mirror of
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
synced 2026-09-18 23:09:29 +02:00
__pcpu_freelist_push() can loop forever when only one CPU is possible
and an NMI re-enters pcpu_freelist_push() while the interrupted context
holds that CPU's freelist lock.
After the current-CPU fast path fails, the fallback loop walks
cpu_possible_mask while skipping the current CPU. With CONFIG_SMP=n, or
when an SMP kernel is limited to one possible CPU with nr_cpus=1 or
possible_cpus=1, there are no other possible CPUs to examine. The loop
therefore makes no lock acquisition attempt and can never make progress.
The following stack was observed on a UP system:
NMI context:
pcpu_freelist_push
free_htab_elem
htab_map_delete_elem
[perf-event BPF program]
__perf_event_overflow
perf_event_nmi_handler
exc_nmi
Interrupted context:
__pcpu_freelist_push
pcpu_freelist_push
free_htab_elem
htab_map_delete_elem
[raw_tp/sys_enter BPF program]
__bpf_trace_sys_enter
do_syscall_64
raw_res_spin_lock() detects the same-CPU recursive acquisition and
returns -EDEADLK, but the subsequent fallback loop has no candidate head
on a system with one possible CPU.
Restore the extra fallback head that existed before the rqspinlock
conversion. Keep the current-CPU fast path, then try the other possible
CPUs and finally the extra head. The additional head lets a push, which
cannot fail without losing a preallocated element, make progress when the
only per-CPU head is held by the interrupted context.
Also check the extra head from the pop path so that nodes placed there
can be reused.
Fixes: f2ac0e5d1c ("bpf: Convert percpu_freelist.c to rqspinlock")
Signed-off-by: Hui Su <sh_def@163.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/bpf/20260806175600.1993595-1-sh_def@163.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
157 lines
3.5 KiB
C
157 lines
3.5 KiB
C
// SPDX-License-Identifier: GPL-2.0-only
|
|
/* Copyright (c) 2016 Facebook
|
|
*/
|
|
#include "percpu_freelist.h"
|
|
|
|
int pcpu_freelist_init(struct pcpu_freelist *s)
|
|
{
|
|
int cpu;
|
|
|
|
s->freelist = alloc_percpu(struct pcpu_freelist_head);
|
|
if (!s->freelist)
|
|
return -ENOMEM;
|
|
|
|
for_each_possible_cpu(cpu) {
|
|
struct pcpu_freelist_head *head = per_cpu_ptr(s->freelist, cpu);
|
|
|
|
raw_res_spin_lock_init(&head->lock);
|
|
head->first = NULL;
|
|
}
|
|
raw_res_spin_lock_init(&s->extralist.lock);
|
|
s->extralist.first = NULL;
|
|
return 0;
|
|
}
|
|
|
|
void pcpu_freelist_destroy(struct pcpu_freelist *s)
|
|
{
|
|
free_percpu(s->freelist);
|
|
}
|
|
|
|
static inline void pcpu_freelist_push_node(struct pcpu_freelist_head *head,
|
|
struct pcpu_freelist_node *node)
|
|
{
|
|
node->next = head->first;
|
|
WRITE_ONCE(head->first, node);
|
|
}
|
|
|
|
static inline bool ___pcpu_freelist_push(struct pcpu_freelist_head *head,
|
|
struct pcpu_freelist_node *node)
|
|
{
|
|
if (raw_res_spin_lock(&head->lock))
|
|
return false;
|
|
pcpu_freelist_push_node(head, node);
|
|
raw_res_spin_unlock(&head->lock);
|
|
return true;
|
|
}
|
|
|
|
void __pcpu_freelist_push(struct pcpu_freelist *s,
|
|
struct pcpu_freelist_node *node)
|
|
{
|
|
struct pcpu_freelist_head *head;
|
|
int cpu, this_cpu;
|
|
|
|
if (___pcpu_freelist_push(this_cpu_ptr(s->freelist), node))
|
|
return;
|
|
|
|
this_cpu = raw_smp_processor_id();
|
|
while (true) {
|
|
for_each_cpu_wrap(cpu, cpu_possible_mask, this_cpu) {
|
|
if (cpu == this_cpu)
|
|
continue;
|
|
|
|
head = per_cpu_ptr(s->freelist, cpu);
|
|
if (___pcpu_freelist_push(head, node))
|
|
return;
|
|
}
|
|
|
|
/*
|
|
* Push cannot fail. Use the extra list when none of the
|
|
* per-CPU freelists can accept the node.
|
|
*/
|
|
if (___pcpu_freelist_push(&s->extralist, node))
|
|
return;
|
|
}
|
|
}
|
|
|
|
void pcpu_freelist_push(struct pcpu_freelist *s,
|
|
struct pcpu_freelist_node *node)
|
|
{
|
|
unsigned long flags;
|
|
|
|
local_irq_save(flags);
|
|
__pcpu_freelist_push(s, node);
|
|
local_irq_restore(flags);
|
|
}
|
|
|
|
void pcpu_freelist_populate(struct pcpu_freelist *s, void *buf, u32 elem_size,
|
|
u32 nr_elems)
|
|
{
|
|
struct pcpu_freelist_head *head;
|
|
unsigned int cpu, cpu_idx, i, j, n, m;
|
|
|
|
n = nr_elems / num_possible_cpus();
|
|
m = nr_elems % num_possible_cpus();
|
|
|
|
cpu_idx = 0;
|
|
for_each_possible_cpu(cpu) {
|
|
head = per_cpu_ptr(s->freelist, cpu);
|
|
j = n + (cpu_idx < m ? 1 : 0);
|
|
for (i = 0; i < j; i++) {
|
|
/* No locking required as this is not visible yet. */
|
|
pcpu_freelist_push_node(head, buf);
|
|
buf += elem_size;
|
|
}
|
|
cpu_idx++;
|
|
}
|
|
}
|
|
|
|
static struct pcpu_freelist_node *___pcpu_freelist_pop(struct pcpu_freelist *s)
|
|
{
|
|
struct pcpu_freelist_node *node = NULL;
|
|
struct pcpu_freelist_head *head;
|
|
int cpu;
|
|
|
|
for_each_cpu_wrap(cpu, cpu_possible_mask, raw_smp_processor_id()) {
|
|
head = per_cpu_ptr(s->freelist, cpu);
|
|
if (!READ_ONCE(head->first))
|
|
continue;
|
|
if (raw_res_spin_lock(&head->lock))
|
|
continue;
|
|
node = head->first;
|
|
if (node) {
|
|
WRITE_ONCE(head->first, node->next);
|
|
raw_res_spin_unlock(&head->lock);
|
|
return node;
|
|
}
|
|
raw_res_spin_unlock(&head->lock);
|
|
}
|
|
|
|
/* Per-CPU lists are empty or unavailable, try the extra list. */
|
|
head = &s->extralist;
|
|
if (!READ_ONCE(head->first))
|
|
return NULL;
|
|
if (raw_res_spin_lock(&head->lock))
|
|
return NULL;
|
|
node = head->first;
|
|
if (node)
|
|
WRITE_ONCE(head->first, node->next);
|
|
raw_res_spin_unlock(&head->lock);
|
|
return node;
|
|
}
|
|
|
|
struct pcpu_freelist_node *__pcpu_freelist_pop(struct pcpu_freelist *s)
|
|
{
|
|
return ___pcpu_freelist_pop(s);
|
|
}
|
|
|
|
struct pcpu_freelist_node *pcpu_freelist_pop(struct pcpu_freelist *s)
|
|
{
|
|
struct pcpu_freelist_node *ret;
|
|
unsigned long flags;
|
|
|
|
local_irq_save(flags);
|
|
ret = __pcpu_freelist_pop(s);
|
|
local_irq_restore(flags);
|
|
return ret;
|
|
}
|