>> Sean maintains a separate set of patches applied to IB diags in order to
>address issues like the x86 requirement for __cdecl on main().
>> Since OSM_CDECL was already defined in OFED opensm, it did not seem to be a
>major concession to prefix main() with it.
>> Options:
>> 1) utilize OSM_CDECL
New release of libnes library (0.9.0) is available at:
http://www.openfabrics.org/downloads/nes/
sha1sum: 6e9374ea9ace5e052c00aa868eea6793839d1ae8 libnes-0.9.0.tar.gz
Changes since last release:
Chien Tung (3):
libnes: fix warning in src/nes_uverbs.c:756
libnes: add support for sq_
On Wed, 2009-10-14 at 15:47 -0700, Roland Dreier wrote:
> > > > First it does not make sense for user to set it below 60; therefore,
> > > > it is forced to have 60 and above
>
> > > Why not? A minute seems to be a really long time given the point of
> > > these patches is supposed to be fai
From: Ira Weiny
Date: Wed, 14 Oct 2009 17:05:53 -0700
Subject: [PATCH] opensm/opensm/osm_subnet.c: adjust buffer to ensure a '\n' is
printed
When printing cached options strings which fill the print buffer.
Adjust the length so the final snprintf(..., "\n"); can succeed.
Signed
Roland Dreier wrote:
> > > First it does not make sense for user to set it below 60; therefore,
> > > it is forced to have 60 and above
> > Why not? A minute seems to be a really long time given the point of
> > these patches is supposed to be failing over faster. Surely we can tell
> >
> > > First it does not make sense for user to set it below 60; therefore,
> > > it is forced to have 60 and above
> > Why not? A minute seems to be a really long time given the point of
> > these patches is supposed to be failing over faster. Surely we can tell
> > if a path really fail
Heap memory consumption by the unicast and multicast routing tables can be
reduced.
This patch is analagous to the previous patch doing this for the unicast
routing tables (LFTs).
Using valgrind --tool=massif (for heap profiling), there are couple of place
->38.75% (11,206,656B) 0x43267E: osm_sw
Bart Van Assche wrote:
+
+ switch (event->event) {
+ case IB_EVENT_PORT_ERR:
+ list_for_each_entry_safe(host, tmp_host,
+&srp_dev->dev_list, list) {
+ if (event->element.port_num == host->port) {
+
Roland Dreier wrote:
> First it does not make sense for user to set it below 60; therefore,
> it is forced to have 60 and above
Why not? A minute seems to be a really long time given the point of
these patches is supposed to be failing over faster. Surely we can tell
if a path really failed
Roland Dreier wrote:
> > > - wait_for_completion(&target->done);
> >
> > How do you avoid leaking connection on module unload etc? Don't we have
> > to wait for the disconnect to finish somewhere?
> >
> > - R.
> >
> Are you talking about cm_id?
> I think that we wait because
Roland Dreier wrote:
> +static int srp_dev_loss_tmo = 60;
I don't think the name needs to be this abbreviated. We don't
necessarily need the srp_ prefix, but probably "device_loss_timeout" is
much clearer without being too much longer.
OK
> +
> +module_param(srp_dev_loss_tmo, int, 0444
> First it does not make sense for user to set it below 60; therefore,
> it is forced to have 60 and above
Why not? A minute seems to be a really long time given the point of
these patches is supposed to be failing over faster. Surely we can tell
if a path really failed sooner than 60 seconds
On 16:01 Wed 14 Oct , Hal Rosenstock wrote:
>
> Signed-off-by: Hal Rosenstock
Applied. Thanks.
Sasha
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.h
> > > - wait_for_completion(&target->done);
> >
> > How do you avoid leaking connection on module unload etc? Don't we have
> > to wait for the disconnect to finish somewhere?
> >
> > - R.
> >
> Are you talking about cm_id?
> I think that we wait because we want to reuse cq/qp
Roland Dreier wrote:
> - wait_for_completion(&target->done);
How do you avoid leaking connection on module unload etc? Don't we have
to wait for the disconnect to finish somewhere?
- R.
Are you talking about cm_id?
