Roland Dreier wrote:
>OK, just a dumb oversight on my part. The change below (already
>checked in) fixes it for me:
>
>--- infiniband/core/uverbs_cmd.c (revision 8055)
>+++ infiniband/core/uverbs_cmd.c (working copy)
>@@ -1123,6 +1123,12 @@ ssize_t ib_uverbs_create_qp(struct ib_uv
>
Here are the kernel driver changes that go with the user library
changes just posted.
Index: src/linux-kernel/infiniband/hw/ipath/ipath_qp.c
===
--- src/linux-kernel/infiniband/hw/ipath/ipath_qp.c (revision 8021)
+++ src/linux-ker
Here are the diffs Roland requested for the ipath driver changes
to mmap the completion and receive queues into the user library.
This isn't quite the final version though since I need to implement
QP receive queue resizing and some version checking/handling.
Index: src/userspace/libipathverbs/sr
On Thu, 2006-06-15 at 14:56 -0700, Roland Dreier wrote:
> Ralph> libmthca uses a single shared page which is created at
> Ralph> driver open time. I'm mmaping vmalloc memory created at
> Ralph> ibv_create_cq(), qp, srq time so I need a way to return the
> Ralph> offset to ipathverb
OK, I rolled this into the cma patch in for-2.6.18 branch.
Unfortunately this means that all of the patches in that branch are
rebased, so you'll have to repull my tree if you're tracking it.
- R.
___
openib-general mailing list
openib-general@openib.o
Roland wrote,
>OK, just a dumb oversight on my part. The change below (already
>checked in) fixes it for me:
Great thanks for the quick response,
woody
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openi
Roland, can you pick up this patch for 2.6.18?
Thanks - committed in 8057.
- Sean
>If you listen on 0.0.0.0, then the backlog isn't passed down to the
>devices because its not stored in the id_priv struct before calling
>cma_listen_on_all(). See cma_list_on_dev() which uses
>id_priv->backlog...
OK, just a dumb oversight on my part. The change below (already
checked in) fixes it for me:
--- infiniband/core/uverbs_cmd.c(revision 8055)
+++ infiniband/core/uverbs_cmd.c(working copy)
@@ -1123,6 +1123,12 @@ ssize_t ib_uverbs_create_qp(struct ib_uv
goto err_copy
>The cma/verbs consumer can't just ignore the event since its qp state is
>still RTR which means an attempt to tx replying the rx would fail.
In most cases, I would expect that the IB CM will eventually receive the RTU,
which will generate an event to the RDMA CM to transition the QP into RTS. Th
Thanks, reproduced it locally.
___
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
This is the normal output for rping...
The status error on the completion is 5 (FLUSHED), which is normal.
Steve.
On Thu, 2006-06-15 at 17:24 -0400, amith rajith mamidala wrote:
> Hi,
>
> With the latest rping code (Revision: 8055) I am still able to see this
> race condition.
>
> server side
Roland Dreier wrote:
>Roland> Hmm, any further clue where in ibv_destroy_cq() it's
>Roland> stuck? Is it doing down_write() or something?
>
>Can you send me full sysrq-t output when it gets stuck?
>
>Thanks...
>
>
>
I just added ibv_destroy_cq() to ibv_rc_pingpong test.
Here's the outp
> > Now that I've looked more into this, I'm not sure there's a
> > simple way for the IWCM to copy the pdata on the upcall.
> > Currently, the IWCM's event upcall, cm_event_handler(),
> > simply queues the work for processing on a workqueue thread.
> > So there's no per-event logic at all there.
>
Ralph> libmthca uses a single shared page which is created at
Ralph> driver open time. I'm mmaping vmalloc memory created at
Ralph> ibv_create_cq(), qp, srq time so I need a way to return the
Ralph> offset to ipathverbs.so to then pass to mmap().
