Hi,
A long time ago, the openib driver used a hack to call sys_mlock() to
lock down a user space memory region. This was because get_user_pages()
wasn't completely locking the region like it was supposed to.
I haven't paid much attention to the openib stuff since then, but now I
want to know w
On Mon, Mar 07, 2005 at 01:26:19PM -0800, Sean Hefty wrote:
> Libor Michalek wrote:
> > Sean, I'm not sure there's an easy way to perform a port source GID
> > to device lookup, did you have something specific in mind for the
> > lookup? Unless there's an easy way to do this, I was going to go ahea
Libor Michalek wrote:
Sean, I'm not sure there's an easy way to perform a port source GID
to device lookup, did you have something specific in mind for the
lookup? Unless there's an easy way to do this, I was going to go ahead
with #1...
There's a device_list maintained in device.c that's used when
On Sun, Mar 06, 2005 at 12:38:40PM +0200, Michael S. Tsirkin wrote:
> Quoting r. Libor Michalek <[EMAIL PROTECTED]>:
> > Subject: Re: [openib-general] SDP_CONN_LOCK
> > > > They do implement exclusive access to the socket, but they implement
> > > > exclusive access from both process and irq cont
On Mon, Mar 07, 2005 at 11:14:07AM -0800, Sean Hefty wrote:
> Roland Dreier wrote:
> > 1. Don't worry about checking. There's nothing too evil a CM user
> > can do with a QP beyond getting another QP to connect to it, since
> > the CM user can't modify a QP unless it legitimately owns it.
On Wed, Mar 02, 2005 at 10:11:34AM -0800, Tom Duffy wrote:
> Now that 2.6.11 is out, need to make sdp compile with 2.6.11.
>
Thanks Tom, applied and commited.
-Libor
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/li
Roland Dreier wrote:
1. Don't worry about checking. There's nothing too evil a CM user
can do with a QP beyond getting another QP to connect to it, since
the CM user can't modify a QP unless it legitimately owns it. And
an evil user can always guess the QPN instead of the QP handle a
Just to make this clear:
- There are only two QP that are defined with specific intention - QP0
and QP1. All other QP may vary throughout the entire QP
space.
- All ULP built on top of IB must assume that the QP are variant and must
discover these through various protocol such as the service
Shirley> I am starting to test mthca driver on PPC64. I want to
Shirley> know whether someone has tested mthca on any PPC platform?
Yes, I have used it successfully on IBM p630 and JS20 systems.
- Roland
___
openib-general mailing list
openib-g
I've just committed support for UD address handles to libibverbs and
libmthca on the roland-uverbs branch. This makes it possible to use
UD from userspace. libibverbs/examples has a new ud-pingpong.c demo
program that shows how this works.
There are no new changes to the kernel uverbs support on
I am starting to test mthca driver on PPC64. I want to know whether someone has tested mthca on any PPC platform?
Thanks
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone(Fax): (503) 578-7638
___
openib-general
On Mon, 2005-03-07 at 11:55, Roland Dreier wrote:
> OK, it won't be hard to fill out those entries. What application is
> using this info?
uDAPL (and kDAPL) use these device attributes currently.
-- Hal
___
openib-general mailing list
openib-general@o
Hal> Certainly the following ones:
OK, it won't be hard to fill out those entries. What application is
using this info?
- R.
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscrib
On Mon, 2005-03-07 at 11:32, Roland Dreier wrote:
> Hal> It appears that mthca_provider.c::mthca_query_device does not
> Hal> fill in many of the device attributes in struct
> Hal> ib_device_attr. When can the remainder of these be completed?
>
> Any time... which ones are needed now?
Hal> It appears that mthca_provider.c::mthca_query_device does not
Hal> fill in many of the device attributes in struct
Hal> ib_device_attr. When can the remainder of these be completed?
Any time... which ones are needed now?
- R.
___
open
Hi Roland,
It appears that mthca_provider.c::mthca_query_device does not fill in
many of the device attributes in struct ib_device_attr. When can the
remainder of these be completed ? Thanks.
-- Hal
___
openib-general mailing list
openib-general@openib
Quoting r. Michael S. Tsirkin <[EMAIL PROTECTED]>:
> Subject: rev 1954 - new flint uploaded
>
> With revision 1954 I have uploaded new flint code,
> synched with current mellanox code.
I made some last minute fixes and uploaded rev 1957.
Enjoy!
> There are lots of changes, since this adds suppor
17 matches
Mail list logo