On 10/16/05 08:26, Steve Wooding wrote:
> Hi there,
>
> I'm trying to make a QP connection using the CM, but the active side
> cannot get to the RTR state. ibv_modify_qp returns errorno -22, invalid
> argument.
>
> What would the best way to find out exactly what the error is and help
> me fix my
On 09/27/2005 09:01 PM, Roland Dreier wrote:
> Linus, please pull from
>
> master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband.git
> for-linus
>
> This tree is also available from kernel.org mirrors at:
>
> rsync://rsync.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.gi
On 06/27/05 13:32, Michael S. Tsirkin wrote:
> Does this mean the bandwidth is 200-300 MByte/sec?
[EMAIL PROTECTED]:~# ./fast_test.pl 10
starting sends
0 messages/sec (0 megabits/sec)
21845 messages/sec (2730 megabits/sec)
26214 messages/sec (3276 megabits/sec)
30247 messages/sec (3780 megabits/s
On 06/27/05 13:32, Michael S. Tsirkin wrote:
> Quoting r. Jeff Carr <[EMAIL PROTECTED]>:
>
>>Here is an updated version and a simple perl script that tests it's
>>performance. With 2K messages, these were the performance numbers
>>between 2 systems (3.6ghz Xe
On 06/22/05 09:43, Jeff Carr wrote:
> On 06/21/2005 12:50 PM, Roland Dreier wrote:
>
>
>>What happens if you try replacing the send_flags line with the one you
>>have commented out?
>>
>>+ // send_wr.send_flags = IB_SEND_SIGNALED;
>
>
> Thanks,
On 06/23/2005 09:52 AM, Talpey, Thomas wrote:
> At 12:43 PM 6/22/2005, Jeff Carr wrote:
>
>>On 06/21/2005 12:50 PM, Roland Dreier wrote:
>>
>>
>>>What happens if you try replacing the send_flags line with the one you
>>>have commented out?
>>&g
On 06/21/2005 12:50 PM, Roland Dreier wrote:
> What happens if you try replacing the send_flags line with the one you
> have commented out?
>
> + // send_wr.send_flags = IB_SEND_SIGNALED;
Thanks, you are correct. IB_SEND_SIGNALED gives me the behavior I was
expecting.
Jeff
_
On 06/21/05 12:50, Roland Dreier wrote:
> Jeff> However, something still goes wrong after MAX_CQ + 1
> Jeff> transfers, just like the cmtest code. If anyone has any
> Jeff> ideas on what is wrong with what I am doing here that would
> Jeff> be helpful. It must be something really si
d to recieve into.
Enjoy,
Jeff
Signed-off-by: Jeff Carr <[EMAIL PROTECTED]>
diff -Naur spts.old/Kconfig trunk/src/linux-kernel/infiniband/ulp/spts/Kconfig
--- spts.old/Kconfig1969-12-31 16:00:00.0 -0800
+++ trunk/src/linux-kernel/infiniband/ulp/spts/Kconfig 2005-06-21
09:34:03.00
On 06/13/05 15:35, James Lentini wrote:
I don't believe Jeff is using DAPL.
That's true; I just need something simple. There is enough abstraction
and layers already so I don't want to add another one. :)
Jeff
___
openib-general mailing list
openi
On 06/13/05 14:56, Fab Tillier wrote:
From: Jeff Carr [mailto:[EMAIL PROTECTED]
Sent: Monday, June 13, 2005 2:52 PM
On 06/13/05 14:12, Roland Dreier wrote:
Jeff> On the recieving end of a request, when the handler gets
Jeff> called and the cm-id gets passed to your handler (ak
On 06/13/05 14:54, Roland Dreier wrote:
Jeff> OK, thanks. I checked and it does look like that is what is
Jeff> happening. I'll set the context for the cm_id I pass into
Jeff> ib_cm_listen to null. Assuming of course, that the cm_id
Jeff> context for that one doesn't matter.
The
On 06/13/05 14:12, Roland Dreier wrote:
Jeff> On the recieving end of a request, when the handler gets
Jeff> called and the cm-id gets passed to your handler (aka when
Jeff> the handler gets IB_CM_REQ_RECEIVED) it seems cm-id->context
Jeff> isn't null. Is there some useful data in
On the recieving end of a request, when the handler gets called and the
cm-id gets passed to your handler (aka when the handler gets
IB_CM_REQ_RECEIVED) it seems cm-id->context isn't null. Is there some
useful data in there? If not, I think it makes more sense (and makes
code easier to write to
On 06/02/05 15:27, Hal Rosenstock wrote:
On Thu, 2005-06-02 at 18:09, Jeff Carr wrote:
Is there a simple way to discover the lid values of other systems?
