On Tue, 2005-06-28 at 18:10, James Lentini wrote:
Are there any tools available for a network administrator to assign a
GID? Does OpenSM provider this capability?
As far as OpenIB OpenSM, there is currrently no capability to set SM
GUIDs. (Note that the SM assigned GID is comprised of the SM
Caitlin An assigned GID meets all of the requirements for an IA
Caitlin Address. I think taking advantage of that existing
Caitlin capability is just one of many options that can be done
Caitlin by the IB CM rather than forcing IB specific changes up
Caitlin to the application
Greg This is no longer needed with the class device interface in
Greg the kernel today. Please use the new api (basically just
Greg set dev_t in the class_device, and you get this for free.)
Thanks, I've killed that code and just set class_dev.devt instead.
- R.
Andrew What's the thinking behind the VM_DONTCOPY there?
I think that was my paranoia about something like a process doing a
fork, the original process exiting, and the new process having page
still mapped even though the file has been released. This is bad
because then we could map the same
Andrew What is a userspace protection domain?
A protection domain is an abstraction enforced by IB hardware --
loosely put, every resource (work queue, memory region, etc) in put in
a PD when it is created, and different resources can only see each
other if they belong to the same PD.
As an
+ int is_kernel;
Andrew I assume we have one body of code which is capable of
Andrew handling data structures in either kenrel memory of user
Andrew memory? (guess).
Andrew If so, that's a fairly sensitive thing to be doing. Tell
Andrew us more, please.
It's actually
On Tue, Jun 28, 2005 at 11:13:22PM -0500, Troy Benjegerdes wrote:
On Tue, Jun 28, 2005 at 05:27:09PM -0700, Greg KH wrote:
On Tue, Jun 28, 2005 at 04:03:43PM -0700, Roland Dreier wrote:
+++ linux/drivers/infiniband/core/uverbs_main.c 2005-06-28
15:20:04.363963991 -0700
@@ -0,0
Roland Just to be clear, the IBA spec is very clear that a GID
Roland _is_ an IPv6 address.
Hal albeit with additional properties/restrictions on IBA which
Hal do not apply to IPv6 (IBA 1,2 p, 143 lines 11-16).
Right: all GIDs are IPv6 addresses. However, as you point out, the
On Wed, Jun 29, 2005 at 09:12:09AM -0700, Greg KH wrote:
On Tue, Jun 28, 2005 at 11:13:22PM -0500, Troy Benjegerdes wrote:
On Tue, Jun 28, 2005 at 05:27:09PM -0700, Greg KH wrote:
On Tue, Jun 28, 2005 at 04:03:43PM -0700, Roland Dreier wrote:
+++
On Tue, 28 Jun 2005, Roland Dreier wrote:
James First off, here [are the] requirement we are trying to satisfy:
JamesOn the passive side of a connection, a InfiniBand kDAPL
James provider must determine a source IB address for an
James InfiniBand connection request. This
James I want to make sure I understand your solution. If we choose this
James option:
James - IB services (both kernel and user space) will be configured using
James IP addresses. By IB services, I'm referring to protocols that are
James layered directly on top of
Greg This is no longer needed with the class device interface in
Greg the kernel today. Please use the new api (basically just
Greg set dev_t in the class_device, and you get this for free.)
Here's a patch that applies on top of this patch set that fixes this:
Greg KH pointed out
On Wed, Jun 29, 2005 at 10:01:53AM -0700, Roland Dreier wrote:
Greg This is no longer needed with the class device interface in
Greg the kernel today. Please use the new api (basically just
Greg set dev_t in the class_device, and you get this for free.)
Here's a patch that
On Tue, Jun 28, 2005 at 01:59:01PM -0700, Arlin Davis wrote:
Hi Libor,
I have a couple of uCM questions regarding create_id and events...
