[PATCH xserver] xfree86: Immediately handle failure to set HW cursor

[Michel Dänzer]

On 28.03.2016 18:35, Alexandre Courbot wrote:
> On 03/23/2016 06:14 PM, Michael Thayer wrote:
>> Looks good at first quick glance (I am away from the computer a bit over
>> the Easter weeks, I will take another look when I am back).  One thing I
>> can say immediately though: when I submitted similar patches a while
>> back, Keith had the idea of keeping the old hooks untouched and adding
>> in the additional "checked" variants so that existing driver code does
>> not have to be changed to keep working.  The other thing is that this
>> might need a bump of ABI_XINPUT_VERSION.
> 
> Thanks for the comments. It has actually been suggested to me to define
> a new _checked() variant of the set_cursor hooks, but since this patch
> changes all the call sites I refrained from doing it. However your
> message sounds like there are users of these hooks outside the xserver
> package, is this correct? (sorry, I'm a complete X n00b).

Yes, these hooks are basically used by all drivers which support a HW
cursor.


> If so I will resend as you suggested, and also bump ABI_XINPUT_VERSION.

I don't think ABI_XINPUT_VERSION has anything to do with your patch, and
ABI_VIDEODRV_VERSION has already been bumped for this release cycle, so
you don't need to worry about that.


P.S. Please consider subscribing acourbot at nvidia.com to this list, so
your posts don't go through the moderation queue. If you don't want to
receive list posts on that address, you can disable delivery in the
mailman preferences.

-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer