Quoting Hal Rosenstock <[EMAIL PROTECTED]>:
> > > > What does SDP use this for?
> > >
> > > Same thing as AT right now.
> >
> > Except SDP drops netdevice and route reference counts after sending an
> > arp :)
>
> So SDP wouldn't take advantage of a pointer to access IPoIB private data
> if this
On Mon, 2005-09-26 at 23:16, Michael S. Tsirkin wrote:
> Quoting Hal Rosenstock <[EMAIL PROTECTED]>:
> > > What does SDP use this for?
> >
> > Same thing as AT right now.
>
> Except SDP drops netdevice and route reference counts after sending an arp :)
So SDP wouldn't take advantage of a pointer
ScanMail for Microsoft Exchange has detected virus-infected attachment(s).
Sender = [EMAIL PROTECTED]
Recipient(s) = openib-general@openib.org
Subject = [openib-general] IMPORTANT NOTIFICATION
Scanning time = 9/26/2005 8:50:25 PM
Engine/Pattern = 7.510-1002/2.857.00
Action on virus found:
The att
ScanMail for Microsoft Exchange has detected virus-infected attachment(s).
Sender = [EMAIL PROTECTED]
Recipient(s) = openib-general@openib.org
Subject = [openib-general] IMPORTANT NOTIFICATION
Scanning time = 9/26/2005 10:50:05 PM
Engine/Pattern = 7.510-1002/2.857.00
Action on virus found:
The at
On Mon, 2005-09-26 at 19:57, Viswanath Krishnamurthy wrote:
> I have an exerciser in the IB network. The exerciser seems to be
> faulty/buggy. When opensm starts I do not
> see 'SUBNET UP" message. It says "Entering MASTER" and waits there.
> Any new node inserted in this state is not assigned any
Quoting Roland Dreier <[EMAIL PROTECTED]>:
> Michael> As a side note, the spec says: "Two types of CQ errors
> Michael> can occur: the CQ can overrun or it can become
> Michael> inaccessible": I wander whether this should be
> Michael> interpreted in a sense that that there should b
Quoting Hal Rosenstock <[EMAIL PROTECTED]>:
> > What does SDP use this for?
>
> Same thing as AT right now.
Except SDP drops netdevice and route reference counts after sending an arp :)
--
MST
___
openib-general mailing list
openib-general@openib.org
Quoting r. Jerome Pioux <[EMAIL PROTECTED]>:
> Subject: Re: FW: SDP problems with 64K page size
>
> > The best way to fix this appears to be to bump the counters up to u32 or
> > s32.
>
> Just an open question: Do you think that we could get better performance if
> we would go with u32 instead
Quoting r. Roland Dreier <[EMAIL PROTECTED]>:
> Subject: Re: FW: SDP problems with 64K page size
>
> Jerome> Just an open question: Do you think that we could get
> Jerome> better performance if we would go with u32 instead of
> Jerome> reducing the buffer to 16K?
>
> Not sure. One e
I have an exerciser in the IB network. The exerciser seems to be faulty/buggy. When opensm starts I do not
see 'SUBNET UP" message. It says "Entering MASTER" and waits there.
Any new node inserted in this state is not assigned any LID. Anybody seen such behavior ?
-Viswa
__
Guy German wrote:
I also added a small test, _for bring up purposes only_, which is not
for real use (has memory leaks and uses global vars etc), but it does
connect and passes private data from side to side.
I would like to port the cmpost, for a real test module, but it is maybe
possible to f
Hal Rosenstock wrote:
examples/cmpost.c: Update to new ibv_create_cq API
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
Thanks - committed.
- Sean
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib
Tried to find out the "default superslotes" for an OpenPower 720. Please try either slot 2 or 5. I delayed my response to make certain that these were indeed the superslots. I am still not a 100% certain -no point waiting beyond a certain stage.
If you can please go ahead and try these and let u
examples/cmpost.c: Update to new ibv_create_cq API
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
Index: examples/cmpost.c
===
--- examples/cmpost.c (revision 3552)
+++ examples/cmpost.c (working copy)
@@ -315,7 +315,7 @@ stat
On Mon, 2005-09-26 at 18:05, Todd Bowman wrote:
> I am having a problem with uDAPL accessing
> /dev/infiniband/{uat,ucm0}. I am running 3549, 2.6.12 kernel with
> backport. Here is a snippet of the uDAPL debug messages running
> dtest. The dat.conf file seems to be correct, the correclty named
On 26.09.2005 [10:15:21 -0700], Grant Grundler wrote:
> On Sat, Sep 24, 2005 at 10:19:53AM -0700, Roland Dreier wrote:
> ...
> > I just checked in a fix for this -- the pci_pretty_name() API has gone
> > away, so I removed our use of it in svn. I don't understand how your
> > other builds of git +
I am having a problem with uDAPL accessing
/dev/infiniband/{uat,ucm0}. I am running 3549, 2.6.12
kernel with backport. Here is a snippet of the uDAPL debug
messages running dtest. The dat.conf file seems to be correct,
the correclty named providers are being loaded.
26248 Running as server
DAT
Roland Dreier wrote:
I just checked a change into subversion to implement the completion
channel API I described last week. This also cleans up some of the
problems in the kernel error paths.
These changes break both the kernel ABI and the userspace library
API, so to use the new code, you wil
On Mon, 26 Sep 2005, Hal Rosenstock wrote:
> On Mon, 2005-09-26 at 13:15, James Lentini wrote:
> > What is the kdapltest command you are using?
>
>
> kdapltest -T T -s -D mthca0a -d -i 1 -w 8 client SR server SR
>
> seems to fail
>
> but
>
> kdapltest -T T -s -D mthca0a -d -t 2 -w 8
James Lentini wrote:
I expect that the IB_CM_REQ_RECEIVED callback will be confusing to
ULPs. The ULP will receive a new cma_id with an old context value.
If the ULP wanted to make an adjustments to the cma_id that
received the request, it would need to store a reference to it in
the old cma_i
On Mon, 26 Sep 2005, Hal Rosenstock wrote:
> On Thu, 2005-09-22 at 13:46, Sean Hefty wrote:
> > I've checked this into svn under
> > svn/gen2/users/mshefty/linux-kernel/infiniband,
> > so that changes can be tracked easier.
>
> I haven't had a chance to look at this as yet but have a couple o
[It's somewhat annoying to have to subscribe to [EMAIL PROTECTED]
just to be able to send patches, but oh well...]
This patch updates Open MPI for the new ibv_create_cq() API.
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
--- ompi/mca/btl/openib/btl_openib.c(revision 7507)
+++ ompi/mca/b
gcc version 4 doesn't like the extern declaration of free_vbuf_head to
followed by a static declaration in vbuf.c. To fix this, we can just
get rid of the declaration in vbuf.h, since free_vbuf_head is not used
outside of vbuf.c.
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
--- mpid/ch_gen2/
This patch updates MVAPICH for the new ibv_create_cq() API.
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
--- mvapich-gen2/mpid/ch_gen2/viainit.c (revision 3534)
+++ mvapich-gen2/mpid/ch_gen2/viainit.c (working copy)
@@ -115,7 +115,7 @@ static void get_lid(void)
static void create_cq(void)
{
On Mon, 26 Sep 2005, Sean Hefty wrote:
> James Lentini wrote:
> > Why would this module be a ULP and not part of the core? Especially since
> > the rdma_cma.h include file is intended for the core include area,
> > include/rdma.
>
> It can be a separately loaded module, so a ULP from the viewp
Jerome> Just an open question: Do you think that we could get
Jerome> better performance if we would go with u32 instead of
Jerome> reducing the buffer to 16K?
