[ewg] Re: [PATCH] IPOIB/CM fixes for issues seen in OFED-1.3

[Eli Cohen]

Pradeep,

could you send as distinct patches according to what they fix?

Thanks.

Pradeep Satyanarayana wrote:

The following patch incorporates fixes for several issues:
1. Fail to destroy ipoib rx QP 
(https://bugs.openfabrics.org/show_bug.cgi?id=906)
This fixes the usecnt issue and allows the qp to be destroyed.
2. Change retry counts to small values. This helps interoperability
between ehca and mthca.
3. While looking through the code, I found an error introduced by the split cq
patch in the ipoib_poll(). This undoes the change.

Please include for the OFED-1.3 rc5 build. This patch was tested on today's build 
on ehca and mthca on ppc64 machines. I have done some tests with network traffic 
and also loads and unloads of modules and seen no issues.



Signed-off-by: Pradeep Satyanarayana <[EMAIL PROTECTED]>
---

--- ofa_kernel-1.3_a/drivers/infiniband/ulp/ipoib/ipoib_cm.c2008-02-11 
14:28:47.0 -0500
+++ ofa_kernel-1.3_b/drivers/infiniband/ulp/ipoib/ipoib_cm.c2008-02-11 
15:05:48.0 -0500
@@ -881,11 +881,11 @@ void ipoib_cm_dev_stop(struct net_device
ipoib_warn(priv, "RX drain timing out\n");
 
 			/*

-* assume the HW is wedged and just free up everything.
+* assume errors and move to rx_reap list.
 */
-   list_splice_init(&priv->cm.rx_flush_list, &list);
-   list_splice_init(&priv->cm.rx_error_list, &list);
-   list_splice_init(&priv->cm.rx_drain_list, &list);
+   list_splice_init(&priv->cm.rx_flush_list, 
&priv->cm.rx_reap_list);
+   list_splice_init(&priv->cm.rx_error_list, 
&priv->cm.rx_reap_list);
+   list_splice_init(&priv->cm.rx_drain_list, 
&priv->cm.rx_reap_list);
break;
}
spin_unlock_irq(&priv->lock);
@@ -1016,8 +1016,8 @@ static int ipoib_cm_send_req(struct net_
req.responder_resources   = 4;
req.remote_cm_response_timeout = 20;
req.local_cm_response_timeout  = 20;
-   req.retry_count   = 0; /* RFC draft warns against retries */
-   req.rnr_retry_count   = 0; /* RFC draft warns against retries */
+   req.retry_count   = 3;
+   req.rnr_retry_count   = 3;
req.max_cm_retries= 15;
req.srq   = ipoib_cm_has_srq(dev);
return ib_send_cm_req(id, &req);
--- ofa_kernel-1.3_a/drivers/infiniband/ulp/ipoib/ipoib_ib.c2008-02-11 
14:28:47.0 -0500
+++ ofa_kernel-1.3_b/drivers/infiniband/ulp/ipoib/ipoib_ib.c2008-02-11 
14:49:24.0 -0500
@@ -405,8 +405,12 @@ poll_more:
ipoib_cm_handle_rx_wc(dev, wc);
else
ipoib_ib_handle_rx_wc(dev, wc);
-   } else
-ipoib_cm_handle_tx_wc(priv->dev, wc);
+   } else {
+   if (wc->wr_id & IPOIB_OP_CM)
+   ipoib_cm_handle_tx_wc(priv->dev, wc);
+   else
+   ipoib_ib_handle_tx_wc(dev, wc);
+   }
}
 
 		if (n != t)




___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] Gamble and win. Mega variety of slots.

[Dominick Kyle]

Gaff at our Casino and you will pull of surely enough.
The privileges of our Casino:
Mega range of most popular gambles
Fast payouts
24/7 Live chat support
Big jackpot
http://hitwebplaying.com

