lkml.org 
[lkml]   [2013]   [Feb]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 5/6] uprobes/tracing: Kill uprobe_trace_consumer, embed uprobe_consumer into trace_uprobe
On 02/04, Srikar Dronamraju wrote:
>
> * Oleg Nesterov <oleg@redhat.com> [2013-01-31 20:18:29]:
>
> > trace_uprobe->consumer and "struct uprobe_trace_consumer" add the
> > unnecessary indirection and complicate the code for no reason.
> >
> > This patch simply embeds uprobe_consumer into "struct trace_uprobe",
> > all other changes only fix the compilation errors.
>
> I know this patch doesnt change the current behaviour.

Yes, and it makes the code simpler.

> We dont handle two concurrent perf record sessions for the same user
> space probe. Since both sessons share the same trace_uprobe and hence
> share the same consumer.

We do? I am testing the patches I am going to send, and I specially
tried to verify that 2 concurent sessions with different/same filtering
constraints work fine.

Or I misunderstood what you meant...

> Initially I had thought of having a chain in
> uprobe_trace_consumer. However we dont get have enough information at
> the probe_event_disable() time to detect which consumer to delete Hence
> I dropped the idea of having a list of consumers attached to the
> trace_uprobe.

You know, until recently I knew absolutely nothing about kernel/events/
and kernel/trace/. Not that I really understand this code now, I can
be easily wrong.

But so far I think that a chain of multiple consumers makes no sense.
Each consumer->handler() will use the same call->perf_events list, this
will only complicate the code for no reason.

> However with allowing prefiltering, we need to have ability to
> distinguish consumers.

Certainly not. Please see the patches I am going to send.

Oleg.



\
 
 \ /
  Last update: 2013-02-04 19:03    [W:0.054 / U:0.096 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site