> > --- a/kernel/fork.c
> > +++ b/kernel/fork.c
> > @@ -74,6 +74,7 @@
> > #include <linux/uprobes.h>
> > #include <linux/aio.h>
> > #include <linux/compiler.h>
> > +#include <linux/livepatch.h>
> >
> > #include <asm/pgtable.h>
> > #include <asm/pgalloc.h>
> > @@ -1538,6 +1539,7 @@ static struct task_struct *copy_process(unsigned long clone_flags,
> > total_forks++;
> > spin_unlock(&current->sighand->siglock);
> > syscall_tracepoint_update(p);
> > + klp_update_task_universe(p);
> > write_unlock_irq(&tasklist_lock);
> >
> > proc_fork_connector(p);
> > diff --git a/kernel/livepatch/Makefile b/kernel/livepatch/Makefile
> > index e136dad..2b8bdb1 100644
> > --- a/kernel/livepatch/Makefile
> > +++ b/kernel/livepatch/Makefile
> > @@ -1,3 +1,3 @@
> > obj-$(CONFIG_LIVEPATCH) += livepatch.o
> >
> > -livepatch-objs := core.o patch.o
> > +livepatch-objs := core.o patch.o transition.o
> > diff --git a/kernel/livepatch/core.c b/kernel/livepatch/core.c
> > index 85d4ef7..790dc10 100644
> > --- a/kernel/livepatch/core.c
> > +++ b/kernel/livepatch/core.c
> > @@ -28,14 +28,17 @@
> > #include <linux/kallsyms.h>
> >
> > #include "patch.h"
> > +#include "transition.h"
> >
> > /*
> > - * The klp_mutex protects the global lists and state transitions of any
> > - * structure reachable from them. References to any structure must be obtained
> > - * under mutex protection (except in klp_ftrace_handler(), which uses RCU to
> > - * ensure it gets consistent data).
> > + * The klp_mutex is a coarse lock which serializes access to klp data. All
> > + * accesses to klp-related variables and structures must have mutex protection,
> > + * except within the following functions which carefully avoid the need for it:
> > + *
> > + * - klp_ftrace_handler()
> > + * - klp_update_task_universe()
> > */
> > -static DEFINE_MUTEX(klp_mutex);
> > +DEFINE_MUTEX(klp_mutex);
> >
> > static LIST_HEAD(klp_patches);
> >
> > @@ -67,7 +70,6 @@ static void klp_find_object_module(struct klp_object *obj)
> > mutex_unlock(&module_mutex);
> > }
> >
> > -/* klp_mutex must be held by caller */
> > static bool klp_is_patch_registered(struct klp_patch *patch)
> > {
> > struct klp_patch *mypatch;
> > @@ -285,18 +287,17 @@ static int klp_write_object_relocations(struct module *pmod,
> >
> > static int __klp_disable_patch(struct klp_patch *patch)
> > {
> > - struct klp_object *obj;
> > + if (klp_transition_patch)
> > + return -EBUSY;
> >
> > /* enforce stacking: only the last enabled patch can be disabled */
> > if (!list_is_last(&patch->list, &klp_patches) &&
> > list_next_entry(patch, list)->enabled)
> > return -EBUSY;
> >
> > - pr_notice("disabling patch '%s'\n", patch->mod->name);
> > -
> > - for (obj = patch->objs; obj->funcs; obj++)
> > - if (obj->patched)
> > - klp_unpatch_object(obj);
> > + klp_init_transition(patch, KLP_UNIVERSE_NEW);
> > + klp_start_transition(KLP_UNIVERSE_OLD);
> > + klp_try_complete_transition();
> >
> > patch->enabled = 0;
> >
> > @@ -340,6 +341,9 @@ static int __klp_enable_patch(struct klp_patch *patch)
> > struct klp_object *obj;
> > int ret;
> >
> > + if (klp_transition_patch)
> > + return -EBUSY;
> > +
> > if (WARN_ON(patch->enabled))
> > return -EINVAL;
> >
> > @@ -351,7 +355,7 @@ static int __klp_enable_patch(struct klp_patch *patch)
> > pr_notice_once("tainting kernel with TAINT_LIVEPATCH\n");
> > add_taint(TAINT_LIVEPATCH, LOCKDEP_STILL_OK);
> >
> > - pr_notice("enabling patch '%s'\n", patch->mod->name);
> > + klp_init_transition(patch, KLP_UNIVERSE_OLD);
> >
> > for (obj = patch->objs; obj->funcs; obj++) {
> > klp_find_object_module(obj);
> > @@ -360,17 +364,24 @@ static int __klp_enable_patch(struct klp_patch *patch)
> > continue;
> >