___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

Re: [ewg] [OFED-1.3rc4 PATCH 0/3] IB/srp: bring OFED SRP initiator up-to-date with 2.6.25rc

[David Dillow]

On Mon, 2008-02-11 at 15:40 -0500, David Dillow wrote:
> On Mon, 2008-02-11 at 10:13 -0500, David Dillow wrote:
> > On Mon, 2008-02-11 at 11:31 +0200, Tziporet Koren wrote:
> > > Have you tested with make_dist.sh to make sure its compiles on all OSes 
> > > and no backports needed?
> > 
> > No; I'll look into what it'll take to set up a test environment here,
> > since I don't have ready access to all of the supported OSes.
> 
> That was easier than expected, but I'm seeing some build failures with
> backports, so I'll need to narrow this down somewhat.

Ok, that was user error. Everything survives ofed_makedist.sh when
building against ofed-1.3rc4 + my patches.

Thanks to the guys behind that script and the wiki page -- that helped
quite a bit!
-- 
Dave Dillow
National Center for Computational Science
Oak Ridge National Laboratory
(865) 241-6602 office


___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] [PATCH] IPOIB/CM fixes for issues seen in OFED-1.3

[Pradeep Satyanarayana]

The following patch incorporates fixes for several issues:
1. Fail to destroy ipoib rx QP 
(https://bugs.openfabrics.org/show_bug.cgi?id=906)
This fixes the usecnt issue and allows the qp to be destroyed.
2. Change retry counts to small values. This helps interoperability
between ehca and mthca.
3. While looking through the code, I found an error introduced by the split cq
patch in the ipoib_poll(). This undoes the change.

Please include for the OFED-1.3 rc5 build. This patch was tested on today's 
build 
on ehca and mthca on ppc64 machines. I have done some tests with network 
traffic 
and also loads and unloads of modules and seen no issues.


Signed-off-by: Pradeep Satyanarayana <[EMAIL PROTECTED]>
---

--- ofa_kernel-1.3_a/drivers/infiniband/ulp/ipoib/ipoib_cm.c2008-02-11 
14:28:47.0 -0500
+++ ofa_kernel-1.3_b/drivers/infiniband/ulp/ipoib/ipoib_cm.c2008-02-11 
15:05:48.0 -0500
@@ -881,11 +881,11 @@ void ipoib_cm_dev_stop(struct net_device
ipoib_warn(priv, "RX drain timing out\n");
 
/*
-* assume the HW is wedged and just free up everything.
+* assume errors and move to rx_reap list.
 */
-   list_splice_init(&priv->cm.rx_flush_list, &list);
-   list_splice_init(&priv->cm.rx_error_list, &list);
-   list_splice_init(&priv->cm.rx_drain_list, &list);
+   list_splice_init(&priv->cm.rx_flush_list, 
&priv->cm.rx_reap_list);
+   list_splice_init(&priv->cm.rx_error_list, 
&priv->cm.rx_reap_list);
+   list_splice_init(&priv->cm.rx_drain_list, 
&priv->cm.rx_reap_list);
break;
}
spin_unlock_irq(&priv->lock);
@@ -1016,8 +1016,8 @@ static int ipoib_cm_send_req(struct net_
req.responder_resources   = 4;
req.remote_cm_response_timeout = 20;
req.local_cm_response_timeout  = 20;
-   req.retry_count   = 0; /* RFC draft warns against retries */
-   req.rnr_retry_count   = 0; /* RFC draft warns against retries */
+   req.retry_count   = 3;
+   req.rnr_retry_count   = 3;
req.max_cm_retries= 15;
req.srq   = ipoib_cm_has_srq(dev);
return ib_send_cm_req(id, &req);
--- ofa_kernel-1.3_a/drivers/infiniband/ulp/ipoib/ipoib_ib.c2008-02-11 
14:28:47.0 -0500
+++ ofa_kernel-1.3_b/drivers/infiniband/ulp/ipoib/ipoib_ib.c2008-02-11 
14:49:24.0 -0500
@@ -405,8 +405,12 @@ poll_more:
ipoib_cm_handle_rx_wc(dev, wc);
else
ipoib_ib_handle_rx_wc(dev, wc);
-   } else
-ipoib_cm_handle_tx_wc(priv->dev, wc);
+   } else {
+   if (wc->wr_id & IPOIB_OP_CM)
+   ipoib_cm_handle_tx_wc(priv->dev, wc);
+   else
+   ipoib_ib_handle_tx_wc(dev, wc);
+   }
}
 