Not sure. One easy test you could try would be increasing 16384 to
32768 in my patch. If that works and improves performance
I just checked a change into subversion to implement the completion
channel API I described last week. This also cleans up some of the
problems in the kernel error paths.
These changes break both the kernel ABI and the userspace library
API, so to use the new code, you will have to update your ke
The best way to fix this appears to be to bump the counters up to u32 or
s32.
Just an open question: Do you think that we could get better performance if
we would go with u32 instead of reducing the buffer to 16K?
Jerome
- Original Message -
From: "Michael S. Tsirkin" <[EMAIL PROTE
James Lentini wrote:
- move listen declaration closer to accept and reject
Accepted - will be pushed in with next version.
- add private data and new cma_id fields to event structure
Added private_data_len field to cma_id structure. Would like to get some
additional feedback before adding
Hal> Also, would IPv6 be extensions to the current API parameters
Hal> or additional APIs ?
I think the API works unchanged for IPv6, since addresses are
specified using struct sockaddr.
- R.
___
openib-general mailing list
openib-general@openi
Hal Rosenstock wrote:
On Thu, 2005-09-22 at 13:46, Sean Hefty wrote:
I've checked this into svn under svn/gen2/users/mshefty/linux-kernel/infiniband,
so that changes can be tracked easier.
What would be done for uDAPL ? Would there be uCMA ?
I've considered uDAPL, but haven't thought throug
On Thu, 2005-09-22 at 13:46, Sean Hefty wrote:
> I've checked this into svn under
> svn/gen2/users/mshefty/linux-kernel/infiniband,
> so that changes can be tracked easier.
I haven't had a chance to look at this as yet but have a couple of
questions:
What would be done for uDAPL ? Would there b
James Lentini wrote:
Why would this module be a ULP and not part of the core? Especially
since the rdma_cma.h include file is intended for the core include
area, include/rdma.
It can be a separately loaded module, so a ULP from the viewpoint of verbs, SA
query, IB CM, etc.
I expect that th
On Mon, 2005-09-26 at 15:52, Sean Hefty wrote:
> Hal Rosenstock wrote:
> > On Mon, 2005-09-26 at 15:29, Sean Hefty wrote:
> >
> >>>Is this ib_ptr or ipoib_ptr ? I would think iWARP devices would need
> >>>this.
> >>
> >>I think that we can implement the CMA interface without adding this pointer
>
Hal Rosenstock wrote:
On Mon, 2005-09-26 at 15:29, Sean Hefty wrote:
Is this ib_ptr or ipoib_ptr ? I would think iWARP devices would need
this.
I think that we can implement the CMA interface without adding this pointer by
using ARP and the private data in the CM REQ.
True if that is the a
On Mon, 2005-09-26 at 15:29, Sean Hefty wrote:
> > Is this ib_ptr or ipoib_ptr ? I would think iWARP devices would need
> > this.
>
> I think that we can implement the CMA interface without adding this pointer
> by
> using ARP and the private data in the CM REQ.
True if that is the approach tak
Hal> Oops. I meant to write "iWARP devices wouldn't need this".
I think they do. For example, an iWARP device driver would want to
get from a struct net_device to a struct rdma_device when using the
route tables.
- R.
___
openib-general mailing li
Jack> I'm starting to fix ib_query_device/ibv_query_device --
Jack> adding missing fields, correcting values in current fields.
Great!
Jack> Enclosed is a patch for the atomic_cap field. Please
Jack> review. Thanks.
Thanks, applied and queued for 2.6.15, with the following fix
On Mon, 2005-09-26 at 15:19, Hal Rosenstock wrote:
> On Fri, 2005-09-16 at 20:15, Roland Dreier wrote:
> > Sean> I continue to hit this same issue, so I've started looking
> > Sean> at the ib_at code. Ib_at accesses struct ipoib_dev_priv to
> > Sean> get information about the related po
Hal Rosenstock wrote:
Sean> I continue to hit this same issue, so I've started looking
Sean> at the ib_at code. Ib_at accesses struct ipoib_dev_priv to
Sean> get information about the related port that IPoIB is using.
