djprobes status

[Satoshi Oshima]

Ingo Molnar wrote:
> On Fri, 2006-09-15 at 15:24 -0400, Satoshi Oshima wrote:
>> We proposed such an idea, but tons of disagreement had come at that
>> time;-) If you support the plan, it is worth to try it again.
> 
> i like this because it creates synergy between the sw-suspend code and
> kprobes. Shared infrastructure is always good. What type of disagreement
> was there, was it on lkml?

We had pushed djprobes to systemtap mailing list first, and lkml once.
And we have talked djprobes idea at OLS and other conferences.
We always recieved disagreement. Even Andrew Morton disagreed this 
approach, If I could remember correctly.

Almost all diagreements were concerned with performance of probe 
insertion.


> i dont like the performance aspect of it though: we'd need to schedule
> nr_tasks times to insert probes into the kernel. So i'd consider this an
> easy "plan B" approach, to be used if "plan A" fails :-)

Agreed. We push "plan A" first.

Satoshi