On Wed, Nov 30, 2011 at 15:37, Jiri Kosina <jkosina@xxxxxxx> wrote:
> On Tue, 29 Nov 2011, Benjamin Tissoires wrote:
>
>> Use the input_register mechanism to avoid a race in which the device
>> may emit events before the slots are ready.
>>
>> Signed-off-by: Benjamin Tissoires <benjamin.tissoires@xxxxxxxxx>
>> ---
>> drivers/hid/hid-multitouch.c | 37 ++++++++++++++++++++++++-------------
>> 1 files changed, 24 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/hid/hid-multitouch.c b/drivers/hid/hid-multitouch.c
>> index 91c13bc..bd6e410 100644
>> --- a/drivers/hid/hid-multitouch.c
>> +++ b/drivers/hid/hid-multitouch.c
>> @@ -574,6 +574,29 @@ static void mt_set_input_mode(struct hid_device *hdev)
>> }
>> }
>>
>> +static int mt_input_register(struct hid_device *hdev,
>> + struct hid_input *hidinput)
>> +{
>> + int ret;
>> + struct mt_device *td = hid_get_drvdata(hdev);
>> +
>> + td->slots = kzalloc(td->maxcontacts * sizeof(struct mt_slot),
>> + GFP_KERNEL);
>> + if (!td->slots) {
>> + dev_err(&hdev->dev, "cannot allocate multitouch slots\n");
>> + return -ENOMEM;
>> + }
>> +
>> + ret = sysfs_create_group(&hdev->dev.kobj, &mt_attribute_group);
>> + if (ret)
>> + hid_warn(hdev,
>> + "can't create sysfs quirks attribute err: %d\n", ret);
>> +
>> + mt_set_input_mode(hdev);
>> +
>> + return 0;
>> +}
>> +
>> static int mt_probe(struct hid_device *hdev, const struct hid_device_id *id)
>> {
>> int ret, i;
>> @@ -610,19 +633,6 @@ static int mt_probe(struct hid_device *hdev, const struct hid_device_id *id)
>> if (ret)
>> goto fail;
>>
>> - td->slots = kzalloc(td->maxcontacts * sizeof(struct mt_slot),
>> - GFP_KERNEL);
>> - if (!td->slots) {
>> - dev_err(&hdev->dev, "cannot allocate multitouch slots\n");
>> - hid_hw_stop(hdev);
>> - ret = -ENOMEM;
>> - goto fail;
>> - }
>> -
>> - ret = sysfs_create_group(&hdev->dev.kobj, &mt_attribute_group);
>> -
>> - mt_set_input_mode(hdev);
>> -
>> return 0;
>>
>> fail:
>> @@ -843,6 +853,7 @@ static struct hid_driver mt_driver = {
>> .input_mapping = mt_input_mapping,
>> .input_mapped = mt_input_mapped,
>> .feature_mapping = mt_feature_mapping,
>> + .input_register = mt_input_register,
>
> Umm, where does this come from? My memory is famous for not being really
> excellent, but I don't immediately remember any patchset adding such
> callback to hid_driver struct.
>
> And it's not anywhere in my applied queue anyway.
Indeed, your memory is perfectly working: I had some hallucinations.
I was sure you accepted the following patch:
http://article.gmane.org/gmane.linux.kernel.input/21692
It was when Jaikumar Ganesh tried to fix a race in the init of hid-magicmouse.
As the input_register callback is not queued, you can drop this.
Cheers,
Benjamin
>
>> .usage_table = mt_grabbed_usages,
>> .event = mt_event,
>> #ifdef CONFIG_PM
>> --
>> 1.7.4.4
>>
>
> --
> Jiri Kosina
> SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