Sean> Is there some other way for AT to get to the same
Sean> in
On Wed, 21 Sep 2005, Sean Hefty wrote:
> Here's the updated implementation. It compiles, but that's it.
Sean,
Overall this looks very good.
A few comments:
Why would this module be a ULP and not part of the core? Especially
since the rdma_cma.h include file is intended for the core includ
On Fri, 2005-09-16 at 20:15, Roland Dreier wrote:
> Sean> I continue to hit this same issue, so I've started looking
> Sean> at the ib_at code. Ib_at accesses struct ipoib_dev_priv to
> Sean> get information about the related port that IPoIB is using.
> Sean> Is there some other way
On Mon, 2005-09-26 at 14:42, Suresh Shelvapille wrote:
> Folks:
>
> I am trying to add Infiniband drivers to a 2.6.10 kernel, and the
> docs/readme-kernel.txt says to apply drivers.diff patch so that the
> Infiniband support can be enabled in configuration. Where do I find this
> patch?
That READ
On Mon, Sep 26, 2005 at 01:13:10PM -0400, James Lentini wrote:
...
> > kdapl_ib: disagrees about version of symbol dat_registry_add_provider
> > kdapl_ib: Unknown symbol dat_registry_add_provider
> >
> > I've rebuilt the kdapl, kdapl_ib, and dat modules and rebooted but this
> > still occurs. Any
I think that's going to far to stick everything into one return
statement. I committed the change below, which I think is a lot
clearer, and queued it for 2.6.15.
- R.
--- linux-kernel/infiniband/core/verbs.c(revision 3544)
+++ linux-kernel/infiniband/core/verbs.c(working copy)
Folks:
I am trying to add Infiniband drivers to a 2.6.10 kernel, and the
docs/readme-kernel.txt says to apply drivers.diff patch so that the
Infiniband support can be enabled in configuration. Where do I find this
patch?
Thanks a lot.
Suri
___
openib-g
On Mon, 26 Sep 2005, Hal Rosenstock wrote:
> On Mon, 2005-09-26 at 12:55, Sean Hefty wrote:
> > Guy German wrote:
> > > I believe that ib_at is still a valuable module even if ATS
> > > reverse ARP is broken, and I think we should discuss this.
> >
> > Here's my thinking on this. ATS is broke
On Mon, 2005-09-26 at 12:55, Sean Hefty wrote:
> Guy German wrote:
> > I believe that ib_at is still a valuable module even if ATS reverse ARP
> > is broken, and I think we should discuss this.
>
> Here's my thinking on this. ATS is broken as you mentioned for reverse
> lookups.
> However, i
On Mon, 2005-09-26 at 13:15, James Lentini wrote:
> What is the kdapltest command you are using?
kdapltest -T T -s -D mthca0a -d -i 1 -w 8 client SR server SR
seems to fail
but
kdapltest -T T -s -D mthca0a -d -t 2 -w 8 -i 20 client SR server SR
works on that machine
___
On 26.09.2005 [10:15:21 -0700], Grant Grundler wrote:
> On Sat, Sep 24, 2005 at 10:19:53AM -0700, Roland Dreier wrote:
> ...
> > I just checked in a fix for this -- the pci_pretty_name() API has gone
> > away, so I removed our use of it in svn. I don't understand how your
> > other builds of git +
On Mon, 2005-09-26 at 13:27, James Lentini wrote:
> The dat module is the problem. When we moved the code to the trunk, we
> called the dat registry module kdapl. There shouldn't be a "dat"
> module.
>
> If you rmmod dat, I expect that this will fix your problem.
