Re: f_hid.c conversion to the new function interface and crash due to race

From: Felipe Balbi
Date: Thu Aug 10 2017 - 05:03:21 EST



Hi,

(please break your lines at 80 columns)

Variksla <variksla@xxxxxxxxx> writes:
>> On Aug 9, 2017, at 2:56 AM, Felipe Balbi <felipe.balbi@xxxxxxxxxxxxxxx> wrote:
>
> Thanks for replying.
>>
>>
>> Hi,
>>
>> noman pouigt <variksla@xxxxxxxxx> writes:
>>> Hello,
>>>
>>> I am currently using 3.18 linux kernel and getting below
>>
>> return -EKERNELTOOOLD;
>>
>> Sorry, you're on your own. You should ask for support from whoever
>> forces you to use 3.18 kernel.
>
> I have updated f_hid.c file to latest kernel though as I knew it would
> be too much to ask to get support for old code.

it's not really enough. In fact, it's even worse as the two codebases
don't really match.

> I am just not sure about how to instantiate the function driver as
> earlier f_fs was instantiating the hid driver. However, after hid got
> converted to new function driver who is responsible for calling setup,
> bind and unbind functions in f_hid.c ?
>
> Below is the glue code in android.c.

there's no android.c upstream.

> drivers/usb/gadget/android.c
> static struct android_usb_function hid_function = {
> .name = "hid",
> .init = hid_function_init,
> .cleanup = hid_function_cleanup,
> .bind_config = hid_function_bind_config,
> .attributes = hid_function_attributes,
> };
>
> Is there something similar in new function driver in upstream kernel?

you can check for yourself ;-)

--
balbi

Attachment: signature.asc
Description: PGP signature