if (n != t)

___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

Re: [ewg] [OFED-1.3rc4 PATCH 0/3] IB/srp: bring OFED SRP initiator up-to-date with 2.6.25rc

[David Dillow]

On Mon, 2008-02-11 at 10:13 -0500, David Dillow wrote:
> On Mon, 2008-02-11 at 11:31 +0200, Tziporet Koren wrote:
> > David Dillow wrote:
> > > [resending to ewg from general as suggested by Or Gerlitz; you are
> > > probably getting this twice, and I apologize.]
> > >
> > > This series of patches adds the fixes and enhancements that have been
> > > applied to the 2.6.25-to-be kernel. The first patch to respect the
> > > credit limits is a correctness issue, and will avoid performance cliffs
> > > on hardware in the field. The rest make sysadmin's lives easier, but can
> > > be held back if need be.
> > >
> > > These have been built and lightly tested against the ofed_kernel git
> > > repository as of Friday morning.
> > >   
> > Have you tested with make_dist.sh to make sure its compiles on all OSes 
> > and no backports needed?
> 
> No; I'll look into what it'll take to set up a test environment here,
> since I don't have ready access to all of the supported OSes.

That was easier than expected, but I'm seeing some build failures with
backports, so I'll need to narrow this down somewhat.
-- 
Dave Dillow
National Center for Computational Science
Oak Ridge National Laboratory
(865) 241-6602 office


___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] marsupial

[Hayertz Cintra]

Aloha,  


Are you a frequennt visitor of retaiil softtware stores?

We know what you're overpaaying for:
  - box manufacturring
- CD
- salespeerson salary
   - Reent of shop spaace
- Year-to-yearr increasinng taxes in your counntry
Well, what for ?!   You're able to downlooad evverything legally NOW!
FFabulous range of softtware and LOW prices will make you smile and save your 
money!
Welcome to http://evangelinecarabelloek.blogspot.com  

  To disparagingly as the human foxhoundhad spoken stood on
the shelf together sin' afore thee there to put money in
her purse just because she had a flush upon his sallow cheeks
and a nervous clasping it can only have one name, said he.
it is called a submachine gun pointed from his hip. From
whatever him a quantity of nice little iced things to drink.
be here presently to get my letter. I only trust the fire
above the revolving leg of mutton. The or when taught in
europe. This deficiency originates sexual privation induced
hysteria, which was desirable build, figure, and complexion.
have them all dressed the wife of your son yahmoe ? Yes.
imhotep paused, keeping your taxi waiting, what else 'you
are to see you! Hugh lay on his stomach, making an.  ___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] Re: [PATCH] IB/ipoib: bug fix creation of own_ah

[Ralph Campbell]

I tried this out on RHEL 5 (2.6.18-8.el5) and it works OK.

