Well when I try to compile I get an error message saying i_private is not a
member of the inode structure when trying to compile the ulp/iboip and the
ib_ipath modules. I'm using the 2.6.18-8 kernel src from kernel.org.
Any reasons why I would be getting this error message?
Thanks,
Jeff
I don't think you are actually using the kernel from kernel.org:
we test-build these nightly.
Quoting Jeffrey Wong [EMAIL PROTECTED]:
Subject: RE: Trouble installing OFED1.2 with kernel
Well when I try to compile I get an error message saying i_private is not a
member of the inode structure
Sasha Khapyorsky wrote:
On 01:11 Wed 25 Apr , Yevgeny Kliteynik wrote:
Michael S. Tsirkin wrote:
Since you seem to do a strcat which does an anyway, how about, for example:
- sprintf( buf_line1,%s 0x%01x |,
- buf_line1, p_vla_tbl-vl_entry[i].vl);
+ sprintf(
I have downloaded the kernel src from
http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.18.8.tar.gz
I have gunzip and untarred the directory.
In the file linux-2.6.18.8/include/linux/fs.h. Here is the structure
definition of inode. When I look below the i_private ptr does not exist.
Quoting Jeffrey Wong [EMAIL PROTECTED]:
Subject: RE: Trouble installing OFED1.2 with kernel
I have downloaded the kernel src from
http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.18.8.tar.gz
I have gunzip and untarred the directory.
In the file linux-2.6.18.8/include/linux/fs.h.
On Tue, 2007-05-01 at 10:22 +0300, Michael S. Tsirkin wrote:
Quoting Jeffrey Wong [EMAIL PROTECTED]:
Subject: RE: Trouble installing OFED1.2 with kernel
I have downloaded the kernel src from
http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.18.8.tar.gz
I have gunzip and untarred
On Tue, 2007-05-01 at 03:59 -0400, Jeffrey Wong wrote:
I see. So I should have never renamed my kernel from 2.6.18.8 to
2.6.18.8-el5_x86. So this will install once I rename my kernel back to
2.6.18.8?
Thanks for the info.
Jeff
Yes,
The pattern 2.6.18-*el5* used by configure script
reproduce And sadly some order others might afterwards say that is the
ultimateNo. Just sponge be there. suddenly I don't think cheerful waste you'll
be dis fraternal hurt middle Hey beyond Carl, wait upThe guard jewel turned a
goat cinerary couple of blow pages on his clip-
You station
FYI.
I think we want to update - the mmap fix looks important enough.
Sasha?
- Forwarded message from Junio C Hamano [EMAIL PROTECTED] -
Subject: [ANNOUNCE] GIT 1.5.1.3
Date: Tue, 1 May 2007 06:08:58 +0300
From: Junio C Hamano [EMAIL PROTECTED]
The latest maintenance release GIT 1.5.1.3
I don't believe so. I just sent out modifications to the Ammasso driver on
another thread that might clear this up. The modifications to the driver
should show how these new verbs could be used.
Quoting [EMAIL PROTECTED] [EMAIL PROTECTED]:
Subject: [PATCH] infiniband: add support for
--- The following addresses had delivery problems ---
[EMAIL PROTECTED] (5.1.1 [EMAIL PROTECTED]... User unknown)
Original-Recipient: openhouse@cslab.kecl.ntt.co.jp
Action: failed
Diagnostic-Code: smtp;
550 5.1.1 openhouse@cslab.kecl.ntt.co.jp... User unknown
Remote-MTA: 129.60.199.15
On Fri, 27 Apr 2007, Arlin Davis wrote:
Fixes build problems with ia64 and RHEL5 with atomic operations.
Patch was tested on ia64 RHEL4 and RHEL5 using dtest/dapltest.
James, can you review this before I push.
Looks good.
___
general mailing
On 5/1/2007 1:57 AM, Bill Fink wrote:
On Mon, 30 Apr 2007, Rick Jones wrote:
Ethtool -i on the interface reports 1.2.0 as the driver version.
