[openib-general] [PATCH] osm: handle local events

[Eitan Zahavi]

Hi MST,

Hotplug event of HCA FATAL is supported and this is the #1 issue we are
troubled with.

Regarding other OpenSM vendors implementations (like the switch stack or
gen1 stacks)
This patch makes OpenSM agnostic for their support of events. If they do
support it OpenSM will
Register and behave according to the events if they do not - then no
harm is done.

Regarding ibumad implementation of propagating of events. Once it is
supported I will vote for
Re-writing the events feature to that interface. But it does not exist
now and ibuverbs do exist.

If required we could make the OSM_VENDOR_SUPPORT_EVENTS depend on the
availability of the libibuverbs.
But today it is common practice to have that lib compiled in all
distributions.

Eitan Zahavi
Senior Engineering Director, Software Architect
Mellanox Technologies LTD
Tel:+972-4-9097208
Fax:+972-4-9593245
P.O. Box 586 Yokneam 20692 ISRAEL


> -----Original Message-----
> From: Michael S. Tsirkin
> Sent: Thursday, August 24, 2006 4:29 PM
> To: Yevgeny Kliteynik
> Cc: OPENIB; Roland Dreier; Hal Rosenstock; Eitan Zahavi
> Subject: Re: [PATCH] osm: handle local events
> 
> Quoting r. Yevgeny Kliteynik <kliteyn at mellanox.co.il>:
> > Index: libvendor/osm_vendor_ibumad.c
> >
> ================================================================
> ===
> > --- libvendor/osm_vendor_ibumad.c	(revision 8998)
> > +++ libvendor/osm_vendor_ibumad.c	(working copy)
> > @@ -72,6 +72,7 @@
> >  #include <opensm/osm_log.h>
> >  #include <opensm/osm_mad_pool.h>
> >  #include <vendor/osm_vendor_api.h>
> > +#include <infiniband/verbs.h>
> >
> >  /****s* OpenSM: Vendor AL/osm_umad_bind_info_t
> >   * NAME
> 
> NAK.
> 
> This means that the SM becomes dependent on the uverbs module.  I
don't
> think this is a good idea.  Let's not go there - SM should depend just
on the
> umad module and libc.  In particular, SM should work even on embedded
> platforms where uverbs do not necessarily work.
> 
> Further, hotplug events still do not seem to be handled, even with
this patch.
> 
> For port events, it seems sane that umad module could provide a way to
listen
> for them.
> 
> A recent patch to mthca converts fatal events to hotplug events, so
fatal
> events can and should be handled as part of general hotplug support.
> 
> --
> MST