Re: [PATCH v2 ] devcoredump : Serialize devcd_del work

[Greg KH <gregkh@linuxfoundation.org>]

On Mon, Apr 25, 2022 at 06:39:53PM +0530, Mukesh Ojha wrote:
> In following scenario(diagram), when one thread X running dev_coredumpm() adds devcd
> device to the framework which sends uevent notification to userspace
> and another thread Y reads this uevent and call to devcd_data_write()
> which eventually try to delete the queued timer that is not initialized/queued yet.
> 
> So, debug object reports some warning and in the meantime, timer is initialized
> and queued from X path. and from Y path, it gets reinitialized again and
> timer->entry.pprev=NULL and try_to_grab_pending() stucks.

Nit, please wrap your lines at 72 columns like git asked you to when you
made the commit

> 
> To fix this, introduce mutex to serialize the behaviour.
> 
>  	cpu0(X)			                      cpu1(Y)
> 
>     dev_coredump() uevent sent to userspace
>     device_add()  =========================> userspace process Y reads the uevents
>                                              writes to devcd fd which
>                                              results into writes to
> 
>                                             devcd_data_write()
>                                               mod_delayed_work()
>                                                 try_to_grab_pending()
>                                                   del_timer()
>                                                     debug_assert_init()
>    INIT_DELAYED_WORK
>    schedule_delayed_work
>                                                      debug_object_fixup()
>                                                       timer_fixup_assert_init()
>                                                        timer_setup()
>                                                          do_init_timer()   ==> reinitialized the
>                                                                                  timer to
>                                                                                  timer->entry.pprev=NULL
> 
>                                                   timer_pending()
>                                                    !hlist_unhashed_lockless(&timer->entry)
>                                                      !h->pprev  ==> del_timer checks
>                                                                   and finds it to be NULL
>  								  try_to_grab_pending() stucks.

Mix of tabs and spaces?  This can all go left a bit as well.

> 
> Link: https://lore.kernel.org/lkml/2e1f81e2-428c-f11f-ce92-eb11048cb271@quicinc.com/
> Signed-off-by: Mukesh Ojha <quic_mojha@quicinc.com>
> ---
> v1->v2:
>  - Added del_wk_queued to serialize the race between devcd_data_write()
>    and disabled_store().
> 
>  drivers/base/devcoredump.c | 21 ++++++++++++++++++---
>  1 file changed, 18 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/base/devcoredump.c b/drivers/base/devcoredump.c
> index f4d794d..3e6fd6b 100644
> --- a/drivers/base/devcoredump.c
> +++ b/drivers/base/devcoredump.c
> @@ -25,6 +25,8 @@ struct devcd_entry {
>  	struct device devcd_dev;
>  	void *data;
>  	size_t datalen;
> +	struct mutex mutex;

Document what this lock is for here please.  I think checkpatch asks you
for that, right?

> +	bool del_wk_queued;

Please spell this out better, you can use vowels :)

>  	struct module *owner;
>  	ssize_t (*read)(char *buffer, loff_t offset, size_t count,
>  			void *data, size_t datalen);
> @@ -84,7 +86,12 @@ static ssize_t devcd_data_write(struct file *filp, struct kobject *kobj,
>  	struct device *dev = kobj_to_dev(kobj);
>  	struct devcd_entry *devcd = dev_to_devcd(dev);
>  
> -	mod_delayed_work(system_wq, &devcd->del_wk, 0);
> +	mutex_lock(&devcd->mutex);
> +	if (!devcd->del_wk_queued) {
> +		devcd->del_wk_queued = true;
> +		mod_delayed_work(system_wq, &devcd->del_wk, 0);
> +	}
> +	mutex_unlock(&devcd->mutex);
>  
>  	return count;
>  }
> @@ -112,7 +119,12 @@ static int devcd_free(struct device *dev, void *data)
>  {
>  	struct devcd_entry *devcd = dev_to_devcd(dev);
>  
> +	mutex_lock(&devcd->mutex);
> +	if (!devcd->del_wk_queued)
> +		devcd->del_wk_queued = true;
> +
>  	flush_delayed_work(&devcd->del_wk);
> +	mutex_unlock(&devcd->mutex);
>  	return 0;
>  }
>  
> @@ -278,13 +290,15 @@ void dev_coredumpm(struct device *dev, struct module *owner,
>  	devcd->read = read;
>  	devcd->free = free;
>  	devcd->failing_dev = get_device(dev);
> -
> +	mutex_init(&devcd->mutex);

Why drop the blank line?

>  	device_initialize(&devcd->devcd_dev);
>  
>  	dev_set_name(&devcd->devcd_dev, "devcd%d",
>  		     atomic_inc_return(&devcd_count));
>  	devcd->devcd_dev.class = &devcd_class;
>  
> +	mutex_lock(&devcd->mutex);

Why lock this here?

> +	devcd->del_wk_queued = false;

This was already set to false above, right?  And if you want to
explicitly initialize it, do it where the other variables are
initialized up by mutex_init() please.

thanks,

greg k-h