0
0
mirror of https://mirrors.bfsu.edu.cn/git/linux.git synced 2024-12-21 18:14:48 +08:00
linux/kernel/rcu/srcutiny.c
Paul E. McKenney f5ad399149 srcu: Remove cleanup_srcu_struct_quiesced()
The cleanup_srcu_struct_quiesced() function was added because NVME
used WQ_MEM_RECLAIM workqueues and SRCU did not, which meant that
NVME workqueues waiting on SRCU workqueues could result in deadlocks
during low-memory conditions.  However, SRCU now also has WQ_MEM_RECLAIM
workqueues, so there is no longer a potential for deadlock.  Furthermore,
it turns out to be extremely hard to use cleanup_srcu_struct_quiesced()
correctly due to the fact that SRCU callback invocation accesses the
srcu_struct structure's per-CPU data area just after callbacks are
invoked.  Therefore, the usual practice of using srcu_barrier() to wait
for callbacks to be invoked before invoking cleanup_srcu_struct_quiesced()
fails because SRCU's callback-invocation workqueue handler might be
delayed, which can result in cleanup_srcu_struct_quiesced() being invoked
(and thus freeing the per-CPU data) before the SRCU's callback-invocation
workqueue handler is finished using that per-CPU data.  Nor is this a
theoretical problem: KASAN emitted use-after-free warnings because of
this problem on actual runs.

In short, NVME can now safely invoke cleanup_srcu_struct(), which
avoids the use-after-free scenario.  And cleanup_srcu_struct_quiesced()
is quite difficult to use safely.  This commit therefore removes
cleanup_srcu_struct_quiesced(), switching its sole user back to
cleanup_srcu_struct().  This effectively reverts the following pair
of commits:

f7194ac32c ("srcu: Add cleanup_srcu_struct_quiesced()")
4317228ad9 ("nvme: Avoid flush dependency in delete controller flow")

Reported-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Tested-by: Bart Van Assche <bvanassche@acm.org>
2019-03-26 14:39:24 -07:00