Thanks. That pointed me in the
On Mon, 26 Sep 2005, Hal Rosenstock wrote:
> On Mon, 2005-09-26 at 13:13, James Lentini wrote:
> > On Mon, 26 Sep 2005, Hal Rosenstock wrote:
> >
> > > Hi James,
> > >
> > > When loading kdapl built on 2.6.11.6, I keep getting the following:
> > >
> > > kdapl_/sbin/modprobe kdapl_ib
> > > FAT
On Mon, 2005-09-26 at 13:13, James Lentini wrote:
> On Mon, 26 Sep 2005, Hal Rosenstock wrote:
>
> > Hi James,
> >
> > When loading kdapl built on 2.6.11.6, I keep getting the following:
> >
> > kdapl_/sbin/modprobe kdapl_ib
> > FATAL: Error inserting kdapl_ib
> > (/lib/modules/2.6.11.6/kernel/d
I ran into another opensm bug which caused opensm to stop functioning. This happened only once.
Here is the test case
1. Run opensm on Machine A
2. Run the following script on M/c B
a. Check ibstatus
b. Ping machine A
c. Run osmtest
d. reboot
The test case is to make sure opensm
On Fri, 2005-09-23 at 15:33, James Lentini wrote:
> On Fri, 23 Sep 2005, Hal Rosenstock wrote:
>
> > > Is there a better way to enumerate all of the network inferaces? I
> > > believe that is what this for loop is attempting to accomplish.
> >
> > Yes. I think that the net_device list from dev_
On Sat, Sep 24, 2005 at 10:19:53AM -0700, Roland Dreier wrote:
...
> I just checked in a fix for this -- the pci_pretty_name() API has gone
> away, so I removed our use of it in svn. I don't understand how your
> other builds of git + svn succeeded though, since pci_pretty_name is
> completely gon
What is the kdapltest command you are using?
On Mon, 26 Sep 2005, Hal Rosenstock wrote:
> Hi James,
>
> I keep getting the following when running kdapltest. This is similar to
> what I saw before and reported a couple of times but now seems more
> consistent in occurring.
>
> -- Hal
>
> Sep
On Mon, 26 Sep 2005, Hal Rosenstock wrote:
> Hi James,
>
> When loading kdapl built on 2.6.11.6, I keep getting the following:
>
> kdapl_/sbin/modprobe kdapl_ib
> FATAL: Error inserting kdapl_ib
> (/lib/modules/2.6.11.6/kernel/drivers/infiniband/ulp/kdapl/ib/kdapl_ib.ko):
> Unknown symbol in
On Mon, 2005-09-26 at 12:52, Michael S. Tsirkin wrote:
> Hmm. I do need the source address for the path record query, do I not?
Yes, SGID is a required component for a SA GetTable request of
PathRecord.
-- Hal
___
openib-general mailing list
openib-gen
On Sat, Sep 24, 2005 at 02:58:10PM +0800, QiWang, Chen wrote:
> Hi, grant
>
> On node c01-14, I installed openib-gen2, kernel 2.6.13.2
> but I have IRQ problem.
Yes, I'm not surprised.
> ---
>
> ib_mthca: Mellanox InfiniBand HCA driver v0.06 (June 23, 2005)
> ib_
Hi James,
When loading kdapl built on 2.6.11.6, I keep getting the following:
kdapl_/sbin/modprobe kdapl_ib
FATAL: Error inserting kdapl_ib
(/lib/modules/2.6.11.6/kernel/drivers/infiniband/ulp/kdapl/ib/kdapl_ib.ko):
Unknown symbol in module, or unknown parameter (see dmesg)
ib: disagrees about v
Hi Eitan,
I see that message in the log.
-Viswa
On 9/24/05, Eitan Zahavi <[EMAIL PROTECTED]> wrote:
Hi Viswa and Hal,I have read through the thread and have few comments.But first let me see if I understand the test run correctly. The test is as follows:1. OpenSM starts up configuring the subnet
Guy German wrote:
I believe that ib_at is still a valuable module even if ATS reverse ARP
is broken, and I think we should discuss this.