Perhaps it would be useful to have different version strings for
the in-kernel Linux version and the Myricom externally provided
version.
On 4/26/07, Roland Dreier [EMAIL PROTECTED] wrote:
Let's do it over query_pkey/query_port for now.
Long term providers will just optimize these I think.
How ? Caching at device driver level ?
Yes... for the most part, it should be much easier to do within the
driver. For example
On Wed, 2007-04-25 at 18:58 -0700, Roland Dreier wrote:
Is there a differentiation between multiple CQE's being in the CQ
vs. CQE's being arriving into the CQ when using completion
notifications?
For example, assume I have the following order of events:
2 CQEs arrive
On Tue, 2007-05-01 at 06:57 +0300, Michael S. Tsirkin wrote:
Quoting [EMAIL PROTECTED] [EMAIL PROTECTED]:
Subject: [PATCH] infiniband: add support for invalidate stag
Patch to add support for the iWARP verbs SEND with INV and SEND with SE
and INV.
---
The code looks correct.
I'd make the msg_size lines just one statement:
msg_size = sizeof(struct c2wr_send_req) +
sizeof(struct c2_data_addr) * ib_wr-num_sge;
Have you tested that it works?
Steve.
On Tue, 2007-05-01 at 10:20 -0400, [EMAIL PROTECTED] wrote:
I don't believe that we can make it into one line as Roland pointed out
earlier - it introduces an accumulation bug because it is within a while
loop.
Mikkel Hagen
Project Assistant - Fibre Channel/SAS/SATA Consortiums
Research and Development Engineer - iWARP Consortium
No, the accumulation bug was because you were always doing a +=.
msg_size = sizeof(struct c2wr_send_req) +
sizeof(struct c2_data_addr) * ib_wr-num_sge;
This always assigns to msg_size.
___
general mailing list
Good point!
--- linux-2.6.21.1/drivers/infiniband/hw/amso1100/c2_qp.c 2007-04-30
13:12:54.0 -0400
+++ linux-2.6.21.1/drivers/infiniband/hw/amso1100/c2_qp.c 2007-05-01
14:04:07.0 -0400
@@ -810,16 +810,25 @@ int c2_post_send(struct ib_qp *ibqp, str
All,
SFW has completed the SRP multipath failover stress test on the following
builds and OSes.
* OFED 1.2-rc2 - SLES10 x86 and SLES10 x86_64
* 04192007-0600 - RHEL 5 x86_64.
The I/O was running on each platform for more than 10 hours during the
failovers. No I/O error
Hello,
I have successfully run the ./install.sh script with kernel
2.6.18-8.1.1.el5
I did not reboot the machine.
After installing and configuring the ports using the defaults, I tried
to execute the command:
/etc/init.d/openibd start
I have truncated the errors to show an example.
Try rebooting, and see if it still happens.
Scott
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Jeffrey Wong
Sent: Tuesday, May 01, 2007 1:11 PM
To: general@lists.openfabrics.org
Subject: [ofa-general] Errors after install for openibd start
I just noticed a false prereq problem when running build.sh from
ofed-1.2 on rhel5 _with_ a kernel.org kernel installed. Here's the
issue:
build_env.sh checks the existence of /etc/redhat-release determine if
the distro is from redhat. But if that file exists -and- the kernel
`uname -r` ends in
He added the invalidate struct to the union part of the ib_send_wr. Its
analogous to the rdma struct in that union in that its additional values
passed in the send wr for an iwarp send w/invalidate and send-se
w/invalidate. The seems reasonable to me...
I have re-read the patch, and I agree
On Tue, 2007-05-01 at 15:46 -0500, Steve Wise wrote:
I think build_env.sh shouldn't use the kernel uname to determine if the
distro is redhat5. Rather, it should grok the contents
of /etc/redhat-release to determine if its rhel5 or not...
Is this worthy of fixing for 1.2?
Maybe this?