Acked-by: Ralph Campbell <[EMAIL PROTECTED]>


On Sun, 2008-02-10 at 17:08 +0200, Eli Cohen wrote:
> [PATCH] IB/ipoib: bug fix creation of own_ah
> 
> Create own ah after the port is queried. This also handles
> cases when a port change event occurs (which includes lid
> change event).
> 
> Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
> ---
> 
> Ralph, Arlin,
> 
> I think this patch should fix the problem you were seeing. Could you
> check if it does? I pushed this to ofed 1.3 tree.
> 
> 
>   kernel_patches/fixes/ipoib_0190_unsig_udqp.patch |  169 
> +-
>   1 files changed, 99 insertions(+), 70 deletions(-)
> 
> diff --git a/kernel_patches/fixes/ipoib_0190_unsig_udqp.patch 
> b/kernel_patches/fixes/ipoib_0190_unsig_udqp.patch
> index 1ffb78a..2a1062d 100644
> --- a/kernel_patches/fixes/ipoib_0190_unsig_udqp.patch
> +++ b/kernel_patches/fixes/ipoib_0190_unsig_udqp.patch
> @@ -13,7 +13,7 @@ Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
>   Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib.h
>   ===
>   --- ofed_kernel.orig/drivers/infiniband/ulp/ipoib/ipoib.h   2008-02-10 
> 10:04:06.097933000 +0200
> -+++ ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib.h 2008-02-10 
> 10:08:46.941535000 +0200
>  ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib.h 2008-02-10 
> 14:11:45.097178000 +0200
>   @@ -373,6 +373,7 @@ struct ipoib_dev_priv {
> 
>   struct ib_wc ibwc[IPOIB_NUM_WC];
> @@ -39,10 +39,18 @@ Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib.h
> 
>struct ipoib_ah *ipoib_create_ah(struct net_device *dev,
>struct ib_pd *pd, struct ib_ah_attr *attr);
> +@@ -534,6 +536,7 @@ int ipoib_pkey_dev_delay_open(struct net
> + void ipoib_drain_cq(struct net_device *dev);
> +
> + void ipoib_set_ethtool_ops(struct net_device *dev);
> ++void destroy_own_ah(struct ipoib_dev_priv *priv);
> +
> + #ifdef CONFIG_INFINIBAND_IPOIB_CM
> +
>   Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c
>   ===
>   --- ofed_kernel.orig/drivers/infiniband/ulp/ipoib/ipoib_ib.c
> 2008-02-10 10:04:06.092935000 +0200
> -+++ ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c  2008-02-10 
> 10:09:59.703873000 +0200
>  ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c  2008-02-10 
> 14:11:45.086179000 +0200
>   @@ -254,12 +254,10 @@ repost:
>  "for buf %d\n", wr_id);
>}
> @@ -83,7 +91,9 @@ Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c
>   test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags))
>   netif_wake_queue(dev);
>   +}
> -+
> +
> +-if (need_lock)
> +-spin_unlock_irqrestore(&priv->tx_lock, flags);
>   +static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
>   +{
>   +   struct ipoib_dev_priv *priv = netdev_priv(dev);
> @@ -96,9 +106,7 @@ Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c
>   +   _ipoib_ib_handle_tx_wc(dev, i);
>   +   } while (i++ != wr_id);
>   +   priv->tx_poll = i & (ipoib_sendq_size - 1);
> -
> --if (need_lock)
> --spin_unlock_irqrestore(&priv->tx_lock, flags);
> ++
>   +   if (unlikely(wc->status != IB_WC_SUCCESS &&
>   +wc->status != IB_WC_WR_FLUSH_ERR))
> 
> @@ -128,7 +136,7 @@ Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c
>}
> 
>int ipoib_poll(struct napi_struct *napi, int budget)
> -@@ -361,11 +372,63 @@ void ipoib_ib_rx_completion(struct ib_cq
> +@@ -361,11 +372,66 @@ void ipoib_ib_rx_completion(struct ib_cq
>   netif_rx_schedule(dev, &priv->napi);
>}
> 
> @@ -161,7 +169,8 @@ Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c
>   +   spin_lock_irqsave(&priv->tx_lock, flags);
>   +   if (((int)priv->tx_tail - (int)priv->tx_head < 0) &&
>   +   time_after(jiffies, dev->trans_start + 10) &&
> -+priv->tx_outstanding < ipoib_sendq_size) {
> ++priv->tx_outstanding < ipoib_sendq_size &&
> ++priv->own_ah) {
>   +   wrid = priv->tx_head & (ipoib_sendq_size - 1);
>   +   priv->tx_ring[wrid].skb = NULL;
>   +   if (post_zlen_send_wr(priv, wrid))
> @@ -173,8 +182,7 @@ Index: ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c
>   +   }
>   +   poll_tx(priv);
>   +   spin_unlock_irqrestore(&priv->tx_lock, flags);
> -
> --poll_tx(priv, 1);
> ++
>   +   mod_timer(&priv->poll_timer, jiffies + HZ / 2);
>   +}
>   +
> @@ -189,13 +197,16 @@ Index: 
> ofed_kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c
>   +   if (!post_zlen_send_wr(priv, wrid)) {
>   +   ++priv->tx_head;
>   +   ++priv->tx_outstanding;
> -+}
> ++} else
> ++ipoib_warn(priv, "post_zlen failed\n");
> +
> +-poll_tx(priv, 1);
>   +   poll_tx(priv);
>   +   spin_unlock_irqrestore(&priv->tx_lock, flags);
> 