Here's my thinking on this. ATS is broken as you mentioned for reverse lookups.
However, if we want to keep ATS, I think that ATS registration/deregistrati
Quoting r. Roland Dreier <[EMAIL PROTECTED]>:
> Subject: Re: [openib-general] [PATCH] mthca: mthca_map_cmd off by one
>
> I applied my version of the patch -- let me know if you think it's buggy.
>
> - R.
>
Looks good to me.
--
MST
___
openib-gener
Quoting r. Sean Hefty <[EMAIL PROTECTED]>:
> Subject: Re: 2.6.14 heads up: ip_dev_find() not exported
>
> Michael S. Tsirkin wrote:
> >>>I noticed while compiling against an up-to-date kernel tree that SDP
> >>>and IBAT both use the function ip_dev_find(). The EXPORT_SYMBOL for
> >>>this function
I applied my version of the patch -- let me know if you think it's buggy.
- R.
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/o
Michael S. Tsirkin wrote:
I noticed while compiling against an up-to-date kernel tree that SDP
and IBAT both use the function ip_dev_find(). The EXPORT_SYMBOL for
this function was removed during the 2.6.14 devel cycle.
I'm calling using this routine in the rdam_cma code as well. I call it to
Yes, good catch, but I think the fix is not quite right. When we fill
up the table, we'll set nent = 0, and then do ++nent at the end of the
loop. So the second time around we'll start with nent = 1.
Something like this is better, right?
--- linux-kernel/infiniband/hw/mthca/mthca_cmd.c(
Quoting r. Sean Hefty <[EMAIL PROTECTED]>:
> Subject: Re: 2.6.14 heads up: ip_dev_find() not exported
>
> Roland Dreier wrote:
> > I noticed while compiling against an up-to-date kernel tree that SDP
> > and IBAT both use the function ip_dev_find(). The EXPORT_SYMBOL for
> > this function was rem
Looks good, applied and queued for 2.6.14.
- R.
___
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
Quoting Roland Dreier <[EMAIL PROTECTED]>:
> Subject: 2.6.14 heads up: ip_dev_find() not exported
>
> I noticed while compiling against an up-to-date kernel tree that SDP
> and IBAT both use the function ip_dev_find(). The EXPORT_SYMBOL for
> this function was removed during the 2.6.14 devel cycl
On Sep 25, 2005, at 3:45 AM, Michael S. Tsirkin wrote:
Roland, I might check in the patch that you posted to work around
this problem for 64K page users, until I have a final fix ready.
Is that OK with everyone?
It looks harmless in the 4K or 16K case. Go for it.
-tduffy
_
Roland Dreier wrote:
I noticed while compiling against an up-to-date kernel tree that SDP
and IBAT both use the function ip_dev_find(). The EXPORT_SYMBOL for
this function was removed during the 2.6.14 devel cycle.
I'm calling using this routine in the rdam_cma code as well. I call it to get
Sean,
I've committed a modified version of the CMA to the svn repository, under
gen2/users/guyg/
This version takes most of your changes to my original posting, except for the
"connect" method - it still uses ib_at and does not discover the path to the
consumer.
I also added a small test,
Hi Eitan,
I finally got a chance to read this over. Here are some comments:
On Tue, 2005-09-06 at 15:18, Eitan Zahavi wrote:
> Hi All,
>
> As we are about to start working on the fast routing algorithms,
> here is the writeup about proposed algorithms for your review.
This appears to be an upd
Hi James,
I keep getting the following when running kdapltest. This is similar to
what I saw before and reported a couple of times but now seems more
consistent in occurring.
-- Hal
Sep 26 10:29:29 hal kernel: DT_Mdep_Thread_: page allocation failure. order:0,
mode:0x20
Sep 26 10:29:29 hal kern
Quoting James Lentini <[EMAIL PROTECTED]>:
> > I think we could get by with just dev_get_by_index, I'll have to
> > investigate this.