Is it possible for a consumer of uCM to provide a context with the
create_id that could be returned with the event? I will have some scale
up issues
On Tue, Jun 28, 2005 at 01:48:55PM -0700, Nishanth Aravamudan wrote:
Using schedule_timeout() without setting the state first is broken and
causes schedule_timeout() to return immediately (effectively you call
schedule() without changing your state and are thus going to run again).
In each of
Libor Michalek wrote:
Is it possible for a consumer of uCM to provide a context with the
create_id that could be returned with the event? I will have some scale
up issues if I have to walk a list looking for a uCM provided connection
ID instead of a context that could point directly to the
Quoting r. Libor Michalek [EMAIL PROTECTED]:
Subject: Re: [PATCH] sdp_inet: fix schedule_timeout() usage
On Tue, Jun 28, 2005 at 01:48:55PM -0700, Nishanth Aravamudan wrote:
Using schedule_timeout() without setting the state first is broken and
causes schedule_timeout() to return
Hello,
Caitlin Bestler wrote:
On 6/28/05, Roland Dreier [EMAIL PROTECTED] wrote:
James First off, here [are the] requirement we are trying to satisfy:
JamesOn the passive side of a connection, a InfiniBand kDAPL
James provider must determine a source IB address for an
Libor Michalek wrote:
On Tue, Jun 28, 2005 at 01:59:01PM -0700, Arlin Davis wrote:
Hi Libor,
I have a couple of uCM questions regarding create_id and events...
Is it possible for a consumer of uCM to provide a context with the
create_id that could be returned with the event? I will have
Committed in revision 2751.
On Tue, 28 Jun 2005, Hal Rosenstock wrote:
halr I was able to build udapltest for both x86 and x86-64 with the following
halr change:
halr
halr uDAPL: Fix compile of dapl_bpool.c on x86_64
halr
halr Signed-off-by: Hal Rosenstock [EMAIL PROTECTED]
halr
halr ---
On 29.06.2005 [21:22:18 +0300], Michael S. Tsirkin wrote:
Quoting r. Libor Michalek [EMAIL PROTECTED]:
Subject: Re: [PATCH] sdp_inet: fix schedule_timeout() usage
On Tue, Jun 28, 2005 at 01:48:55PM -0700, Nishanth Aravamudan wrote:
Using schedule_timeout() without setting the state
Quoting r. Nishanth Aravamudan [EMAIL PROTECTED]:
@@ -710,11 +711,12 @@
if (!accept_conn) {
DECLARE_WAITQUEUE(wait, current);
add_wait_queue(listen_sk-sk_sleep, wait);
- set_current_state(TASK_INTERRUPTIBLE);
On 30.06.2005 [00:24:10 +0300], Michael S. Tsirkin wrote:
Quoting r. Nishanth Aravamudan [EMAIL PROTECTED]:
@@ -710,11 +711,12 @@
if (!accept_conn) {
DECLARE_WAITQUEUE(wait, current);
On Tue, Jun 28, 2005 at 03:39:59PM -0400, James Lentini wrote:
Hi Bernhard,
The changes look fine. Why the additional copyright? I need to be able
to explain it to my legal department.
My legaleeze states that whatever i do during work-time is contributed
to work and whatever is related to
On Wed, Jun 29, 2005 at 05:18:53PM -0400, James Lentini wrote:
Committed in revision 2751.
On Tue, 28 Jun 2005, Hal Rosenstock wrote:
halr I was able to build udapltest for both x86 and x86-64 with the following
halr change:
halr
halr uDAPL: Fix compile of dapl_bpool.c on x86_64
halr
halr
The following patch (partially) fixes an issue extracting the packet
lifetime from the local ACK timeout when processing a received CM
REQ or LAP message. The local ACK timeout carries twice the packet
lifetime, plus CA ACK delay - which I'm currently ignoring as being
too painful to deal with.
This patch fixes building ucm against 2.6.13-rc1. Please apply when
2.6.13 has arrived.
Signed-off-by: Tom Duffy [EMAIL PROTECTED]
Index: linux-2.6.13-rc1/drivers/infiniband/core/ucm.c
===
---
This patch renames the sdp_opt struct sdp_sock as this is what all the
other socket classes do. Against r2751.
