[BUG] etnaviv: broken timeouts

[Arnd Bergmann]

On Mon, Nov 6, 2017 at 2:40 PM, Lucas Stach <l.stach at pengutronix.de> wrote:
> Am Montag, den 06.11.2017, 13:42 +0100 schrieb Arnd Bergmann:
>
>> I have prototyped a patch to address all of the above, will send
>> that later today after doing a bit of build testing. I hope you can
>> give it some testing, and clarify what the timebase is that
>> user space expects to put into the timeout (MONOTONIC,
>> MONOTONIC_RAW, or REALTIME).
>
> Thanks. I don't think we spelled it out anywhere, but the timeouts are
> supposed to be based on CLOCK_MONOTONIC.

Ok, I'll use that in my patch then. Have you ever observed a timeout
in practice? Since CLOCK_MONOTONIC differs from jiffies in both
the INITIAL_JIFFIES 5 minute offset and the NTP clock drift, I wonder
if the timeouts have simply always ended up being too far in the future
to ever hit ;-)

      Arnd