>
> FYI: I found a bug in IBAT's use of dev_get_by_index(). See:
>
> http://openib.org/pipermail/openib-general/2005-September/011668.html
>
Right. We should jus
Woodruff, Robert J wrote:
Seems to hang around the time of the modify QP.
I just pulled the latest (3541) and ran on my 2.6.13 systems and saw no
problems with MPI.
-arlin
ibv_rc_pingpong seems to work OK and also your
DAPL-socket CM version that you gave me yesterday seems
to work, but
On Mon, 2005-09-26 at 11:16, yipee wrote:
> Is there any proposal you want me to review? Where can I read it?
It hasn't been written yet.
-- Hal
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-genera
Title: RE: [openib-general] Re: OpenSM & pkeys
Hi Yipee,
A Tcl API for sending and receiving MADs is available under
https://openib.org/svn/gen2/utils/src/linux-user/ibis
(IBIS == IB Inband Services)
If you follow on the autogen.sh && configure && make && make install
you should get /usr/l
ScanMail for Microsoft Exchange has detected virus-infected attachment(s).
Sender = [EMAIL PROTECTED]
Recipient(s) = openib-general@openib.org
Subject = [openib-general] *DETECTED* Online User Violation
Scanning time = 9/26/2005 8:31:06 AM
Engine/Pattern = 7.510-1002/2.857.00
Action on virus foun
ScanMail for Microsoft Exchange has detected virus-infected attachment(s).
Sender = [EMAIL PROTECTED]
Recipient(s) = openib-general@openib.org
Subject = [openib-general] *DETECTED* ONLINE USER VIOLATION
Scanning time = 9/26/2005 8:30:57 AM
Engine/Pattern = 7.510-1002/2.857.00
Action on virus foun
On Mon, 26 Sep 2005, Michael S. Tsirkin wrote:
> Quoting r. Roland Dreier <[EMAIL PROTECTED]>:
> > Subject: 2.6.14 heads up: ip_dev_find() not exported
> >
> > I noticed while compiling against an up-to-date kernel tree that SDP
> > and IBAT both use the function ip_dev_find(). The EXPORT_SYMB
Hal Rosenstock voltaire.com> writes:
[cut]
> Can you share any requirements relative to this (or alternatively review
> a proposal) ? The policy on PKey management (partition management) is
> beyond the IB spec.
I'm looking for a way to set the PKey tables for some hosts ports in my fabric.
Th
Hal Rosenstock voltaire.com> writes:
[cut]
> The PKey tables can also be obtained by the diagnostic tools: smpdump in
> particular.
OK, But currently there's no way (at least that I know of) to set PKeys for the
hosts ports.
y
___
openib-genera
Title: RE: core and ipoib questions and oops
Problem is at ipoib_multicast.c:223 :
if (!memcmp(mcast->mcmember.mgid.raw, priv->dev->broadcast + 4,
a. r14 contains mcast->dev:
drivers/infiniband/ulp/ipoib/ipoib_multicast.c:216
41a8: 4c 8b b7 f0 00 00 00 mov 0xf0(%rdi)
- Forwarded message from Leonid Keller <[EMAIL PROTECTED]> -
Subject: a bug ?
Date: Mon, 26 Sep 2005 17:17:05 +0300
From: "Leonid Keller" <[EMAIL PROTECTED]>
look at the following code in mthca_map_cmd():
---
pages[ne
- Forwarded message from Leonid Keller <[EMAIL PROTECTED]> -
Subject: a bug ?