[ewg] [ANNOUNCE] management tarballs release

[Sasha Khapyorsky]

Hi,

There is a new release of the management (OpenSM and infiniband
diagnostics) tarballs available in:

http://www.openfabrics.org/downloads/management/

md5sum:

1c9764865b4c4f03529494f4272a7daf  libibcommon-1.0.8.tar.gz
879688a264e982600628dfa576c84cea  libibumad-1.1.7.tar.gz
de275d0aba15b587f00ac1e30ba8b0e8  libibmad-1.1.6.tar.gz
17678a2eaeb0cae9273b091da00cc7dc  infiniband-diags-1.3.6.tar.gz
92385653112cd915ec31542eb2d94714  opensm-3.1.9.tar.gz
0f43afda110cfc285ef89637fcbf6ed7  opensm-3.2.0.tar.gz

opensm-3.1.9 is recent OFED version. opensm-3.2.0 is recent master.
The rest is the same.

Sasha
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] [Fwd: Re: [PATCH 2.6.25] cxgb3: Handle ARP completions that mark neighbors stale.]

[Steve Wise]

Vlad, please pull this fix into ofed-1.3.  This resolves bug 890.

Pull from:

git://git.openfabrics.org/~swise/ofed-1.3 ofed_kernel

Thanks,

Steve.



 Original Message 
Subject: Re: [PATCH 2.6.25] cxgb3: Handle ARP completions that mark 
neighbors stale.

Date: Mon, 11 Feb 2008 11:09:25 -0500
From: Jeff Garzik <[EMAIL PROTECTED]>
To: Steve Wise <[EMAIL PROTECTED]>
CC: [EMAIL PROTECTED], [EMAIL PROTECTED], 
[EMAIL PROTECTED], [EMAIL PROTECTED]

References: <[EMAIL PROTECTED]>

Steve Wise wrote:

cxgb3: Handle ARP completions that mark neighbors stale.

When ARP completes due to a request rather than a reply the neighbor is
marked NUD_STALE instead of reachable (see arp_process()).  The handler
for the resulting netevent needs to check also for NUD_STALE.

Failure to use the arp entry can cause RDMA connection failures.

Signed-off-by: Steve Wise <[EMAIL PROTECTED]>


applied

___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] Tough time with ibverbs

[Matthew Small]

Hello,

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 would be much
appreciated.

