openfabrics-general mailing list

FromSubjectsort iconDate
Felix Marti
RE: [ofa-general] cmatose fails whereas rping passes on iWarp

Yes, Chelsio supports peer2peer mode but I'm not sure if it is enabled
by default; I'm sure Steve will respond to this but he's out for a few
days. Maybe modinfo on the iw_cxgb3 module will reveal a peer2peer
related parameters?

-----Original Message-----
From: general-bounces@lists.openfabrics.org
[mailto:general-bounces@lists.openfabrics.org] On Behalf Of Sean Hefty
Sent: Thursday, July 02, 2009 11:44 PM
To: Davis, Arlin R; 'pandit ib'; general@lists.openfabrics.org

This was my thought as wel...

Jul 3, 12:47 pm 2009
Lars Ellenberg
[ofa-general] upstream stable merge request for OFED 1.4

maybe I'm doing things wrong.
but the last merge with upstream for
git://git.openfabrics.org/ofed_1_4/linux-2.6.git
appears to have been 2.6.27.

there is a reason for the kernel.org stable branches.

if you do not plan to upgrade 2.6.28 -> 2.6.31 in the near future,
please consider to merge with
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.27.y.git

current upstream 2.6.27 stable is v2.6.27.26
958 files changed, 13481 insertions(+), 6784 deletions(-)

Thanks,

DRBD...

Jul 3, 10:48 am 2009
Bart Van Assche
Re: [ofa-general] upstream stable merge request for OFED 1.4

Hello Lars,

As you probably know any patch submitted for inclusion in a kernel.org
stable branch must have been accepted first for inclusion in the
latest vanilla kernel. So the above question is equivalent to whether
the development of the OFED kernel components should continue as a
separate distribution or whether all patches should be submitted for
inclusion in the latest Linux kernel. This topic has been discussed
several times on this mailing list. You can find the most recent
discussion of...

Jul 3, 1:34 pm 2009
Yossi Etigin
[ofa-general] A question about tx lock in ipoib_flush_paths

In ipoib_flush_paths(), we take the netif_tx_lock to remove a path
My question is - what data does this lock protect?
It isn't path->list and path->rb_node, because priv->lock is enough to protect them.

It might be neigh and neigh->ah, to avoid freeing the neighbour and its address
handle while ipoib_start_xmit() is using it, but this particular part is done *outside*
the tx lock.

Unless I'm missing something - shouldn't the code:

spin_unlock_irqrestore(&priv->lock, flags);...

Jul 3, 10:43 am 2009
Nicolas Morey-Chaise...
[ofa-general] [PATCH] opensm: Added support for select count...

Support for xmit_wait counters was missing in the perfmgr though it was read from the mad and event plugin interface already handles it.
This patch adds support for it (tested and working with an event plugin)

Signed-off-by: Nicolas Morey-Chaisemartin <nicolas@morey-chaisemartin.com>
---
opensm/include/opensm/osm_perfmgr_db.h | 23 ++++++-
opensm/opensm/osm_perfmgr.c | 30 +++++++-
opensm/opensm/osm_perfmgr_db.c | 124 +++++++++++++++++++++++++++++--
3 files changed, ...

Jul 3, 4:03 am 2009
Hal Rosenstock
[ofa-general] [PATCHv2] opensm/osm_mesh.c: Improve VL utiliz...

opensm/osm_mesh.c: Improve VL utilization

Add steps to the algorithm to further condition the netlist of switches
that is handed to lash to analyze.

(1) The previous version sorted out all the links out of switches into
dimension order’in the order the dimensions were found which is driven by
the arbitrary choice of links out of the seed’switch. This takes the
additional step of waiting until after the lengths along each dimension
are measured then reordering the links out of each switch s...

Jul 3, 9:34 am 2009
Hal Rosenstock
[ofa-general] [PATCH] opensm/osm_ucast_ftree.c: Made error n...

Signed-off-by: Hal Rosenstock <hal.rosenstock@gmail.com>
---
diff --git a/opensm/opensm/osm_ucast_ftree.c b/opensm/opensm/osm_ucast_ftree.c
index 6c695de..bd61002 100644
--- a/opensm/opensm/osm_ucast_ftree.c
+++ b/opensm/opensm/osm_ucast_ftree.c
@@ -3617,7 +3617,7 @@ static int fabric_read_guid_files(IN ftree_fabric_t * p_ftree)

if (!cl_qmap_count(&p_ftree->cn_guid_tbl)) {
OSM_LOG(&p_ftree->p_osm->log, OSM_LOG_ERROR,
- "ERR AB23: "
+ "ERR AB27: "
"Com...

Jul 3, 9:18 am 2009
Hal Rosenstock
[ofa-general] [PATCH] opensm/osm_sa_(path multipath)_record....

Signed-off-by: Hal Rosenstock <hal.rosenstock@gmail.com>
---
diff --git a/opensm/opensm/osm_sa_multipath_record.c b/opensm/opensm/osm_sa_multipath_record.c
index 737d892..6ea5ba0 100644
--- a/opensm/opensm/osm_sa_multipath_record.c
+++ b/opensm/opensm/osm_sa_multipath_record.c
@@ -660,7 +660,7 @@ static ib_api_status_t mpr_rcv_get_path_parms(IN osm_sa_t * sa,
if (p_qos_level && p_qos_level->sl_set &&
p_qos_level->sl != required_sl) {
OSM_LOG(sa->p_log,...

Jul 2, 11:23 pm 2009
Max Lapan
[ofa-general] Problem with ib_dma_map_single

Hello,

currently, I'm trying to write small sample kernel module which
performs verbs communication in UD mode. Almost everything is
complete, but when I am trying to map allocated memory for DMA, I get
GPF:

general protection fault: 0000 [1] SMP
last sysfs file: /sys/devices/pci0000:80/0000:80:11.0/0000:86:01.1/irq
CPU 0
Modules linked in: verbs(N+) bridge stp netbk blkbk blktap xenbus_be
rdma_ucm(N) ib_ucm(N) ib_srp(N) scsi_transport_srp scsi_tgt ib_sdp(N)
rdma_cm(N) iw_cm(N) ib_addr(N) ib_i...

Jul 3, 7:47 am 2009
Vladimir Sokolovsky ...
[ofa-general] ofa_1_5_kernel 20090703-0200 daily build status

This email was generated automatically, please do not reply

git_url: git://git.openfabrics.org/ofed_1_5/linux-2.6.git
git_branch: ofed_kernel_1_5

Common build parameters:

Passed:

Failed:
Build failed on i686 with linux-2.6.19
Build failed on i686 with linux-2.6.21.1
Build failed on i686 with linux-2.6.18
Build failed on i686 with linux-2.6.22
Build failed on i686 with linux-2.6.24
Build failed on i686 with linux-2.6.26
Build failed on i686 with linux-2.6.27
Build failed on x86_64 wit...

Jul 3, 5:21 am 2009
Nicolas Morey-Chaise...
[ofa-general] [PATCH] opensm: Added support for select count...

Support for xmit_wait counters was missing in the perfmgr though it was read from the mad and event plugin interface already handles it.
This patch adds support for it (tested and working with an event plugin)

Signed-off-by: Nicolas Morey-Chaisemartin <nicolas@morey-chaisemartin.com>
---
opensm/include/opensm/osm_perfmgr_db.h | 23 ++++++-
opensm/opensm/osm_perfmgr.c | 30 +++++++-
opensm/opensm/osm_perfmgr_db.c | 124 +++++++++++++++++++++++++++++--
3 files changed, ...

Jul 3, 4:04 am 2009
Davis, Arlin R
RE: [ofa-general] cmatose fails whereas rping passes on iWarp

If this test sends data from server side first you could
be running into the iWARP requirement of sending from
client first.

-arlin
_______________________________________________
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Jul 2, 9:03 pm 2009
Sean Hefty
RE: [ofa-general] cmatose fails whereas rping passes on iWarp

This was my thought as well. I think Chelsio supports sending from the server
side first, but I'm not sure, or if it's enabled by default.

- Sean

_______________________________________________
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Jul 3, 2:44 am 2009
Davis, Arlin R
RE: [ofa-general] cmatose fails whereas rping passes on iWarp

I don't believe it is enabled by default. To enable:

echo 1 > /sys/module/iw_cxgb3/parameters/peer2peer

or

add following /etc/modprobe.conf driver option:

options iw_cxgb3 peer2peer=1_______________________________________________
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Jul 3, 3:03 pm 2009
previous daytodaynext day
July 2, 2009July 3, 2009July 4, 2009