Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 13/21] ofed_1_2 Added gso_size for cxgb3
Added gso_size for cxgb3
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
.../2.6.5_sles9_sp3/include/linux/skbuff.h |3 +++
1 files changed, 3 insertions(+), 0
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 17/21] ofed_1_2 Backport
infiniband/hw/cxgb3/iwch_provider.c to sles9sp3
Backport infiniband/hw/cxgb3/iwch_provider.c to sles9sp3
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
.../iwch_provider_to_2_6_5-7_244.patch
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 18/21] ofed_1_2 Backport
infiniband/hw/cxgb3/core/cxio_resource.c to sles9sp3
Backport infiniband/hw/cxgb3/core/cxio_resource.c to sles9sp3
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 16/21] ofed_1_2 Backport infinband/hw/cxgb3/iwch_cm.[ch]
to sles9sp3
Backport infinband/hw/cxgb3/iwch_cm.[ch] to sles9sp3
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
.../2.6.5_sles9_sp3/iwch_cm_to_2_6_5-7_244.patch |
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 19/21] ofed_1_2 Add LINUXINCLUDE to
drivers/net/cxgb3/Makefile
Add LINUXINCLUDE to drivers/net/cxgb3/Makefile
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
.../cxgb3_makefile_to_2_6_5-7_244.patch| 12
Quoting Michael S. Tsirkin [EMAIL PROTECTED]:
Subject: Re: [PATCH] 2.6.20 ib_cm: limit cm message timeouts
Quoting Roland Dreier [EMAIL PROTECTED]:
Subject: Re: [PATCH] 2.6.20 ib_cm: limit cm message timeouts
We really need this in 2.6.20 I think.
Why is this urgent? This is
Quoting Woodruff, Robert J [EMAIL PROTECTED]:
Subject: Re: [openfabrics-ewg] Reminder: OFED 1.2
Sasha wrote,
On 17:25 Mon 15 Jan , Tziporet Koren wrote:
Dhabaleswar Panda wrote:
Shaun Roland from my group (cc'ed in this e-mail) will be in charge of
this. Vlad and Shaun can
All, please run git-update-server-info on your repos if you didn't already.
Quoting Krishna Kumar2 [EMAIL PROTECTED]:
Subject: GIT fails to clone librdmacm and libamso
Hi,
When I run the following commands :
git clone http://staging.openfabrics.org/pub/scn/~swise/libamso.git libamso
git clone
Quoting Steve WIse [EMAIL PROTECTED]:
Subject: problem building user libs with build_ofa_user.sh
Vlad,
I'm trying to build the ofed 1.2 user package on sles9sp3 using
ofabuild/build_ofa_user.sh and running into problems configuring
libmthca and others. The config of libmthca fails
Try replacing 'echo libibverbs.a' with 'ar qc liby.a' - does this help?
Sorry, should have been ar qc libibverbs.a.
--
MST
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To
Added checks to memory allocation failure when using asprintf.
Signed-off-by: Dotan Barak [EMAIL PROTECTED]
---
Index: gen2_devel_user/src/userspace/libibcm/src/cm.c
===
--- gen2_devel_user.orig/src/userspace/libibcm/src/cm.c
Added checks to memory allocation failure when using asprintf.
Signed-off-by: Dotan Barak [EMAIL PROTECTED]
---
Index: gen2_devel_user/src/userspace/perftest/rdma_bw.c
===
--- gen2_devel_user.orig/src/userspace/perftest/rdma_bw.c
Added checks to memory allocation failure when using asprintf.
Signed-off-by: Dotan Barak [EMAIL PROTECTED]
---
Index: gen2_devel_user/src/userspace/libibverbs/src/device.c
===
---
Fix some memory leaks in case of error flows during driver init.
Signed-off-by: Dotan Barak [EMAIL PROTECTED]
---
Index: gen2_devel_user/src/userspace/libibverbs/src/init.c
===
---
Hi Michael,
I have just realized that
a) this patch was not pushed upstream
and
b) the --same-- instance of it is kept on all backports of both OFED 1.1 1.2
staging
It also does not have a changelog comment and Signed-Off-By signature...
Can you shed some light on what's going on here?