Date: Wed, 21 Sep 2005 21:22:05 +0300
From: "Leonid Keller" <[EMAIL PROTECTED]>
in mthca_init_eq_table() there is code:
dev->eq_table.clr_int = dev->clr_base + (dev->eq_table.inta_pin < 31 ? 4 :
0);
In VAP
Quoting r. Roland Dreier <[EMAIL PROTECTED]>:
> Subject: 2.6.14 heads up: ip_dev_find() not exported
>
> I noticed while compiling against an up-to-date kernel tree that SDP
> and IBAT both use the function ip_dev_find(). The EXPORT_SYMBOL for
> this function was removed during the 2.6.14 devel c
I noticed while compiling against an up-to-date kernel tree that SDP
and IBAT both use the function ip_dev_find(). The EXPORT_SYMBOL for
this function was removed during the 2.6.14 devel cycle.
I haven't looked yet at what this function does, how SDP and IBAT use
it or what it could be replaced b
I'm starting to fix ib_query_device/ibv_query_device -- adding missing fields,
correcting values in current fields.
Enclosed is a patch for the atomic_cap field. Please review.
Thanks.
Jack
Signed-off-by: Jack Morgenstein <[EMAIL PROTECTED]>
Index: linux-kernel/infiniband/core/uverbs_cmd.c
===
Quoting Michael S. Tsirkin <[EMAIL PROTECTED]>:
> #ifconfig ib0 down
>
> Unable to handle kernel NULL pointer dereference at 0388
> RIP:
> Code: 49 8b 7d 08 48 81 c7 b4 00 00 00 f3 a6 75 17 49 8b 45 70 8b
> RIP {:ib_ipoib:ipoib_mcast_join_finish+100} RSP
>
> CR2: 0388
Her
Two questions:
1. Roland, looking at ipoib_multicast, I see
if (mcast->query) {
ib_sa_cancel_query(mcast->query_id, mcast->query);
mcast->query = NULL;
ipoib_dbg_mcast(priv, "waiting for MGID " IPOIB_GID_FMT
"\
On Sun, 2005-09-25 at 03:20, Yael Kalka wrote:
> During our windows checks we noticed an issue in
> the __osm_lid_mgr_init_sweep function under osm_lid_mgr.c.
> The initializing of max_persistent_lid and max_discovered_lid is
> correct only if the vector is not empty.
> Attached is a patch resolvi
On Sun, 2005-09-25 at 03:14, Yael Kalka wrote:
> Sorry, this mail was sent with wrong subject
The previous one. I could see :-) No big deal...
> This patch resolves the ERROR numbering issue.
> As you mentioned - there was a problem with the cl_event_wheel.c. The
> patch fixes the error ther
On Sun, 2005-09-25 at 09:02, Eitan Zahavi wrote:
> Current OpenSM implementation does not support PKey assignment.
> There is no programmatic way for you to get the PKey tables on the
> HCAs from within the OpenSM. However, if you turn on verbose mode (-V)
> you should be able to see the
>
> PKeyT
Hi,
On Mon, 2005-09-26 at 04:12, yipee wrote:
> Eitan Zahavi mellanox.co.il> writes:
> [cut]
>
> > We had PKey manager support optionally planned for implementation this
> quarter.
> > However, the pace of stabilizing OpenSM on OpenIB stack is slower then
> expected and we might not be able t
On Sun, 2005-09-25 at 07:16, yipee wrote:
> How do I tell the current OpenSM to set specific pkeys to hca ports of some
> of
> my host nodes?
As Eitan indicated, OpenSM does not currently support this.
> Can OpenSM tell me what are the pkeys for all the host nodes in the fabric?
There is diagn
Title: RE: [openib-general] Re: OpenSM & pkeys
Hi Yipee,
I am sorry to disappoint you again. The level of interest for PKey support is not very high so I am not sure it will not shift gain. But I hope it will not.
Are there any other people interested in PKey management feature in OpenSM???
Eitan Zahavi mellanox.co.il> writes:
[cut]
> We had PKey manager support optionally planned for implementation this
quarter.
> However, the pace of stabilizing OpenSM on OpenIB stack is slower then
expected and we might not be able to make it this quarter.
Thanks for the response,
Can I coun
100 matches
Mail list logo
