Re: [PATCH] ion_test: Add compat_ioctl support

[Colin Cross]

On Thu, Dec 19, 2013 at 3:30 PM, John Stultz <john.stultz@xxxxxxxxxx> wrote:
> On 12/19/2013 02:52 PM, Colin Cross wrote:
>> On Thu, Dec 19, 2013 at 2:40 PM, John Stultz <john.stultz@xxxxxxxxxx> wrote:
>>> Prior to subitting this, Colin reworked the compat_ioctl support
>>> for the ion_test driver, moving the structure to be the same size
>>> on both 32 and 64 bit architectures.
>>>
>>> Two small things were left out. The compat_ioctl ptr assignment,
>>> and the fact that despite having uniform sized types in the
>>> structure, the structure pads out to different sizes on different
>>> arches.
>>>
>>> This patch resolves this issue by setting the write flag as
>>> a __u64, and adding the compat_ioctl ptr.
>>>
>>> While this does affect the ABI for 32bit users, its only
>>> the ABI for the ion_test driver, not ion itself.
>>>
>>> Cc: Colin Cross <ccross@xxxxxxxxxxx>
>>> Cc: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>
>>> Cc: Android Kernel Team <kernel-team@xxxxxxxxxxx>
>>> Signed-off-by: John Stultz <john.stultz@xxxxxxxxxx>
>>> ---
>>>  drivers/staging/android/ion/ion_test.c  | 1 +
>>>  drivers/staging/android/uapi/ion_test.h | 2 +-
>>>  2 files changed, 2 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/staging/android/ion/ion_test.c b/drivers/staging/android/ion/ion_test.c
>>> index 3e20349..654acb5 100644
>>> --- a/drivers/staging/android/ion/ion_test.c
>>> +++ b/drivers/staging/android/ion/ion_test.c
>>> @@ -231,6 +231,7 @@ static int ion_test_release(struct inode *inode, struct file *file)
>>>  static const struct file_operations ion_test_fops = {
>>>         .owner = THIS_MODULE,
>>>         .unlocked_ioctl = ion_test_ioctl,
>>> +       .compat_ioctl = ion_test_ioctl,
>> Setting the compat_ioctl to the same thing as unlocked_ioctl shouldn't
>> be necessary, compat_sys_ioctl will fall through to do_vfs_ioctl if
>> compat_ioctl is not set, which will call unlocked_ioctl.
>
> Hrm. I've not looked into the implementation, but doesn't seem to be the
> case on x86_64. Without it I get -1 back when calling the ioctl from a
> 32bit application. With it the ioctls work.

I missed that compat_sys_ioctl returns -ENOTTY if do_ioctl_trans
fails, so the compat_ioctl is necessary.

>>
>>>         .open = ion_test_open,
>>>         .release = ion_test_release,
>>>  };
>>> diff --git a/drivers/staging/android/uapi/ion_test.h b/drivers/staging/android/uapi/ion_test.h
>>> index 614d1e3..f1727f5 100644
>>> --- a/drivers/staging/android/uapi/ion_test.h
>>> +++ b/drivers/staging/android/uapi/ion_test.h
>>> @@ -31,7 +31,7 @@ struct ion_test_rw_data {
>>>         __u64 ptr;
>>>         __u64 offset;
>>>         __u64 size;
>>> -       int write;
>>> +       __u64 write;
>>>  };
>> Whoops, missed that one.  Can you add "int padding" after "int write"
>> instead?  That at least lets us use the 32 bits later.
>
> Will do.
>
>
> thanks
> -john
>
>
> To unsubscribe from this group and stop receiving emails from it, send an email to kernel-team+unsubscribe@xxxxxxxxxxxx
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/