I think that we wait because we want to reuse cq/qp associate with the
c
On Wed, Oct 14, 2009 at 12:33:17PM -0700, David J. Wilder wrote:
>
> On Wed, 2009-10-14 at 11:01 -0600, Jason Gunthorpe wrote:
> > On Wed, Oct 14, 2009 at 09:23:57AM -0700, David J. Wilder wrote:
> >
> > > This new patch should closely emulate tcp_ipv6.c. when both source and
> > > destination sc
Signed-off-by: Hal Rosenstock
---
diff --git a/opensm/opensm/osm_mcast_mgr.c b/opensm/opensm/osm_mcast_mgr.c
index 77e0b94..0ee689c 100644
--- a/opensm/opensm/osm_mcast_mgr.c
+++ b/opensm/opensm/osm_mcast_mgr.c
@@ -351,7 +351,7 @@ static int mcast_mgr_set_mft_block(osm_sm_t * sm, IN
osm_switch_t
On 15:51 Wed 14 Oct , Hal Rosenstock wrote:
>
> Signed-off-by: Hal Rosenstock
Applied. Thanks.
Sasha
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.h
Roland Dreier wrote:
One meta comment: when sending a series of 4 patches, please choose a
descriptive subject for each one. We don't want the same headline in
the kernel log for all 4 patches.
I'll remember next time
> - qp_attr.qp_state = IB_QPS_RESET;
> - ret = ib_modify_qp(target->qp
On 15:46 Wed 14 Oct , Hal Rosenstock wrote:
>
> Seems to me that the equivalent (to LFT) is to invoke the table
> (re)allocation from osm_mcast_mgr_process/process_mgroups. Is that
> what you mean ?
Yes, exactly.
Sasha
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma"
Signed-off-by: Hal Rosenstock
---
diff --git a/opensm/include/opensm/osm_subnet.h
b/opensm/include/opensm/osm_subnet.h
index 9488225..b63c97e 100644
--- a/opensm/include/opensm/osm_subnet.h
+++ b/opensm/include/opensm/osm_subnet.h
@@ -634,6 +634,10 @@ typedef struct osm_subn {
* Th
On Wed, Oct 14, 2009 at 12:25 PM, Sasha Khapyorsky wrote:
> On 11:52 Wed 14 Oct , Hal Rosenstock wrote:
>>
>> Heap memory consumption by the unicast and multicast routing tables can be
>> reduced.
>>
>> This patch is analagous to the previous patch doing this for the unicast
>> routing tables
On Wed, 2009-10-14 at 11:01 -0600, Jason Gunthorpe wrote:
> On Wed, Oct 14, 2009 at 09:23:57AM -0700, David J. Wilder wrote:
>
> > This new patch should closely emulate tcp_ipv6.c. when both source and
> > destination scope_ids are given with link-local address.
>
> Maybe like this:
>
>
> +static int srp_dev_loss_tmo = 60;
I don't think the name needs to be this abbreviated. We don't
necessarily need the srp_ prefix, but probably "device_loss_timeout" is
much clearer without being too much longer.
> +
> +module_param(srp_dev_loss_tmo, int, 0444);
> +MODULE_PARM_DESC(srp_de
> -wait_for_completion(&target->done);
How do you avoid leaking connection on module unload etc? Don't we have
to wait for the disconnect to finish somewhere?
- R.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.or
One meta comment: when sending a series of 4 patches, please choose a
descriptive subject for each one. We don't want the same headline in
the kernel log for all 4 patches.
> -qp_attr.qp_state = IB_QPS_RESET;
> -ret = ib_modify_qp(target->qp, &qp_attr, IB_QP_STATE);
> -if (ret)
>
Thanks, applied.
> Resending the patch based on Roland's "for-linus" branch.
By the way, just to be clear, my master branch is just the state of
Linus's tree at the time I branched. So it's not a particularly good
thing to base work on. My "for-next" branch is probably the best if you
are worr
On Wed, Oct 14, 2009 at 10:30:05AM -0700, David J. Wilder wrote:
> This looks good. Once concern, it may be obtuse, but if both the src and
> dst are link-local addresses should only one need to be scoped? This
> patch will required the src to always be scoped when using link local.
The TCPv6
On Wed, 2009-10-14 at 11:01 -0600, Jason Gunthorpe wrote:
> On Wed, Oct 14, 2009 at 09:23:57AM -0700, David J. Wilder wrote:
>
> > This new patch should closely emulate tcp_ipv6.c. when both source and
> > destination scope_ids are given with link-local address.
>
> Maybe like this:
>
>
On Wed, Oct 14, 2009 at 09:23:57AM -0700, David J. Wilder wrote:
> This new patch should closely emulate tcp_ipv6.c. when both source and
> destination scope_ids are given with link-local address.
