breakpoint assistance (uprobes)

[Roland McGrath]

> Uprobes folks have talked a fair amount amongst ourselves about
> single-stepping in multithreaded apps.  This approach was on our list,
> but we haven't tried it because (1) it didn't seem very promising
> performance-wise, 

I agree, but reiterate that it's never been tried in the in-kernel utrace
context to see what the performance overhead is.  It certainly doesn't
scale well in the number of threads.

> and (2) it could really drive up the number of
> SIGSTOPs utrace injects into the probed process.

SIGSTOP per se is never involved.

> > Fancy VM tricks.  [...]
> 
> This is not currently on our radar, at least as far as code probepoints
> are concerned.

Like I said, it's a long-term prospect in my mind.


Thanks,
Roland