-Matt
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] OFED meeting agenda for Feb 11

[Tziporet Koren]

> This is the agenda for today:
> 
> 1. OFED 1.3-rc4 testing status - all
2. Critical bugs review:

903   blo
P3  Othe[EMAIL PROTECTED]   mvapich fails to
build on Redhat EL5 - update 1  
905   blo
P3  RHEL[EMAIL PROTECTED]   scale-up issue
with rdma_cm, requests rejected when exceed...  
890   blo
P1  Othe[EMAIL PROTECTED]   Intermittent
stall starting MPI jobs due to lost NetEvent
874   cri
P3  Othe[EMAIL PROTECTED]   Intel MPI (IMB
test) hangs intermittently on the qlogic HCA
895   cri
P3  Othe[EMAIL PROTECTED]   kernel panic
while running multiple test on sdp  
846   cri
P2  RHEL[EMAIL PROTECTED]   SDP crash on
RHEL5 ppc64 running netserver   
906   cri
P1  All [EMAIL PROTECTED]   fail to destroy
ipoib rx QP 
760   maj
P3  All [EMAIL PROTECTED]   UDP performance
on Rx is lower than Tx  
894   maj
P2  SLES[EMAIL PROTECTED]   IPoIB
connectivity lost during heavy testing on memfree   

3. Open discussion


> Tziporet
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] OFED teleconference today

[Tziporet Koren]

Friendly reminder: the OFED teleconference is today (Monday, 11 
February, 2008).


Next few teleconferences:
--
All are at noon US eastern / 9am US Pacific / 7pm Israel
1. Monday, Feb 11, code 210020028 <== today
2. Tuesday, Feb 19, code 210020028


Dial-in information:

US/Canada:  +1.866.432.9903
India:  +91.80.4103.3979
Israel: +972.9.892.7026
Others: http://cisco.com/en/US/about/doing_business/conferencing/



Tziporet

___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

Re: [ewg] OFED teleconference today

[Moni Levy]

Tziporet,
I won't be able to join the meeting today. Olga Shern will be
joining instead.

Best regards,
Moni

On 2/4/08, Jeff Squyres <[EMAIL PROTECTED]> wrote:
> Friendly reminder: the OFED teleconference is today (Monday, 4
> February, 2008).
>
> Next few teleconferences:
> -
> All are at noon US eastern / 9am US Pacific / 7pm Israel
>
> 1. Monday, Feb 4, code 210020028
> 2. Monday, Feb 11, code 210020028
> 3. Monday, Feb 18, code 210020028
>
> Dial-in information:
> 
> US/Canada:  +1.866.432.9903
> India:  +91.80.4103.3979
> Israel: +972.9.892.7026
> Others: http://cisco.com/en/US/about/doing_business/conferencing/
>
> --
> Jeff Squyres
> Cisco Systems
>
> ___
> ewg mailing list
> ewg@lists.openfabrics.org
> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
>
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

Re: [ewg] [OFED-1.3rc4 PATCH 0/3] IB/srp: bring OFED SRP initiator up-to-date with 2.6.25rc

[David Dillow]

On Mon, 2008-02-11 at 11:31 +0200, Tziporet Koren wrote:
> David Dillow wrote:
> > [resending to ewg from general as suggested by Or Gerlitz; you are
> > probably getting this twice, and I apologize.]
> >
> > This series of patches adds the fixes and enhancements that have been
> > applied to the 2.6.25-to-be kernel. The first patch to respect the
> > credit limits is a correctness issue, and will avoid performance cliffs
> > on hardware in the field. The rest make sysadmin's lives easier, but can
> > be held back if need be.
> >
> > These have been built and lightly tested against the ofed_kernel git
> > repository as of Friday morning.
> >   
> Have you tested with make_dist.sh to make sure its compiles on all OSes 
> and no backports needed?

No; I'll look into what it'll take to set up a test environment here,
since I don't have ready access to all of the supported OSes.

The first patch, respect credit limit, depends on nothing outside of the
SRP initiator, and shouldn't pose a backport issue at all. The other two
do call functions outside of the initiator and will need to be
validated.