Quoting Or Gerlitz [EMAIL PROTECTED]:
Subject: OFED ipoib_8111_to_2_6_16.patch
Hi Michael,
I have just realized that
a) this patch was not pushed upstream
and
b) the --same-- instance of it is kept on all backports of both OFED 1.1
1.2 staging
It also does not have a
Roland Dreier wrote:
Did you push the static library link support in the last fixes/changes
that you made?
Yes, it should be working better now.
I don't know what i tried to do several days ago (or on which driver
version), but now it works without any problem.
I'm adding this test
Michael S. Tsirkin wrote:
Quoting Or Gerlitz [EMAIL PROTECTED]:
Subject: OFED ipoib_8111_to_2_6_16.patch
Isn't it obvious from the name?
sure, thanks for the clarification.
It's a backport for kernels = 2.6.16.
Can you please send (and add to OFED 1.2) a changelog comment explaining
the
Quoting Or Gerlitz [EMAIL PROTECTED]:
Subject: Re: OFED ipoib_8111_to_2_6_16.patch
Michael S. Tsirkin wrote:
Quoting Or Gerlitz [EMAIL PROTECTED]:
Subject: OFED ipoib_8111_to_2_6_16.patch
Isn't it obvious from the name?
sure, thanks for the clarification.
It's a backport for
Michael S. Tsirkin wrote:
It's a backport for kernels = 2.6.16.
Can you please send (and add to OFED 1.2) a changelog comment explaining
the problem and how it is solved in 2.6.17 and above ?!
We are looking on some code around ipoib_neigh_destructor() and friends
and the changelog would
That iSER target code is just a reference. It requires an unrelease
iSCSI module and as a result is not usable.
On Tue, 16 Jan 2007, Karthik Sarangan wrote:
Target: Compiling open-iser-target (Voltaire) and kdapl target for gen2
openib
I have downloaded kdapl for openib-gen2 from
Quoting Or Gerlitz [EMAIL PROTECTED]:
Subject: Re: OFED ipoib_8111_to_2_6_16.patch
Michael S. Tsirkin wrote:
It's a backport for kernels = 2.6.16.
Can you please send (and add to OFED 1.2) a changelog comment explaining
the problem and how it is solved in 2.6.17 and above ?!
We
We don't support creating UC connections at the moment, so we should
fall back to UD if remote side only supports UC, but not RC.
Signed-off-by: Michael S. Tsirkin [EMAIL PROTECTED]
---
Roland,
This is an incremental patch, please apply in -mm on top of the last one.
Alternatively, I can repost
Surely random32 can be a macro in kernel_addons?
I started out trying to back-port random32(). But it was too hairy and
not worth the effort IMO. I don't know how make random32() use
get_random_bytes() in a macro and return the value. Show me how.
On Tue, 2007-01-16 at 10:06 +0200, Michael S. Tsirkin wrote:
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 16/21] ofed_1_2 Backport
infinband/hw/cxgb3/iwch_cm.[ch] to sles9sp3
Backport infinband/hw/cxgb3/iwch_cm.[ch] to sles9sp3
Signed-off-by: Steve Wise [EMAIL
On Tue, 2007-01-16 at 10:08 +0200, Michael S. Tsirkin wrote:
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 19/21] ofed_1_2 Add LINUXINCLUDE to
drivers/net/cxgb3/Makefile
Add LINUXINCLUDE to drivers/net/cxgb3/Makefile
Signed-off-by: Steve Wise [EMAIL PROTECTED]
On Tue, 2007-01-16 at 11:04 +0200, Michael S. Tsirkin wrote:
Quoting Steve WIse [EMAIL PROTECTED]:
Subject: problem building user libs with build_ofa_user.sh
Vlad,
I'm trying to build the ofed 1.2 user package on sles9sp3 using
ofabuild/build_ofa_user.sh and running into problems
Quoting Steve WIse [EMAIL PROTECTED]:
Subject: Re: [PATCH RFC 18/21] ofed_1_2
Backportinfiniband/hw/cxgb3/core/cxio_resource.c to sles9sp3
Surely random32 can be a macro in kernel_addons?