> > ret = klp_patch_object(obj);
> > - if (ret)
> > - goto unregister;
> > + if (ret) {
> > + pr_warn("failed to enable patch '%s'\n",
> > + patch->mod->name);
> > +
> > + klp_unpatch_objects(patch);
> > + klp_complete_transition();
> > +
> > + return ret;
> > + }
> > }
> >
> > + klp_start_transition(KLP_UNIVERSE_NEW);
> > +
> > + klp_try_complete_transition();
> > +
> > patch->enabled = 1;
> >
> > return 0;
> > -
> > -unregister:
> > - WARN_ON(__klp_disable_patch(patch));
> > - return ret;
> > }
> >
> > /**
> > @@ -407,6 +418,7 @@ EXPORT_SYMBOL_GPL(klp_enable_patch);
> > * /sys/kernel/livepatch
> > * /sys/kernel/livepatch/<patch>
> > * /sys/kernel/livepatch/<patch>/enabled
> > + * /sys/kernel/livepatch/<patch>/transition
> > * /sys/kernel/livepatch/<patch>/<object>
> > * /sys/kernel/livepatch/<patch>/<object>/<func>
> > */
> > @@ -435,7 +447,9 @@ static ssize_t enabled_store(struct kobject *kobj, struct kobj_attribute *attr,
> > goto err;
> > }
> >
> > - if (val) {
> > + if (klp_transition_patch == patch) {
> > + klp_reverse_transition();
> > + } else if (val) {
> > ret = __klp_enable_patch(patch);
> > if (ret)
> > goto err;
> > @@ -463,9 +477,21 @@ static ssize_t enabled_show(struct kobject *kobj,
> > return snprintf(buf, PAGE_SIZE-1, "%d\n", patch->enabled);
> > }
> >
> > +static ssize_t transition_show(struct kobject *kobj,
> > + struct kobj_attribute *attr, char *buf)
> > +{
> > + struct klp_patch *patch;
> > +
> > + patch = container_of(kobj, struct klp_patch, kobj);
> > + return snprintf(buf, PAGE_SIZE-1, "%d\n",
> > + klp_transition_patch == patch);
> > +}
> > +
> > static struct kobj_attribute enabled_kobj_attr = __ATTR_RW(enabled);
> > +static struct kobj_attribute transition_kobj_attr = __ATTR_RO(transition);
> > static struct attribute *klp_patch_attrs[] = {
> > &enabled_kobj_attr.attr,
> > + &transition_kobj_attr.attr,
> > NULL
> > };
> >
> > @@ -543,6 +569,7 @@ static int klp_init_func(struct klp_object *obj, struct klp_func *func)
> > {
> > INIT_LIST_HEAD(&func->stack_node);
> > func->patched = 0;
> > + func->transition = 0;
> >
> > return kobject_init_and_add(&func->kobj, &klp_ktype_func,
> > obj->kobj, func->old_name);
> > @@ -725,7 +752,7 @@ static void klp_module_notify_coming(struct klp_patch *patch,
> > if (ret)
> > goto err;
> >
> > - if (!patch->enabled)
> > + if (!patch->enabled && klp_transition_patch != patch)
> > return;
> >
> > pr_notice("applying patch '%s' to loading module '%s'\n",
> > @@ -746,7 +773,7 @@ static void klp_module_notify_going(struct klp_patch *patch,
> > struct module *pmod = patch->mod;
> > struct module *mod = obj->mod;
> >
> > - if (!patch->enabled)
> > + if (!patch->enabled && klp_transition_patch != patch)
> > goto free;
> >
> > pr_notice("reverting patch '%s' on unloading module '%s'\n",
> > diff --git a/kernel/livepatch/patch.c b/kernel/livepatch/patch.c
> > index 281fbca..f12256b 100644
> > --- a/kernel/livepatch/patch.c
> > +++ b/kernel/livepatch/patch.c
> > @@ -24,6 +24,7 @@
> > #include <linux/slab.h>
> >
> > #include "patch.h"
> > +#include "transition.h"
> >
> > static LIST_HEAD(klp_ops);
> >
> > @@ -38,14 +39,34 @@ static void notrace klp_ftrace_handler(unsigned long ip,
> > ops = container_of(fops, struct klp_ops, fops);
> >
> > rcu_read_lock();
> > +
> > func = list_first_or_null_rcu(&ops->func_stack, struct klp_func,
> > stack_node);
> > - rcu_read_unlock();
> >
> > if (WARN_ON_ONCE(!func))
> > - return;
> > + goto unlock;
> > +
> > + if (unlikely(func->transition)) {
> > + /* corresponding smp_wmb() is in klp_init_transition() */
> > + smp_rmb();
> > +
> > + if (current->klp_universe == KLP_UNIVERSE_OLD) {
> > + /*
> > + * Use the previously patched version of the function.
