[Mesa-dev] [PATCH 1/5] anv: add ETIME fallback define

[Greg V]

On 10/26/2017 17:02, Emil Velikov wrote:

> On 24 October 2017 at 18:21, Greg V <greg at unrelenting.technology> wrote:
>>
>> On 10/24/2017 20:16, Dylan Baker wrote:
>>> Quoting Greg V (2017-10-23 13:20:39)
>>>> FreeBSD only has ETIMEDOUT, not ETIME
>>>> ---
>>>>    src/intel/vulkan/anv_device.c | 4 ++++
>>>>    src/intel/vulkan/anv_gem.c    | 4 ++++
>>>>    src/intel/vulkan/anv_queue.c  | 4 ++++
>>>>    3 files changed, 12 insertions(+)
>>>>
>>>> diff --git a/src/intel/vulkan/anv_device.c
>>>> b/src/intel/vulkan/anv_device.c
>>>> index 67028e8da9..fab2a394ca 100644
>>>> --- a/src/intel/vulkan/anv_device.c
>>>> +++ b/src/intel/vulkan/anv_device.c
>>>> @@ -39,6 +39,10 @@
>>>>      #include "genxml/gen7_pack.h"
>>>>
>>> I would add a comment here that this is for FreeBSD. Alternatively you
>>> might
>>> consider #ifdef __FreeBSD__ since it's self documenting.
>>>
>>> Dylan
>> It's not just for FreeBSD, pretty much any non-Linux OS doesn't have ETIME.
>>
> Indeed the lack of ETIME (and ERESTART) is pretty wide-spread.
> Greg, do you have any references why others (or FreeBSD at least)
> haven't implemented it - say wiki page, discussion thread, etc?
It's listed on https://wiki.freebsd.org/FreeBSD_and_Standards as "not 
implemented" without any other comments…

Every time it comes up (which is pretty rare — only a couple 
non-duplicate results for googling '"etime" site:bugs.freebsd.org') it's 
dismissed as a Linuxism, e.g. 
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=209101

I can't find why it *was* introduced :)
> I have no opinion on the patch, although it's strange to go through
> patching both ends -  userspace components and kernel(?).
>
> Thanks
> Emil