I started out trying to back-port random32(). But it was too hairy and
not worth the effort
Quoting Steve WIse [EMAIL PROTECTED]:
Subject: Re: [PATCH RFC 19/21] ofed_1_2 Add LINUXINCLUDE
todrivers/net/cxgb3/Makefile
On Tue, 2007-01-16 at 10:08 +0200, Michael S. Tsirkin wrote:
Quoting Steve Wise [EMAIL PROTECTED]:
Subject: [PATCH RFC 19/21] ofed_1_2 Add LINUXINCLUDE to
The URLs are incorrect. That's scm, not scn.
On Tue, 2007-01-16 at 10:59 +0200, Michael S. Tsirkin wrote:
All, please run git-update-server-info on your repos if you didn't already.
Quoting Krishna Kumar2 [EMAIL PROTECTED]:
Subject: GIT fails to clone librdmacm and libamso
Hi,
When I
On Tue, 2007-01-16 at 11:10 +0200, Michael S. Tsirkin wrote:
Try replacing 'echo libibverbs.a' with 'ar qc liby.a' - does this help?
Sorry, should have been ar qc libibverbs.a.
That solved configuring the provider libs. I'm also failing in tvflash:
configure:3507: gcc -c -g -O2
Quoting Steve WIse [EMAIL PROTECTED]:
Subject: Re: problem building user libs with build_ofa_user.sh
On Tue, 2007-01-16 at 11:10 +0200, Michael S. Tsirkin wrote:
Try replacing 'echo libibverbs.a' with 'ar qc liby.a' - does this help?
Sorry, should have been ar qc libibverbs.a.
Sean Hefty wrote:
OK, got you at last (sorry but i have somehow ignored the call to
ib_addr_get_mgid() at the rdmacm code). So to achieve interop with
IPoIB all we need to do is remove the rdmacm signature bit and not to
over-write the rdmacm qkey on the the qkey of the ipoib ipv4 broadcast
This patch checks if the kmalloc in match_strdup was successful.
Signed-off-by: Ishai Rabinovitz [EMAIL PROTECTED]
---
Index: gen2_devel_kernel/drivers/infiniband/ulp/srp/ib_srp.c
===
---
When there is a call to send_tsk_mgmt it posts a send and waits for 5 seconds
to get a response.
When the QP is in an error state it is obvious that there will be no response
so it is quite useless to wait.
This timeout causes SRP to wait a long time to reconnect. (Each abort and
each
15.2.5.17.1 GROUP MEMBERSHIP
An endport must specify the type of multicast subscription or deletion that
it wants. The MCMemberRecord:JoinState component indicates the
membership qualities a port wishes to add (in joining or creating a group)
or remove (in leaving a group). The meanings of
Or Gerlitz wrote:
15.2.5.17.1 GROUP MEMBERSHIP
An endport must specify the type of multicast subscription or deletion that
it wants. The MCMemberRecord:JoinState component indicates the
membership qualities a port wishes to add (in joining or creating a group)
or remove (in leaving
, Vladimir Sokolovsky wrote:
Hi Steve,
I got the following failure on 2.6.17 and 2.6.18 kernels:
/tmp/ofa_kernel-20070116-1806_linux-2.6.18.6_check/drivers/net/cxgb3/Module.symvers:
No such file or directory
make[2]: *** [__modpost] Error 1
make[1]: *** [modules] Error 2
make[1]: Leaving
right now trying to get them to build from 2.6.19 back to 2.6.11 as
michael suggested...
On Tue, 2007-01-16 at 18:50 +0200, Vladimir Sokolovsky wrote:
Hi Steve,
I got the following failure on 2.6.17 and 2.6.18 kernels:
/tmp/ofa_kernel-20070116-1806_linux-2.6.18.6_check/drivers/net
On Tue, 2007-01-16 at 18:50 +0200, Vladimir Sokolovsky wrote:
Hi Steve,
I got the following failure on 2.6.17 and 2.6.18 kernels:
/tmp/ofa_kernel-20070116-1806_linux-2.6.18.6_check/drivers/net/cxgb3/Module.symvers:
No such file or directory
make[2]: *** [__modpost] Error 1
make[1
On Tue, 2007-01-16 at 19:01 +0200, Vladimir Sokolovsky wrote:
It breaks compilation without selecting cxgb3...
