>
> This looks very nice to me, one nitpick: I think
> ibonly_verbs.h should be called just ib_verbs.h
>
>
Agreed. I should have commented on that. I used
the 'ibonly_verbs.h' as a stand-in so I could
borrow the existing make when I tested. I didn't
want to pre-empt 'ib_verbs.h' yet.
Title: RE: [Fwd: [openib-commits] r2826 - in gen2/utils/src/linux-user/IBMgtSim: . config src tests utils]
Hi Hal
I plan to do the development of IBDM and IBMgtSim in OpenIB.
I also think they should be moved to the main gen2 trunk.
I will deal with the errors today.
Eitan Zahavi
>
Title: RE: [openib-general] Re: Gen2 build/update process
I propose to build a top level auto-tools project that will call the sub directories recursively. In the meanwhile it is just a convinience issue
Eitan Zahavi
Design Technology Director
Mellanox Technologies LTD
Tel:+972-4-9097208
Fax:
Please acknowledge that you understand how inappropriate
and problem causing your huge patch bomb was today to this
mailing list.
It is nearly 8 hours later, and vger.kernel.org is still
trying mightily to spit out all of your patches to the
5000+ people subscribed to linux-kernel.
There is abou
Hal Rosenstock <[EMAIL PROTECTED]> wrote:
>
> > I'll tentatively consider this material to be not-for-2.6.13?
>
> Presuming that "this material" refers to the patch to add the kernel CM
> implementation, if kernel CM does not make 2.6.13, then user CM should
> not either as it is dependent on i
On Mon, Jul 11, 2005 at 09:47:04AM -0700, Caitlin Bestler wrote:
> The attached .h files follow the alternate approach where each
> transport can define its own specifics in a supplemental .h file
> and avoid the massive union approach.
>
> This is mostly cosmetic, except for the work request. Bec
On Mon, 2005-07-11 at 20:05, Andrew Morton wrote:
> Hal Rosenstock <[EMAIL PROTECTED]> wrote:
> >
> > This is version 2 of a patch series to get the Infiniband core up to
> > date.
>
> Well that was interesting.
>
> - All the patches had mangled headers:
>
> -- linux-2.6.13-rc2-mm1-16/...
> +++
Hal Rosenstock <[EMAIL PROTECTED]> wrote:
>
> This is version 2 of a patch series to get the Infiniband core up to
> date.
Well that was interesting.
- All the patches had mangled headers:
-- linux-2.6.13-rc2-mm1-16/...
+++ linux-2.6.13-rc2-mm1-17/...
instead of
--- linux-2.6.13-rc2-mm1-16/.
Hal,
>-Original Message-
>From: Hal Rosenstock [mailto:[EMAIL PROTECTED]
>Sent: Monday, July 11, 2005 4:21 PM
>To: Davis, Arlin R
>Cc: openib-general@openib.org
>Subject: Re: ib_uat kernel panic without async callback
>
>Hi Arlin,
>
>On Mon, 2005-07-11 at 18:37, Davis, Arlin R wrote:
>> I
[ un-cc lkml ]
On Mon, 2005-07-11 at 19:16 -0400, Hal Rosenstock wrote:
> > [ also, in future patch bombs, could you please set the references
> > header so that the message thread properly, thanks ]
>
> How do I do this ?
Attached is the script I use to sent patchbombs. Thanks to Greg Kroah
an
Hi Arlin,
On Mon, 2005-07-11 at 18:37, Davis, Arlin R wrote:
> I was attempting to call ib_at_route_by_ip() in sync mode and get the
> following oops:
>
> You can change the att.c (line 193) example as follow to duplicate:
>
> r = ib_at_route_by_ip(0x1400a8c0, 0, 0, 0,
>
On Mon, 2005-07-11 at 17:19, Tom Duffy wrote:
> On Mon, 2005-07-11 at 16:59 -0400, Hal Rosenstock wrote:
> > Add kernel portion of user CM implementation
>
> Hal, does this compile? As it doesn't seem to include the patch I sent
> to openib-general changing class_simple to class, I don't think it
On Thu, 2005-06-30 at 10:44 -0400, James Lentini wrote:
> I'm going to be on vacation next week. I'll make the conversion when I
> return on the 11th. If anyone else has comments, please send them by
> July 10th.