Maybe like this:
fl.oif = 0;
if (ipv6_addr_type(&src_in->sin6_addr) & IPV6_ADDR_L
Sean-
This patch should fix the behavior of rdma_resolve_addr when using
link-local addressing.
On Tue, 2009-10-13 at 17:12 -0600, Jason Gunthorpe wrote:
> On Tue, Oct 13, 2009 at 03:09:40PM -0700, David J. Wilder wrote:
> > Here is a patch to addr6_resolve_remote() to correctly handle link-loc
On 11:52 Wed 14 Oct , Hal Rosenstock wrote:
>
> Heap memory consumption by the unicast and multicast routing tables can be
> reduced.
>
> This patch is analagous to the previous patch doing this for the unicast
> routing tables (LFTs).
>
> Using valgrind --tool=massif (for heap profiling), t
Heap memory consumption by the unicast and multicast routing tables can be
reduced.
This patch is analagous to the previous patch doing this for the unicast
routing tables (LFTs).
Using valgrind --tool=massif (for heap profiling), there are couple of place
->38.75% (11,206,656B) 0x43267E: osm_sw
On 10:45 Wed 14 Oct , Hal Rosenstock wrote:
>
> Signed-off-by: Hal Rosenstock
Applied. Thanks.
Sasha
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.h
Signed-off-by: Hal Rosenstock
---
Changes since v1:
Removed change of malloc/memset into calloc
diff --git a/opensm/opensm/osm_ucast_updn.c b/opensm/opensm/osm_ucast_updn.c
index bb9ccda..ced076a 100644
--- a/opensm/opensm/osm_ucast_updn.c
+++ b/opensm/opensm/osm_ucast_updn.c
@@ -2,6 +2,7 @@
*
On 09:33 Wed 14 Oct , Hal Rosenstock wrote:
> On Wed, Oct 14, 2009 at 9:29 AM, Sasha Khapyorsky wrote:
> > On 09:14 Wed 14 Oct , Hal Rosenstock wrote:
> >>
> >> Also, combine malloc/memset into calloc
> >
> > Any special reason for malloc/calloc change?
>
> It's a little more concise and
On Wed, Oct 14, 2009 at 10:34 AM, Sasha Khapyorsky wrote:
> On 09:33 Wed 14 Oct , Hal Rosenstock wrote:
>> On Wed, Oct 14, 2009 at 9:29 AM, Sasha Khapyorsky
>> wrote:
>> > On 09:14 Wed 14 Oct , Hal Rosenstock wrote:
>> >>
>> >> Also, combine malloc/memset into calloc
>> >
>> > Any specia
Yevgeny Kliteynik wrote:
Hal Rosenstock wrote:
On Wed, Oct 14, 2009 at 9:30 AM, Aaron Knister
wrote:
I searched the lists and can't really find an answer to my question-
When I run saquery -s on my cluster with ConnectX HCAs (fw ver
2.6.648) all I get is the following output
IsSM ports
IsSM
Hal Rosenstock wrote:
On Wed, Oct 14, 2009 at 9:30 AM, Aaron Knister wrote:
I searched the lists and can't really find an answer to my question-
When I run saquery -s on my cluster with ConnectX HCAs (fw ver
2.6.648) all I get is the following output
IsSM ports
IsSMdisabled ports
I thought
On Wed, Oct 14, 2009 at 9:30 AM, Aaron Knister wrote:
> I searched the lists and can't really find an answer to my question-
>
> When I run saquery -s on my cluster with ConnectX HCAs (fw ver
> 2.6.648) all I get is the following output
>
> IsSM ports
>
> IsSMdisabled ports
>
> I thought this shou
On Wed, Oct 14, 2009 at 9:29 AM, Sasha Khapyorsky wrote:
> On 09:14 Wed 14 Oct , Hal Rosenstock wrote:
>>
>> Also, combine malloc/memset into calloc
>
> Any special reason for malloc/calloc change?
It's a little more concise and efficient to calloc rather than
malloc/memset. Any reason not to
I searched the lists and can't really find an answer to my question-
When I run saquery -s on my cluster with ConnectX HCAs (fw ver
2.6.648) all I get is the following output
IsSM ports
IsSMdisabled ports
I thought this should show ports that are running a subnet manager
(there are 2 running on
On 09:14 Wed 14 Oct , Hal Rosenstock wrote:
>
> Also, combine malloc/memset into calloc
Any special reason for malloc/calloc change?
Please don't mix two things in one patch.
Sasha
>
> Signed-off-by: Hal Rosenstock
> ---
> diff --git a/opensm/opensm/osm_ucast_updn.c b/opensm/opensm/osm_
Also, combine malloc/memset into calloc
Signed-off-by: Hal Rosenstock
---
diff --git a/opensm/opensm/osm_ucast_updn.c b/opensm/opensm/osm_ucast_updn.c
index bb9ccda..a2acdbd 100644
--- a/opensm/opensm/osm_ucast_updn.c
+++ b/opensm/opensm/osm_ucast_updn.c
@@ -2,6 +2,7 @@
* Copyright (c) 2004-2
It will be difficult to get you those logs until the older cluster is
decommissioned (which in theory should be soon), but as soon as I am
able I will get them too you.
On Wed, Oct 14, 2009 at 4:46 AM, Yevgeny Kliteynik
wrote:
> Aaron,
>
> Aaron Knister wrote:
As I said, the older opens
Hi Hal,
On 07:14 Wed 14 Oct , Hal Rosenstock wrote:
>
> Heap memory consumption by the unicast and multicast routing tables can be
> reduced.
>
> This patch is analagous to the previous patch doing this for the unicast
> routing tables (LFTs).
>
> Using valgrind --tool=massif (for heap prof
On 07:28 Wed 14 Oct , Hal Rosenstock wrote:
>
> Signed-off-by: Hal Rosenstock
Applied. Thanks.
Sasha
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.h
Sergey Zhumatiy wrote:
Hi Sergey,
Try to add '--without-depcheck' to install.pl command:
./install.pl -c ofed.conf --without-depcheck
In this case it fails on unistallation procedure. It seems, that some
old dependency was forgotten or something like this. If this script can
simply print in
Signed-off-by: Hal Rosenstock
---
diff --git a/opensm/doc/opensm_release_notes-3.2.txt
b/opensm/doc/opensm_release_notes-3.2.txt
index 3356e95..5a83092 100644
--- a/opensm/doc/opensm_release_notes-3.2.txt
+++ b/opensm/doc/opensm_release_notes-3.2.txt
@@ -271,7 +271,7 @@ information regarding eac
Sumeet Lahorani wrote:
We are using OFED 1.4.2
Please note that the bonding driver provided by the latest distros
supports IPoIB. So if your distro happen to be RHEL 5.4 (or its OEL 5.4
derivative), or SLES11 you can and should use the distro provided
bonding. Moving forward, OTOH customers wo
Sumeet Lahorani wrote:
We are [...] trying to simulate the effect of a bonding failover
initiated by a switch failure using echo commands in parallel to the
/sys/class/net/bond0/bonding/active_slave file on a few of the nodes
attached to the switch. Is this an acceptable technique?
yes
We are
Heap memory consumption by the unicast and multicast routing tables can be
reduced.
This patch is analagous to the previous patch doing this for the unicast
routing tables (LFTs).
Using valgrind --tool=massif (for heap profiling), there are couple of place
->38.75% (11,206,656B) 0x43267E: osm_sw
> Hi Sergey,
> Try to add '--without-depcheck' to install.pl command:
>
> ./install.pl -c ofed.conf --without-depcheck
>
In this case it fails on unistallation procedure. It seems, that some
old dependency was forgotten or something like this. If this script can
simply print in one line all pac
> OFED debian packages (up to OFED 1.4.2) are available here:
>
> http://pkg-ofed.alioth.debian.org/
>
Ok! Thank you very much! I'll try them.
But main question is still open: how to install OFED on AltLinux or
other distributive (not Fedora/Suse). Older installer simply did
compilation on
Aaron,
Aaron Knister wrote:
As I said, the older opensms
on the older mellanox model HCAs failsover and failsback instantly.
The instant failback is expected, and this is the bug that
we're discussing. As for the instant failover - I'll check
how the things supposed to work and get back to you.
-Original Message-
From: Jeffrey Scott [mailto:j...@splitrockpr.com]
Sent: Tuesday, October 13, 2009 9:09 AM
To: bb...@systemfabricworks.com; 'Sujal Das'
Subject: FW: OFED installation
Who should this inquiry go to?
-Original Message-
From: Sergey Zhumatiy [mailto:s...@paralle
56 matches
Mail list logo