Re: [PATCH 20/22] gpio: protect the pointer to gpio_chip in gpio_device with SRCU

[Bartosz Golaszewski <brgl@xxxxxxxx>]

On Thu, Feb 1, 2024 at 6:03 AM Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote:
>
> Hi Bartosz,
>
> kernel test robot noticed the following build warnings:
>
> https://git-scm.com/docs/git-format-patch#_base_tree_information]
>
> url:    https://github.com/intel-lab-lkp/linux/commits/Bartosz-Golaszewski/gpio-protect-the-list-of-GPIO-devices-with-SRCU/20240130-205537
> base:   https://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux.git gpio/for-next
> patch link:    https://lore.kernel.org/r/20240130124828.14678-21-brgl%40bgdev.pl
> patch subject: [PATCH 20/22] gpio: protect the pointer to gpio_chip in gpio_device with SRCU
> config: i386-randconfig-141-20240131 (https://download.01.org/0day-ci/archive/20240201/202402010641.idtEaO24-lkp@xxxxxxxxx/config)
> compiler: clang version 17.0.6 (https://github.com/llvm/llvm-project 6009708b4367171ccdbf4b5905cb6a803753fe18)
>
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <lkp@xxxxxxxxx>
> | Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
> | Closes: https://lore.kernel.org/r/202402010641.idtEaO24-lkp@xxxxxxxxx/
>
> New smatch warnings:
> drivers/gpio/gpiolib.c:4776 gpiolib_dbg_show() error: we previously assumed 'gc' could be null (see line 4773)
>
>
> vim +/gc +4776 drivers/gpio/gpiolib.c
>
> fdeb8e1547cb9d Linus Walleij       2016-02-10  4762  static void gpiolib_dbg_show(struct seq_file *s, struct gpio_device *gdev)
> d2876d08d86f22 David Brownell      2008-02-04  4763  {
> 0338f6a6fb659f Bartosz Golaszewski 2023-12-21  4764     bool active_low, is_irq, is_out;
> 0338f6a6fb659f Bartosz Golaszewski 2023-12-21  4765     unsigned int gpio = gdev->base;
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4766     struct gpio_desc *desc;
> 2796d5332f8ac8 Bartosz Golaszewski 2024-01-30  4767     struct gpio_chip *gc;
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4768     int value;
> d2876d08d86f22 David Brownell      2008-02-04  4769
> 2796d5332f8ac8 Bartosz Golaszewski 2024-01-30  4770     guard(srcu)(&gdev->srcu);
> 2796d5332f8ac8 Bartosz Golaszewski 2024-01-30  4771
> 2796d5332f8ac8 Bartosz Golaszewski 2024-01-30  4772     gc = rcu_dereference(gdev->chip);
> 2796d5332f8ac8 Bartosz Golaszewski 2024-01-30 @4773     if (!gc)
>                                                             ^^^
> The patch adds a NULL check
>
> 2796d5332f8ac8 Bartosz Golaszewski 2024-01-30  4774             seq_puts(s, "Underlying GPIO chip is gone\n");
> 2796d5332f8ac8 Bartosz Golaszewski 2024-01-30  4775
> 3de69ae1c407da Andy Shevchenko     2022-04-08 @4776     for_each_gpio_desc(gc, desc) {
>                                                                            ^^
> But this dereference isn't checked...  Probably it should return after
> the seq_puts().
>

Of course it should. Thanks. I fixed it for v2.

Bart

> bedc56b1695b27 Bartosz Golaszewski 2024-01-30  4777             guard(srcu)(&desc->srcu);
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4778             if (test_bit(FLAG_REQUESTED, &desc->flags)) {
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4779                     gpiod_get_direction(desc);
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4780                     is_out = test_bit(FLAG_IS_OUT, &desc->flags);
> 234c52097ce416 Andy Shevchenko     2022-04-08  4781                     value = gpio_chip_get_value(gc, desc);
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4782                     is_irq = test_bit(FLAG_USED_AS_IRQ, &desc->flags);
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4783                     active_low = test_bit(FLAG_ACTIVE_LOW, &desc->flags);
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4784                     seq_printf(s, " gpio-%-3d (%-20.20s|%-20.20s) %s %s %s%s\n",
> 32648f473c7f46 Bartosz Golaszewski 2024-01-30  4785                                gpio, desc->name ?: "", gpiod_get_label(desc),
> d2876d08d86f22 David Brownell      2008-02-04  4786                                is_out ? "out" : "in ",
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4787                                value >= 0 ? (value ? "hi" : "lo") : "?  ",
> 90fd227029a25b Linus Walleij       2018-10-01  4788                                is_irq ? "IRQ " : "",
> 90fd227029a25b Linus Walleij       2018-10-01  4789                                active_low ? "ACTIVE LOW" : "");
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4790             } else if (desc->name) {
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4791                     seq_printf(s, " gpio-%-3d (%-20.20s)\n", gpio, desc->name);
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4792             }
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4793
> 3de69ae1c407da Andy Shevchenko     2022-04-08  4794             gpio++;
> d2876d08d86f22 David Brownell      2008-02-04  4795     }
> d2876d08d86f22 David Brownell      2008-02-04  4796  }
>
> --
> 0-DAY CI Kernel Test Service
> https://github.com/intel/lkp-tests/wiki
>