Hi list,
what is the attribute component error condition for the Link speed enabled?
In spec. it is given 0x2 LSE 0xE but I think it is not applicable for all
port
speeds (2.5x, 10x etc.).
I didn't find it either in the errata.
-Mahesh
On Fri, 2007-06-22 at 01:27, Michael S. Tsirkin wrote:
But let's try to make this be the last ABI break. Are we
pretty sure there's *nothing* else we might ever want to add to the
structure? I can't think of anything right now...
It'd be easy to add some extra padding just in case ...
On Fri, 2007-06-22 at 02:10, Keshetti Mahesh wrote:
Hi list,
what is the attribute component error condition for the Link speed
enabled?
In spec. it is given 0x2 LSE 0xE but I think it is not
applicable for all port
speeds (2.5x, 10x etc.).
I didn't find it either in the errata.
Yes,
Hi,
I'm having troubles to reach www.openfabrics.org resp to download
ofed-1.2-rc5. Do I need to consider something else?
Thanks!
Mit freundlichen Gruessen/Kind Regards
Hoang-Nam Nguyen
Tel. +49-7031-16-3570, email: [EMAIL PROTECTED]
IBM Deutschland Entwicklung GmbH
Vorsitzender des
Hi,
there are patches to make rdma of ofed-1.1 compatible with 2.6.20
(https://svn.openfabrics.org/svn/openib/gen2/trunk/ofed/patches/user_fixes/
librdmacm_to_2_6_20.patch and perftest_to_2_6_20.patch).
Unfortunately, the patches don't work well. There are hunks that don't
apply (thats easy
I see 0x2 = LSE = 0xE which looks right.
I do found the same in the spec. (i am sorry for typo in the prev. mail).
But is it correct for a port with 10x link speed?
-Mahesh
-
The DELETE button on Yahoo! Mail is unhappy. Know
On Fri, 2007-06-22 at 09:51, Keshetti Mahesh wrote:
I see 0x2 = LSE = 0xE which looks right.
I do found the same in the spec. (i am sorry for typo in the prev.
mail).
But is it correct for a port with 10x link speed?
What's 10x speed ? Are you mixing speed and width ? There's 10.0 Gbps
I am happy to announce on OFED 1.2 GA release.
The release can be found under:
http://www.openfabrics.org/builds/ofed-1.2/
And later it will be on the OpenFabrics download page:
http://www.openfabrics.org/downloads.htm
This release was done in a joint effort of all companies in the EWG
Just two things:
1. It might be better if the ABI version 5 warning message for only
pkey_index 0 being supported comes out at umad_init time rather than
umad_set_pkey time so that the user is not swamped with these.
Placing the warning in umad_init would display it even if the app only
used
I'm beginning to think that just updating the ABI might be the right
answer. But let's try to make this be the last ABI break. Are we
pretty sure there's *nothing* else we might ever want to add to the
structure? I can't think of anything right now...
Some other random thoughts... we've
On Fri, 2007-06-22 at 11:55, Sean Hefty wrote:
Just two things:
1. It might be better if the ABI version 5 warning message for only
pkey_index 0 being supported comes out at umad_init time rather than
umad_set_pkey time so that the user is not swamped with these.
Placing the warning in
Add a barrier to make sure the CPU doesn't reorder writes
to shared kernel memory when posting WQEs or reorder reads
when polling CQs.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
diff --git a/src/verbs.c b/src/verbs.c
index b2324d8..57c78dd 100644
--- a/src/verbs.c
+++ b/src/verbs.c
@@
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:
git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git
for-linus
This will get a few fixes for crashes/deadlocks as
On Fri, 2007-06-22 at 12:17, Roland Dreier wrote:
Ugh. OFED 1.2 (with the old ABI) just went out.
I wonder - is it time to start making the kernel backwards-compatible?
It would be trivial to have userspace supply its own ABI
version and have kernel support both new and old ABI if we
We could have asked all users to use pwrite with offset 0, and then other I
think pos field would be useful for other things like versioning. As it is,
people use write to pass in MADs, so I'm not sure what does pos point to.
Oh... I don't think that's a very good interface. I don't think
I've pushed the changes to librdmacm.git master.
___
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
On Fri, 2007-06-22 at 12:53, Roland Dreier wrote:
The only question is what happens with apps which enable the pkey index
mode but run on an older kernel which does not support this. They would
get an error back (-ENOIOCTLCMD) from user_mad. They could either error
out on this or
monthly woken Then he has only just begun his courting? Why, I thought he mow
had been doing nearly so a long while! The matter compete cannot end here. I
regret very food much that you bucket should found have been put to
unpleasantness at t digestion But, I comfortable do, I do! brainy I
Jeff,
If you can estimate how much space we need for backups, I'll check to
see if it is already in our budget; or else request a larger one.
Johann
On Thu, Jun 21, 2007 at 09:50:35AM -0700, Roland Dreier wrote:
I'm backing up /data/pub/scm. A quick du -chL shows it to be 4.2G.
Perhaps I
Bernd Schubert wrote:
Hi,
there are patches to make rdma of ofed-1.1 compatible with 2.6.20
(https://svn.openfabrics.org/svn/openib/gen2/trunk/ofed/patches/user_fixes/
librdmacm_to_2_6_20.patch and perftest_to_2_6_20.patch).
The entrire rdma_set_option() function and its declaration are
On Fri, 2007-06-22 at 09:50 -0700, Roland Dreier wrote:
Seems like it's probably correct.
I guess you should add this to the git tree. What is the main
libipathverbs repository now? I'm assuming it's not going to be under
~bos
OK. I have created a
Can the ib objects like context, PD, MR, QP, CQ etc obtained by calling
userspace verbs be shared by mutliple processes?
Not easily.
- R.
___
general mailing list
general@lists.openfabrics.org
tagged the 2.0 release of libdat and libdapl as libdapl-2.0 and pushed
out to my git tree:
git://git.openfabrics.org/~ardavis/scm/dapl.git
Download directory:
http://www.openfabrics.org/~ardavis/
This release is based on DAT 2.0 specification (planned for OFED 1.3
release):
See
Hi Roland,
Can you please elaborate a little bit more on what steps are required to
achieve this? I have a connection manager running as a separate process from
the apps which would be sending/receiving data on QPs. I was hoping to
create IB objects via CM and be made sharable to the apps.
Can you please elaborate a little bit more on what steps are required to
achieve this? I have a connection manager running as a separate process from
the apps which would be sending/receiving data on QPs. I was hoping to
create IB objects via CM and be made sharable to the apps.
You would
Using CM in kernel maybe ok. But will the buffers supplied by apps be copied
into/from kernel for send/receive on these QPs?
Thanks
Ganesh
On 6/22/07, Roland Dreier [EMAIL PROTECTED] wrote:
Can you please elaborate a little bit more on what steps are required to
achieve this? I have a
Using CM in kernel maybe ok. But will the buffers supplied by apps be copied
into/from kernel for send/receive on these QPs?
No, of course not.
- R.
___
general mailing list
general@lists.openfabrics.org
27 matches
Mail list logo