Do you have a patch that implements the errno changes? Are you going to
add in a n
Hal,
I was attempting to call ib_at_route_by_ip() in sync mode and get the following
oops:
You can change the att.c (line 193) example as follow to duplicate:
r = ib_at_route_by_ip(0x1400a8c0, 0, 0, 0,
att_rt + i, NULL, NULL);
I will go back to
On Mon, 2005-07-11 at 16:59 -0400, Hal Rosenstock wrote:
> Add kernel portion of user CM implementation
Hal, does this compile? As it doesn't seem to include the patch I sent
to openib-general changing class_simple to class, I don't think it
should work on 2.6.13-rc.
[ also, in future patch bomb
Add core locking documentation to Infiniband
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
--
core_locking.txt | 114 +++
1 files changed, 114 insertions(+)
diff -uprN linux-2.6.13-rc2-mm1-
Eliminate sparse warnings in SA client
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 20/29.
--
core/sa_query.c |6 +++--
include/ib_sa.h | 10 +--
2 files changed, 8 insertions(+), 8 deletions(-)
diff -uprN linux-2.6.13-rc2-mm1-27/drivers/infiniband/co
Hook up userspace CM to the make system
Signed-off-by: Libor Michalek <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 26/29.
--
Makefile |4 +++-
1 files changed, 3 insertions(+), 1 deletion(-)
diff -uprN linux-2.6.13-rc2-mm1-26/drivers/infi
Add kernel portion of user CM implementation
Signed-off-by: Libor Michalek <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 25/29.
--
ucm.c | 1396 ++
ucm.h | 89
2 files cha
Add the header file for user space CM. This file defines
the ABI used by the CM for kernel/user communication.
Signed-off-by: Libor Michalek <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
--
ib_user_cm.h | 328 +++
Implementation for RMPP support in user MAD
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 23/29.
--
user_mad.c | 300 +++--
1 files changed, 194 insertions(+), 106 deletions(-)
diff -uprN linux-2.6.13-rc2-mm1-23/drivers/infini
User MAD ABI changes to support RMPP
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 22/29.
--
Documentation/infiniband/user_mad.txt| 51 +++--
drivers/infiniband/include/ib_user_mad.h | 28 +--
2 files changed, 61 insertions
Add the header file for kernel CM (Communications Manager)
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 20/29.
--
ib_cm.h | 568
1 files changed, 568
Add Service Record support to SA client
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 19/29.
--
core/sa_query.c | 166 +++-
include/ib_sa.h | 75 -
2 files changed, 236 insertions(+),
Add RMPP implementation
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 18/29.
--
Makefile |2
mad.c | 167 +--
mad_priv.h | 28 +-
mad_rmpp.c | 765 +
Introduce RMPP APIs
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 17/29.
--
core/mad.c |4 +
core/sa_query.c | 20 --
include/ib_mad.h | 132 ++--
include
Replace be32_to_cpup with be32_to_cpu and fix bug referencing pointer
rather than value in ib_create_ah_from_wc().
Signed-off-by: Tom Duffy <[EMAIL PROTECTED]>
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 16/29.
--
Added new call: ib_create_ah_from_wc. Call will allocate an
address handle given work completion information, including any
received GRH.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 15/29.
--
core/verbs.c |
Fixed locking to handle error posting MAD send work requests.
Fixed handling canceling a MAD with an active work request.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 14/29.
--
mad.c | 28 ++--
1 file
On Mon, 2005-07-11 at 22:57 +0300, Michael S. Tsirkin wrote:
> Sure, I have such scripts.
> Where's the good place to put them?
wiki
-tduffy
signature.asc
Description: This is a digitally signed message part
___
openib-general mailing list
openib-gene
Optimize canceling a MAD.
- Eliminate searching timeout list in cancel case.
- Remove duplicate calls to queue work item.
- Eliminate resending a MAD before MAD is completed.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on pat
Add new MAD layer call to modify (ib_modify_mad) the timeout of a sent
MAD, and simplify cancel code.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 12/29.
--
core/mad.c | 83 ++--
core/mad_pr
Eliminate MAD cache leak associated with local completions.
Also, when canceling MAD, empty local completion list as well.
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 11/29.
--
mad.c |7 ++-
1 files changed, 6 insertions(+), 1 deletion(-)
diff -uprN l
Simplify calling of list_del.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 10/29.
--
mad.c |3 +--
1 files changed, 1 insertion(+), 2 deletions(-)
diff -uprN linux-2.6.13-rc2-mm1-10/drivers/infiniband/core/mad.c
Add automatic retries to MAD layer.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 9/29.
--
core/mad.c | 26 +-
core/mad_priv.h|2 ++
core/sa_query.c|3 ++-
core/user_m
Quoting r. Bernhard Fischer <[EMAIL PROTECTED]>:
> Subject: Re: Gen2 build/update process
>
> On Mon, Jul 11, 2005 at 12:10:48PM -0700, Tom Duffy wrote:
> >Sayantan Sur wrote:
> >
> >>Hi,
> >>
> >>Is there any plan to have any formal install procedure instead of the
> >>current process of going in
Add implementation for ib_coalesce_recv_mad.
Also, clear allocated MAD data buffer in ib_create_send_mad.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 8/29.
--
core/mad.c |9 +--
include/ib_mad.h |3 +-
On Mon, Jul 11, 2005 at 12:10:48PM -0700, Tom Duffy wrote:
>Sayantan Sur wrote:
>
>>Hi,
>>
>>Is there any plan to have any formal install procedure instead of the
>>current process of going into subdirectories and building?
>>
>>If someone has build scripts to share, that'll be real nice.
>>
>>
>
Minor cleanup during startup and shutdown
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 7/29.
--
mad.c | 44 +--
1 files changed, 9 insertions(+), 35 deletions(-)
diff -uprN linux-2.6.13-rc2-mm1-7/drivers/infiniband/core/mad.c
linux-2.6.13-rc2-mm1-8/dr
Fixes an issue processing a sent MAD after it has timed out or
been canceled. The race occurs when a response MAD matches with
the send request. The request could time out or be canceled after
the response MAD matches with the request, but before the request
completion can be processed.
Signed-
Have ib_mad_send_wr_private reference the private agent structure
directly, rather than the exposed agent definition. Remove unneeded
parameters to functions and simplify code were possible from this
change.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROT
Quoting r. Sayantan Sur <[EMAIL PROTECTED]>:
> Subject: Gen2 build/update process
>
> Hi,
>
> Is there any plan to have any formal install procedure instead of the
> current process of going into subdirectories and building?
>
> If someone has build scripts to share, that'll be real nice.
>
> T
Move saving of user's send wr_id to better match layering of received
response handling.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 4/29.
--
mad.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff
NO No more huge patch bombs to linux-kernel@vger.kernel.org
people! Condense your changes into more manageable pieces
or only submit smaller groups of changes at a time.
You people doing this are absolutely killing vger.kernel.org.
Every patch you post has to go to 5000+ subscribers, so ple
On Mon, 2005-07-11 at 15:46 -0400, Sayantan Sur wrote:
> I am building the user level support. Currently, that requires manual
> autogen, config, make, make install in lots of subdirectories. I was
> wondering if someone has a script to automate the build process. This
> will also ensure that new a
Combine response_mad() and solicited_mad() routines into a single
function and simplify/encapsulate its usage.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 3/29.
--
mad.c | 105 --
1 files changed,
Add new helper routines for allocating MADs for sending and formatting
a send WR.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 2/29.
--
core/mad.c | 76 +++
* On Jul,2 Tom Duffy<[EMAIL PROTECTED]> wrote :
> Sayantan Sur wrote:
>
> >Hi,
> >
> >Is there any plan to have any formal install procedure instead of the
> >current process of going into subdirectories and building?
> >
> >If someone has build scripts to share, that'll be real nice.
> >
> >
>
Automatically allocate a MR when registering a MAD agent.
MAD clients are modified to use this updated API.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
--
core/agent.c | 16 +--
core/agent_priv.h |3 +--
core/mad.c| 31 +++
Change some functions to return void rather than an int since they are
always returning 0, thus making checking return values rather pointless.
Signed-off-by: Tom Duffy <[EMAIL PROTECTED]>
Signed-off-by: Libor Michalek <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
---
co
This is version 2 of a patch series to get the Infiniband core up to
date. Aside from bug fixes, the following new functionality is also
introduced:
MAD (Management Datagram) support for RMPP
(Reliable MultiPacket Protocol)
CM (Communications Manager) suppo
Sayantan Sur wrote:
Hi,
Is there any plan to have any formal install procedure instead of the
current process of going into subdirectories and building?
If someone has build scripts to share, that'll be real nice.
