Re: [PATCH 7/7] time: remove timespec64 hack

[Arnd Bergmann]

On Tue, Oct 24, 2017 at 10:49 PM, John Stultz <john.stultz@xxxxxxxxxx> wrote:
> On Thu, Oct 19, 2017 at 4:14 AM, Arnd Bergmann <arnd@xxxxxxxx> wrote:

>
> This one doesn't seem to build for me...  In the meantime, I'm going
> to go ahead testing with patches 1-6.
>
> jstultz@buildbox:~/projects/linux/time$ make -j24 bzImage > /dev/null
> fs/select.c: In function âcompat_core_sys_selectâ:
> fs/select.c:1244:27: error: passing argument 3 of âdo_selectâ from
> incompatible pointer type [-Werror=incompatible-pointer-types]
>   ret = do_select(n, &fds, end_time);

I looked into this again, as I have some more patches on top now that
I wan to submit soon. The problem you ran into was fixed by Deepa's
"io_getevents: Use timespec64 to represent timeouts" patch that Al
Viro merged into vfs/for-next. It's currently in linux-net but not in
mainline, and I didn't notice the problem because I was building on
linux-next.

If you agree with this patch in principle, I'll resend it after the merge
window, applying this one will help with my "x86: convert x86_platform_ops
to timespec64" that needed a rather ugly workaround otherwise.

      Arnd