Re: [PATCH] usbhid: Fix lockdep unannotated irqs-off warning

[Sedat Dilek]

On 3/2/16, Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote:
> On 3/1/16, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote:
>> On Tue, 1 Mar 2016, Sedat Dilek wrote:
>>
>>> On Tue, Oct 13, 2015 at 2:57 AM, Steven Rostedt <rostedt@xxxxxxxxxxx>
>>> wrote:
>>> > On Sat, 3 Oct 2015 12:05:42 +0200
>>> > Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote:
>>> >
>>> >> So, at the beginning... dunno WTF is causing the problems - no
>>> >> workaround for CLANG.
>>> >
>>> > Probably need to compile with gcc and with clang and look at the
>>> > binary
>>> > differences. Or at least what objdump shows.
>>> >
>>>
>>> [ Hope to address this issue to the correct people - CCed some people
>>> I taped on their nerves ]
>>>
>>> Not sure if I should open a new thread?
>>> Please, some clear statements on this.
>>> Thanks.
>>>
>>> The issue is still visible and alive.
>>
>> ...
>>
>>> [ FACT #3: TEST-CASE #2 ]
>>>
>>> The most reliable test-case is to simply unplug my external Logitech
>>> USB mouse - saw this by accident.
>>> YES, it was so simple.
>>>
>>> --- dmesg_4.5.0-rc6-2-llvmlinux-amd64.txt       2016-02-29
>>> 21:23:56.399691975 +0100
>>> +++ dmesg_4.5.0-rc6-2-llvmlinux-amd64_usbmouse-unplugged.txt
>>> 2016-02-29 21:28:14.401832240 +0100
>>> @@ -832,3 +832,75 @@
>>>  [   66.529779] PPP BSD Compression module registered
>>>  [   66.563013] PPP Deflate Compression module registered
>>>  [   66.978977] usb 2-1.5: USB disconnect, device number 4
>>> +[  321.937369] usb 2-1.4: USB disconnect, device number 3
>>> +[  321.950810] BUG: sleeping function called from invalid context at
>>> kernel/workqueue.c:2785
>>> +[  321.950816] in_atomic(): 0, irqs_disabled(): 1, pid: 44, name:
>>> kworker/2:1
>>
>>> +[  321.950885] hardirqs last  enabled at (47769):
>>> [<ffffffff819426a2>] _raw_spin_unlock_irq+0x32/0x60
>>> +[  321.950889] hardirqs last disabled at (47770):
>>> [<ffffffff81115bdc>] del_timer_sync+0x3c/0x110
>>> +[  321.950894] softirqs last  enabled at (47112):
>>> [<ffffffff810871a2>] __do_softirq+0x5a2/0x610
>>> +[  321.950898] softirqs last disabled at (47097):
>>> [<ffffffff8108747c>] irq_exit+0x9c/0x120
>>> +[  321.950903] CPU: 2 PID: 44 Comm: kworker/2:1 Not tainted
>>> 4.5.0-rc6-2-llvmlinux-amd64 #1
>>> +[  321.950905] Hardware name: SAMSUNG ELECTRONICS CO., LTD.
>>> 530U3BI/530U4BI/530U4BH/530U3BI/530U4BI/530U4BH, BIOS 13XK 03/28/2013
>>> +[  321.950908] Workqueue: usb_hub_wq hub_event
>>> +[  321.950911]  ffff8800d3326948 0000000000000092 0000000000000000
>>> ffff8800d4347568
>>> +[  321.950915]  ffffffff814ba7d4 ffff8800d43474f8 ffff8800d4340cc0
>>> ffff8800d4347568
>>> +[  321.950919]  ffffffff810e28fd 0000000000000092 0000000000000096
>>> ffff8800d43475a8
>>> +[  321.950923] Call Trace:
>>> +[  321.950927]  [<ffffffff814ba7d4>] dump_stack+0xa4/0xf0
>>> +[  321.950931]  [<ffffffff810e28fd>] ? print_irqtrace_events+0xcd/0xe0
>>> +[  321.950936]  [<ffffffff810bf495>] ___might_sleep+0x295/0x2b0
>>> +[  321.950939]  [<ffffffff810bf18f>] __might_sleep+0x4f/0xc0
>>> +[  321.950943]  [<ffffffff810a12ef>] start_flush_work+0x2f/0x2a0
>>> +[  321.950946]  [<ffffffff810a129c>] flush_work+0x5c/0x80
>>> +[  321.950949]  [<ffffffff810a125a>] ? flush_work+0x1a/0x80
>>> +[  321.950953]  [<ffffffff810e247d>] ? trace_hardirqs_off+0xd/0x10
>>> +[  321.950956]  [<ffffffff810a032a>] ? try_to_grab_pending+0x4a/0x260
>>> +[  321.950960]  [<ffffffff810a1717>] __cancel_work_timer+0x197/0x290
>>> +[  321.950963]  [<ffffffff81115b8d>] ? try_to_del_timer_sync+0xad/0xc0
>>> +[  321.950966]  [<ffffffff810a1578>] cancel_work_sync+0x18/0x20
>>
>> It's possible that this could be a compiler-related error connected
>> with local_irq_save().  __cancel_work_timer() calls
>>
>> 		ret = try_to_grab_pending(work, is_dwork, &flags);
>>
>> and try_to_grab_pending() starts this way:
>>
>> static int try_to_grab_pending(struct work_struct *work, bool is_dwork,
>> 			       unsigned long *flags)
>> {
>> 	struct worker_pool *pool;
>> 	struct pool_workqueue *pwq;
>>
>> 	local_irq_save(*flags);
>>
>> Then later on, __cancel_work_timer() does
>>
>> 	local_irq_restore(flags);
>>
>> Maybe CLANG doesn't like local_irq_save() applied to a pointer as
>> opposed to a stack variable.
>>
>> As a quick test, try applying the patch below.  (Note that there is a
>> similar construction in kernel/signal.c.)
>>
>> Alan Stern
>>
>>
>>
>> Index: usb-4.4/kernel/workqueue.c
>> ===================================================================
>> --- usb-4.4.orig/kernel/workqueue.c
>> +++ usb-4.4/kernel/workqueue.c
>> @@ -1175,12 +1175,14 @@ out_put:
>>   * This function is safe to call from any context including IRQ handler.
>>   */
>>  static int try_to_grab_pending(struct work_struct *work, bool is_dwork,
>> -			       unsigned long *flags)
>> +			       unsigned long *pflags)
>>  {
>>  	struct worker_pool *pool;
>>  	struct pool_workqueue *pwq;
>> +	unsigned long flags;
>>
>> -	local_irq_save(*flags);
>> +	local_irq_save(flags);
>> +	*pflags = flags;
>>
>>  	/* try to steal the timer if it exists */
>>  	if (is_dwork) {
>> @@ -1241,7 +1243,7 @@ static int try_to_grab_pending(struct wo
>>  	}
>>  	spin_unlock(&pool->lock);
>>  fail:
>> -	local_irq_restore(*flags);
>> +	local_irq_restore(flags);
>>  	if (work_is_canceling(work))
>>  		return -ENOENT;
>>  	cpu_relax();
>>
>>
>
> might_sleep() is invoked in start_flush_work(), but there are no
> "flags" passed to local_irq_{disable,enable}.
>
> What about something like this?
>
> @@ -2781,13 +2783,14 @@ static bool start_flush_work(struct
> work_struct *work, struct wq_barrier *barr)
>         struct worker *worker = NULL;
>         struct worker_pool *pool;
>         struct pool_workqueue *pwq;
> +       unsigned long flags;
>
>         might_sleep();
>
> -       local_irq_disable();
> +       local_irq_disable(flags);
>         pool = get_work_pool(work);
>         if (!pool) {
> -               local_irq_enable();
> +               local_irq_enable(flags);
>                 return false;
>         }
>

start_flush_work() is the only place where no flags are passed.
That stuff above does not work - no clue how to do that.
Any help?

Then I see there...
...
spin_lock(&pool->lock);
...
spin_unlock_irq(&pool->lock);
...

Should that be spin_lock_irq()?

static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr)
{
        struct worker *worker = NULL;
        struct worker_pool *pool;
        struct pool_workqueue *pwq;

        might_sleep();

        local_irq_disable();
        pool = get_work_pool(work);
        if (!pool) {
                local_irq_enable();
                return false;
        }

        spin_lock(&pool->lock); <--- XXX: spin_lock_irq() ???
        /* see the comment in try_to_grab_pending() with the same code */
        pwq = get_work_pwq(work);
        if (pwq) {
                if (unlikely(pwq->pool != pool))
                        goto already_gone;
        } else {
                worker = find_worker_executing_work(pool, work);
                if (!worker)
                        goto already_gone;
                pwq = worker->current_pwq;
        }

        check_flush_dependency(pwq->wq, work);

        insert_wq_barrier(pwq, barr, work, worker);
        spin_unlock_irq(&pool->lock);

        /*
         * If @max_active is 1 or rescuer is in use, flushing another work
         * item on the same workqueue may lead to deadlock.  Make sure the
         * flusher is not running on the same workqueue by verifying write
         * access.
         */
        if (pwq->wq->saved_max_active == 1 || pwq->wq->rescuer)
                lock_map_acquire(&pwq->wq->lockdep_map);
        else
                lock_map_acquire_read(&pwq->wq->lockdep_map);
        lock_map_release(&pwq->wq->lockdep_map);

        return true;
already_gone:
        spin_unlock_irq(&pool->lock);
        return false;
}

- Sedat -