Sean> Is there any interest among people to reuse receive MADs?
Sean> I.e. once allocated and mapped, the receive MAD and work
Sean> request would be re-posted to the QP when freed.
I'm not sure this is that useful... MAD processing is not such a
super-hot path that we need to keep pe
Is there any interest among people to reuse receive MADs? I.e. once
allocated and mapped, the receive MAD and work request would be
re-posted to the QP when freed.
I ask because if people are interested in such an optimization at some
point in the future, it will affect how I structure send qu
On Thu, 2004-11-04 at 16:31, Krishna Kumar wrote:
> Also, updated a comment so that it is known that the recv_handler
> (when it is implemented) is in charge of freeing up recv during it's
> processing. Applies to gen2/trunk.
Thanks. Applied the commentary part of the change.
The memory leak "fi
Also, updated a comment so that it is known that the recv_handler
(when it is implemented) is in charge of freeing up recv during it's
processing. Applies to gen2/trunk.
Thanks,
- KK
diff -ruNp 1/mad.c 2/mad.c
--- 1/mad.c 2004-11-04 10:38:30.0 -0800
+++ 2/mad.c 2004-11-04 13:26:3
mad: Removed print of "No client for received MAD" as this can be a
normal case
Index: mad.c
===
--- mad.c (revision 1139)
+++ mad.c (working copy)
@@ -798,9 +798,7 @@
&mad_agent->agent, port
On Thu, 2004-11-04 at 15:44, Shirley Ma wrote:
> Please review this patch.
Thanks. Applied.
-- Hal
___
openib-general mailing list
[EMAIL PROTECTED]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailm
Please review this patch.
diff -urN infiniband/core/agent.c infiniband.patch/core/agent.c
--- infiniband/core/agent.c 2004-11-04 10:35:20.0 -0800
+++ infiniband.patch/core/agent.c 2004-11-04 12:35:55.916027072 -0800
@@ -480,6 +480,7 @@
/* Release allocated memory */
On Thu, 2004-11-04 at 13:14, Shirley Ma wrote:
> So everybody should start working on this tree. What's the difference
> between openib-candidate and trunk under gen2?
trunk is much more complete with IPoIB. This is what is heading towards
being pushed to the 2.6 kernel. Yes, you should use this t
So everybody should start working on this tree. What's the difference between openib-candidate and trunk under gen2?
Thanks
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone(Fax): (503) 578-7638
Roland Dreier <[EMAIL PROTECTED]>
Roland Dreier <[
On Thu, 4 Nov 2004, Roland Dreier wrote:
> Not sure what the goal is here, but I should point out that current
> mthca code does not implement resizing either CQs or QPs.
Yes, I agree on that. Infact the verbs layer will return ENOSYS for
mthca driver. But I was assuming that any other driver by
I have just copied the roland-merge branch to
https://openib.org/svn/gen2/trunk
This tree will become the main development tree and will be used to
create the tree we will submit to the kernel for inclusion. Please
use this tree for testing and as the base for all patches.
I will be cleanin
OK, I merged the MAD code in my branch up to r1135 and applied this
patch (there was one missing chunk in ib_verbs.h to remove the
snoop_mad method from struct ib_device, which I added by hand).
Thanks,
Roland
___
openib-general mailing list
[EMAIL PRO
Hello!
Quoting r. Ronald G. Minnich ([EMAIL PROTECTED]) "Re: [openib-general] announcement:
mstflint flash burning package uploaded":
>
>
> On Thu, 4 Nov 2004, Michael S. Tsirkin wrote:
>
> > I have uploaded an mstflint flash burning package to openib.org.
> > You can find it here: https://open
Hello!
Quoting r. Roland Dreier ([EMAIL PROTECTED]) "Re: [openib-general] announcement:
mstflint flash burning package uploaded":
> Ronald> neat. How does this differ from tvflash that Roland wrote?
>
> correction: I just cleaned up the code.
Something mstflint could benefit from, too :)
MS
Hello!
Quoting r. Ronald G. Minnich ([EMAIL PROTECTED]) "Re: [openib-general] announcement:
mstflint flash burning package uploaded":
>
>
> On Thu, 4 Nov 2004, Michael S. Tsirkin wrote:
>
> > I have uploaded an mstflint flash burning package to openib.org.
