Re: [PATCH 0/2] printk, Add printk.clock kernel parameter [v2]

[Prarit Bhargava]

On 01/14/2016 07:52 AM, Petr Mladek wrote:
> On Wed 2016-01-13 18:28:50, Thomas Gleixner wrote:
>> You can solve the whole business by changing the timestamp in printk_log to
>>
>>    u64	    mono;
>>    u64	    offset_real;
> 
> This is not so easy because the structure is proceed by userspace tool,
> e.g. crash, see log_buf_kexec_setup(). We would need to update all
> the tools as well.
> 
> 
>> and have a function which does:
>>
>> u64 ktime_get_log_ts(u64 *offset_real)
>> {
>> 	*offset_real = tk_core.timekeeper.offs_real;
>>
>> 	if (timekeeping_active)
>> 		return ktime_get_mono_fast_ns();
>> 	else
>> 		return local_clock();
>> }
> 
> A solution would be to apply the offset_real immediately. I wonder if
> any tool expects the messages to be sorted by a monotonic clock. In

/var/log/messages from systemd will have to be fixed, but that's something that
was brought up previously (and IMO should be trivial based on the value in
/sys/modules/printk/parameters/time).

> fact, it might be useful to see that some messages are disordered
> against the real time, e.g. because of the leaf second.

I kicked a leap seconds during my testing (I have been running tests from
/tools/tests/selftests/timers) and didn't see anything strange with both a stock
leap-a-day.c and a modified leap-a-day.c which only does leap insertions.

P.

> 
> Best Regards,
> Petr
>