Re: [RFC PATCH 3/7] timekeeping: Fix cross-timestamp interpolation for non-x86

[Peter Hilber]

On 08.07.23 01:31, John Stultz wrote:
> On Fri, Jun 30, 2023 at 10:12 AM Peter Hilber
> <peter.hilber@xxxxxxxxxxxxxxx> wrote:
>>
>> So far, get_device_system_crosststamp() unconditionally passes
>> system_counterval.cycles to timekeeping_cycles_to_ns(). But when
>> interpolating system time (do_interp == true), system_counterval.cycles is
>> before tkr_mono.cycle_last, contrary to the timekeeping_cycles_to_ns()
>> expectations.
>>
>> On x86, CONFIG_CLOCKSOURCE_VALIDATE_LAST_CYCLE will mitigate on
>> interpolating, setting delta to 0. With delta == 0, xtstamp->sys_monoraw
>> and xtstamp->sys_realtime are then set to the last update time, as
>> implicitly expected by adjust_historical_crosststamp(). On other
>> architectures, the resulting nonsense xtstamp->sys_monoraw and
>> xtstamp->sys_realtime corrupt the xtstamp (ts) adjustment in
>> adjust_historical_crosststamp().
>>
>> Fix this by always setting the delta to 0 when interpolating.
>>
>> Fixes: 2c756feb18d9 ("time: Add history to cross timestamp interface supporting slower devices")
>> Signed-off-by: Peter Hilber <peter.hilber@xxxxxxxxxxxxxxx>
>> ---
>>  kernel/time/timekeeping.c | 13 +++++++++----
>>  1 file changed, 9 insertions(+), 4 deletions(-)
>>
>> diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c
>> index 7e86d5cd784d..7ccc2377c319 100644
>> --- a/kernel/time/timekeeping.c
>> +++ b/kernel/time/timekeeping.c
>> @@ -1259,10 +1259,15 @@ int get_device_system_crosststamp(int (*get_time_fn)
>>                                       tk_core.timekeeper.offs_real);
>>                 base_raw = tk->tkr_raw.base;
>>
>> -               nsec_real = timekeeping_cycles_to_ns(&tk->tkr_mono,
>> -                                                    system_counterval.cycles);
>> -               nsec_raw = timekeeping_cycles_to_ns(&tk->tkr_raw,
>> -                                                   system_counterval.cycles);
>> +               if (do_interp) {
>> +                       nsec_real = timekeeping_delta_to_ns(&tk->tkr_mono, 0);
>> +                       nsec_raw = timekeeping_delta_to_ns(&tk->tkr_raw, 0);
>> +               } else {
>> +                       nsec_real = timekeeping_cycles_to_ns(
>> +                               &tk->tkr_mono, system_counterval.cycles);
>> +                       nsec_raw = timekeeping_cycles_to_ns(
>> +                               &tk->tkr_raw, system_counterval.cycles);
>> +               }
> 
> Rather than adding another conditional branch here to go through, why
> not just use "cycles" instead of system_counterval.cycles as it seems
> to be set properly already?

OK. Thanks for the review and suggestion!