[PATCH 5/5] input: support onkey in 88pm860x
Haojian Zhuang
haojian.zhuang at gmail.com
Thu Feb 4 06:08:42 EST 2010
On Wed, Feb 3, 2010 at 12:36 PM, Dmitry Torokhov
<dmitry.torokhov at gmail.com> wrote:
> Hi Haojian,
>
> On Wed, Feb 03, 2010 at 08:04:30AM -0500, Haojian Zhuang wrote:
>
> Looks most;y good with the exceptin of the following:
>
>> +
>> + info = kzalloc(sizeof(struct pm860x_onkey_info), GFP_KERNEL);
>> + if (!info)
>> + return -ENOMEM;
>> + info->chip = chip;
>> + info->i2c = (chip->id == CHIP_PM8607) ? chip->client : chip->companion;
>> + info->dev = &pdev->dev;
>> + info->irq = irq + chip->irq_base;
>> +
>> + ret = request_threaded_irq(info->irq, NULL, pm860x_onkey_handler,
>> + IRQF_ONESHOT, "onkey", info);
>> + if (ret < 0) {
>> + dev_err(chip->dev, "Failed to request IRQ: #%d: %d\n",
>> + info->irq, ret);
>> + goto out;
>> + }
>> +
>> + info->idev = input_allocate_device();
>
> You can't enable IRQ first and allocate the device later. If IRQ comes
> here it will go "BOOM".
>
>> + if (!info->idev) {
>> + dev_err(chip->dev, "Failed to allocate input dev\n");
>> + ret = -ENOMEM;
>> + goto out_input;
>> + }
>> +
>> + info->idev->name = "88pm860x_on";
>> + info->idev->phys = "88pm860x_on/input0";
>> + info->idev->id.bustype = BUS_I2C;
>> + info->idev->dev.parent = &pdev->dev;
>> + info->irq = irq;
>> + info->idev->evbit[0] = BIT_MASK(EV_KEY);
>> + info->idev->keybit[BIT_WORD(KEY_POWER)] = BIT_MASK(KEY_POWER);
>> +
>> + ret = input_register_device(info->idev);
>> + if (ret) {
>> + dev_err(chip->dev, "Can't register input device: %d\n", ret);
>> + goto out_reg;
>> + }
>> + platform_set_drvdata(pdev, info);
>> +
>> + return 0;
>> +
>> +out_reg:
>> + input_free_device(info->idev);
>> +out_input:
>> + free_irq(info->irq, info);
>
> So this needs to be in reverse order as well.
>
>> +out:
>> + kfree(info);
>> + return ret;
>> +}
>> +
>> +static int __devexit pm860x_onkey_remove(struct platform_device *pdev)
>> +{
>> + struct pm860x_onkey_info *info = platform_get_drvdata(pdev);
>> +
>> + if (info) {
>
> "info" will never be NULL, your probe routine makes sure of it. Also it
> is considered a good manners to set drvdata back to NULL after
> unbinding.
>
>> + free_irq(info->irq, info);
>> + input_unregister_device(info->idev);
>> + kfree(info);
>> + }
>> + return 0;
>> +}
>> +
>
Fix and format the new patch.
Thanks
Haojian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0005-input-support-onkey-in-88pm860x.patch
Type: text/x-patch
Size: 6025 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20100204/9345e64e/attachment-0001.bin>
More information about the linux-arm-kernel
mailing list