Michael S. Tsirkin wrote:
Erez, add_open_iscsi_h currently does:
-#include
+#include "iscsi_if.h"
why is ths bit needed?
Michael, Erez
This is strict OFED issue, this cross-posting to both ewg and the
general list is annoying and inefficient, please handle this at the ewg
list.
Or.
_
Tziporet Koren wrote:
1.2.c-10 supports FMRs and it will be available tomorrow
thats very important progress re iser support since unlike srp we can't
work without fmr. When are you planning to send the mlx4 FMR code to
review on the general list? I guess this is code candidate for 2.6.24,
c
Tziporet Koren wrote:
Main changes from OFED 1.2.c-10:
- Merged all 1.2 branch patches to 1.2.c
- mlx4
- Added support for interrupt moderation for create cq to support ipoib.
- Display the following device information via sysfs: board_id, fw_ver,
hw_rev and hca_type
Tziporet Koren wrote:
The patches that does not include sign-off and explanation were posted
by Eli, and he forgot this by mistake - to be fixed.
All the changes are going to be send for the list for review next week.
thanks for the clarification, however, I don't see a reason not to post
a p
Michael S. Tsirkin wrote:
In 1.2.c most patches are backported from 2.6.23.
Some have been submitted for 2.6.24.
Some have been carried from 1.2.
I see mistakes in the process here: for example patches that solve IB
bugs in 2.6.22 should have been submitted to -stable and this ofed
version ne
Tziporet Koren wrote:
OFED Aug 13 meeting summary (short meeting):
Hi Tziporet,
1. OFED 1.2.5 (was 1.2.c) is ready for release:
Can Michael and/or Eli enhance the ipoib patches to include change log
comments and signature?
Can the official policy of OFED be changed such that --only-- pat
Tziporet Koren wrote:
OFED Sep 10 meeting summary on OFED 1.3 development status
Meeting summary:
1. We reviewed OFED 1.3 features status:
IPoIB: stateless offloads - 90%
Tziporet, Michael
Based on the fact that OFED 1.3 is now based on 2.6.23-rc5, it simply
does not make --any-- sense to h
issues
which I work on now and expect to be able to deliver in the coming days.
Please let me know if you see any problem with that,
thanks,
Or.
On Thu, 20 Sep 2007, Or Gerlitz wrote:
> changes from v1 -
> http://lists.openfabrics.org/pipermail/general/2007-September/040250.html
>
Tziporet Koren wrote:
Or Gerlitz wrote:
As Voltaire communicated to the ewg in the past, we want this patch to be
in OFED 1.3 even if it misses the upstream accepatance for 2.6.24
The patch is working fine for 2.6.23-rc5 and I have a working version
also
for OFED 1.2. However, under OFED 1.3
-off-by: Or Gerlitz <[EMAIL PROTECTED]>
Index: ofa_kernel-1.3-work/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
===
--- ofa_kernel-1.3-work.orig/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
2007-09-25 19:49:33.0
kernels
Signed-off-by: Or Gerlitz <[EMAIL PROTECTED]>
Index: ofa_kernel-1.3-work/drivers/infiniband/ulp/ipoib/ipoib_main.c
===
--- ofa_kernel-1.3-work.orig/drivers/infiniband/ulp/ipoib/ipoib_main.c
2007-09-25 19:54:15.0
Michael S. Tsirkin wrote:
Quoting Or Gerlitz <[EMAIL PROTECTED]>:
Subject: [for OFED 1.3 PATCH 2/2] IB/ipoib: enable IGMP for userpsace multicast
IB apps
Michael,
This patch needs to go to all the directories under kernel_patches/backport
that contain the
ipoib_class_device_to_2_6_20
Hi Vlad,
Opening a bug on the ofa bugzilla whose component is RDS results in
[EMAIL PROTECTED] being assigned to it ...
I understand it should be assigned to you? if yes, can you fix it?
For the mean time I have re-assigned the two RDS problems reported by
Voltaire (https://bugs.openfabrics.
Tziporet Koren wrote:
OFED October 8 meeting summary on OFED 1.3 alpha readiness
Meeting summary:
1. Alpha release is planed for this week (Wed or Thursday)
2. Requests for the beta release
Hi Tziporet,
As of the centrality and importance of IPoIB, my take is that at this
point t
Tziporet Koren wrote:
OFED October 8 meeting summary on OFED 1.3 alpha readiness
3. We discussed some ideas for talks in the developer's summit. The
following ideas were raised: sa caching (Intel), QoS support (Sean),
Extended RC (MPI team)
We can discuss the connectx ipoib stateless offload
Eli Cohen wrote:
As for the interrupt mitigation patches, I have sent some comments and
Eli only replied on some of them, nothing was changed or fixed yet, and
other then this, no review has been done.
I am not aware that you have more questions about interrupt mitigation.
We even discussed
On 10/24/07, Johann George <[EMAIL PROTECTED]> wrote:
>
> Thanks for your feedback. We can certainly expand, shorten and change
> sessions around to maximize benefit for the attendees.
>
OK, so lets do it!
In short, the agenda as it is now, has a great bias to "updates and
reporting" && windowz-
On 10/10/07, Or Gerlitz <[EMAIL PROTECTED]> wrote:
>
> Hi Vlad,
>
> Opening a bug on the ofa bugzilla whose component is RDS results in
> [EMAIL PROTECTED] being assigned to it ...
>
> I understand it should be assigned to you? if yes, can you fix it?
>
> For the
(Assuming that the allocation of slots within the schedule to have
enough time for Linux IB developers to discuss what ever they decide
they need to would be taken care of) I'd like to check with people what
we want to be on the agenda of these slots. My thinking for issues to
discuss was:
1)
On 10/29/07, Or Gerlitz <[EMAIL PROTECTED]> wrote:
>
> (Assuming that the allocation of slots within the schedule to have
> enough time for Linux IB developers to discuss what ever they decide
> they need to would be taken care of) I'd like to check with people what
> w
On 10/29/07, Sean Hefty <[EMAIL PROTECTED]> wrote:
>
> > 1) the long time and endless threads related to the SA caching thing
> > need to be there. Sean - I saw that you prepare a session, correct? will
> > you presenting few possible designs?
>
> I was asked to prepare a session and will mention s
On 10/29/07, Johann George <[EMAIL PROTECTED]> wrote:
>
> (2) If we are able to, would you prefer to see simultaneous
> tracks and lengthen some of the sessions.
It makes some sense to make a poll if people prefer simultaneous tracks or
not, however if the answer is "no", still you can't allo
Sean Hefty wrote:
7) the inform info code. Sean - you have implemented and attempted to
push it through the sa caching push, but since the cache was rejected
so did the inform info code. So the questions here - how do we make
this push happen? are there any open issues, etc
There either needs
Dror Goldenberg wrote:
2) as for IPoIB stateless offload - with Eli and Liran not planned to
be there. Dror - do you intend to actually present the actual ipoib /
core / drivers related design and implementation? Also, personally, I
felt that the 1-2 slides you delivered on Sonoma where way bel
Dror Goldenberg wrote:
Or Gerlitz wrote:
I haven't yet prepared the presentation. I am willing to cover whatever
you think is important. Indeed 20m allotted time is too short. So, I
should either adjust myself to this short time-slot or ask for more.
Given that the other sessions are
On 10/30/07, Roland Dreier <[EMAIL PROTECTED]> wrote:
>
> So with that said:
>
> > 1) the long time and endless threads related to the SA caching thing
> > need to be there. Sean - I saw that you prepare a session, correct?
> > will you presenting few possible designs?
>
> This is the perfect type
Hi Jack,
With OFED 1.3, the printk below in core/verbs.c is present in
my dmesg each time a qp is destroyed, can you please remove it?
Or.
if (qp_type == IB_QPT_XRC)
atomic_dec(&xrcd->usecnt);
else
printk("ib_destroy_qp: type = %d, xrcd = %p\n", qp
Johann George wrote:
There was mixed reaction to coming earlier on Thursday or
staying later on Friday. If a particular group is available
and would like to hold a discussion during either of these
times, let me know and we'll try to make the room available.
I suggest that the Linux IB develop
Appliance
> 18:20 20m Lustre
> Eric Barton, Sun Microsystems
> 18:40 20m Bonding
> Or Gerlitz, Voltaire
>
Johann, I don't need 20 minutes for the bonding update, 10m will be more
then enough.
Or.
On 11/5/07, Roland Dreier <[EMAIL PROTECTED]> wrote:
>
> > 11-12: SA cache session
> > 12-1: IPoIB stateless offload issues
>
> > Sean, Roland, Dror - can you make it?
>
> I guess I can make it but what are the stateless offload issues? If
> we want to spend an hour on it I think we need to figure
Dror Goldenberg wrote:
a) it's an option that is turned off by default and people that are
willing to take the risk can turn it on and get better performance
b) indeed we will need to make it work with the kernel lso, no need to
have something specific for ipoib
Anyway, I am planning to presen
Roland Dreier wrote:
> For (B): the LRO implementation posted is exactly using the generic
> network stack software LRO helpers!
Sorry, I take that back. I just checked again, and the patch that was
posted just used the same function names as the upstream LRO generic
helpers, which is what co
Roland Dreier wrote:
As I said earlier on this thread, the open issues I see with the stateless
offload series are (A) the non interoperable checksum offload patch based on
the IB ICRC sent by Michael (and if it is inter-operable, I'd like to be
educated how)
For (A): as far as I'm concerned
On 11/6/07, Sean Hefty <[EMAIL PROTECTED]> wrote:
>
> > 11-12: SA cache session
> > 12-1: IPoIB stateless offload issues
> >
> > Sean, Roland, Dror - can you make it?
>
> I should be able to make this, but as soon as you start pushing sessions
> before noon, time should probably be made for lunch.
Roland Dreier wrote:
I would suggest trying to figure out how to make OFED 1.4 the last
OFED release. OFED was originally supposed to be a short-term
distribution until mainstream Linux distributions caught up and were
able to distribute IB/iWARP support. So we should assess how close we
are to
Johann George wrote:
I should be able to make this, but as soon as you start pushing sessions
before noon, time should probably be made for lunch.
Perhaps we can include a working lunch for the developers who want to
and are able to meet earlier and discuss these issues?
We can, but I prefer
Hi Jim,
I was handed a benchmark results from which I suspected that the SDP
code that was used might not respect the TCP_NODELAY option, so I wanted
to clarify this with you. I think it was OFED 1.2 based. So is it
supported in OFED 1.2? in OFED 1.3?
thanks,
Or.
__
Jim Mott wrote:
SDP includes support for TCP_NODELAY and TCP_CORK. At a high level,
these options control buffering and they seem to be working. If you
turn off the nagle buffering algorithm using these options, then sends
go right away. This is how TCP_RR testing with netperf run to measur
Tziporet Koren wrote:
OFED-1.2.5.3 is ready:
Changes since OFED 1.2.5
...
- IPOIB CM - tx use unsignalled QP
...
Changes from OFED 1.2.5.2:
==
- IPoIB: Fix kernel oops in Connected mode
...
Looking on the 1_2_c branch of
http://www.openfabr
Tziporet Koren wrote:
Or Gerlitz wrote:
In one of our tests that is doing up/down of the driver we see some
deadlock when this patch is used, and it does not happened when its
removed.
Thus we decided to remove this patch till we get to the root cause of
the problem
Note that we do not think
Tziporet Koren wrote:
Or Gerlitz wrote:
I understand that this patch does not exist in OFED 1.3, correct?
correct - but don't worry for the bugs we see new oops even with IPoIB
vanilla from 2.6.24-rc3 :-(
I see, how about sending the oop trace to the general list
Hi Jim,
Looking on the sdp module info, I see that some of the module params are
used to --report-- on this or that counter value (poll miss/"help"/etc).
Can you change the code such that reporting of counter values be
carried out in more standard (sysfs, netlink, etc) fashion?
Also the only mai
Hi Eli,
The patch below in OFED 1.3 does not check the device capabilities and hence
always fail on non connectx systems. Can you fix it such that we will not
get all those "why I we mthca0: failed to modify CQ params prints in the logs".
Other then that and maybe even more important... I underst
Eli Cohen wrote:
One more thing regarding latency - you can use ethtool to modify these
parameters if your application is latency sensitive.
Indeed, but I want us to think together what should be the default
value, biased towards latency or towards bandwidth.
Or.
___
test please ignore
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Eli Cohen wrote:
On Sun, 2008-01-06 at 10:47 +0200, Or Gerlitz wrote:
The patch below in OFED 1.3 does not check the device capabilities and hence
always fail on non connectx systems. Can you fix it such that we will not
get all those "why I we mthca0: failed to modify CQ params prints i
For some reason this message was not delivered to the list
Eli Cohen wrote:
On Sun, 2008-01-06 at 10:47 +0200, Or Gerlitz wrote:
The patch below in OFED 1.3 does not check the device capabilities and hence
always fail on non connectx systems. Can you fix it such that we will not
get all
OK, I see 10us difference in a udp ping pong latency test between the
case where ethtool sets interrupt moderation to (10,16) vs (10,1)
$ ethtool -C ib1 adaptive-rx on rx-usecs 10 rx-frames 16 --> 9us
$ ethtool -C ib1 adaptive-rx on rx-usecs 10 rx-frames 1 --> 19us
Or.
_
For some reason, the message below was not delivered to the list
Or Gerlitz wrote:
Eli Cohen wrote:
On Sun, 2008-01-06 at 10:47 +0200, Or Gerlitz wrote:
The patch below in OFED 1.3 does not check the device capabilities and hence
always fail on non connectx systems. Can you fix it such that
Tziporet Koren wrote:
We wish to test in this way and also add this explanation to OFED docs
sure, its all documented in the ib-bonding.txt file that comes with the
ib-bonding package, so I suggest just to add a note in the docs pointing
to it.
Or.
# rpm -ql ib-bonding
/etc/sysconfig/net
Johann George wrote:
Oh, and if you're using RDMA - does this happen to be with
qlogic HCAs? If so, I just received a patch from Ralph
Campbell with some fixes to the way we set up out DMA
mapping.
RDS in OFED 1.3 does not currently work on the QLogic HCAs
due to the way you are setting up DMA
Tziporet,
The rdma-cm tavor quirk as it now in ofed 1.3
(kernel_patches/fixes/cma_0030_tavor_quirk.patch)
is useless, I suggest to remove it, unless Moni says he needs it and provides a
fix.
Or.
Tavor systems get better performance with 1K MTU. Since there does
not seem to be any way to
Ralph Campbell wrote:
Attached is the patch I sent to Olaf.
It basically exchanges calls like dma_map_sg() to ib_dma_map_sg()
so that the InfiniPath driver can intercept the DMA mapping
calls and use kernel virtual addresses instead of physical addresses.
The InfiniPath driver uses the host CPU t
Eli Cohen wrote:
Support modifying IPOIB CQ moderation params
This can be used to tune at run time the paramters controlling
the event (interrupt) generation rate and thus reduce the overhead
incurred by handling interrupts resulting in better throughput.
Eli,
I did some testing with the copy
Or Gerlitz wrote:
Index: linux-2.6.18-rc2-devel/drivers/infiniband/core/cma.c
@@ -1123,6 +1127,11 @@ static int cma_query_ib_route(struct rdm
path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(addr));
path_rec.numb_path = 1;
+ if (tavor_quirk
Eli Cohen wrote:
On Sun, 2008-01-20 at 14:33 +0200, Or Gerlitz wrote:
I did not use ethtool through these scripts. Did it work for you on
other distributions or have you not tried that? And did actual
configuration took place (did rx-usecs become 2 and rx-frames 3)?
I did not try other
Tziporet Koren wrote:
Release information:
Specific module changes:
Utilities:
Performance tests:
* Added multicast support to ib_send_bw and ib_send_lat tests
Tziporet,
Looking on the commit, it seems that the perftest tools were
instrumented to send/receive multicast on a hard coded MGID/
Hoang-Nam Nguyen wrote:
Hello Tziporet and Vlad!
This is a set of upstream patches, which we sent previously to Roland and
have been queued by him for 2.6.25, and backport patches for rhel4.5 and
rhel4.6 due to the introduced changes.
1/3: upstream patches, see its heading for more detail
2/3:
Steve Wise wrote:
Vlad,
Please pull this patch from
git://www.openfabrics.org/~swise/ofed-1.3 ofed_kernel
This has been accepted upstream and is needed for ofed-1.3 to support
new device types.
Hi Steve,
can you please send such emails only to ewg? else we see each of your
patches twice on t
Jonathan L. Perkins wrote:
I've uploaded a new SRPM for MVAPICH2 to the openfabrics
server. This is located in ~perkinjo/ofed_1_3/ and is identified by the
latest.txt file.
This SRPM includes a fix that should solve the problem reported by Scott
Weitzenkamp as well as some other improvements.
Tziporet Koren wrote:
Or Gerlitz wrote:
Jonathan, can you please exclude the general list from such postings?
it has nothing to do with that list.
I do not agree with you
Not everybody is registered to ewg but many people wish to know on a new
MPI package
So they should subscribe to the
Stijn De Smet wrote:
I'm trying to get IPOIB bonding to work with the hw_csum enabled.
...
When I disable hw_csums, I can start iperf's, pull and replug all cables
and the iperf's run uninterrupted.
This is interesting report, however, since currently the hw checksum
patch in not being submi
Koen Segers wrote:
Do you mean that bonding with hw_csum enabled will never work?
no, I meant to say that I am not enough into the details and mechanics
of the hw_csum approach/patch and since I understand it is going to be
removed, I will not look now on going into this report.
Why is hw_
On 1/30/08, Koen Segers <[EMAIL PROTECTED]> wrote:
> On Wed, 2008-01-30 at 16:26 +0200, Or Gerlitz wrote:
>> you should ask that the individual/s that are signed on the patch
> Is this Michael S. Tsirkin?
> I don't know where else to find this information.
In the Lin
This patch was tested against upstream kernel and ofed_kernel.
+
+Signed-off-by: Or Gerlitz <[EMAIL PROTECTED]>
+Signed-off-by: Moni Shoua <[EMAIL PROTECTED]>
+
Index: ofa_kernel-1.3/drivers/infiniband/ulp/
On 1/31/08, Tziporet Koren <[EMAIL PROTECTED]> wrote:
> Due to feedback we got regarding Redhat inclusion of OFED 1.3 in RHEL 5.2
> and 4.7 we are changing the release schedule
> RC4 - Feb 6
> RC5 - Feb 18 <== Gold
> GA - Feb 25
> RC4 should include:
> Critical and major bug fixes
does this mean
On 1/31/08, Oren Meron <[EMAIL PROTECTED]> wrote:
> perftest: cancel inline default for Hermon to optimize BW performance in send
> and write
Can you please explain what was the problem and what is the fix?
Or.
___
ewg mailing list
ewg@lists.openfabric
Eli Cohen wrote:
On Thu, 2008-01-31 at 21:57 -0800, Shirley Ma wrote:
On Fri, 2008-02-01 at 15:40 +0200, Eli Cohen wrote:
On the receive side the code is changed to post to receive queue once in
16 completions. This is done in for both UD and and CM.
Ohmm, have you tested latency? I think
> commit f17ebf3e2099257da244587f1ee33f51745f7cdb
> Author: Eli Cohen <[EMAIL PROTECTED]>
> Date: Tue Feb 5 11:15:46 2008 +0200
>
> Call skb_orphan() after sending an SKB
>
> This will call the destructor of the SKB (but not free the
> memory). It appears that some applications (ttcpv
your thinking,
thanks,
Or.
On Wed, 6 Feb 2008, Or Gerlitz wrote:
> --- /dev/null 2008-02-05 10:18:44.755516936 +0200
> +++ ofed_scripts/openibd 2008-02-06 13:46:50.0 +0200
> @@ -0,0 +1,1375 @@
> +#!/bin/bash
> +
> +#
> +# Copyright (c) 2006 Mellanox Technol
Eli,
You have left somehow too many... debug prints in the last patches,
please clean this up. See for example how the system log after less
then a minute when ipoib debug prints are opened, it has one original
print ("ib0: Send unicast ARP to 0023") and all the rest are yours.
Or
Feb 6 14:39:2
I just opened case #897 on the below, it happens with last night snapshot.
Or
client MT25204 FW 1.2.0 two CPUs, four cores each
server MT25418 FW 2.3.0 two CPUs, four cores each
client : iperf -c $server -P 4 -d -t 3600 -i 1
server : iperf -s -i 1
[ 5] 39.0-40.0 sec 29.4 MBytes246 Mbits/s
> ib_mthca :03:00.0: SQ 000404 full (756910656 head, 756910592 tail, 64
> max, 0 nreq)
> ib0: failed to post zlen send
Eli,
can this be a bug in the send ring accounting wrt to the zlen packet you use in
the unsig-ud-qp patch?
Or.
___
ewg mailing
Eli Cohen wrote:
On Wed, 2008-02-06 at 10:17 +0200, Or Gerlitz wrote:
The problem with ttcpv was that it stopped sending packets since it was
waiting for freeing the memory. The system did not hang, just the
application (ttcpv) stopped sending. Other applications could continue
working over
eld, however, your 3 prints per second addition makes them useless,
at least for me, and I use them a lot where working to debug and help
others, so please do.
Or
>
> On Wed, 2008-02-06 at 10:38 +0200, Or Gerlitz wrote:
> > Eli,
> >
> > You have left somehow too many
>> ib_mthca :03:00.0: SQ 000404 full (756910656 head, 756910592 tail, 64
>> max, 0 nreq)
>> ib0: failed to post zlen send
OK, Eli, taking the kernel bits from OFED-1.3-20080206-0751.tgz I don't
see these prints any more. When probing out the driver inorder to replace
it with the drop, I have
Or Gerlitz wrote:
You have left somehow too many... debug prints in the last patches,
please clean this up. See for example how the system log after less
then a minute when ipoib debug prints are opened, it has one original
print ("ib0: Send unicast ARP to 0023") and all the rest
Eli Cohen wrote:
On Thu, 2008-02-07 at 09:42 +0200, Or Gerlitz wrote:
ib_mthca :03:00.0: SQ 000404 full (756910656 head, 756910592 tail, 64 max,
0 nreq)
ib0: failed to post zlen send
OK, Eli, taking the kernel bits from OFED-1.3-20080206-0751.tgz I don't
see these prints any more.
Hoang-Nam Nguyen wrote:
Hello Eli!
Looked at ipoib code from ofed-1.3-rc4 and the saw the following code snippet
in ipoib_ib_post_receive():
Hi Nam,
Again,
Can you exclude the general list from such postings? in this example you
are treating below patches which are not candidates for upstrea
Ralph Campbell wrote:
Here is a suggested patch for ib_ipoib for fixing the OFED-1.3 RC4
problem with ib_ipath.
Note that I'm not completely familiar with all the ib_ipoib changes.
In particular, I haven't checked that IB_EVENT_LID_CHANGE correctly
updates priv->own_ah.
Good catch, Ralph.
Eli
Hi Eli,
Once in a while when probing out the ipoib driver, I see these prints:
ib0: timing out; will leak address handles
ib0: ib_dealloc_pd failed
is it a known issue? I see that also with RC4
Or.
___
ewg mailing list
ewg@lists.openf
Eli Cohen wrote:
Thanks Or - I will try to reproduce. Were you able to unload the module
eventually?
yes
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
On Mon, 11 Feb 2008, Or Gerlitz wrote:
> Once in a while when probing out the ipoib driver, I see these prints:
> ib0: timing out; will leak address handles
> ib0: ib_dealloc_pd failed
OK, I saw it this morning under RC4 / MT25204 / FW 1.2.0 which had iperf server
running
Eli Cohen wrote:
could you send as distinct patches according to what they fix?
Pradeep Satyanarayana wrote:
2. Change retry counts to small values. This helps interoperability
between ehca and mthca.
Indeed, I sent a note on that now to the general list, lets discuss it
there since its an
Matthew Small wrote:
I am trying to learn how to code in ibverbs and I am having trouble
finding any information on the semantics and structure of the API
(besides the formal implementers specification). If anyone could
suggest some documentation and/or learning aids that could be useful it
w
Did anyone at Cisco, Qlogic, Mellanox, Voltaire noticed the phenomena
reported by Shirley on their testing?
Or.
