Re: drivers/greybus/es2.c:439 message_send() error: double unlocked 'es2->cport_out_urb_lock' (orig line 417)

From: Johan Hovold
Date: Mon Aug 24 2020 - 05:53:54 EST


On Fri, Aug 14, 2020 at 03:26:27AM +0800, kernel test robot wrote:
> Hi Greg,
>
> First bad commit (maybe != root cause):
>
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head: dc06fe51d26efc100ac74121607c01a454867c91
> commit: b81beec9cb2d586412c7166c893894930f19965e staging: greybus: move es2 to drivers/greybus/
> date: 12 months ago
> config: x86_64-randconfig-m001-20200813 (attached as .config)
> compiler: gcc-9 (Debian 9.3.0-15) 9.3.0
>
> If you fix the issue, kindly add following tag as appropriate
> Reported-by: kernel test robot <lkp@xxxxxxxxx>
>
> New smatch warnings:
> drivers/greybus/es2.c:439 message_send() error: double unlocked 'es2->cport_out_urb_lock' (orig line 417)
> drivers/greybus/es2.c:879 cport_out_callback() error: double unlocked 'es2->cport_out_urb_lock' (orig line 871)
> drivers/greybus/es2.c:1018 arpc_sync() error: double unlocked 'es2->arpc_lock' (orig line 992)
>
> Old smatch warnings:
> drivers/greybus/es2.c:441 message_send() error: double unlocked 'es2->cport_out_urb_lock' (orig line 439)

None of these warnings make any sense. Is the robot drunk?

> vim +439 drivers/greybus/es2.c
>
> d29b3d631e57240 drivers/staging/greybus/es2.c Alex Elder 2015-06-13 385
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 386 /*
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 387 * Returns zero if the message was successfully queued, or a negative errno
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 388 * otherwise.
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 389 */
> 2537636abae5b81 drivers/staging/greybus/es2.c Johan Hovold 2015-11-03 390 static int message_send(struct gb_host_device *hd, u16 cport_id,
> 7cf7bca9ec5659e drivers/staging/greybus/es2.c Johan Hovold 2015-04-07 391 struct gb_message *message, gfp_t gfp_mask)
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 392 {
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 393 struct es2_ap_dev *es2 = hd_to_es2(hd);
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 394 struct usb_device *udev = es2->usb_dev;
> 7cf7bca9ec5659e drivers/staging/greybus/es2.c Johan Hovold 2015-04-07 395 size_t buffer_size;
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 396 int retval;
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 397 struct urb *urb;
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 398 unsigned long flags;
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 399
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 400 /*
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 401 * The data actually transferred will include an indication
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 402 * of where the data should be sent. Do one last check of
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 403 * the target CPort id before filling it in.
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 404 */
> 144670c2ae3f9b4 drivers/staging/greybus/es2.c Fabien Parent 2015-09-02 405 if (!cport_id_valid(hd, cport_id)) {
> 100e90000840741 drivers/staging/greybus/es2.c Johan Hovold 2015-12-07 406 dev_err(&udev->dev, "invalid cport %u\n", cport_id);
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 407 return -EINVAL;
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 408 }
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 409
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 410 /* Find a free urb */
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 411 urb = next_free_urb(es2, gfp_mask);
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 412 if (!urb)
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 413 return -ENOMEM;
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 414
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 415 spin_lock_irqsave(&es2->cport_out_urb_lock, flags);
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 416 message->hcpriv = urb;
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 @417 spin_unlock_irqrestore(&es2->cport_out_urb_lock, flags);

How is this a double unlock?

> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 418
> d29b3d631e57240 drivers/staging/greybus/es2.c Alex Elder 2015-06-13 419 /* Pack the cport id into the message header */
> d29b3d631e57240 drivers/staging/greybus/es2.c Alex Elder 2015-06-13 420 gb_message_cport_pack(message->header, cport_id);
> 491e60d63fde0a9 drivers/staging/greybus/es2.c Johan Hovold 2015-04-07 421
> 821c620afa1ad29 drivers/staging/greybus/es2.c Alex Elder 2015-06-13 422 buffer_size = sizeof(*message->header) + message->payload_size;
> 491e60d63fde0a9 drivers/staging/greybus/es2.c Johan Hovold 2015-04-07 423
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 424 usb_fill_bulk_urb(urb, udev,
> 606addd2847ccc5 drivers/staging/greybus/es2.c Alexandre Bailon 2015-06-15 425 usb_sndbulkpipe(udev,
> 403074b50b66f1a drivers/staging/greybus/es2.c Greg Kroah-Hartman 2016-08-17 426 es2->cport_out_endpoint),
> 821c620afa1ad29 drivers/staging/greybus/es2.c Alex Elder 2015-06-13 427 message->buffer, buffer_size,
> 7cf7bca9ec5659e drivers/staging/greybus/es2.c Johan Hovold 2015-04-07 428 cport_out_callback, message);
> 977e209ab41073d drivers/staging/greybus/es2.c Alexandre Bailon 2015-08-31 429 urb->transfer_flags |= URB_ZERO_PACKET;
> 495787a792ac498 drivers/staging/greybus/es2.c Alex Elder 2016-06-03 430
> 495787a792ac498 drivers/staging/greybus/es2.c Alex Elder 2016-06-03 431 trace_gb_message_submit(message);
> 495787a792ac498 drivers/staging/greybus/es2.c Alex Elder 2016-06-03 432
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 433 retval = usb_submit_urb(urb, gfp_mask);
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 434 if (retval) {
> 05e3095563670ab drivers/staging/greybus/es2.c Viresh Kumar 2016-06-23 435 dev_err(&udev->dev, "failed to submit out-urb: %d\n", retval);
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 436
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 437 spin_lock_irqsave(&es2->cport_out_urb_lock, flags);
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 438 message->hcpriv = NULL;
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 @439 spin_unlock_irqrestore(&es2->cport_out_urb_lock, flags);
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 440
> 4b1d82047ebbb10 drivers/staging/greybus/es2.c Alex Elder 2015-10-27 441 free_urb(es2, urb);
> d29b3d631e57240 drivers/staging/greybus/es2.c Alex Elder 2015-06-13 442 gb_message_cport_clear(message->header);
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 443
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 444 return retval;
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 445 }
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 446
> 3e136cc9e05e1a3 drivers/staging/greybus/es2.c Johan Hovold 2015-07-01 447 return 0;
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 448 }
> f587027e793cf89 drivers/staging/greybus/gb-es2.c Greg Kroah-Hartman 2015-01-21 449
>
> :::::: The code at line 439 was first introduced by commit
> :::::: 4b1d82047ebbb108e9961fffd51f8ccd54459221 greybus: es2: rename es2 data structures
>
> :::::: TO: Alex Elder <elder@xxxxxxxxxx>
> :::::: CC: Greg Kroah-Hartman <gregkh@xxxxxxxxxx>
>
> ---
> 0-DAY CI Kernel Test Service, Intel Corporation
> https://lists.01.org/hyperkitty/list/kbuild-all@xxxxxxxxxxxx

Johan