Signed-off-by: Tom Duffy [EMAIL PROTECTED]
Index: linux-2.6.12-openib/drivers/infiniband/ulp/sdp/sdp_write.c
===
---
On Tue, 2005-06-28 at 11:19 -0700, Libor Michalek wrote:
Here's a patch to reorder listen_lookup() in req_handler() and a few
other fixups, such as making a couple int return functions void.
Great, that fixes the panic.
Thanks,
-tduffy
signature.asc
Description: This is a digitally signed
Signed-off-by: Tom Duffy [EMAIL PROTECTED]
Index: linux-2.6.12-rc6-openib/drivers/infiniband/ulp/sdp/TODO
===
--- linux-2.6.12-rc6-openib/drivers/infiniband/ulp/sdp/TODO (revision 2597)
+++
The following patch includes the packet lifetime as part of the delay
waiting for a request after receiving an MRA. Problem pointed out by
Fab Tillier.
Signed-off-by: Sean Hefty [EMAIL PROTECTED]
Index: core/cm.c
===
--- core/cm.c
Title: Message
Dear DAT and OpenIB members,
There isa
debate going on on OpenIB and DAT reflectors which is going around about kDAT
registry for Linux.
I would like to
review the requirements wehad
agreedat DAT
collaborativeand captured in the
kDAT
and uDAT specs and review DAT registry
On Wed, Jun 29, 2005 at 12:42:09PM -0700, Arlin Davis wrote:
Libor Michalek wrote:
On Tue, Jun 28, 2005 at 01:59:01PM -0700, Arlin Davis wrote:
I have a couple of uCM questions regarding create_id and events...
Is it possible for a consumer of uCM to provide a context with the
create_id
On 6/28/05, James Lentini [EMAIL PROTECTED] wrote:
The attached PDF, prepared by Itamar Rabenstein of Mellanox, contains
a proposed mapping from DAT_RETURN values to ERRNO values.
These all look very good. Thanks Itamar for putting this doc together.
-tduffy
On Tue, 28 Jun 2005, Greg KH wrote:
On Tue, Jun 28, 2005 at 04:03:43PM -0700, Roland Dreier wrote:
+++ linux/drivers/infiniband/core/uverbs_main.c 2005-06-28
15:20:04.363963991 -0700
@@ -0,0 +1,708 @@
+/*
+ * Copyright (c) 2005 Topspin Communications. All rights reserved.
+ *
Title: Message
Source compatability for an existingProvider is *not*
maintained by OpenIB
kDAPL because there are fields *missing* from the
Provider Info. That *will*
result in a compilation error.
It also precludes experimentation with future releases of
DAT through
kernel modules. This is
I did netperf test over Mellanox 23108
4X HCA against r2720, after a while the HCA stopped to accept packets.
I am pretty sure it's a bug in driver
not IPoIB. rmmod ib_ipoib didn't help, after removing ib_mthca, and restarted,
the driver worked again. It's easy to reprocude on my 4-way intel
Now I am adding a permanent arp
entry to avoid this problem. It seems to work fine.
Oops, I was wrong, it stopped to receive
any packets.
Thanks
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone(Fax): (503) 578-7638
Shirley Ma/Beaverton/[EMAIL
Shirley wrote,
Now I am adding a permanent arp entry to avoid this problem. It seems
to work fine.
Oops, I was wrong, it stopped to receive any packets.
When this happens, if you reset the switch, does the node resume
accepting packets ?
I have seen a similar problem (only once or twice) where
--
Message: 2
Date: Wed, 29 Jun 2005 09:00:37 -0700
From: Roland Dreier [EMAIL PROTECTED]
Subject: Re: [openib-general] IP addressing on InfiniBand networks
To: Caitlin Bestler [EMAIL PROTECTED]
Cc: Lentini, James [EMAIL PROTECTED], Christoph Hellwig
[EMAIL
40 matches
Mail list logo