Re: [PATCH 2/2][RFC] ipmi: Update timespec usage to timespec64

[John Stultz <john.stultz@linaro.org>]

On Wed, Jan 7, 2015 at 1:12 PM, Arnd Bergmann <arnd@arndb.de> wrote:
> On Wednesday 07 January 2015 12:51:50 John Stultz wrote:
>> As part of the internal y2038 cleanup, this patch removes
>> timespec usage in the ipmi driver, replacing it timespec64
>>
>> Cc: Corey Minyard <minyard@acm.org>
>> Cc: openipmi-developer@lists.sourceforge.net
>> Cc: Arnd Bergmann <arnd@arndb.de>
>> Signed-off-by: John Stultz <john.stultz@linaro.org>
>>
>
> In other drivers, we tended to use ktime_t and monotonic time,
> but your approach is definitely simpler because it doesn't have
> to rework the ipmi_si_is_busy logic and just do a
> s/timespec/timespec64/ conversion.
>
> Do you think it makes sense to use ktime_get_ts64 instead of
> getnstimeofday64 to get a monotonic time? The advantage would
> be to have the code work slightly better when racing against
> settimeofday, the downside would be that the debug printk
> shows the changed time base, but that would hopefully be
> irrelevant to someone debugging the code.

So here they were explicitly printing wall-time in sec/usec, so I was
somewhat hoping not to create any major behavioral changes.

But looking at this again, I see I sent a stale patch which has a
build bug (using tv_usec w/ a timespec64)... so let me redo this
anyway.

thanks
-john