From: Alex Netes <alexne-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
To: Hal Rosenstock <hal-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
Cc: Ira Weiny <weiny2-i2BcT+NCU+M@public.gmane.org>,
"linux-rdma
(linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org)"
<linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
Subject: Re: [PATCH] opensm/libvendor/osm_vendor_ibumad.c: Fix DR path printing on send timeouts
Date: Tue, 20 Dec 2011 11:04:29 +0200 [thread overview]
Message-ID: <20111220090429.GA2747@calypso> (raw)
In-Reply-To: <4EEBAA4F.4070701-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
Sorry for late reply...
On 15:30 Fri 16 Dec , Hal Rosenstock wrote:
> On 12/16/2011 2:51 PM, Ira Weiny wrote:
> > On Fri, 16 Dec 2011 11:24:54 -0800
> > Hal Rosenstock <hal-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org> wrote:
> >
> >>
> >> Also, make log_send_error into a routine and also call when MAD is cancelled.
> >>
> >> This patch is an alternative to Ira's v4 patch entitled
> >> "opensm: Move Error printing to MAD error call back functions"
> >> which pushes this functionality up into the send error callbacks.
> >>
> >> This patch is currently compile tested only.
> >>
> >> Signed-off-by: Hal Rosenstock <hal-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> >
> > I think this is acceptable as well.
>
> Guess it's up to Alex to decide which way to go...
>
I think that having these prints as low as possible is better and that
osm_vendor_ibumad is more suitable for doing that.
> > My main need was to report which request MAD did not get a response.
>
> I'll update to v2 eliminating the hop pointer as you indicated in a
> previous email and maybe look for any send error callback logging
> redundancy and remove that too but I'd prefer to wait to here Alex's
> comments.
>
> > I will test when I get back on the system which was having the issues.
>
> Thanks. That will be most helpful (and save me some time)!
>
> -- Hal
>
> > Acked-by: Ira Weiny <weiny2-i2BcT+NCU+M@public.gmane.org>
> >
> >> ---
> >> diff --git a/libvendor/osm_vendor_ibumad.c b/libvendor/osm_vendor_ibumad.c
> >> index e2ebd8e..3781792 100644
> >> --- a/libvendor/osm_vendor_ibumad.c
> >> +++ b/libvendor/osm_vendor_ibumad.c
> >> @@ -98,6 +98,37 @@ typedef struct _umad_receiver {
> >>
> >> static void osm_vendor_close_port(osm_vendor_t * const p_vend);
> >>
> >> +static void log_send_error(osm_vendor_t * const p_vend, osm_madw_t *p_madw)
> >> +{
> >> + if (p_madw->p_mad->mgmt_class != IB_MCLASS_SUBN_DIR) {
> >> + /* LID routed */
> >> + OSM_LOG(p_vend->p_log, OSM_LOG_ERROR, "ERR 5410: "
> >> + "Send completed with error (%s) -- dropping\n"
> >> + "\t\t\tClass 0x%x, Method 0x%X, Attr 0x%X, "
> >> + "TID 0x%" PRIx64 ", LID %u\n",
> >> + ib_get_err_str(p_madw->status),
> >> + p_madw->p_mad->mgmt_class, p_madw->p_mad->method,
> >> + cl_ntoh16(p_madw->p_mad->attr_id),
> >> + cl_ntoh64(p_madw->p_mad->trans_id),
> >> + cl_ntoh16(p_madw->mad_addr.dest_lid));
> >> + } else {
> >> + ib_smp_t *p_smp;
> >> +
> >> + /* Direct routed SMP */
> >> + p_smp = osm_madw_get_smp_ptr(p_madw);
> >> + OSM_LOG(p_vend->p_log, OSM_LOG_ERROR, "ERR 5411: "
> >> + "DR SMP Send completed with error (%s) -- dropping\n"
> >> + "\t\t\tMethod 0x%X, Attr 0x%X, TID 0x%" PRIx64
