reentrant probes

[Seth, Rohit]

Ananth N Mavinakayanahalli <> wrote on Monday, May 02, 2005 11:13 AM:

> Yes, concurrency imposes that the handlers written be smp-safe.
> However, current kprobe users did not have to bother about making
> their handlers smp safe 'cos the lower layer serializes kprobe
> handling. 

Current usage of kprobe is very limited.  I guess all SW goes through
that period of infancy...  

> However, I'd like to see a generic change in the kprobe
> layer rather than having special casings for users who can/cannot
> handle concurrency . 
>

Oh absolutely.
 
> Wondering if it would it be fair to impose a restriction and say:
> existing kprobe users should make their handlers smp safe if they are
> not already?


This is definitely not a restriction.  But yes, anything that you are
running as part of kernel will have to be at the same level of maturity
as rest of the kernel (when we talk about usage in production
environments).

-rohit