Am Dienstag, den 18.07.2006, 10:09 -0700 schrieb Sean Hefty:
Arne Redlich wrote:
@@ -1354,6 +1354,7 @@ static int cm_req_handler(struct cm_work
id.local_id);
if (IS_ERR(cm_id_priv-timewait_info)) {
ret =
Am Dienstag, den 18.07.2006, 12:21 -0700 schrieb Sean Hefty:
Arne Redlich wrote:
Yep - the Gen1 SRP initiator does. It sends a REQ with an invalid DGID.
If rejected with the correct code (INVALID GID), it will retry after
looking up the GID.
Didn't it have a DGID from a path record that
Quoting r. Sean Hefty [EMAIL PROTECTED]:
Subject: [PATCH] ib_cm: fix REJ due to invalid GID
The following patch should set the reject code properly when
rejecting a request that contains an invalid GID. A suitable
GID is returned by the IB CM in the additional reject information
(ARI).
Am Dienstag, den 18.07.2006, 15:09 -0700 schrieb Sean Hefty:
The following patch should set the reject code properly when
rejecting a request that contains an invalid GID. A suitable
GID is returned by the IB CM in the additional reject information
(ARI).
Signed-off-by: Sean Hefty [EMAIL
Quoting r. Roland Dreier [EMAIL PROTECTED]:
Subject: Re: [PATCH 2/2] ib_cm: fix REJ due to invalid GID
Michael By the way, AFAIK by design cache might be out of sync
Michael with actual hardware. Roland, could you confirm this
Michael pls?
Yes, that's true. But even without
OpenSM: Support configurable SL per partition
The following patch series adds support for configurable SL per
partition in the partition manager as well as the underlying SA (path
and multipath) record support.
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
OpenSM: Support configurable SL per partition
Signed-off-by: Sasha Khapyorsky [EMAIL PROTECTED]
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: include/opensm/osm_partition.h
===
--- include/opensm/osm_partition.h
OpenSM/osm_sa_path_record.c: Support SL in PathRecord requests
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_sa_path_record.c
===
--- opensm/osm_sa_path_record.c (revision 8579)
+++ opensm/osm_sa_path_record.c
OpenSM/osm_sa_multipath_record.c: Support SL in MultiPathRecord requests
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_sa_multipath_record.c
===
--- opensm/osm_sa_multipath_record.c(revision 8574)
+++
Michael S. Tsirkin wrote:
Quoting r. Roland Dreier [EMAIL PROTECTED]:
Subject: Re: OFED 1.1 release - schedule and features
So if the link which ib0 maps to is DOWN you move the ib0 IPv4 address
to another device whose link is UP (eg ib1) and you somehow have ib1
send a gratuitous ARP?
OpenSM/osm_qos.c: In vl_high_limit_update, use
ib_port_info_set_state_no_change routine
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_qos.c
===
--- opensm/osm_qos.c(revision 8584)
+++ opensm/osm_qos.c
OpenSM/QoS: Better SLtoVLMappingTable determination
On a switch, if VLCap 1 indicating 1 data VL, determine whether
SLtoVLMappingTable exists from port 0's capability mask. If more than 1
data VL, SLtoVLMappingTable must be supported. For HCA and router ports,
just check the capability mask.
OpenSM: Simplify some endian conversions in osm_prtn.c and
osm_pkey_mgr.c
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_prtn.c
===
--- opensm/osm_prtn.c (revision 8586)
+++ opensm/osm_prtn.c (working copy)
@@
This patch looks good and worked well here for several days.
An earlier version of this patch (by Jack) has also shipped with OFED 1.0.
As was discussed earlier, I suggest queueing this for 2.6.18 - without this
patch, we observe that RMPP transactions are sometimes aborted because of
duplicate
Cool!
I'm busy now with other things, but it sounds like there's some bugs in
rping...
On Tue, 2006-07-18 at 17:30 -0700, Ravinandan Arakali wrote:
Steve/Pradipta,
Without the -O2 option, rping is now working !
Earlier, I did not realize that the cable was yanked out.
Thanks for all the
Roland, I have put the following patches in my mst-for-2.6.18 tree:
$git log --pretty=short origin..mst-for-2.6.18
commit e882c87183b09edaf827d887d7897d2f73212392
Author: Michael S. Tsirkin [EMAIL PROTECTED]
[PATCH] IB/ipoib: fix packet loss after hardware address update
commit
[PATCH] IB/ipoib: fix packet loss after hardware address update
Hm, OK I guess. Are we sure that the memcmp in the data path is OK?
[PATCH] IB/mad: validate MADs for spec compliance C13-18.1.1
Seems fine but I wonder why the ib_response_mad function moved into a
public header.
Also, how
OpenSM/osm_sa_service_record.c: Some trivial cleanups
Assert on attribute ID where most common
Also, eliminate unused variable in osm_sr_rcv_process
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_sa_service_record.c
Sure. Did you catch the P_Key issues using a test?
Eitan
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Roland Dreier wrote:
[PATCH] IB/mad: validate MADs for spec compliance C13-18.1.1
Seems fine but I wonder why the ib_response_mad function moved into a
public header.
This functionality was implemented / called from several places. Private
implementations have been replaced with a single
Michael S. Tsirkin wrote:
Should we be worried that this also sends IB_CM_REJ_INVALID_GID if
find_cached_pkey fails?
