Re: [PATCH for-next V5 1/5] IB/uverbs: Fix reference counting usage of event files

[Yishai Hadas <yishaih-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>]

On 6/29/2015 8:40 PM, Jason Gunthorpe wrote:
> On Sun, Jun 28, 2015 at 05:33:04PM +0300, Yishai Hadas wrote:
>
>> You are wrong here, we have here balanced put, the first is done as
>> part of fput(filp) -> ib_uverbs_event_close_file ->
>> kref_put(&file->ref, ib_uverbs_release_event_file) and the second at
>> the end of this function as part of the err: label
>> kref_put(&ev_file->ref, ib_uverbs_release_event_file);
>
> Ugh, that is so gross, seriously? Sure, the counts work out, but that
> is *totally unreadable*.
>
> Sigh, just use this:
>
>   struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file,
>                                          int is_async)
>   {
>          struct ib_uverbs_event_file *ev_file;
>          struct file *filp;
> +       int ret;
>
> -       ev_file = kmalloc(sizeof *ev_file, GFP_KERNEL);
> +       ev_file = kzalloc(sizeof(*ev_file), GFP_KERNEL);
>          if (!ev_file)
>                  return ERR_PTR(-ENOMEM);
>
> @@ -555,15 +561,41 @@ struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file,
>          INIT_LIST_HEAD(&ev_file->event_list);
>          init_waitqueue_head(&ev_file->poll_wait);
>          ev_file->uverbs_file = uverbs_file;
> +       kref_get(&ev_file->uverbs_file->ref);
>          ev_file->async_queue = NULL;
> -       ev_file->is_async    = is_async;
>          ev_file->is_closed   = 0;
>
>          filp = anon_inode_getfile("[infinibandevent]", &uverbs_event_fops,
>                                    ev_file, O_RDONLY);
>          if (IS_ERR(filp))
> -               kfree(ev_file);
> +               goto err_put_refs;
> +
> +       if (is_async) {
> +               WARN_ON(uverbs_file->async_file);
> +               uverbs_file->async_file = ev_file;
> +               kref_get(&uverbs_file->async_file->ref);
> +               INIT_IB_EVENT_HANDLER(&uverbs_file->event_handler,
> +                                     uverbs_file->device->ib_dev,
> +                                     ib_uverbs_event_handler);
> +               ret = ib_register_event_handler(&uverbs_file->event_handler);
> +               if (ret)
> +                       goto err_put_file;
> +
> +               /* At that point async file stuff was fully set */
> +               ev_file->is_async = 1;
> +       }
> +
> +       return filp;
> +
> +err_put_file:
> +       fput(filp);
> +       kref_put(&uverbs_file->async_file->ref, ib_uverbs_release_file);

It should be:
kref_put(&uverbs_file->async_file->ref, ib_uverbs_release_event_file)
instead of:
kref_put(&uverbs_file->async_file->ref, ib_uverbs_release_file);

Please note that in that approach we duplicate above line as it appears 
here and in below err_put_refs label as uverbs_file->async_file is 
really ev_file.

In my patch we used one line instead of those duplicated lines, however 
as you think that it clarifies things will go with your suggestion after 
fixing above note, thanks.


> +       uverbs_file->async_file = NULL;
> +       return ERR_PTR(ret);
>
> +err_put_refs:
> +       kref_put(&ev_file->uverbs_file->ref, ib_uverbs_release_file);
> +       kref_put(&ev_file->ref, ib_uverbs_release_event_file);
>          return filp;
>   }
>
> Jason
>

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html