Re: [PATCH] HID: i2c-hid: goodix: Tie the reset line to true state of the regulator

From: Dan Carpenter
Date: Tue Jun 22 2021 - 04:49:48 EST

Hi Douglas,

url: https://github.com/0day-ci/linux/commits/Douglas-Anderson/HID-i2c-hid-goodix-Tie-the-reset-line-to-true-state-of-the-regulator/20210617-102450
base: https://git.kernel.org/pub/scm/linux/kernel/git/hid/hid.git for-next
config: i386-randconfig-m021-20210622 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@xxxxxxxxx>
Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>

smatch warnings:
drivers/hid/i2c-hid/i2c-hid-of-goodix.c:151 i2c_hid_of_goodix_probe() warn: inconsistent returns '&ihid_goodix->regulator_mutex'.

vim +151 drivers/hid/i2c-hid/i2c-hid-of-goodix.c

c1ed18c11bdb80 Douglas Anderson 2021-01-15 95 static int i2c_hid_of_goodix_probe(struct i2c_client *client,
c1ed18c11bdb80 Douglas Anderson 2021-01-15 96 const struct i2c_device_id *id)
c1ed18c11bdb80 Douglas Anderson 2021-01-15 97 {
c1ed18c11bdb80 Douglas Anderson 2021-01-15 98 struct i2c_hid_of_goodix *ihid_goodix;
2b87ff72c696bc Douglas Anderson 2021-06-11 99 int ret;
c1ed18c11bdb80 Douglas Anderson 2021-01-15 100 ihid_goodix = devm_kzalloc(&client->dev, sizeof(*ihid_goodix),
c1ed18c11bdb80 Douglas Anderson 2021-01-15 101 GFP_KERNEL);
c1ed18c11bdb80 Douglas Anderson 2021-01-15 102 if (!ihid_goodix)
c1ed18c11bdb80 Douglas Anderson 2021-01-15 103 return -ENOMEM;
c1ed18c11bdb80 Douglas Anderson 2021-01-15 104
2b87ff72c696bc Douglas Anderson 2021-06-11 105 mutex_init(&ihid_goodix->regulator_mutex);
2b87ff72c696bc Douglas Anderson 2021-06-11 106
c1ed18c11bdb80 Douglas Anderson 2021-01-15 107 ihid_goodix->ops.power_up = goodix_i2c_hid_power_up;
c1ed18c11bdb80 Douglas Anderson 2021-01-15 108 ihid_goodix->ops.power_down = goodix_i2c_hid_power_down;
c1ed18c11bdb80 Douglas Anderson 2021-01-15 109
c1ed18c11bdb80 Douglas Anderson 2021-01-15 110 /* Start out with reset asserted */
c1ed18c11bdb80 Douglas Anderson 2021-01-15 111 ihid_goodix->reset_gpio =
c1ed18c11bdb80 Douglas Anderson 2021-01-15 112 devm_gpiod_get_optional(&client->dev, "reset", GPIOD_OUT_HIGH);
c1ed18c11bdb80 Douglas Anderson 2021-01-15 113 if (IS_ERR(ihid_goodix->reset_gpio))
c1ed18c11bdb80 Douglas Anderson 2021-01-15 114 return PTR_ERR(ihid_goodix->reset_gpio);
c1ed18c11bdb80 Douglas Anderson 2021-01-15 115
c1ed18c11bdb80 Douglas Anderson 2021-01-15 116 ihid_goodix->vdd = devm_regulator_get(&client->dev, "vdd");
c1ed18c11bdb80 Douglas Anderson 2021-01-15 117 if (IS_ERR(ihid_goodix->vdd))
c1ed18c11bdb80 Douglas Anderson 2021-01-15 118 return PTR_ERR(ihid_goodix->vdd);
c1ed18c11bdb80 Douglas Anderson 2021-01-15 119
c1ed18c11bdb80 Douglas Anderson 2021-01-15 120 ihid_goodix->timings = device_get_match_data(&client->dev);
c1ed18c11bdb80 Douglas Anderson 2021-01-15 121
2b87ff72c696bc Douglas Anderson 2021-06-11 122 /*
2b87ff72c696bc Douglas Anderson 2021-06-11 123 * We need to control the "reset" line in lockstep with the regulator
2b87ff72c696bc Douglas Anderson 2021-06-11 124 * actually turning on an off instead of just when we make the request.
2b87ff72c696bc Douglas Anderson 2021-06-11 125 * This matters if the regulator is shared with another consumer.
2b87ff72c696bc Douglas Anderson 2021-06-11 126 * - If the regulator is off then we must assert reset. The reset
2b87ff72c696bc Douglas Anderson 2021-06-11 127 * line is active low and on some boards it could cause a current
2b87ff72c696bc Douglas Anderson 2021-06-11 128 * leak if left high.
2b87ff72c696bc Douglas Anderson 2021-06-11 129 * - If the regulator is on then we don't want reset asserted for very
2b87ff72c696bc Douglas Anderson 2021-06-11 130 * long. Holding the controller in reset apparently draws extra
2b87ff72c696bc Douglas Anderson 2021-06-11 131 * power.
2b87ff72c696bc Douglas Anderson 2021-06-11 132 */
2b87ff72c696bc Douglas Anderson 2021-06-11 133 mutex_lock(&ihid_goodix->regulator_mutex);
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2b87ff72c696bc Douglas Anderson 2021-06-11 134 ihid_goodix->nb.notifier_call = ihid_goodix_vdd_notify;
2b87ff72c696bc Douglas Anderson 2021-06-11 135 ret = regulator_register_notifier(ihid_goodix->vdd, &ihid_goodix->nb);
2b87ff72c696bc Douglas Anderson 2021-06-11 136 if (ret)
2b87ff72c696bc Douglas Anderson 2021-06-11 137 return dev_err_probe(&client->dev, ret,
2b87ff72c696bc Douglas Anderson 2021-06-11 138 "regulator notifier request failed\n");
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Drop the lock before returning?

2b87ff72c696bc Douglas Anderson 2021-06-11 139
2b87ff72c696bc Douglas Anderson 2021-06-11 140 /*
2b87ff72c696bc Douglas Anderson 2021-06-11 141 * If someone else is holding the regulator on (or the regulator is
2b87ff72c696bc Douglas Anderson 2021-06-11 142 * an always-on one) we might never be told to deassert reset. Do it
2b87ff72c696bc Douglas Anderson 2021-06-11 143 * now. Here we'll assume that someone else might have _just
2b87ff72c696bc Douglas Anderson 2021-06-11 144 * barely_ turned the regulator on so we'll do the full
2b87ff72c696bc Douglas Anderson 2021-06-11 145 * "post_power_delay" just in case.
2b87ff72c696bc Douglas Anderson 2021-06-11 146 */
2b87ff72c696bc Douglas Anderson 2021-06-11 147 if (ihid_goodix->reset_gpio && regulator_is_enabled(ihid_goodix->vdd))
2b87ff72c696bc Douglas Anderson 2021-06-11 148 goodix_i2c_hid_deassert_reset(ihid_goodix, true);
2b87ff72c696bc Douglas Anderson 2021-06-11 149 mutex_unlock(&ihid_goodix->regulator_mutex);
2b87ff72c696bc Douglas Anderson 2021-06-11 150
c1ed18c11bdb80 Douglas Anderson 2021-01-15 @151 return i2c_hid_core_probe(client, &ihid_goodix->ops, 0x0001);
c1ed18c11bdb80 Douglas Anderson 2021-01-15 152 }

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@xxxxxxxxxxxx