Messages in this thread | | | Date | Fri, 8 Jan 2016 10:04:44 +0100 (CET) | From | Thomas Gleixner <> | Subject | Re: [PATCH 2/2] printk, Add printk.clock kernel parameter |
| |
On Thu, 7 Jan 2016, Petr Mladek wrote: > On Thu 2016-01-07 10:38:37, Prarit Bhargava wrote: > > Heh :) I didn't even think of testing that. I'll have to think about that a > > bit more. systemd logging must assume boot time/local time. I'll ping those > > guys with a patch if/when the next version of this gets accepted. > > I am not sure how many tools would be affected by this.
I don't see a point in a gazillion of different clocks for printk either.
> An easier solution would be to normalize all times and always save the > relative time since the boot as it is done now.
Why? What's wrong with simply using the existing ktime_get_mono_fast_ns() instead of local_clock() and be done with it?
Thanks,
tglx
| |