[lkml]   [2011]   [Dec]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [RFC PATCH v1 5/7] media: v4l2: introduce two IOCTLs for face detection

On Fri, Dec 2, 2011 at 8:33 PM, Arnd Bergmann <> wrote:
> On Friday 02 December 2011 17:12:56 Ming Lei wrote:
>> +/**
>> + * struct v4l2_fd_result - VIDIOC_G_FD_RESULT argument
>> + * @buf_index: entry, index of v4l2_buffer for face detection
>> + * @face_cnt:  return, how many faces detected from the @buf_index
>> + * @fd:                return, result of faces' detection
>> + */
>> +struct v4l2_fd_result {
>> +       __u32   buf_index;
>> +       __u32   face_cnt;
>> +       __u32   reserved[6];
>> +       struct v4l2_fd_detection *fd;
>> +};
> This data structure is not 32/64 bit safe: running a 64 bit kernel with 32 bit
> user space will see an incompatible layout.

I agree that this is not 32/64 bit safe, but I understand lib32 can handle
this correctly, otherwise many 32bit applications can't run on current
64bit kernel
since many kernel structures used by user space contained pointer,
such as struct v4l2_buffer, struct v4l2_ext_controls in v4l2 ABI.

> One way to solve this is to remove the pointer and just start the array
> directly after the __u32 members. Alternatively, you can use a __u64
> to pass the pointer in an integer representation.

So I think this need not to be solved.

> A nicer interface from the data structure perspective would be to
> get rid of the array altogether and always return exactly one face.

I choose to return array to user space since v4l2 ioctl has provided
this kind of support, see video_usercopy().

Ming Lei
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2011-12-04 12:21    [W:0.074 / U:1.424 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site