Hmm... it seems simpler to have us
[EMAIL PROTECTED] wrote:
> On Thu, 2006-06-15 at 08:41 -0500, Steve Wise wrote:
>> On Wed, 2006-06-14 at 20:35 -0500, Bob Sharp wrote:
>>
+void c2_ae_event(struct c2_dev *c2dev, u32 mq_index) {
+
>>
>>
>>
+ case C2_RES_IND_EP:{
+
+ struct c2wr_ae_connection_re
On Thu, 2006-06-15 at 13:55 -0700, Roland Dreier wrote:
> > I am working on a ipathverbs.so version of ibv_poll_cq(),
> > ibv_post_recv(), and ibv_post_srq_recv() which mmaps the
> > queue into user space. I found that I needed to modify the
> > core libibverbs and kernel uverbs code in order
Roland> Hmm, any further clue where in ibv_destroy_cq() it's
Roland> stuck? Is it doing down_write() or something?
Can you send me full sysrq-t output when it gets stuck?
Thanks...
___
openib-general mailing list
openib-general@openib.org
http
Arlin> I see the same thing running the uDAPL test
Arlin> (dapl/test/dtest). I am running a 2.6.16 kernel and svn8805
Arlin> and it appears to be deadlocked (uninterruptible sleep) in
Arlin> the ibv_destroy_cq() call. This all worked fine on
Arlin> svn7843; my last update on th
Hi,
With the latest rping code (Revision: 8055) I am still able to see this
race condition.
server side:
[EMAIL PROTECTED] examples]$ ./rping -s -vV -C10 -S26 -a 0.0.0.0 -p 9997
server ping data: rdma-ping-0: ABCDEFGHIJKL
server ping data: rdma-ping-1: BCDEFGHIJKLM
server ping data: rdma-ping-2:
Woodruff, Robert J wrote:
> It appears that processes are not exiting cleanly on SVN7946 trunk
> backported to 2.6.9-34 EL.
>
> They seem to be stuck in a state of "DL" and I cannot even attach to them
> wil gdb or kill them with a kill -9.
>
> [EMAIL PROTECTED] core]# ps -uax | grep IMB
> woody
> I am working on a ipathverbs.so version of ibv_poll_cq(),
> ibv_post_recv(), and ibv_post_srq_recv() which mmaps the
> queue into user space. I found that I needed to modify the
> core libibverbs and kernel uverbs code in order to return
> the information I need from ib_ipath to the ipathve
Sean, I think this is a bug, eh?
If you listen on 0.0.0.0, then the backlog isn't passed down to the
devices because its not stored in the id_priv struct before calling
cma_listen_on_all(). See cma_list_on_dev() which uses
id_priv->backlog...
Signed-off-by: Steve Wise <[EMAIL PROTECTED]>
--
I am working on a ipathverbs.so version of ibv_poll_cq(),
ibv_post_recv(), and ibv_post_srq_recv() which mmaps the
queue into user space. I found that I needed to modify the
core libibverbs and kernel uverbs code in order to return
the information I need from ib_ipath to the ipathverbs.so
library.
As one of the authors of IB and iWARP, I can say that both Roland and
Todd's responses are correct and the intent of the specifications.
The number of outstanding RDMA Reads are bounded and that is communicated
during session establishment. The ULP can choose to be aware of
this requirement (c
Hi Eitan,
On 15:19 Thu 15 Jun , Eitan Zahavi wrote:
> >>+/*
> >>+* PARAMETERS
> >>+* p_physp
> >>+* [in] Pointer to an osm_physp_t object.
> >>+*
> >>+* RETURN VALUES
> >>+* The pointer to the P_Key table object.
> >>+*
> >>+* NOTES
> >>+*
> >>+* SEE ALSO
> >>+* Port, Physical Port
> >>
Sounds like maybe the librdma.so that's installed is down-level...
Did you nuke the old one, rerun autgen.sh, configure, make, make install
in the librdmacm directory?
Stevo.