> >> + ", Hop Ptr: 0x%X\n",
> >> + ib_get_err_str(p_madw->status),
> >> + p_madw->p_mad->method,
> >> + cl_ntoh16(p_madw->p_mad->attr_id),
> >> + cl_ntoh64(p_madw->p_mad->trans_id),
> >> + p_smp->hop_ptr);
> >> + osm_dump_smp_dr_path(p_vend->p_log, p_smp, OSM_LOG_ERROR);
> >> + }
> >> +}
> >> +
> >> static void clear_madw(osm_vendor_t * p_vend)
> >> {
> >> umad_match_t *m, *e, *old_m;
> >> @@ -185,6 +216,7 @@ put_madw(osm_vendor_t * p_vend, osm_madw_t * p_madw, ib_net64_t tid)
> >> p_req_madw = old_lru->v;
> >> p_bind = p_req_madw->h_bind;
> >> p_req_madw->status = IB_CANCELED;
> >> + log_send_error(p_vend, p_req_madw);
> >> pthread_mutex_lock(&p_vend->cb_mutex);
> >> (*p_bind->send_err_callback) (p_bind->client_context, p_req_madw);
> >> pthread_mutex_unlock(&p_vend->cb_mutex);
> >> @@ -326,32 +358,6 @@ static void *umad_receiver(void *p_ptr)
> >>
> >> /* if status != 0 then we are handling recv timeout on send */
> >> if (umad_status(p_madw->vend_wrap.umad)) {
> >> -
> >> - if (mad->mgmt_class != IB_MCLASS_SUBN_DIR) {
> >> - /* LID routed */
> >> - OSM_LOG(p_vend->p_log, OSM_LOG_ERROR, "ERR 5410: "
> >> - "Send completed with error -- dropping\n"
> >> - "\t\t\tClass 0x%x, Method 0x%X, Attr 0x%X, "
> >> - "TID 0x%" PRIx64 ", LID %u\n",
> >> - mad->mgmt_class, mad->method,
> >> - cl_ntoh16(mad->attr_id),
> >> - cl_ntoh64(mad->trans_id),
> >> - cl_ntoh16(ib_mad_addr->lid));
> >> - } else {
> >> - ib_smp_t *smp;
> >> -
> >> - /* Direct routed SMP */
> >> - smp = (ib_smp_t *) mad;
> >> - OSM_LOG(p_vend->p_log, OSM_LOG_ERROR, "ERR 5411: "
> >> - "DR SMP Send completed with error -- dropping\n"
> >> - "\t\t\tMethod 0x%X, Attr 0x%X, TID 0x%" PRIx64
> >> - ", Hop Ptr: 0x%X\n",
> >> - mad->method, cl_ntoh16(mad->attr_id),
> >> - cl_ntoh64(mad->trans_id), smp->hop_ptr);
> >> - osm_dump_smp_dr_path(p_vend->p_log, smp,
> >> - OSM_LOG_ERROR);
> >> - }
> >> -
> >> if (!(p_req_madw = get_madw(p_vend, &mad->trans_id))) {
> >> OSM_LOG(p_vend->p_log, OSM_LOG_ERROR,
> >> "ERR 5412: "
> >> @@ -361,6 +367,7 @@ static void *umad_receiver(void *p_ptr)
> >> cl_ntoh64(mad->trans_id));
> >> } else {
> >> p_req_madw->status = IB_TIMEOUT;
> >> + log_send_error(p_vend, p_req_madw);
> >> /* cb frees req_madw */
> >> pthread_mutex_lock(&p_vend->cb_mutex);
> >> pthread_cleanup_push(unlock_mutex,
> >
> >
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
-- Alex
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
prev parent reply other threads:[~2011-12-20 9:04 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-12-16 19:24 [PATCH] opensm/libvendor/osm_vendor_ibumad.c: Fix DR path printing on send timeouts Hal Rosenstock
[not found] ` <4EEB9B06.8090908-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2011-12-16 19:51 ` Ira Weiny
[not found] ` <20111216115125.68c52d60.weiny2-i2BcT+NCU+M@public.gmane.org>
2011-12-16 20:30 ` Hal Rosenstock
[not found] ` <4EEBAA4F.4070701-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2011-12-20 9:04 ` Alex Netes [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20111220090429.GA2747@calypso \
--to=alexne-vpraknaxozvwk0htik3j/w@public.gmane.org \
--cc=hal-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org \
--cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=weiny2-i2BcT+NCU+M@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox