Passed on ia64 with linux-2.6.18
Passed on powerpc with linux-2.6.14
Passed on ia64 with linux-2.6.17
Passed on ppc64 with linux-2.6.15
Passed on ia64 with linux-2.6.16
Failed:
Build failed on ia64 with linux-2.6.16.21-0.8-default
Log:
/home/vlad/tmp/ofa_1_2_kernel-20070208-0200_linux-2.6.16.21-0.8
Hi
I've been running netpipe over Infiniband's SDP and uncovered a race
when using the -r option.
The problem is when both sides close their sockets, the listening socket
is closed last, which allows a faster
client to try to connect to it before it closes. When this happens, next
time the client
OpenSM/osm_ucast_lash.c: In osm_get_lash_sl, fix SL when CA ports on same switch
This change resolves an issue with strange SL assignment when
two HCAs communicate with other and are on the same switch.
Since LASH is switch to switch routing, the get_lash_sl
function was casting (the
Or Gerlitz wrote:
Sean Hefty wrote:
Sean Hefty (3):
rdma_cm: Increment port number after close to avoid re-use.
ib_sa: track multicast join/leave requests
rdma_cm: add multicast communication support
Assuming that you haven't look at this yet, I updated the ib_sa patch
Quoting Jeff Squyres [EMAIL PROTECTED]:
Subject: Re: Open MPI rpmbuild fails in OFED-1.2
On Feb 7, 2007, at 2:49 PM, Michael S. Tsirkin wrote:
My $0.02: This is another in a growing list of issues reflecting the
whole build everything in DESTDIR is a problematic approach.
I don't
On Thu, 2007-02-08 at 08:40 +0200, Michael S. Tsirkin wrote:
OK, so I looked at cxgb3 some more.
Thanks!
To summarise my previous comments, I think the cxio hal layer needs to go to
make the code readable - if I understand correctly it is there for historical
reasons only.
I can do this
Michael S. Tsirkin wrote:
Quoting Moni Shoua [EMAIL PROTECTED]:
Subject: Re: [PATCH] IB/ipoib get net_device from ipoib_neigh instead of
linux neighbour
Another concern: assume that one device goes away (e.g. hotplug).
It seems that neighbours whose dev field point to another device, will
[EMAIL PROTECTED] wrote on Wed, 07 Feb 2007 15:50 -0800:
Pete Before I dig into this anymore, do you expect this to work?
Pete Are there fundamental problems with QP sharing between user
Pete and kernel? It would sure be nice not to have to stick the
Pete connection management
The following lightly tested patch addresses Roland's comments on IPoIB CM.
Applies on top of PATCHv6:
- Randomise RQ PSN
- Fix for modular IPv6
- MTU endian-ness fix for ICMPs
- Cosmetics
Signed-off-by: Michael S. Tsirkin [EMAIL PROTECTED]
---
Roland, do you want me to report the full
On Thu, 2007-02-08 at 10:24 -0500, Pete Wyckoff wrote:
[EMAIL PROTECTED] wrote on Wed, 07 Feb 2007 15:50 -0800:
Pete Before I dig into this anymore, do you expect this to work?
Pete Are there fundamental problems with QP sharing between user
Pete and kernel? It would sure be
- It seems that by passing in huge resource sizes, userspace will be able to
drink up unlimited amounts of kernel memory.
mthca handles this by using the mlock rlimit, should something be done
here
as well?
Hmm. That's a good point. I'll put this on the todo as well. So
That is correct.
I am working with Krishna on it.
Expect patches soon.
By the way the problem is not DAPL specific
and so is a proposed solution.
There are 3 aspects of the solution.
One is APIs. We suggest that we do not augment these.
That is a connection requestor sets its QP
RDMA ORD and
- Consider a user that does e.g. create QP, but never calls mmap.
Is there some code that will clean out the unclamed mmap object?
I couldn't find it, and iwch_dealloc_ucontext does not seem to
do anything with it.
BTW: Here is my fix for this.
-
Clean up pending mmaps on ucontext
The windows open IB has decided on using a BSD only license.
The common implementation of pthreads as far as I know is LGPL, which
means that it can not be used in open IB.
The only two ways that I see around this are 1) Change the license of
open IB windows which might be a complicated thing.
On Thu, Feb 08, 2007 at 10:23:11AM -0800, Sean Hefty wrote:
The active side clearly cannot learn what the SLID of the passive
side's router should be.
We don't want to have the routers snoop and alter CM GMPs.
The passive side cannot use information from the LRH to get the router
LID
On 20:31 Thu 08 Feb , Tzachi Dar wrote:
The windows open IB has decided on using a BSD only license.
The common implementation of pthreads as far as I know is LGPL, which
means that it can not be used in open IB.
Why not? AFAIK it works perfectly (see (5,6 and Preamble)):
diff --git a/drivers/infiniband/hw/cxgb3/iwch_provider.c
b/drivers/infiniband/hw/cxgb3/iwch_provider.c
index db2b0a8..98568ee 100644
--- a/drivers/infiniband/hw/cxgb3/iwch_provider.c
+++ b/drivers/infiniband/hw/cxgb3/iwch_provider.c
@@ -99,6 +99,7 @@ static int
Hum, you mean to meet the LID validation rules of 9.6.1.5? That is a
huge PITA..
[IMHO, 9.6.1.5 C9-54 is a mistake, if there is a GRH then the LRH.SLID
should not be validated against the QP context since it makes it
extra hard for multipath routing and QoS to work...]
Yes - this gets messy.
This patch fix libehca build errors if disable-libcheck option is choosen.
Applied
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit
From: Steve Wise [EMAIL PROTECTED]
Lock mmap_lock is never taken from non-process context, so just use
bare spin_lock()/spin_unlock().
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
drivers/infiniband/hw/cxgb3/iwch_provider.h | 10 +-
1 files changed, 5 insertions(+), 5 deletions(-)
Here are some fixes to address various comments from Michael and Roland.
This is _not_ for ofed_1_2, but rather for merging into 2.6.21.
Steve.
___
openib-general mailing list
openib-general@openib.org
From: Steve Wise [EMAIL PROTECTED]
Free all pending mmap structs when the ucontext is deallocated.
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
drivers/infiniband/hw/cxgb3/iwch_provider.c |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git
From: Steve Wise [EMAIL PROTECTED]
Use a liked list.
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
drivers/infiniband/hw/cxgb3/core/cxio_hal.c | 57 +--
drivers/infiniband/hw/cxgb3/core/cxio_hal.h |2 -
2 files changed, 19 insertions(+), 40 deletions(-)
diff
From: Steve Wise [EMAIL PROTECTED]
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
drivers/infiniband/hw/cxgb3/iwch.c |4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)
diff --git a/drivers/infiniband/hw/cxgb3/iwch.c
b/drivers/infiniband/hw/cxgb3/iwch.c
index 0c95f2c..c353a9b
On Thu, 2007-02-08 at 14:54, Sean Hefty wrote:
Hum, you mean to meet the LID validation rules of 9.6.1.5? That is a
huge PITA..
[IMHO, 9.6.1.5 C9-54 is a mistake, if there is a GRH then the LRH.SLID
should not be validated against the QP context since it makes it
extra hard for multipath
New OFED package was uploaded to the OFA server:
http://www.openfabrics.org/builds/ofed-1.2/OFED-1.2-20070208-1508.tgz
Many of the issues reported on the previous version are resolved
(bugzilla will be updated next week).
Since we had lab restructuring we did only basic tests on RHEL up4
See bellow.
Thanks
Tzachi
-Original Message-
From: Sasha Khapyorsky [mailto:[EMAIL PROTECTED]
Sent: Thursday, February 08, 2007 9:47 PM
To: Tzachi Dar
Cc: Yossi Leybovich; Gilad Shainer; Yevgeny Kliteynik;
OPENIB; Michael S. Tsirkin; Hal Rosenstock
Subject: Re: [Fwd: Re: win
Windows Open IB is part of OpenFabrics. OpenFabrics includes Linux and
Windows communities. The Linux code is dual license while the Windows
code is BSD only.
Gilad.
-Original Message-
From: Tzachi Dar
Sent: Thursday, February 08, 2007 1:24 PM
To: Sasha Khapyorsky
Cc: Yossi
At 03:41 PM 2/7/2007, Roland Dreier wrote:
Changqing What I mean is that, is there any performance penalty
Changqing for receiver's overall performance if RNR happens
Changqing continuously on one of the QP ?
Not for the receiver, but the sender will be severely slowed down by
At 07:43 AM 2/8/2007, Kanevsky, Arkady wrote:
That is correct.
I am working with Krishna on it.
Expect patches soon.
By the way the problem is not DAPL specific
and so is a proposed solution.
There are 3 aspects of the solution.
One is APIs. We suggest that we do not augment these.
That is a
At 12:39 PM 2/8/2007, Hal Rosenstock wrote:
On Thu, 2007-02-08 at 14:54, Sean Hefty wrote:
Hum, you mean to meet the LID validation rules of 9.6.1.5? That is a
huge PITA..
[IMHO, 9.6.1.5 C9-54 is a mistake, if there is a GRH then the LRH.SLID
should not be validated against the QP
On 23:24 Thu 08 Feb , Tzachi Dar wrote:
The windows open IB has decided on using a BSD only license.
The common implementation of pthreads as far as I know is
LGPL, which
means that it can not be used in open IB.
Why not? AFAIK it works perfectly (see (5,6 and Preamble)):
This requires that the passive side be able to issue path record queries, but
I
think that it could work for static routes. A point was made to me that the
remote side could be a TCA without query capabilities.
Are you referring to SA query capabilities ? Would such a device just be
expected
Well, the way I see it one can take code from the Linux part under the BSD
licance and use it in The windows part. The otherway around seems fine to me
but
some say that since the windows BSD liscance Reqires that some text will
always
remain there, the other way around is not possibale.
Quoting r. Michael S. Tsirkin [EMAIL PROTECTED]:
Subject: Re: [Fwd: Re: win related [was: Re: [PATCH 1/2] opensm: sigusr1:
syslog() fixes]]
Well, the way I see it one can take code from the Linux part under the BSD
licance and use it in The windows part. The otherway around seems fine to
On Thu, 2007-02-08 at 09:02 +0200, Moni Levy wrote:
Doug,
On 2/7/07, Yosef Etigin [EMAIL PROTECTED] wrote:
7. On RHAS5 beta 2, the setup requires sysfstuils-devel RPM which is not
included in this distro.
Can you please help us with that ?
The value of the sysfsutils is far overshadowed
Quoting Roland Dreier [EMAIL PROTECTED]:
Subject: Re: [PATCH] IB/ipoib_cm: fix up issues from code review
OK, I pulled this in and fixed it to build with the netdevice
class_device-ectomy that just went upstream, and pushed it out on my
for-2.6.21 branch like this.
Thanks!
--
MST
opensm/main.c is not shared by win OpenSM, and #ifdef __WIN__ is not
needed here.
Signed-off-by: Sasha Khapyorsky [EMAIL PROTECTED]
---
osm/opensm/main.c |5 -
1 files changed, 0 insertions(+), 5 deletions(-)
diff --git a/osm/opensm/main.c b/osm/opensm/main.c
index 69c940c..fa09360
Use more descriptive constant names for osmtest flows.
Signed-off-by: Sasha Khapyorsky [EMAIL PROTECTED]
---
osm/osmtest/include/osmtest.h | 12
osm/osmtest/main.c| 20 ++--
osm/osmtest/osmtest.c | 22 +-
3 files
On Wed, 2007-02-07 at 15:57 -0600, Steve Wise wrote:
On Wed, 2007-02-07 at 14:02 -0600, Steve Wise wrote:
Arlin,
The OFED dapl code is assuming the responder_resources and
initiator_depth passed up on a connection request event are from the
remote peer. This doesn't happen for iWARP.
Looking at the problem more, I think that the issue extends to the remote
port
LID as well. My expectation with a local path record query is that the SLID
is
the local port, and the DLID is the local router. This should be sufficient
for
one-way UD traffic, but for connected traffic
On Thu, 2007-02-08 at 17:02, Sean Hefty wrote:
This requires that the passive side be able to issue path record queries,
but I
think that it could work for static routes. A point was made to me that the
remote side could be a TCA without query capabilities.
Are you referring to SA query
I merged the increment port number and remove redundant '_wq'
patches from git.openfabrics.org/~shefty/scm/rdma-dev.git for-roland
I plan to review to multicast stuff next week and I hope to merge it
for 2.6.21. Or, have you or anyone else at Voltaire read over the
code in addition to using it?
OK, I've pulled the cxgb3 stuff into a single commit in my for-2.6.21
branch. I took the liberty of cleaning up some sparse warnings, etc.
There's still a few other obvious things to fix up:
drivers/infiniband/hw/cxgb3/iwch_ev.c:102:6: warning: symbol 'iwch_ev_disp
atch' was not declared.
On Thu, 2007-02-08 at 16:26 -0800, Roland Dreier wrote:
OK, I've pulled the cxgb3 stuff into a single commit in my for-2.6.21
branch. I took the liberty of cleaning up some sparse warnings, etc.
There's still a few other obvious things to fix up:
Oh yeah -- Steve, please keep sending cleanup patches based on my tree
now. I'm planning on asking Linus to merge what's in for-2.6.21 in
the next couple of days, but there's still more than a week before the
merge window closes, and even after the merge window closes I'll still
accept
Michael,
From your email, it sounded like you would regression test this. Is it
ready to pull in?
Thanks!
Steve.
On Tue, 2007-02-06 at 17:39 -0600, Steve Wise wrote:
Here it is (only tested with rping over iWARP on sles9sp3):
xxx_ip_dev_find() must use scope
Hi,
I think I have found the problem. It is the timeout parameter on the
umad_send function. How exactly I have to handle this parameter? It seems to
be that it shoult be zero if there is no response exspected. But what value
should it be if there is a response expected. In a test I used zero
For OFED 1.2, I think I should just have the IWCM set them to 8. The
only RNIC in ofed is cxgb3 and it supports 8...
Steve,
If we can create the new attributes for RNICs, it seems like would be
better to agree on the mapping of IRD/ORD to IB parameters than it would
be to limit these
On Thu, 2007-02-08 at 19:19 -0600, Bob Sharp wrote:
For OFED 1.2, I think I should just have the IWCM set them to 8. The
only RNIC in ofed is cxgb3 and it supports 8...
Steve,
If we can create the new attributes for RNICs, it seems like would be
better to agree on the mapping of
For OFED 1.2, I think I should just have the IWCM set them to 8. The
only RNIC in ofed is cxgb3 and it supports 8...
Steve,
If we can't create the new attributes for RNICs, it seems like it would be
better to agree on the mapping of IRD/ORD to IB parameters than it would
be
Roland,
Yes, we will do some arm wrestling today :)
thanks,
KK
Roland Dreier [EMAIL PROTECTED] wrote on 02/09/2007 05:20:42 AM:
Hmm, Steve likes it, Tom doesn't. Can you guys arm wrestle or
something and tell me if this patch is correct or not?
- R.
On Thu, 2007-02-08 at 19:39, Michael Arndt wrote:
Hi,
I think I have found the problem. It is the timeout parameter on the
umad_send function. How exactly I have to handle this parameter? It seems to
be that it shoult be zero if there is no response exspected. But what value
should it be
BTW, while looking at iwcm.c, I noticed the following highly dubious
code for the first time:
static int iwcm_deref_id(struct iwcm_id_private *cm_id_priv)
{
int ret = 0;
BUG_ON(atomic_read(cm_id_priv-refcount)==0);
if
On Thu, Feb 08, 2007 at 03:43:24PM -0800, Sean Hefty wrote:
Looking at the problem more, I think that the issue extends to the remote
port
LID as well. My expectation with a local path record query is that the
SLID is
the local port, and the DLID is the local router. This should be
Regarding the race - can this and the other problem (of
using internal data-structure) both be taken care of by
changing iw_deref_id to return 1 if atomic_dec_and_test
finds the last reference ? Then the waitqueue_active()
code can be removed, just do the completion (reaching
here implies that
Also I agree with MST, I would like to see the core/ subdirectory die
completely.
ok ok...I'll kill the subdir...
It's not just the directory BTW. Stuff like building completions in
t3_cqe format and then reformatting to ib_wc seems to be much more confusing
(and some of it is actually
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: Re: [openib-general] [PATCH] [RFC] ofed_1_2 - SLES9SP3 Backport-IWCM
workaroundfor ip_dev_find() bug.
Michael,
From your email, it sounded like you would regression test this.
Not yet, we had lab restructuring - hopefully next week.
--
And here's what I have pending in for-2.6.21 so far:
What about the mthca memory registration patches?
I thought they are on their way. Should I repost?
--
MST
___
openib-general mailing list
openib-general@openib.org
59 matches
Mail list logo