--- Begin Message ---
On Tue, 2008-02-12 at 09:00 -0800, Sean Hefty wrote:
> >Saying all that, I don't think we want to have --any RNR retries--, as
> >for retries, I am open to hear w
Pradeep Satyanarayana wrote:
This patch fixes -fail to destroy ipoib rx QP
(https://bugs.openfabrics.org/show_bug.cgi?id=906)
Hence the usecnt issue reported previously on ehca is solved and allows the qp
to be destroyed.
As per Eli's request, I am splitting up the patches. This is first porti
Stefan Roscher wrote:
yes this problem does also exist in 2.6.25-rc1. It was introduced by a patch
from roland:
http://git.kernel.org/?p=linux/kernel/git/roland/infiniband.git;a=commitdiff;h=efcd99717f76c6d19dd81203c60fe198480de522
In function ipoib_cm_dev_stop() the error-,drain- and flush lis
On 2/13/08, Matthew Small <[EMAIL PROTECTED]> wrote:
> Wow, thanks a lot for that reference. I was actually wondering how to set
> up a connection without using the ad hoc socket-exchange method in pingpong.
> You saved me some time. Do you know of any good source examples
> implementing a simpl
On 2/12/08, Shirley Ma <[EMAIL PROTECTED]> wrote:
Hello, where is the change-log comment that explains what the patch is
doing? is it relevant to 2.6.25-rc1? if yes, please send it to the
general list for review.
> Signed-off-by: Shirley Ma <[EMAIL PROTECTED]>
> ---
>
> diff -urpN ofed/drivers/in
Ralph Campbell wrote:
Yes. I see RNR NAKs with IPoIB connected mode quite often.
What is the setup one needs to make it happen? is it only when each node
has different HCA or it happens also on a homogeneous setup?
Or
___
ewg mailing list
ewg@
Tziporet Koren wrote:
We will take this one
Can you please explain what was the problem and what was the fix? I
can't learn that from the patch change-log nor from the thread.
Or
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabr
Eli Cohen wrote:
We took Pradeep's patch which moves the CM rx objects from the
error, flush and drain lists and puts them in the reap list. Putting
them in the reap list will cause the corresponding QPs to be destroyed.
Thanks, this sheds more light on the solution but I still can not
underst
Roland Dreier wrote:
OK, applied, although that link points to an HTML-mangled version of
the patch, and I also had to figure out why we needed that change and
write the patch description myself.
Eli, can you add the upstream change-log description (commit
727f89d27d623f87a637c1868dc8b6af7103f
Tziporet Koren wrote:
OFED meeting summary:
=
2. OFED 1.4:
- Next meeting will be devoted to define the features set
Before going into the features of the next release, I think we want to
hear each of the companies being signed on patches which were not pushed
upstream am
Tziporet Koren wrote:
Or Gerlitz wrote:
Before going into the features of the next release, I think we want
to hear each of the companies being signed on patches which were not
pushed upstream among the ones present under kernel_patches/fixes
(below), stating its plan for pushing them to the
Steve Wise wrote:
Tziporet Koren wrote:
* OFED 1.4: *
1. Kernel base: since we target 1.4 release to Sep we target the
kernel base to be 2.6.27
This is a good target, but we may need to stay with 2.6.26 if the
kernel progress will not be aligned.
2. Suggestions for new features:
* V
On Thu, Apr 3, 2008 at 5:40 PM, Tang, Changqing <[EMAIL PROTECTED]> wrote:
> The problem is, from MPI side, (and by default), we don't know which port is
> on which
> fabric, since the subnet prefix is the same. We rely on system admin to
> config two
> different subnet prefixes for HP-MPI to
On Thu, Apr 3, 2008 at 5:53 PM, Tang, Changqing <[EMAIL PROTECTED]> wrote:
> for example, in MPI, process A know the HCA guid on another node. After
> running for
> some time, the switch is restarted for some reason, and the whole fabric is
> re-configured.
CQ,
If by "the whole fabric is re-
https://bugs.openfabrics.org/show_bug.cgi?id=989
--- Comment #13 from [EMAIL PROTECTED] 2008-04-23 13:23 ---
I think I found the problem and have a fix.
In ipoib_cm_handle_rx_wc(), if the byte_len is < SKB_TSHOLD, a new sk_buff
is allocated. The sk_buff's mac.raw is not initialized.
101 - 200 of 233 matches
Mail list logo