> > You can find it here: https://open
Michael> /dev/ib/hca0/ports/1/mad
I like this idea.
Thanks,
Roland
___
openib-general mailing list
[EMAIL PROTECTED]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Ronald> neat. How does this differ from tvflash that Roland wrote?
correction: I just cleaned up the code. Kamen and Johannes here at
Topspin did most of the real work in writing tvflash...
- R
___
openib-general mailing list
[EMAIL PROTECTED]
htt
Hal> This requires the MAD layer to peek into the outgoing MAD but
Hal> all it has is the DMA address in the sg-list and the last
Hal> time I tried to do this (go from DMA address to a VA), the
Hal> approach used to do this was in the process of being
Hal> deprecated. Last time,
On Mon, 2004-11-01 at 17:15, Roland Dreier wrote:
> I think SMI processing should be applied to all DR SMPs passed to
> ib_post_send_mad().
This requires the MAD layer to peek into the outgoing MAD but all it has
is the DMA address in the sg-list and the last time I tried to do this
(go from DMA
On Thu, 4 Nov 2004, Michael S. Tsirkin wrote:
> I have uploaded an mstflint flash burning package to openib.org.
> You can find it here: https://openib.org/svn/trunk/contrib/mellanox/mstflint/
neat. How does this differ from tvflash that Roland wrote?
thanks
ron
__
Hello!
Quoting r. Roland Dreier ([EMAIL PROTECTED]) "Re: [openib-general] [PATCH 1/2] [RFC]
Implement resize of CQ":
> Michael> If the max. number of QPs is very big, you may want the
> Michael> actual CQ size to grow gradually with demand.
>
> sure but there are only 2 soecial qps per po
agent: Minor modifications to smi_check_local_xxx routines
Index: agent.c
===
-- agent.c (revision 1133)
+++ agent.c (working copy)
@@ -117,8 +117,7 @@
{
/* C14-9:3 -- We're at the end of the DR segment of path */
/* C14-9:4 -- Hop Po
agent: Change calling argument to agent_mad_send
Rather than taking a struct ib_mad_recv_wc *, take a struct ib_wc *
Index: agent.c
===
--- agent.c (revision 1131)
+++ agent.c (working copy)
@@ -296,7 +296,7 @@
static int age
Michael> If the max. number of QPs is very big, you may want the
Michael> actual CQ size to grow gradually with demand.
sure but there are only 2 soecial qps per port.
___
openib-general mailing list
[EMAIL PROTECTED]
http://openib.org/mailman/l
On Wed, 2004-11-03 at 21:23, Krishna Kumar wrote:
> Now messages are printed if either :
>
> 1. mad_agent is not found.
> 2. mad_agent is found but doesn't have a handler.
>
> Printing messages during errors in the process of finding the
> mad_agent has been removed.
Thanks. Applied.
-- Hal
_
On Wed, 2004-11-03 at 18:52, Krishna Kumar wrote:
> Hi Hal,
>
> The same patch on latest bits
Thanks. Applied.
-- Hal
___
openib-general mailing list
[EMAIL PROTECTED]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit
Hello!
I have uploaded an mstflint flash burning package to openib.org.
You can find it here: https://openib.org/svn/trunk/contrib/mellanox/mstflint/
This is an update to the original flint utility that makes it possible to
perform flash burning without loading special kernel level drivers,
perfo
Hello!
Quoting r. Johannes Erdfelt ([EMAIL PROTECTED]) "Re: [openib-general] [PATCH] Initial
checkin of userspace MAD access":
> On Wed, Nov 03, 2004, Roland Dreier <[EMAIL PROTECTED]> wrote:
> > By the way, buried down at the end of the patch is some documentation
> > about creating device files:
If the max. number of QPs is very big, you may want
the actual CQ size to grow gradually with demand.
Quoting r. Roland Dreier ([EMAIL PROTECTED]) "Re: [openib-general] [PATCH 1/2] [RFC]
Implement resize of CQ":
> Not sure what the goal is here, but I should point out that current
> mthca code d
29 matches
Mail list logo