> > + * If no previous patches exist, use the original
> > + * function.
> > + */
> > + func = list_entry_rcu(func->stack_node.next,
> > + struct klp_func, stack_node);
> > +
> > + if (&func->stack_node == &ops->func_stack)
> > + goto unlock;
> > + }
> > + }
> >
> > klp_arch_set_pc(regs, (unsigned long)func->new_func);
> > +unlock:
> > + rcu_read_unlock();
> > }
> >
> > struct klp_ops *klp_find_ops(unsigned long old_addr)
> > @@ -174,3 +195,12 @@ int klp_patch_object(struct klp_object *obj)
> >
> > return 0;
> > }
> > +
> > +void klp_unpatch_objects(struct klp_patch *patch)
> > +{
> > + struct klp_object *obj;
> > +
> > + for (obj = patch->objs; obj->funcs; obj++)
> > + if (obj->patched)
> > + klp_unpatch_object(obj);
> > +}
> > diff --git a/kernel/livepatch/patch.h b/kernel/livepatch/patch.h
> > index bb34bd3..1648259 100644
> > --- a/kernel/livepatch/patch.h
> > +++ b/kernel/livepatch/patch.h
> > @@ -23,3 +23,4 @@ struct klp_ops *klp_find_ops(unsigned long old_addr);
> >
> > extern int klp_patch_object(struct klp_object *obj);
> > extern void klp_unpatch_object(struct klp_object *obj);
> > +extern void klp_unpatch_objects(struct klp_patch *patch);
> > diff --git a/kernel/livepatch/transition.c b/kernel/livepatch/transition.c
> > new file mode 100644
> > index 0000000..2630296
> > --- /dev/null
> > +++ b/kernel/livepatch/transition.c
> > @@ -0,0 +1,300 @@
> > +/*
> > + * transition.c - Kernel Live Patching transition functions
> > + *
> > + * Copyright (C) 2015 Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of the GNU General Public License
> > + * as published by the Free Software Foundation; either version 2
> > + * of the License, or (at your option) any later version.
> > + *
> > + * This program is distributed in the hope that it will be useful,
> > + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > + * GNU General Public License for more details.
> > + *
> > + * You should have received a copy of the GNU General Public License
> > + * along with this program; if not, see <http://www.gnu.org/licenses/>.
> > + */
> > +
> > +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
> > +
> > +#include <linux/cpu.h>
> > +#include <asm/stacktrace.h>
> > +#include "../sched/sched.h"
> > +
> > +#include "patch.h"
> > +#include "transition.h"
> > +
> > +static void klp_transition_work_fn(struct work_struct *);
> > +static DECLARE_DELAYED_WORK(klp_transition_work, klp_transition_work_fn);
> > +
> > +struct klp_patch *klp_transition_patch;
> > +
> > +int klp_universe_goal = KLP_UNIVERSE_UNDEFINED;
> > +
> > +static void klp_set_universe_goal(int universe)
> > +{
> > + klp_universe_goal = universe;
> > +
> > + /* corresponding smp_rmb() is in klp_update_task_universe() */
> > + smp_wmb();
> > +}
> > +
> > +/*
> > + * The transition to the universe goal is complete. Clean up the data
> > + * structures.