216 lines
5.9 KiB
C

// SPDX-License-Identifier: GPL-2.0+
/*
* Sleepable Read-Copy Update mechanism for mutual exclusion,
* tiny version for non-preemptible single-CPU use.
*
* Copyright (C) IBM Corporation, 2017
*
* Author: Paul McKenney <paulmck@linux.ibm.com>
*/
#include <linux/export.h>
#include <linux/mutex.h>
#include <linux/preempt.h>
#include <linux/rcupdate_wait.h>
#include <linux/sched.h>
#include <linux/delay.h>
#include <linux/srcu.h>
#include <linux/rcu_node_tree.h>
#include "rcu_segcblist.h"
#include "rcu.h"
int rcu_scheduler_active __read_mostly;
static LIST_HEAD(srcu_boot_list);
static bool srcu_init_done;
static int init_srcu_struct_fields(struct srcu_struct *ssp)
{
ssp->srcu_lock_nesting[0] = 0;
ssp->srcu_lock_nesting[1] = 0;
init_swait_queue_head(&ssp->srcu_wq);
ssp->srcu_cb_head = NULL;
ssp->srcu_cb_tail = &ssp->srcu_cb_head;
ssp->srcu_gp_running = false;
ssp->srcu_gp_waiting = false;
ssp->srcu_idx = 0;
INIT_WORK(&ssp->srcu_work, srcu_drive_gp);
INIT_LIST_HEAD(&ssp->srcu_work.entry);
return 0;
}
#ifdef CONFIG_DEBUG_LOCK_ALLOC
int __init_srcu_struct(struct srcu_struct *ssp, const char *name,
struct lock_class_key *key)
{
/* Don't re-initialize a lock while it is held. */
debug_check_no_locks_freed((void *)ssp, sizeof(*ssp));
lockdep_init_map(&ssp->dep_map, name, key, 0);
return init_srcu_struct_fields(ssp);
}
EXPORT_SYMBOL_GPL(__init_srcu_struct);
#else /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */
/*
* init_srcu_struct - initialize a sleep-RCU structure
* @ssp: structure to initialize.
*
* Must invoke this on a given srcu_struct before passing that srcu_struct
* to any other function. Each srcu_struct represents a separate domain
* of SRCU protection.
*/
int init_srcu_struct(struct srcu_struct *ssp)
{
return init_srcu_struct_fields(ssp);
}
EXPORT_SYMBOL_GPL(init_srcu_struct);
#endif /* #else #ifdef CONFIG_DEBUG_LOCK_ALLOC */
/*
* cleanup_srcu_struct - deconstruct a sleep-RCU structure
* @ssp: structure to clean up.
*
* Must invoke this after you are finished using a given srcu_struct that
* was initialized via init_srcu_struct(), else you leak memory.
*/
void cleanup_srcu_struct(struct srcu_struct *ssp)
{
WARN_ON(ssp->srcu_lock_nesting[0] || ssp->srcu_lock_nesting[1]);
flush_work(&ssp->srcu_work);
WARN_ON(ssp->srcu_gp_running);
WARN_ON(ssp->srcu_gp_waiting);
WARN_ON(ssp->srcu_cb_head);
WARN_ON(&ssp->srcu_cb_head != ssp->srcu_cb_tail);
}
EXPORT_SYMBOL_GPL(cleanup_srcu_struct);
/*
* Removes the count for the old reader from the appropriate element of
* the srcu_struct.
*/
void __srcu_read_unlock(struct srcu_struct *ssp, int idx)
{
int newval = ssp->srcu_lock_nesting[idx] - 1;
WRITE_ONCE(ssp->srcu_lock_nesting[idx], newval);
if (!newval && READ_ONCE(ssp->srcu_gp_waiting))
swake_up_one(&ssp->srcu_wq);
}
EXPORT_SYMBOL_GPL(__srcu_read_unlock);
/*
* Workqueue handler to drive one grace period and invoke any callbacks
* that become ready as a result. Single-CPU and !PREEMPT operation
* means that we get away with murder on synchronization. ;-)
*/
void srcu_drive_gp(struct work_struct *wp)
{
int idx;
struct rcu_head *lh;
struct rcu_head *rhp;
struct srcu_struct *ssp;
ssp = container_of(wp, struct srcu_struct, srcu_work);
if (ssp->srcu_gp_running || !READ_ONCE(ssp->srcu_cb_head))
return; /* Already running or nothing to do. */
/* Remove recently arrived callbacks and wait for readers. */
WRITE_ONCE(ssp->srcu_gp_running, true);
local_irq_disable();
lh = ssp->srcu_cb_head;
ssp->srcu_cb_head = NULL;
ssp->srcu_cb_tail = &ssp->srcu_cb_head;
local_irq_enable();
idx = ssp->srcu_idx;
WRITE_ONCE(ssp->srcu_idx, !ssp->srcu_idx);
WRITE_ONCE(ssp->srcu_gp_waiting, true); /* srcu_read_unlock() wakes! */
swait_event_exclusive(ssp->srcu_wq, !READ_ONCE(ssp->srcu_lock_nesting[idx]));
WRITE_ONCE(ssp->srcu_gp_waiting, false); /* srcu_read_unlock() cheap. */
/* Invoke the callbacks we removed above. */
while (lh) {
rhp = lh;
lh = lh->next;
local_bh_disable();
rhp->func(rhp);
local_bh_enable();
}
/*
* Enable rescheduling, and if there are more callbacks,
* reschedule ourselves. This can race with a call_srcu()
* at interrupt level, but the ->srcu_gp_running checks will
* straighten that out.
*/
WRITE_ONCE(ssp->srcu_gp_running, false);
if (READ_ONCE(ssp->srcu_cb_head))
schedule_work(&ssp->srcu_work);
}
EXPORT_SYMBOL_GPL(srcu_drive_gp);
/*
* Enqueue an SRCU callback on the specified srcu_struct structure,
* initiating grace-period processing if it is not already running.
*/
void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
rcu_callback_t func)
{
unsigned long flags;
rhp->func = func;
rhp->next = NULL;
local_irq_save(flags);
*ssp->srcu_cb_tail = rhp;
ssp->srcu_cb_tail = &rhp->next;
local_irq_restore(flags);
if (!READ_ONCE(ssp->srcu_gp_running)) {
if (likely(srcu_init_done))
schedule_work(&ssp->srcu_work);
else if (list_empty(&ssp->srcu_work.entry))
list_add(&ssp->srcu_work.entry, &srcu_boot_list);
}
}
EXPORT_SYMBOL_GPL(call_srcu);
/*
* synchronize_srcu - wait for prior SRCU read-side critical-section completion
*/
void synchronize_srcu(struct srcu_struct *ssp)
{
struct rcu_synchronize rs;
init_rcu_head_on_stack(&rs.head);
init_completion(&rs.completion);
call_srcu(ssp, &rs.head, wakeme_after_rcu);
wait_for_completion(&rs.completion);
destroy_rcu_head_on_stack(&rs.head);
}
EXPORT_SYMBOL_GPL(synchronize_srcu);
/* Lockdep diagnostics. */
void __init rcu_scheduler_starting(void)
{
rcu_scheduler_active = RCU_SCHEDULER_RUNNING;
}
/*
* Queue work for srcu_struct structures with early boot callbacks.
* The work won't actually execute until the workqueue initialization
* phase that takes place after the scheduler starts.
*/
void __init srcu_init(void)
{
struct srcu_struct *ssp;
srcu_init_done = true;
while (!list_empty(&srcu_boot_list)) {
ssp = list_first_entry(&srcu_boot_list,
struct srcu_struct, srcu_work.entry);
list_del_init(&ssp->srcu_work.entry);
schedule_work(&ssp->srcu_work);
}
}