Locally, you can run /usr/local/ib/bin/ibstatus or ibstat.
Remotely, you would need to know the remote GID (subnet prefix + GUID)
and
On 06/01/05 14:43, William Jordan wrote:
Has anyone seen ib_send_cm_req() return -22?
I'm not sure what you are testing with, Jeff, but I ran into the same
problem the first time I tried to use ucm_simple. The source and
destination lid and guid are embedded in the source, and need to be
modif
On 05/31/05 16:30, Sean Hefty wrote:
Has anyone seen ib_send_cm_req() return -22?
I believe that this is a timeout error, possibly indicating that the
server
side of the connection wasn't running. You may also want to verify the
slid
and dlid are correct for your configuration.
Don't
Has anyone seen ib_send_cm_req() return -22?
I seem to be stuck with this error today. I've seen it in the past; and
it seems to have gone away without my being able to discern a conclusive
reason. Usually rebooting the systems will cause it to work again.
This go around, It's really refusing
On 05/25/05 10:18, Sean Hefty wrote:
I've been able to hit an issue in the MAD layer that might be related.
I still do not know the root cause, however.
Just to chim in with the same. I've been able to generate a potentially
similar error on bootup. I put the openib svn on top of 2.6.12-rc4-g
On 05/26/05 14:35, Sean Hefty wrote:
If the code is based on cmpost, this is likely the case. Cmpost creates
the QP with IB_SIGNAL_REQ_WR, but does not set the IB_SEND_SIGNALED flag.
Since people are refering to this as a sample, I will look at changing
the code to avoid confusing people.
A
On 05/26/05 14:35, Sean Hefty wrote:
Roland Dreier wrote:
Are you seeing any send completions on the CQ attached to the send
queue?
No.
One possibility is that you're not generating any send
completions because you created your CQ with a sq_sig_type of
IB_SIGNAL_REQ_WR and then posting se
On 05/26/05 14:17, Jeff Carr wrote:
When I run my test, the output for the server and client look similar.
For the server, it first runs: ib_post_recv() for N times (N=5 here)
sorry s/ib_post_recv/ib_post_send/
ib_mthca :09:00.0: SQ 050406 (1 head, 0 tail, 5 max)
ib_mthca :09:00.0
In my code (based on cmtest) I was using a single CQ for both send and
recv. I broke them into seperate queues to try to figure out why I could
only do N transfers for each CQ that I created.
I added some debugging output to mthca_cq.c to dump out the SQ and RQ
for each QP.
When I run my tes
Sean Hefty wrote:
In this case you won't know what's on a cqe until you poll the
completion.
The cmpost test only poll's on the completion on the client side. I'll
change it to run ib_poll_cq() on the server also. Now I'm so confused
I'm not sure why I wanted this in the first place though :)
Sean Hefty wrote:
Jeff Carr wrote:
May 5 16:31:50 localhost kernel: ib_mthca :09:00.0: 1a0084/0:
error CQE -> QPN 1a0406, WQE @ 0042
May 5 16:31:50 localhost kernel: [ 0] 001a0406
May 5 16:31:50 localhost kernel: [ 4] 1aed
May 5 16:31:50 localhost kernel: [ 8] 0004
Libor Michalek wrote:
OK. Well I would rather make something generic. Besides, wasn't there
some MS patent issue? The last thread on the subject that I read kinda
made it sound like you were going to look into the issue and respond.
Maybe I missed the response; there's a lot of mail in the archi
Grant Grundler wrote:
4K -> 1.8 GB/s
16k -> 3.3 GB/s
64k -> 3.8 GB/s
This seems reasonable.
IIRC the ZX1 chipset has 6GB/s backplane but one CPU can only drive ~4GB/s.
I have a E7501. Thanks for running this test. I'd not looked so closely
at this before or been up to the wall against it where it
Roland Dreier wrote:
The most interesting optimization available is implementing the IPoIB
connected mode draft, although I don't think it's as easy as Vivek
indicated -- for example, I'm not sure how to deal with having
different MTUs depending on the destination.