> > + */
> > +void klp_complete_transition(void)
> > +{
> > + struct klp_object *obj;
> > + struct klp_func *func;
> > +
> > + for (obj = klp_transition_patch->objs; obj->funcs; obj++)
> > + for (func = obj->funcs; func->old_name; func++)
> > + func->transition = 0;
> > +
> > + klp_transition_patch = NULL;
> > +}
> > +
> > +static int klp_stacktrace_address_verify_func(struct klp_func *func,
> > + unsigned long address)
> > +{
> > + unsigned long func_addr, func_size;
> > +
> > + if (klp_universe_goal == KLP_UNIVERSE_OLD) {
> > + /* check the to-be-unpatched function (the func itself) */
> > + func_addr = (unsigned long)func->new_func;
> > + func_size = func->new_size;
> > + } else {
> > + /* check the to-be-patched function (previous func) */
> > + struct klp_ops *ops;
> > +
> > + ops = klp_find_ops(func->old_addr);
> > +
> > + if (list_is_singular(&ops->func_stack)) {
> > + /* original function */
> > + func_addr = func->old_addr;
> > + func_size = func->old_size;
> > + } else {
> > + /* previously patched function */
> > + struct klp_func *prev;
> > +
> > + prev = list_next_entry(func, stack_node);
> > + func_addr = (unsigned long)prev->new_func;
> > + func_size = prev->new_size;
> > + }
> > + }
> > +
> > + if (address >= func_addr && address < func_addr + func_size)
> > + return -1;
> > +
> > + return 0;
> > +}
> > +
> > +/*
> > + * Determine whether the given return address on the stack is within a
> > + * to-be-patched or to-be-unpatched function.
> > + */
> > +static void klp_stacktrace_address_verify(void *data, unsigned long address,
> > + int reliable)
> > +{
> > + struct klp_object *obj;
> > + struct klp_func *func;
> > + int *ret = data;
> > +
> > + if (*ret)
> > + return;
> > +
> > + for (obj = klp_transition_patch->objs; obj->funcs; obj++) {
> > + if (!obj->patched)
> > + continue;
> > + for (func = obj->funcs; func->old_name; func++) {
> > + if (klp_stacktrace_address_verify_func(func, address)) {
> > + *ret = -1;
> > + return;
> > + }
> > + }
> > + }
> > +}
> > +
> > +static int klp_stacktrace_stack(void *data, char *name)
> > +{
> > + return 0;
> > +}
> > +
> > +static const struct stacktrace_ops klp_stacktrace_ops = {
> > + .address = klp_stacktrace_address_verify,
> > + .stack = klp_stacktrace_stack,
> > + .walk_stack = print_context_stack_bp,
> > +};
> > +
> > +/*
> > + * Try to safely transition a task to the universe goal. If the task is
> > + * currently running or is sleeping on a to-be-patched or to-be-unpatched
> > + * function, return false.
> > + */
> > +static bool klp_transition_task(struct task_struct *t)
> > +{
> > + struct rq *rq;
> > + unsigned long flags;
> > + int ret;
> > + bool success = false;
> > +
> > + if (t->klp_universe == klp_universe_goal)
> > + return true;
> > +
> > + rq = task_rq_lock(t, &flags);
> > +
> > + if (task_running(rq, t) && t != current) {
> > + pr_debug("%s: pid %d (%s) is running\n", __func__, t->pid,
> > + t->comm);
> > + goto done;
> > + }
> > +
> > + ret = 0;
> > + dump_trace(t, NULL, NULL, 0, &klp_stacktrace_ops, &ret);
> > + if (ret) {
> > + pr_debug("%s: pid %d (%s) is sleeping on a patched function\n",
> > + __func__, t->pid, t->comm);
> > + goto done;
> > + }
> > +
> > + klp_update_task_universe(t);
> > +
> > + success = true;
> > +done:
> > + task_rq_unlock(rq, t, &flags);
> > + return success;
> > +}
> > +
> > +/*
> > + * Try to transition all tasks to the universe goal. If any tasks are still
> > + * stuck in the original universe, schedule a retry.