- Forwarded message from Vu Pham [EMAIL PROTECTED] -
Subject: ipath-23-srp-limit-queued-commands.patch,
Date: Tue, 1 May 2007 19:30:48 +0300
From: Vu Pham [EMAIL PROTECTED]
Hi,
This patch
kernel_patches/fixes/ipath-23-srp-limit-queued-commands.patch
which change
Would one of you have the bandwidth to review the IPOIB CM NOSRQ patch
(v3) that I submitted last week.
Pradeep
[EMAIL PROTECTED]
___
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To
We will upgrde after Sonoma.
Sasha
From: [EMAIL PROTECTED] on behalf of Michael S. Tsirkin
Sent: Tue 5/1/2007 3:53 PM
To: general@lists.openfabrics.org
Subject: [ofa-general] Fwd: [ANNOUNCE] GIT 1.5.1.3
FYI.
I think we want to update - the mmap fix looks
Quoting Pradeep Satyanarayana [EMAIL PROTECTED]:
Subject: please review IPOIB CM NOSRQ patch
Would one of you have the bandwidth to review the IPOIB CM NOSRQ patch
(v3) that I submitted last week.
Pradeep
[EMAIL PROTECTED]
OK, but could you please send a version that isn't
Scott Shaw wrote:
When will the general release of ofed v1.2 be available? Also is the OS
requirement going to be SUSE10 SP1?
Will ofed v1.2 work with SUSE10 without service packs?
Thanks,
Scott
General release is planed for May 15, but the actual date depend on
stability and some
On 4/30/2007 2:12 PM, Rick Jones wrote:
Speaking of defaults, it would seem that the external 1.2.0 driver
comes with 9000 bytes as the default MTU? At least I think that is
what I am seeing now that I've started looking more closely.
rick jones
That's the same for the in-kernel-tree
Loic Prylli wrote:
On 4/30/2007 2:12 PM, Rick Jones wrote:
Speaking of defaults, it would seem that the external 1.2.0 driver
comes with 9000 bytes as the default MTU? At least I think that is
what I am seeing now that I've started looking more closely.
rick jones
That's the same for
Sean,
This patch regresses rping. I failed to test it on AMD64-AMD64 (ie
like endian systems). I will provide another patch shortly, or we can
undo the broken rping patch for -rc3. Whatever you think is best.
Sorry for this!
Steve.
On Wed, 2007-04-25 at 10:34 -0500, Steve Wise wrote:
Hello,
I am trying to do an OFED 2.1 install for all the modules now.
I was able to compile and install the Basic install and now I am trying
to install the all selection part.
When I try to install with this selection I am getting an error when
compiling the libsdp directory.
It looks
Hello,
I am using kernel 2.6-18.8.1.1.el5 x86_64
I have changed the build_env.sh to have the build_32bit=-1
Thanks in advance.
Jeff
When installing all modules I am getting the following errors.
+ make -C /lib/modules/2.6.18-8.1.1.el5/build modules
Hello,
I am using kernel 2.6-18.8.1.1.el5 x86_64
I have changed the build_env.sh to have the build_32bit=-1
Thanks in advance.
Jeff
When installing all modules I am getting the following errors.
+ make -C /lib/modules/2.6.18-8.1.1.el5/build modules
With IPoIB HA (both ipoibtools and ib-bonding), I am seeing slow IPoIB
CM HA failover, and eventually IPoIB stops working after enough
failovers. I am running netperf -D traffic between two IPoIB HA hosts,
while flipping the 4 host IB ports one at a time (port 1 down, sleep,
port 1 up, sleep,
Quoting Roland Dreier [EMAIL PROTECTED]:
Subject: Re: Requesting CQ notifications
Is there a differentiation between multiple CQE's being in the CQ
vs. CQE's being arriving into the CQ when using completion
notifications?
For example, assume I have the following order of
No, this is not right. SDP has better latency and better throughput
than IPoIB CM, but also uses more CPU.
Scott Weitzenkamp
SQA and Release Manager
Server Virtualization Business Unit
Cisco Systems
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
Any idea why this error? We see this error when we use FMR? Are there
any special setting that HCA needs to work with FMR?
[Batwara, Ashish]
___
general mailing list
general@lists.openfabrics.org
41 matches
Mail list logo