What are you trying to build? Kernel build should not need to do this.
-t
Caitlin Bestler wrote:
Given both of those factors, I believe it makes more sense
to preserve the full DAT_RETURN value (including detail error
data) as an out parameter. kdat.h could then simply turn it
into the functional return for compatability clients and there
would be no loss of informa
Hi,
Is there any plan to have any formal install procedure instead of the
current process of going into subdirectories and building?
If someone has build scripts to share, that'll be real nice.
TIA,
Sayantan.
--
http://www.cse.ohio-state.edu/~surs
_
On Monday 11 July 2005 21:52, Tom Duffy wrote:
> Alexey Dobriyan wrote:
>
> >unsigned int __nocast gfp_mask, please. 430 or so infiniband sparse warnings
> >is not a reason to add more.
> >
> >
> Can you please elaborate on the sparse warnings that you are seeing
> throughout the rest of infini
Alexey Dobriyan wrote:
unsigned int __nocast gfp_mask, please. 430 or so infiniband sparse warnings
is not a reason to add more.
Can you please elaborate on the sparse warnings that you are seeing
throughout the rest of infiniband?
Thanks,
-tduffy
_
At 04:20 PM 7/8/2005, Kevin Reilly wrote:
Mike,
Ideally your right the ULP shouldn't care what HCA it's
running on. There are some practical reasons why an ULP might want to
know the vendor and part number it was using like for debug or taking
advantage a perform nuance of a particular HCA.
At 06:37 AM 7/11/2005, James Lentini wrote:
On Tue, 5 Jul 2005, Michael
Krause wrote:
The intention was to allow one
to manage the fabric by having mapping functions from traditional IP
management applications to IB GID to minimize the amount of work to
enable IB within a solution.
I was unaware
On Mon, 11 Jul 2005, Sean Hefty wrote:
James Lentini wrote:
Thanks for writing this up Roland. I would recommend sending it upstream.
Will this document be expanded to include the functions declared in
ib_cm.h?
I can either expand the documentation in ib_cm.h directly or add to this
fil
On Mon, 2005-07-11 at 12:29, Alexey Dobriyan wrote:
> On Monday 11 July 2005 19:30, Hal Rosenstock wrote:
> > On Mon, 2005-07-11 at 10:39, Alexey Dobriyan wrote:
> > > On Monday 11 July 2005 17:48, Hal Rosenstock wrote:
> > > > Add new helper routines for allocating MADs for sending and formatting
On Mon, 2005-07-11 at 12:35, James Lentini wrote:
> Thanks for writing this up Roland. I would recommend sending it
> upstream.
Since it seems like I will be resubmitting the core patches again, I
could include this. Any objections ?
-- Hal
___
openib
James Lentini wrote:
Thanks for writing this up Roland. I would recommend sending it upstream.
Will this document be expanded to include the functions declared in
ib_cm.h?
I can either expand the documentation in ib_cm.h directly or add to this
file. The CM seems separate from what's conta
Thanks for writing this up Roland. I would recommend sending it
upstream.
Will this document be expanded to include the functions
declared in ib_cm.h?
On Fri, 8 Jul 2005, Roland Dreier wrote:
I've tried to write up the locking rules used by our kernel IB layer,
so that they're made explic
$B(.(,(/(.(,(/(.(,(/(.(,(/(B
$B(-AG(-(-?M(-(-=P(-(-D%(-(B
$B(1(,(0(1(,(0(1(,(0(1(,(0(B
http://awg.webchu.com/?freehost
$B"(%[%9%HBgJg=8!J$*;n$740A4L5NAEPO?#O#K!*!K(B
$B"($9$Y$F%([EMAIL PROTECTED];W$$$r$7$F>.8/$$2T$.!*(B
$B"($*Ajn!S>n!M>n!&(B
$B%J!<%9!&%9%C%A!<$J$I(B
$B(.(,(/(.(,(
On Monday 11 July 2005 19:30, Hal Rosenstock wrote:
> On Mon, 2005-07-11 at 10:39, Alexey Dobriyan wrote:
> > On Monday 11 July 2005 17:48, Hal Rosenstock wrote:
> > > Add new helper routines for allocating MADs for sending and formatting
> > > a send WR.
> >
> > > -- linux-2.6.13-rc2-mm1/drivers/
On Mon, 2005-07-11 at 12:05, Nishanth Aravamudan wrote:
> On 11.07.2005 [11:30:23 -0400], Hal Rosenstock wrote:
> > On Mon, 2005-07-11 at 10:39, Alexey Dobriyan wrote:
> > > On Monday 11 July 2005 17:48, Hal Rosenstock wrote:
> > > > -- linux-2.6.13-rc2-mm1/drivers/infiniband2/core/mad.c
> > > > ++
There exists a NAL that supports the old gen1 IBAL stack and
I thought that there was one for the gen1 topspin stack, but
there is not one yet for gen2. I suspect that someone will write
one pretty soon, but do not know if anyone has started this effort
yet or not.
-Original Message-
Fr
On 11.07.2005 [11:30:23 -0400], Hal Rosenstock wrote:
> On Mon, 2005-07-11 at 10:39, Alexey Dobriyan wrote:
> > On Monday 11 July 2005 17:48, Hal Rosenstock wrote:
> > > Add new helper routines for allocating MADs for sending and formatting
> > > a send WR.
> >
> > > -- linux-2.6.13-rc2-mm1/driver
On 7/10/05, Talal jaafar <[EMAIL PROTECTED]> wrote:
> hi all,
>
> I am fairly new to OpenIB and I have downloaded and built the
> 2.6.12 kernel that has OpenIB support. My question now is how to
> compile the OpenIB modules in that kernel; There is no documentation,
> and all what I need to do i
On Mon, 2005-07-11 at 10:39, Alexey Dobriyan wrote:
> On Monday 11 July 2005 17:48, Hal Rosenstock wrote:
> > Add new helper routines for allocating MADs for sending and formatting
> > a send WR.
>
> > -- linux-2.6.13-rc2-mm1/drivers/infiniband2/core/mad.c
> > +++ linux-2.6.13-rc2-mm1/drivers/infi
Hi Eitan,
A few nits with the latest:
1. Still get:
g++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/local/include
-I/home/hal/openib/trunk/src/userspace/management/osm/include -I/usr/include
-I../datamodel -O6 -Wall -Werror -c -o sim.o `test -f 'sim.cpp' || echo
'./'`sim.cpp
sim.cpp:342: parse erro
On Sun, 2005-07-10 at 20:23, Talal jaafar wrote:
> hi all,
>
>I am fairly new to OpenIB and I have downloaded and built the
> 2.6.12 kernel that has OpenIB support. My question now is how to
> compile the OpenIB modules in that kernel; There is no documentation,
> and all what I need to do is
Hook up userspace CM to the make system
Signed-off-by: Libor Michalek <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 26/27.
--
Makefile |4 +++-
1 files changed, 3 insertions(+), 1 deletion(-)
diff -uprN linux-2.6.13-rc2-mm1/drivers/infin
Add kernel portion of user CM implementation
Signed-off-by: Libor Michalek <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 25/27.
--
ucm.c | 1396 ++
ucm.h | 89
2 files cha
Add the header file for user space CM. This file defines
the ABI used by the CM for kernel/user communication.
Signed-off-by: Libor Michalek <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
--
ib_user_cm.h | 328 +++
Implementation for RMPP support in user MAD
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 23/27.
--
user_mad.c | 300 +++--
1 files changed, 194 insertions(+), 106 deletions(-)
diff -uprN linux-2.6.13-rc2-mm1/drivers/infiniba
User MAD ABI changes to support RMPP
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 22/27.
--
ib_user_mad.h | 28 ++--
1 files changed, 22 insertions(+), 6 deletions(-)
user_mad.txt | 51 +++--
1 files
Add the header file for kernel CM (Communications Manager)
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
--
ib_cm.h | 568
1 files changed, 568 insertions(+)
diff -uprN linux-2.6.
Add Service Record support to SA client
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 19/27.
--
core/sa_query.c | 166 +++-
include/ib_sa.h | 75 -
2 files changed, 236 insertions(+),
Add RMPP implementation
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 18/27.
--
Makefile |2
mad.c | 167 +--
mad_priv.h | 28 +-
mad_rmpp.c | 765 +++
Introduce RMPP APIs
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 17/27.
--
core/mad.c |4 +
core/sa_query.c | 20 --
include/ib_mad.h | 132 ++--
include
Replace be32_to_cpup with be32_to_cpu and fix bug referencing pointer
rather than value in ib_create_ah_from_wc().
Signed-off-by: Tom Duffy <[EMAIL PROTECTED]>
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 16/27.
--
FYI, you can find a kDAPL implementation for OpenIB at
svn/gen2/users/jlentini/linux-kernel
and a uDAPL implementation for OpenIB at
svn/gen2/users/jlentini/userspace
I can answer any questions you might have about these ULPs.
I'm not familiar with NAL, so I can't comment on which ULP is rig
Added new call: ib_create_ah_from_wc. Call will allocate an
address handle given work completion information, including any
received GRH.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 15/27.
--
core/verbs.c |
Fixed locking to handle error posting MAD send work requests.
Fixed handling canceling a MAD with an active work request.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 14/27.
--
mad.c | 28 ++--
1 file
On Monday 11 July 2005 17:48, Hal Rosenstock wrote:
> Add new helper routines for allocating MADs for sending and formatting
> a send WR.
> -- linux-2.6.13-rc2-mm1/drivers/infiniband2/core/mad.c
> +++ linux-2.6.13-rc2-mm1/drivers/infiniband3/core/mad.c
^^
Optimize canceling a MAD.
- Eliminate searching timeout list in cancel case.
- Remove duplicate calls to queue work item.
- Eliminate resending a MAD before MAD is completed.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on pat
Add new MAD layer call to modify (ib_modify_mad) the timeout of a sent
MAD, and simplify cancel code.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 12/27.
--
core/mad.c | 83 ++--
core/mad_pr
Eliminate MAD cache leak associated with local completions.
Also, when canceling MAD, empty local completion list as well.
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 11/27.
--
mad.c |7 ++-
1 files changed, 6 insertions(+), 1 deletion(-)
diff -uprN l
Simplify calling of list_del.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 10/27.
--
mad.c |3 +--
1 files changed, 1 insertion(+), 2 deletions(-)
diff -uprN linux-2.6.13-rc2-mm1/drivers/infiniband10/core/mad.c
Add automatic retries to MAD layer.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 9/27.
--
core/mad.c | 26 +-
core/mad_priv.h|2 ++
core/sa_query.c|3 ++-
core/user_
http://awg.webchu.com/nanpara/?present1
$B=i$a$^$7$F!*!*FMA3$N%a!<%k$4LBOG$G$7$?$i$*5v$72<$5$$!#(B
$B?75,%5%$%HN)$A>e$2$KH<$$!"0lK|1_(B(1000$B%]%$%s%H(B)$BJ,%W%l%<%s%H$N$*CN$i$;$G$9!#(B
http://awg.webchu.com/nanpara/?present1
$B!X$^$?=P2q$$7O$+$!!Y$H;W$o$J$$$G2<$5$$!*!*!#(B
$B;dC#$b3'
Add implementation for ib_coalesce_recv_mad.
Also, clear allocated MAD data buffer in ib_create_send_mad.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 8/27.
--
core/mad.c |9 +--
include/ib_mad.h |3 +
Minor cleanup during startup and shutdown
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 7/27.
--
mad.c | 44 +--
1 files changed, 9 insertions(+), 35 deletions(-)
diff -uprN linux-2.6.13-rc2-mm1/drivers/infiniband7/core/mad.c
linux-2.6.13-rc2-mm1/driv
Fixes an issue processing a sent MAD after it has timed out or
been canceled. The race occurs when a response MAD matches with
the send request. The request could time out or be canceled after
the response MAD matches with the request, but before the request
completion can be processed.
Signed-
Have ib_mad_send_wr_private reference the private agent structure
directly, rather than the exposed agent definition. Remove unneeded
parameters to functions and simplify code were possible from this
change.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTE
Move saving of user's send wr_id to better match layering of received
response handling.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 4/27.
--
mad.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff
Combine response_mad() and solicited_mad() routines into a single
function and simplify/encapsulate its usage.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 3/27.
--
mad.c | 105 --
1 files changed,
Add new helper routines for allocating MADs for sending and formatting
a send WR.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
This patch depends on patch 2/27.
--
core/mad.c | 76 +++
Automatically allocate a MR when registering a MAD agent.
MAD clients are modified to use this updated API.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
Signed-off-by: Hal Rosenstock <[EMAIL PROTECTED]>
---
core/agent.c | 16 +--
core/agent_priv.h |3 +--
core/mad.c| 31 +
1 - 100 of 104 matches
Mail list logo