Thank you for that reference. I'l
Grant Grundler wrote:
We..Looks like I'm wrong. Previous email on this thread
suggested it's possible by people who know alot more about
it than I do. But I'm still concerned it's going to affect
latency.
Maybe that's one reason why NAPI was made a compile time option? Anyway,
it might be just
Grant Grundler wrote:
[EMAIL PROTECTED]:/# dd if=/dev/shm/test of=/dev/null bs=4K
196608+0 records in
196608+0 records out
805306368 bytes transferred in 0.628504 seconds (1281306571 bytes/sec)
Yeah. Sounds like there is. Should be able to do several GB/s like that.
I suppose it's possibly an issu
Grant Grundler wrote:
If it's NAPI that means nothing, here's probably the best summary:
http://lwn.net/Articles/30098/
Cool; I see now.
But IPoIB can't really implement NAPI since it's sending work to
a shared HCA.
Hmm. I'm not knowledgeable to know why; I'll have to take your word for
i
Grant Grundler wrote:
vmstat doesn't tell you where the time is being spent.
>
Get a profile or try out the beta Pentium M or AMD64 perfmon support
Yes, I would but I didn't think I could because I have Xeon's.
You also want to explore "netperf -C" option that' available with
netperf 2.4.0-rc1 (See
Roland Dreier wrote:
Jeff> (side note: it would seem IPoIB could be re-written to
Jeff> dramatically improve it's performance).
Out of curiousity, what would the rewrite change to obtain better
performance?
Could (or would it help if) the MTU was increased to something much
larger than 204
Roland Dreier wrote:
Jeff> (side note: it would seem IPoIB could be re-written to
Jeff> dramatically improve it's performance).
Out of curiousity, what would the rewrite change to obtain better
performance?
I'm just speculating that it could be rewritten to improve performance.
There were m
Libor Michalek wrote:
On Mon, May 16, 2005 at 03:26:57PM -0700, Jeff Carr wrote:
It seems to me it would be useful to have a simple ib_send_page() function.
This is essentially what I'm going to end up writing for what I need IB
to do. If there is anyone else that has similar needs or inte
It seems to me it would be useful to have a simple ib_send_page() function.
This is essentially what I'm going to end up writing for what I need IB
to do. If there is anyone else that has similar needs or interests I'd
be happy to work with them.
The CM works well enough to allow me to initiate
Sean Hefty wrote:
You're looking in the right place.
I'll add some code to format the data and verify that it is making it to
the remote side, and see if I can track down any problems. (This will
need to wait until tomorrow for me though...)
- Sean
May 5 16:31:50 localhost kernel: ib_mthca 000
Sean Hefty wrote:
In theory, yes. Though I've never verified the data. Note that the
code uses a single data buffer for the sends and receives, so it can get
overwritten.
Yep, noticed that. I was going to set about changing it to add a few
more (I don't need that many) but then noticed that no
Sean Hefty wrote:
Within the SVN repository, your best bet for finding things is staying
withing the gen2 branch. For a relatively simple example that does
what you mention above, try:
https://openib.org/svn/gen2/utils/src/linux-kernel/infiniband/util/cmpost/
This is a simple CM test program f
In trying to build the modules from:
https://openib.org/svn/gen2/utils/src/linux-kernel/infiniband/util/
I would like (need in this case) to use the newest IB kernel code. There
seem to be three places in the gen2 tree for this code.
Is this correct: ?
the newest code is here:
https://openib.org/
Sean Hefty wrote:
Within the SVN repository, your best bet for finding things is staying
withing the gen2 branch.
I figured that, but when I check out the repository with svn it doesn't
seem to keep the correct dates on the files (it doesn't preserve the
mtime). That makes it hard to figure o
zed for a
new user; so if there is some starting code that may be of use, that is
what I was looking for.
Also, if there is anyone that would be willing to work on this problem
on a contractual basis (code will be GPL'd) then please contact me.
Thanks,
Jeff Carr
Some choice quotes from
James Lentini wrote:
I see it too.
On Tue, 26 Apr 2005, Tom Duffy wrote:
Has anybody else noticed that openib.org doesn't seem to render
properly on Firefox/Linux? Check out this screenshot.
Not for me. Perhaps reload fixes it?
Jeff
(using debian sid - firefox 1.0.1)
__
43 matches
Mail list logo