Messages in this thread | | | Subject | Re: [PATCH] ASoC: Intel: bytcht_es8316: Add missed put_device() | From | Pierre-Louis Bossart <> | Date | Thu, 16 Jul 2020 11:48:06 -0500 |
| |
>>>> snd_byt_cht_es8316_mc_probe() misses to call put_device() in an error >>>> path. Add the missed function call to fix it. >>>> >>>> Fixes: ba49cf6f8e4a ("ASoC: Intel: bytcht_es8316: Add quirk for >>>> inverted jack detect") >>>> Signed-off-by: Jing Xiangfeng <jingxiangfeng@huawei.com> >>> >>> Patch looks good to me: >>> >>> Reviewed-by: Hans de Goede <hdegoede@redhat.com> >> >> Actually the existing code looks quite odd to me: >> >> if (cnt) { >> ret = device_add_properties(codec_dev, props); >> if (ret) >> return ret; >> } >> >> devm_acpi_dev_add_driver_gpios(codec_dev, byt_cht_es8316_gpios); >> priv->speaker_en_gpio = >> gpiod_get_index(codec_dev, "speaker-enable", 0, >> /* see comment in byt_cht_es8316_resume */ >> GPIOD_OUT_LOW | GPIOD_FLAGS_BIT_NONEXCLUSIVE); >> put_device(codec_dev); >> >> if the first branch is not taken the put_device() will not be balanced. > > The get_device() does not come from the device_add_properties, it comes > from the earlier: > > codec_dev = bus_find_device_by_name(&i2c_bus_type, NULL, codec_name); > > call.
I probably needed more coffee when I wrote this, indeed this is fine...
Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
| |