I'm not sure that an invalid pkey necessarily maps better to any other reject
code...
- Sean
___
openib-general mailing list
Arne Redlich wrote:
Won't this lead to a second REJ (IB_CM_REJ_CONSUMER_DEFINED) being
issued by cm_destroy_id()? That's actually what I tried to prevent using
the reject param.
Hmm... it does. I missed that in my traces.
- Sean
___
openib-general
Quoting r. Arne Redlich [EMAIL PROTECTED]:
Subject: Re: [openib-general] [PATCH 2/2] ib_cm: fix REJ due to invalid GID
Am Dienstag, den 18.07.2006, 12:21 -0700 schrieb Sean Hefty:
Arne Redlich wrote:
Yep - the Gen1 SRP initiator does. It sends a REQ with an invalid DGID.
If rejected
I see this messages while booting my system:
Jul 17 10:35:37 alekhine kernel: ib_mthca :03:00.0: HCA FW version
4.6.0 is old (4.7.0 is current).
Jul 17 10:35:37 alekhine kernel: ib_mthca :03:00.0: If you have
problems, try updating your HCA FW.
where can I find the latest firmware for
only documentation I find on this API is:
/**
* ib_get_dma_mr - Returns a memory region for system memory that is
* usable for DMA.
* @pd: The protection domain associated with the memory region.
* @mr_access_flags: Specifies the memory access rights.
*/
I see there is no address
Arlin Davis wrote:
Fix for Bug 158. Add support for dat_evd_set_unwaitable on a DTO EVD.
Committed revision 8592.
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe,
Quoting r. Roland Dreier [EMAIL PROTECTED]:
Subject: Re: [openib-general] 2 patches in mst-for-2.6.18
[PATCH] IB/ipoib: fix packet loss after hardware address update
Hm, OK I guess. Are we sure that the memcmp in the data path is OK?
I did some measurements with netperf and did not
Hi,
We are using kernel 2.6.16.3 and installed the latest code/drivers from
the iwarp branch. Currently, we are seeing an address resolution problem
while using rping.
[EMAIL PROTECTED] examples]$ ./rping -c -VvC100 -a 150.10.11.200 -p
libibverbs: Warning: couldn't load driver
Hi Hal,
The minor comment is below.
On 07:11 Fri 14 Jul , Hal Rosenstock wrote:
OpenSM: Support configurable SL per partition
Signed-off-by: Sasha Khapyorsky [EMAIL PROTECTED]
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: include/opensm/osm_partition.h
Hi Sasha,
On Wed, 2006-07-19 at 13:40, Sasha Khapyorsky wrote:
Hi Hal,
The minor comment is below.
[snip...]
Index: opensm/osm_prtn_config.c
===
--- opensm/osm_prtn_config.c(revision 8520)
+++
What does your network configuration look like? IE send ifconfig output
on both systems.
Send the output of lsmod on both systems.
Does krping work?
Can you ssh between the 2 systems using the T3 ip addresses?
Do the following libs exist in /usr/local/lib:
libibverbs.*
Ravinandan Arakali wrote:
Steve/Pradipta,
Without the -O2 option, rping is now working !
Good to know that!! But still we need to find out the root cause of the problem.
Thanks,
Pradipta
Earlier, I did not realize that the cable was yanked out.
Thanks for all the help..
Ravi
On Wed, 2006-07-19 at 12:25, Eitan Zahavi wrote:
Sure. Did you catch the P_Key issues using a test?
No; those part of the changes were to make the multipath record code
look more like the path record code.
-- Hal
Eitan
___
openib-general
Let's try this again... Once we're past cm_match_req(), we can just
call into the main API to reject the connection, which will set the
state of the cm_id correctly. I looked at changing cm_destroy_id(),
but didn't like the idea of trying to have the destroy call set the
ARI data.
- Sean
---
OpenSM/osm_helper.c: Support InformInfoRecord attribute ID decode
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_helper.c
===
--- opensm/osm_helper.c (revision 8596)
+++ opensm/osm_helper.c (working copy)
@@
OpenSM/osm_sa_informinfo.c: In osm_infr_rcv_process_set_method, if
sending error, don't call osm_infr_rcv_process_set_method
Also, fix typo in osm_log message
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_sa_informinfo.c
OpenSM/ib_types.h: Pad ib_inform_info_record_t to be modulo 8 in size so
that attribute offset is calculated properly
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: include/iba/ib_types.h
===
--- include/iba/ib_types.h
OpenSM/osm_sa_informinfo.c: Remove unneeded line of code in
__osm_infr_rcv_respond
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_sa_informinfo.c
===
--- opensm/osm_sa_informinfo.c (revision 8606)
+++
OpenSM/osm_sa_class_port_info.c: Eliminate unneeded line of code in
__osm_cpi_rcv_respond
Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
Index: opensm/osm_sa_class_port_info.c
===
--- opensm/osm_sa_class_port_info.c (revision
Thanks to those who sent suggestions for topics,
and to Bill Boas, Matt Leininger, and Helen Chen for helping come up
with this draft list.
At this point not considering scheduling or presenters,
here's the tentative list of topics, in no particular order, for the
next OFA Developers'
40 matches
Mail list logo
