On 2017-02-26 21:47, Dan Zach wrote:
> Made the root cell to be 4.8.2-rt as well. With Jailhouse on, it runs
> the same test with typical jitter of 2-3uS and worse case of 35uS.
> 
> 
> COUNTER                              SUM   PER
> SEC                                                                           
>                                        
> 
> vmexits_total                     445327      1136
> vmexits_virt_irq                  430071      1107
> vmexits_mmio                       10073        15
> vmexits_virt_sgi                    4552         8
> vmexits_hypercall                    623         7
> vmexits_management                    14         0
> vmexits_cp15                           0         0
> vmexits_maintenance                    0         0
> vmexits_psci                           0         0
> 
> 
> 
> - Is the root cell handled differently on IRQ handling then non-root?

Nope.

> - Why aren't there any CP15 exits, while the non-root cell with the same
> kernel, runs 3.5K/sec on timer rearming?

The root cell booted with a hardware-provided broadcast timer and, thus,
can switch to hires timer mode (it actually did that prior to loading
Jaihouse). The non-root cell needs the patch Ralf pointed you to already.

> 
> 
> And now the interesting part:
> 
> _*If the root cell itself*_ runs real-time 1ms task, the non-root cell
> average jitter drops to 12uS (from 50) and worth case to 70uS.
> How about that for a puzzle?

That can be caching effects (keeping a core busy with cache-bound tasks
means it can request less from RAM and, thus, cause less pressure on
that shared link), or it is related to power management (the idle root
cell may enable power savings that have side effects on other cores).

Jan

-- 
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux

-- 
You received this message because you are subscribed to the Google Groups 
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to