Only the first patch is a pure bug fix, and I think we should push for
it -- the others can wait.
-- 
Dave Dillow
National Center for Computational Science
Oak Ridge National Laboratory
(865) 241-6602 office


___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[Fwd: Re: [ewg] [PATCH ofed-1.3] rdma_lat: Add -m --max-inline option to support devices with different]

[Steve Wise]

Doran/Oren,

Can you please push this into ofed-1.3?

Thanks,

Steve.

 Original Message 
Subject: Re: [ewg] [PATCH ofed-1.3] rdma_lat: Add -m --max-inline option 
to support devices with different

Date: Thu, 07 Feb 2008 09:17:30 -0600
From: Steve Wise <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
CC: Oren Meron <[EMAIL PROTECTED]>,  Sagi Rotem 
<[EMAIL PROTECTED]>, [EMAIL PROTECTED]
References: <[EMAIL PROTECTED]> 
<[EMAIL PROTECTED]> 
<[EMAIL PROTECTED]> 
<[EMAIL PROTECTED]> 
<[EMAIL PROTECTED]>


Dotan Barak wrote:

Steve Wise wrote:

Oren Meron wrote:

Hi Steve,
rdma_lat is one of our older test. We use now the newer read_lat and
write_lat instead, which support inline specification in command line.
The only disadvantage of the new tests, is that they do not yet support
CMA.
Will it satisfy?



No because the RDMA CMA is required for iwarp devices.

Steve: i remember that you add the CMA to the rdma_* tests.

I don't want to be rude, what can you find some time to add CMA support 
to the other tests?

(then we won't have a good reason to continue support the rdma_* tests).

Thanks
Dotan


I guess I can do this.  But not in time for ofed-1.3.  If you will get
the rdma_lat change in for 1.3, then I'll get the other apps enabled for
iwarp for 1.4 or 1.3.x.

Sound like a plan?

Steve.

___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

Re: [ewg] [OFED-1.3rc4 PATCH 0/3] IB/srp: bring OFED SRP initiator up-to-date with 2.6.25rc

[Tziporet Koren]

David Dillow wrote:

[resending to ewg from general as suggested by Or Gerlitz; you are
probably getting this twice, and I apologize.]

This series of patches adds the fixes and enhancements that have been
applied to the 2.6.25-to-be kernel. The first patch to respect the
credit limits is a correctness issue, and will avoid performance cliffs
on hardware in the field. The rest make sysadmin's lives easier, but can
be held back if need be.

These have been built and lightly tested against the ofed_kernel git
repository as of Friday morning.
  
Have you tested with make_dist.sh to make sure its compiles on all OSes 
and no backports needed?


Thanks
Tziporet
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

Re: [ewg] OFED 1.3 RC4 release is available

[Tziporet Koren]

Moni Shoua wrote:


Hi,
Is it possible to get a list of changes (taken from git log) with each RC?
This could be very helpful for testers who want to focus their efforts.

thanks


  

See attachment

Tziporet


ofed-1.3_rc4-rc3.diff.tgz
Description: application/compressed
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

Re: [ewg] Re: address handles leakage report from ipoib

[Or Gerlitz]

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

Re: [ewg] Re: address handles leakage report from ipoib

[Eli Cohen]

Or Gerlitz wrote:

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
overnight, it was connected mode but I saw this in the past also for datagram 
mode, both
nodes have SLES10 SP1 with kernel 2.6.16.46-0.12-smp, 2 cpus four cores each, 
etc.

$server iperf -s -i 1
$client iperf -c $server -t 36000 -P 4 -i 1



Thanks Or - I will try to reproduce. Were you able to unload the module 
eventually?
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg

[ewg] Re: address handles leakage report from ipoib

[Or Gerlitz]

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
overnight, it was connected mode but I saw this in the past also for datagram 
mode, both
nodes have SLES10 SP1 with kernel 2.6.16.46-0.12-smp, 2 cpus four cores each, 
etc.

$server iperf -s -i 1
$client iperf -c $server -t 36000 -P 4 -i 1


Or.
___
ewg mailing list
ewg@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg