There is at least:
* pthread_t, which one cannot portably format in a sane way.
* gettid(), which is Linux specific.
* ovsthread_id_self() from ovs-thread.h.
and I didn't know which one you meant.
gettid() is Linux specific.
On Thu, Sep 05, 2013 at 12:46:23PM -0700, Ethan Jackson wrote:
> Basically the result of gettid(). There's more than one thread ID?
>
> Ethan
>
> On Thu, Sep 5, 2013 at 12:43 PM, Ben Pfaff <[email protected]> wrote:
> > Which thread ID do you mean?
> >
> > On Thu, Sep 05, 2013 at 10:55:26AM -0700, Ethan Jackson wrote:
> >> Does it make sense to use the thread ID instead? That way it'd be
> >> clear if some reconfiguration happens that destroys and recreates
> >> them.
> >>
> >> Ethan
> >>
> >> On Thu, Sep 5, 2013 at 10:38 AM, Ben Pfaff <[email protected]> wrote:
> >> > This may occasionally make debugging easier.
> >> >
> >> > Suggested-by: Keith Amidon <[email protected]>
> >> > Signed-off-by: Ben Pfaff <[email protected]>
> >> > ---
> >> > ofproto/ofproto-dpif-upcall.c | 3 ++-
> >> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >> >
> >> > diff --git a/ofproto/ofproto-dpif-upcall.c
> >> > b/ofproto/ofproto-dpif-upcall.c
> >> > index 54f441b..1605c63 100644
> >> > --- a/ofproto/ofproto-dpif-upcall.c
> >> > +++ b/ofproto/ofproto-dpif-upcall.c
> >> > @@ -448,7 +448,8 @@ udpif_miss_handler(void *arg)
> >> > struct list misses = LIST_INITIALIZER(&misses);
> >> > struct handler *handler = arg;
> >> >
> >> > - set_subprogram_name("miss_handler");
> >> > + set_subprogram_name("miss_handler_%td",
> >> > + handler - handler->udpif->handlers);
> >> > for (;;) {
> >> > size_t i;
> >> >
> >> > --
> >> > 1.7.10.4
> >> >
> >> > _______________________________________________
> >> > dev mailing list
> >> > [email protected]
> >> > http://openvswitch.org/mailman/listinfo/dev
_______________________________________________
dev mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/dev