Oh. I see. Alright...lemme rework this...
Thanks,
Steve.
___
openib-general mailing list
openib-general@openib.org
Sean Hefty wrote:
mimic IPoIB qkey flow:
+3 on rdma_create_qp do modify qp with some def qkey (eg zero)
+4 on the join completion path before attaching a qp to the associated
mgid, do modify qp with this mrec qkey (=ipv4 broadcast one)
The rdma cm allows UD QP communication, which requires
sure, it can use the rdmacm qkey (0x1234567 etc) when it creates the QP
and later --if-- the user joins a multicast group modify the qp state
with the group qkey and report it in the cma event such that the
consumer of the rdmacm would set this into his IB UD TX WR
Changing the qkey would
-Original Message-
From: Steve WIse [mailto:[EMAIL PROTECTED]
Sent: Tuesday, January 16, 2007 8:56 AM
To: Davis, Arlin R; Hefty, Sean
Cc: openib-general
Subject: building dapl for ofed
I'm having problems building dapl for ofed 1.2.
I'm using the dapl rdma_ucm branch and still getting
On Tue, 2007-01-16 at 12:31, Or Gerlitz wrote:
Eitan Zahavi wrote:
Or Gerlitz wrote:
Eitan Zahavi wrote:
So you are saying that the GW **has** to listen on IGMP at the Eth
side and **has** to do IB SA join in the only way that forces the SA
to create the group -- FullMember ?
On Tue, 2007-01-16 at 09:43 -0800, Arlin Davis wrote:
-Original Message-
From: Steve WIse [mailto:[EMAIL PROTECTED]
Sent: Tuesday, January 16, 2007 8:56 AM
To: Davis, Arlin R; Hefty, Sean
Cc: openib-general
Subject: building dapl for ofed
I'm having problems building dapl for
There is no rdma_ucm branch in librdmacm.git. Which should be used:
master
rdma_ucm-abi3
multicast
1.1
1.0
rdma_ucm-abi3
___
openib-general mailing list
openib-general@openib.org
FWIW, having git's for the MPI implementations was asked for on the
call yesterday (by Tziporet, IIRC?). The rationale, as I understood
it, was threefold:
1. Putting the MPI release in git provides a level of OFED-specific
history and version control. This was explicitly stated on the
On Jan 16, 2007, at 1:34 PM, Jeff Squyres wrote:
FWIW, having git's for the MPI implementations was asked for on the
call yesterday (by Tziporet, IIRC?). The rationale, as I understood
it, was threefold:
Threefold, fourfold, tenfold... who's counting? :-)
--
Jeff Squyres
Server
Changes to kernel_patches/ for Chelsio T3 Support.
- qp_num - qp ptr patch for cxgb3.
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
kernel_patches/fixes/cxgb3_wc_qpn_to_qp.patch | 23 +++
1 files changed, 23 insertions(+), 0 deletions(-)
diff --git
Michael/Vlad,
Here is version 5.
Changes since version 4:
- Redid cxgb3 support in ofed_scripts/Makefile ala the iscsi rules.
Signed-off-by: Steve Wise [EMAIL PROTECTED]
___
openib-general mailing list
openib-general@openib.org
Changes to ofed scripts for Chelsio T3 Support.
- added cxgb3 and iw_cxgb3 config stuff
- visit and build driver/net/cxgb3 to get the cxgb3 driver
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
ofed_scripts/Makefile | 73 +++-
Provide generic allocator backport to 2.6.20.
This is needed even on kernels that might have genalloc because it
may not be configured into the running kernel. Right now the only
subsystem that turns on the generic allocator config option is IA64.
So chances are good that 2.6.20 kernels will
On Tue, 2007-01-16 at 19:01 +0200, Vladimir Sokolovsky wrote:
It breaks compilation without selecting cxgb3...
Regards,
Vladimir
I just reposted the series (version 5) with a new Makefile scheme
similar to iscsi.
Steve.
___
openib-general
Thanks, applied to master and stable branches of libibverbs.
___
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
Good catch, I pushed out a patch fixing this in a slightly different way.
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit
Looks pretty good, but could you please:
- split off the asprintf() checks into a separate patch (one idea
per patch).
- change the code so it keeps a count of cq events and then just
does one ibv_ack_cq_events(cq, num_cq_events) at the end. If
people are going to use this code as an
OK, I added this patch to the for-mm branch, and rolled it into the
patch in the ipoib-cm branch. (I did it that way to avoid redoing the
NAPI/CM merge).
- R.
___
openib-general mailing list
openib-general@openib.org
Thanks, applied.
___
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
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Roland Dreier
Sent: Tuesday, January 16, 2007 1:11 PM
To: Tziporet Koren
Cc: EWG; OPENIB
Subject: Re: [openib-general] Minutes for January 15, 2007
teleconference about OFED 1.2 development
To what extent Fork is supported ? Is this a complete solution
that any style fork() code should work, or do we still need some sync
between parent and child so that child won't touch pinned page before
call exec() ?
Just about anything should work. However child processes may not
+ libibverbs 1.0 to be included as well
I believe I have a way to make libibverbs 1.1
binary-compatible with applications (but not drivers) built
against libibverbs 1.0. So it will probably not be necessary
to include libibverbs 1.0.
We need this compatible feature
Thanks Steve, it works now. Copy-n-paste was not working and I had wrongly
typed the link.
Thanks for the help.
- KK
Steve WIse [EMAIL PROTECTED] wrote on 01/16/2007 08:01:15 PM:
The URLs are incorrect. That's scm, not scn.
On Tue, 2007-01-16 at 10:59 +0200, Michael S. Tsirkin wrote:
OSM Simulation Regression Summary
OpenSM rev = Mon_Jan_15_23:49:50_2007 6ff53d
ibutils rev = Wed_Jan_3_11:42:12_2007 913448
Total=410 Pass=410 Fail=0
Pass:
30 Stability IS1-16.topo
30 Pkey IS1-16.topo
30 OsmTest IS1-16.topo
30 OsmStress IS1-16.topo
30 Multicast IS1-16.topo
30 LidMgr IS1-16.topo
Quoting Sean Hefty [EMAIL PROTECTED]:
Subject: Re: building dapl for ofed
There is no rdma_ucm branch in librdmacm.git. Which should be used:
master
rdma_ucm-abi3
multicast
1.1
1.0
rdma_ucm-abi3
OFED is packaging master, BTW.
--
MST
Michael S. Tsirkin wrote:
There is no rdma_ucm branch in librdmacm.git. Which should be used:
master
rdma_ucm-abi3
multicast
1.1
1.0
rdma_ucm-abi3
OFED is packaging master, BTW.
Hi Sean,
Can you reply which branch we should use for OFED 1_2?
Quoting Tziporet Koren [EMAIL PROTECTED]:
Subject: Re: [openib-general] building dapl for ofed
Michael S. Tsirkin wrote:
There is no rdma_ucm branch in librdmacm.git. Which should be used:
master
rdma_ucm-abi3
multicast
1.1
1.0
rdma_ucm-abi3
OFED
Applied.
Regards,
Vladimir
On Tue, 2007-01-16 at 13:51 -0600, Steve Wise wrote:
Changes to ofed scripts for Chelsio T3 Support.
- added cxgb3 and iw_cxgb3 config stuff
- visit and build driver/net/cxgb3 to get the cxgb3 driver
Signed-off-by: Steve Wise [EMAIL PROTECTED]
---
Applied.
Regards,
Vladimir
On Tue, 2007-01-16 at 13:52 -0600, Steve Wise wrote:
Provide generic allocator backport to 2.6.20.
This is needed even on kernels that might have genalloc because it
may not be configured into the running kernel. Right now the only
subsystem that turns on the
On Mon, 15 Jan 2007 10:46:19 +0200
Dotan Barak [EMAIL PROTECTED] wrote:
Joerg Zinke wrote:
Hi,
is there a non-blocking alternative for ib_cm_get_event()
available? Any hints on how to check for events without blocking
and waiting for the next one?
Here is a response that Or Gerlitz
70 matches
Mail list logo