On Thu, 2006-06-15 at 12:10 -0400, amith rajith mamidala wrote:
> Hi,
>
> I have installed the latest infiniband stac
Hi,
I have installed the latest infiniband stack with 2.6.16.20 kernel.
I tested the installation using ibv_rc_pingpong and it works fine.
But, while trying to use rping, I get the following error:
librdmacm: couldn't open rdma_cm ABI version.
rdma_create_event_channel error 2
Any clues as to wh
On Thu, 15 Jun 2006, Or Gerlitz wrote:
> Sean Hefty wrote:
> > James Lentini wrote:
> >> The IBTA spec (volume 1, version 1.2) describes a communication
> >> established affiliated asynchronous event.
> >> We've seen this event delivered to our NFS-RDMA server and aren't sure
> >> what to do w
On Thu, 2006-06-15 at 08:41 -0500, Steve Wise wrote:
> On Wed, 2006-06-14 at 20:35 -0500, Bob Sharp wrote:
>
> > > +void c2_ae_event(struct c2_dev *c2dev, u32 mq_index)
> > > +{
> > > +
>
>
>
> > > + case C2_RES_IND_EP:{
> > > +
> > > + struct c2wr_ae_connection_request *req =
> > > +
I'll be on vacation through June 24. I won't be online most of the time.
--
MST
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo
On Wed, 2006-06-14 at 20:35 -0500, Bob Sharp wrote:
> > +void c2_ae_event(struct c2_dev *c2dev, u32 mq_index)
> > +{
> > +
> > + case C2_RES_IND_EP:{
> > +
> > + struct c2wr_ae_connection_request *req =
> > + &wr->ae.ae_connection_request;
> > + struct iw
On Thu, 2006-06-15 at 08:19, Eitan Zahavi wrote:
> >>+ p_pkey_tbl = osm_physp_get_mod_pkey_tbl( p_physp );
> >>+ if (! p_pkey_tbl)
> >
> >^
> > Is it possible?
> Yes it is ! I run into it during testing. The port did not have any pkey
> table.
PKey tables are optional
Quoting r. Sean Hefty <[EMAIL PROTECTED]>:
> Subject: RE: [openib-general] RFC: detecting duplicate MAD requests
>
> >Well the ACK for the direction switch is special, isn't it?
> >All I'm saying, let's pass it up to the application.
>
> I really don't think that this is the direction that we wan
Sasha Khapyorsky wrote:
> Hi Eitan,
>
> Some comments about the patch.
Thanks for the review.
The major point you bring up is the fact I intentionally impose the result of
the
pkey settings on the SMDB and not wait for the GetResp to do that for me.
The idea I had was that once the Pkey Manage
Hi
We were thinking of using Linux Bond0 driver for Load balancing and
Fault tolerance for IB, any thoughts on this would be welcome
Regards
Bhuvi
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-gener
Hi Eitan,
Some comments about the patch.
Personally I'm glad to see that you are using tab instead of spaces as
identaion character. But it would be nice if next time you will not mix
the functional changes and identaion fixes in the same patch, but instead
will provide two different patches. Als
Usually you can see the owners in bugzilla
Tziporet
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Rimmer, Todd
Sent: Wednesday, June 14, 2006 4:55 PM
To: openib-general@openib.org
Subject: [openib-general] Maintainers List
Is there a convenient list of
Or Gerlitz wrote:
> I suggest the following design: the CMA would replace the event handler
> provided with the qp_init_attr struct with a callback of its own and
> keep the original handler/context on a private structure.
>
> On the delivery of IB_EVENT_COMM_EST event, the CMA would call down t
Sean Hefty wrote:
> James Lentini wrote:
>> The IBTA spec (volume 1, version 1.2) describes a communication
>> established affiliated asynchronous event.
>> We've seen this event delivered to our NFS-RDMA server and aren't sure
>> what to do with it.
> This event is delivered to the verbs consum
39 matches
Mail list logo