连续调用usleep_range导致的lockup
现象
crng_finalize_init后出现死锁
[ 3.610168] random: crng init done
[ 13.615447] suspected lockup task = kworker/0:3, pid: 42
[ 13.620788] pc = 0x0
[ 23.615447] suspected lockup task = kworker/0:3, pid: 42
[ 23.620788] pc = 0x0
[ 33.615447] suspected lockup task = kworker/0:3, pid: 42
[ 33.620788] pc = 0x0
[ 43.615447] suspected lockup task = kworker/0:3, pid: 42
[ 43.620788] pc = 0x0分析
最后一句打印的地方，添加栈回溯看什么触发的
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -829,6 +829,7 @@ static void crng_finalize_init(struct crng_state *crng)
wake_up_interruptible(&crng_init_wait);
kill_fasync(&fasync, SIGIO, POLL_IN);
pr_notice("crng init done\n");
+ dump_stack();
if (unseeded_warning.missed) {
pr_notice("random: %d get_random_xx warning(s) missed "
"due to ratelimiting\n",也即初始化的时候，发送mmc的request的时候，触发了中断
[ 3.636016] Workqueue: events_freezable mmc_rescan
[ 3.640838] [<c0013245>] (unwind_backtrace) from [<c00109b7>] (show_stack+0xb/0xc)
[ 3.648468] [<c00109b7>] (show_stack) from [<c0208e49>] (crng_finalize_init.constprop.18+0xc9/0xfc)
[ 3.657592] [<c0208e49>] (crng_finalize_init.constprop.18) from [<c0209285>] (crng_reseed.constprop.14+0xe9/0x174)
[ 3.668029] [<c0209285>] (crng_reseed.constprop.14) from [<c020981b>] (add_interrupt_randomness+0x10b/0x160)
[ 3.677917] [<c020981b>] (add_interrupt_randomness) from [<c004e5e7>] (handle_irq_event_percpu+0x1b/0x3c)
[ 3.687561] [<c004e5e7>] (handle_irq_event_percpu) from [<c004e62f>] (handle_irq_event+0x27/0x3c)
[ 3.696502] [<c004e62f>] (handle_irq_event) from [<c0050ae3>] (handle_level_irq+0x67/0xb8)
[ 3.704833] [<c0050ae3>] (handle_level_irq) from [<c004dea3>] (generic_handle_irq+0x17/0x20)
[ 3.713317] [<c004dea3>] (generic_handle_irq) from [<c004e0c9>] (__handle_domain_irq+0x35/0x70)
[ 3.722076] [<c004e0c9>] (__handle_domain_irq) from [<c0009ad3>] (__irq_svc+0x53/0x94)
[ 3.730072] Exception stack(0xc5de5d00 to 0xc5de5d48)
[ 3.735137] 5d00: 00008001 ffff8c9d ffffff9c 00000000 c5c3d000 c066ac40 00000001 c5c3d810
[ 3.743377] 5d20: c5c3d640 00000013 00000000 c5de5dc0 10624dd3 c5de5d50 c042a2b9 c02dbc96
[ 3.751617] 5d40: a0000033 ffffffff
[ 3.755126] [<c0009ad3>] (__irq_svc) from [<c02dbc96>] (sdhci_request+0xd2/0x270)
[ 3.762664] [<c02dbc96>] (sdhci_request) from [<c02c97bb>] (mmc_start_request+0x5f/0x7c)
[ 3.770812] [<c02c97bb>] (mmc_start_request) from [<c02c990d>] (mmc_wait_for_req+0x45/0x8c)
[ 3.779235] [<c02c990d>] (mmc_wait_for_req) from [<c02d05a3>] (mmc_app_send_scr+0x8f/0xb8)
[ 3.787567] [<c02d05a3>] (mmc_app_send_scr) from [<c02cfbfb>] (mmc_sd_setup_card+0x2f/0x250)
[ 3.796051] [<c02cfbfb>] (mmc_sd_setup_card) from [<c02cfec9>] (mmc_sd_init_card+0xad/0x2fc)
[ 3.804565] [<c02cfec9>] (mmc_sd_init_card) from [<c02d023f>] (mmc_attach_sd+0x77/0xf4)
[ 3.812622] [<c02d023f>] (mmc_attach_sd) from [<c02caba5>] (mmc_rescan+0x21d/0x290)
[ 3.820312] [<c02caba5>] (mmc_rescan) from [<c0034293>] (process_one_work+0x127/0x2e4)
[ 3.828308] [<c0034293>] (process_one_work) from [<c003454b>] (worker_thread+0xfb/0x3e4)
[ 3.836456] [<c003454b>] (worker_thread) from [<c0037a01>] (kthread+0x105/0x108)
[ 3.843902] [<c0037a01>] (kthread) from [<c00090e9>] (ret_from_fork+0x11/0x28)
[ 3.851165] Exception stack(0xc5de5fb0 to 0xc5de5ff8)
[ 3.856262] 5fa0: 00000000 00000000 00000000 00000000
[ 3.864471] 5fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[ 3.872711] 5fe0: 00000000 00000000 00000000 00000000 00000013 00000000
[ 13.875732] suspected lockup task = kworker/0:2, pid: 39
[ 13.881072] pc = 0x0
[ 23.875732] suspected lockup task = kworker/0:2, pid: 39
[ 23.881072] pc = 0x0添加打印，看看触发了什么中断，发现是不同的中断触发的，那就不是某个中断处理引发的问题
--- a/kernel/irq/irqdesc.c
+++ b/kernel/irq/irqdesc.c
@@ -664,7 +664,7 @@ int __handle_domain_irq(struct irq_domain *domain, unsigned int hwirq,
struct pt_regs *old_regs = set_irq_regs(regs);
unsigned int irq = hwirq;
int ret = 0;
-
+ printk("__handle_domain_irq irq = %d", irq);
irq_enter();add_interrupt_randomness
在中断处理中，add_interrupt_randomness()函数利用设备两次中断的间隔时间作为噪声源将随机数据加入熵池。这个产生的随机值，可供usleep_range使用
handle_irq
分析下这个中断随机值的产生流程
int generic_handle_irq(unsigned int irq)
{
struct irq_desc *desc = irq_to_desc(irq);
if (!desc)
return -EINVAL;
generic_handle_irq_desc(desc);
return 0;
}
static inline void generic_handle_irq_desc(struct irq_desc *desc)
{
desc->handle_irq(desc);
}irq_set_chip_and_handler
中断控制器通过irq_set_chip_and_handler将上面的desc->handle_irq(desc)赋值为handle_level_irq；类似的处理看《这篇》
static int mmp_irq_domain_map(struct irq_domain *d, unsigned int irq,
irq_hw_number_t hw)
{
irq_set_chip_and_handler(irq, &icu_irq_chip, handle_level_irq);
return 0;
}
//
void handle_level_irq(struct irq_desc *desc)
{
raw_spin_lock(&desc->lock);
mask_ack_irq(desc);
if (!irq_may_run(desc))
goto out_unlock;
desc->istate &= ~(IRQS_REPLAY | IRQS_WAITING);
/*
* If its disabled or no action available
* keep it masked and get out of here
*/
if (unlikely(!desc->action || irqd_irq_disabled(&desc->irq_data))) {
desc->istate |= IRQS_PENDING;
goto out_unlock;
}
kstat_incr_irqs_this_cpu(desc);
handle_irq_event(desc);
cond_unmask_irq(desc);
out_unlock:
raw_spin_unlock(&desc->lock);
}
//
irqreturn_t handle_irq_event(struct irq_desc *desc)
{
irqreturn_t ret;
desc->istate &= ~IRQS_PENDING;
irqd_set(&desc->irq_data, IRQD_IRQ_INPROGRESS);
raw_spin_unlock(&desc->lock);
ret = handle_irq_event_percpu(desc);
raw_spin_lock(&desc->lock);
irqd_clear(&desc->irq_data, IRQD_IRQ_INPROGRESS);
return ret;
}handle_irq_event_percpu
__handle_irq_event_percpu执行中断处理函数，这个add_interrupt_randomness函数利用设备两次中断的间隔时间作为噪声源将随机数据加入熵池，供产生随机数使用
irqreturn_t handle_irq_event_percpu(struct irq_desc *desc)
{
irqreturn_t retval;
unsigned int flags = 0;
retval = __handle_irq_event_percpu(desc, &flags);
add_interrupt_randomness(desc->irq_data.irq, flags);
if (!noirqdebug)
note_interrupt(desc, retval);
return retval;
}action->handler执行request_irq注册的中断处理函数handler
irqreturn_t __handle_irq_event_percpu(struct irq_desc *desc, unsigned int *flags)
{
irqreturn_t retval = IRQ_NONE;
unsigned int irq = desc->irq_data.irq;
struct irqaction *action;
record_irq_time(desc);
for_each_action_of_desc(desc, action) {
irqreturn_t res;
trace_irq_handler_entry(irq, action);
res = action->handler(irq, action->dev_id);
trace_irq_handler_exit(irq, action, res);
if (WARN_ONCE(!irqs_disabled(),"irq %u handler %pS enabled interrupts\n",
irq, action->handler))
local_irq_disable();
switch (res) {
case IRQ_WAKE_THREAD:
/*
* Catch drivers which return WAKE_THREAD but
* did not set up a thread function
*/
if (unlikely(!action->thread_fn)) {
warn_no_thread(irq, action);
break;
}
__irq_wake_thread(desc, action);
/* Fall through - to add to randomness */
case IRQ_HANDLED:
*flags |= action->flags;
break;
default:
break;
}
retval |= res;
}
return retval;
}结论
正常的打印，后面该接着打印pcie相关的信息，最终发现是pcie主机驱动里，加复位操作的时候，连续调用usleep_range引起的死锁
原文地址：https://blog.csdn.net/weixin_44586903/article/details/140695931
免责声明：本站文章内容转载自网络资源，如本站内容侵犯了原著者的合法权益，可联系本站删除。更多内容请关注自学内容网（zxcms.com）！