> > + */
> > +void klp_try_complete_transition(void)
> > +{
> > + unsigned int cpu;
> > + struct task_struct *g, *t;
> > + bool complete = true;
> > +
> > + /* try to transition all normal tasks */
> > + read_lock(&tasklist_lock);
> > + for_each_process_thread(g, t)
> > + if (!klp_transition_task(t))
> > + complete = false;
> > + read_unlock(&tasklist_lock);
> > +
> > + /* try to transition the idle "swapper" tasks */
> > + get_online_cpus();
> > + for_each_online_cpu(cpu)
> > + if (!klp_transition_task(idle_task(cpu)))
> > + complete = false;
> > + put_online_cpus();
> > +
> > + /* if not complete, try again later */
> > + if (!complete) {
> > + schedule_delayed_work(&klp_transition_work,
> > + round_jiffies_relative(HZ));
> > + return;
> > + }
> > +
> > + /* success! unpatch obsolete functions and do some cleanup */
> > +
> > + if (klp_universe_goal == KLP_UNIVERSE_OLD) {
> > + klp_unpatch_objects(klp_transition_patch);
> > +
> > + /* prevent ftrace handler from reading old func->transition */
> > + synchronize_rcu();
> > + }
> > +
> > + pr_notice("'%s': %s complete\n", klp_transition_patch->mod->name,
> > + klp_universe_goal == KLP_UNIVERSE_NEW ? "patching" :
> > + "unpatching");
> > +
> > + klp_complete_transition();
> > +}
> > +
> > +static void klp_transition_work_fn(struct work_struct *work)
> > +{
> > + mutex_lock(&klp_mutex);
> > +
> > + if (klp_transition_patch)
> > + klp_try_complete_transition();
> > +
> > + mutex_unlock(&klp_mutex);
> > +}
> > +
> > +/*
> > + * Start the transition to the specified universe so tasks can begin switching
> > + * to it.
> > + */
> > +void klp_start_transition(int universe)
> > +{
> > + if (WARN_ON(klp_universe_goal == universe))
> > + return;
> > +
> > + pr_notice("'%s': %s...\n", klp_transition_patch->mod->name,
> > + universe == KLP_UNIVERSE_NEW ? "patching" : "unpatching");
> > +
> > + klp_set_universe_goal(universe);
> > +}
> > +
> > +/*
> > + * Can be called in the middle of an existing transition to reverse the
> > + * direction of the universe goal. This can be done to effectively cancel an
> > + * existing enable or disable operation if there are any tasks which are stuck
> > + * in the original universe.
> > + */
> > +void klp_reverse_transition(void)
> > +{
> > + struct klp_patch *patch = klp_transition_patch;
> > +
> > + klp_start_transition(!klp_universe_goal);
> > + klp_try_complete_transition();
> > +
> > + patch->enabled = !patch->enabled;
> > +}
> > +
> > +/*
> > + * Reset the universe goal and all tasks to the starting universe, and set all
> > + * func->transition's to 1 to prepare for patching.
> > + */
> > +void klp_init_transition(struct klp_patch *patch, int universe)
> > +{
> > + struct task_struct *g, *t;
> > + unsigned int cpu;
> > + struct klp_object *obj;
> > + struct klp_func *func;
> > +
> > + klp_transition_patch = patch;
> > +
> > + /*
> > + * If the previous transition was in the opposite direction, we may
> > + * already be in the requested initial universe.
> > + */
> > + if (klp_universe_goal == universe)
> > + goto init_funcs;
> > +
> > + klp_set_universe_goal(universe);
> > +
> > + /* init all normal task universes */
> > + read_lock(&tasklist_lock);
> > + for_each_process_thread(g, t)
> > + klp_update_task_universe(t);
> > + read_unlock(&tasklist_lock);
> > +
> > + /* init all idle "swapper" task universes */
> > + get_online_cpus();
> > + for_each_online_cpu(cpu)
> > + klp_update_task_universe(idle_task(cpu));
> > + put_online_cpus();
> > +
> > +init_funcs:
> > + /* corresponding smp_rmb() is in klp_ftrace_handler() */
> > + smp_wmb();
> > +
> > + for (obj = patch->objs; obj->funcs; obj++)
> > + for (func = obj->funcs; func->old_name; func++)
> > + func->transition = 1;
>
> So I finally got to review of this one. I have only two concerns:
> 1) it removes the ability for the user to use 'no consistency model'.
> But you don't need to worry about this, I plan to implement this as soon
> as you send v2 of these.

Ok, sounds good.

> 2) How is this 'transition = 1' store above guaranteed to reach other
> CPUs before you start registering ftrace handlers? The CPUs need not see
> the update when some handler is already invoked before start_transition
> AFAICS.

Yeah, I think the order of the 'transition = 1' and adding the func to
the ops stack list should be enforced. Also I'll probably rework the
barriers a little bit in v2 so that they're more explicit and better
commented.

--
Josh
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/