On 4/15/2015 4:07 PM, Doug Ledford wrote:
On Apr 15, 2015, at 9:02 AM, Sagi Grimberg sa...@dev.mellanox.co.il wrote:
On 4/15/2015 3:39 PM, Doug Ledford wrote:
On Apr 15, 2015, at 7:43 AM, Pradeep Kankipati pradeep.kankip...@emulex.com
wrote:
We at Emulex, have been working with Doug on
On Apr 2, 2015, at 6:39 AM, Or Gerlitz ogerl...@mellanox.com wrote:
From: Erez Shitrit ere...@mellanox.com
As the result of a completion error the QP can moved to SQE state by
the hardware. Since it's not the Error state, there are no flushes
and hence the driver doesn't know about that.
On Apr 15, 2015, at 9:02 AM, Sagi Grimberg sa...@dev.mellanox.co.il wrote:
On 4/15/2015 3:39 PM, Doug Ledford wrote:
On Apr 15, 2015, at 7:43 AM, Pradeep Kankipati
pradeep.kankip...@emulex.com wrote:
We at Emulex, have been working with Doug on the RH inboxing effort for a
while and
Hi Doug
(now with reply to all :) )
On Wed, Apr 15, 2015 at 3:08 PM, Doug Ledford dledf...@redhat.com wrote:
On Apr 2, 2015, at 6:39 AM, Or Gerlitz ogerl...@mellanox.com wrote:
From: Erez Shitrit ere...@mellanox.com
As the result of a completion error the QP can moved to SQE state by
the
On 4/15/2015 3:39 PM, Doug Ledford wrote:
On Apr 15, 2015, at 7:43 AM, Pradeep Kankipati pradeep.kankip...@emulex.com
wrote:
We at Emulex, have been working with Doug on the RH inboxing effort for a while
and feel this just makes complete sense for him to take on this role for all
the
On Apr 15, 2015, at 7:43 AM, Pradeep Kankipati pradeep.kankip...@emulex.com
wrote:
We at Emulex, have been working with Doug on the RH inboxing effort for a
while and feel this just makes complete sense for him to take on this role
for all the reasons mentioned. Personally too I feel he
On Mon, Apr 13, 2015 at 10:38:08AM -0600, Jason Gunthorpe wrote:
On Mon, Apr 13, 2015 at 07:34:23PM +0530, Varun Prakash wrote:
define struct ulptx_idata in common header file t4_msg.h
to remove duplicate definitions.
The Infiniband side of this patch looks OK.
Reviewed-By: Jason
On 4/15/2015 6:17 PM, Guy Shapiro wrote:
While unmapping an ODP writable page, the dirty bit of the page is set. In
order to do so, the head of the compound page is found.
Currently, the compound head is found even on non-writable pages, where it is
never used, leading to unnecessary cpu barrier
This patchset contains small changes that improve the performance of memory
map updates.
The patches reduce the execution time of ib_umem_odp_unmap_dma_pages when
invalidating 4MB of memory from 46us to 13us. This amounts to a reduction of
about 30% in the total execution time of the invalidation
While unmapping an ODP writable page, the dirty bit of the page is set. In
order to do so, the head of the compound page is found.
Currently, the compound head is found even on non-writable pages, where it is
never used, leading to unnecessary cpu barrier that impacts performance.
This patch
Currently, while mapping or unmapping pages for ODP, the umem mutex is locked
and unlocked once for each page. Such lock/unlock operation take few tens to
hundreds of nsecs. This makes a significant impact when mapping or unmapping few
MBs of memory.
To avoid this, the mutex should be locked only
On 4/15/2015 6:17 PM, Guy Shapiro wrote:
Currently, while mapping or unmapping pages for ODP, the umem mutex is locked
and unlocked once for each page. Such lock/unlock operation take few tens to
hundreds of nsecs. This makes a significant impact when mapping or unmapping few
MBs of memory.
To
It does not break every app, the choice of which GID type to use is made
by the RDMA-CM based on network topology hint obtained from the IP stack.
Please refer to patch 15/33: IB/Core: Changes to the IB Core
infrastructure for RoCEv2 support.
Of course, if the user does not want to go with
IMHO, it would be good to have a physical layer representation in the naming
convention.
-Original Message-
From: linux-rdma-ow...@vger.kernel.org
[mailto:linux-rdma-ow...@vger.kernel.org] On Behalf Of Hefty, Sean
Sent: Wednesday, April 15, 2015 12:08 PM
To: Somnath Kotur; Matan Barak;
On Wed, Apr 15, 2015 at 5:39 AM, Doug Ledford dledf...@redhat.com wrote:
Regardless of what we end up doing long term, Roland is out at the moment, so
4.1 is happening more or less without him :-/
Sorry. I am back at work now.
I went ahead and put together a 4.1 branch and pulled the
On Wed, Apr 15, 2015 at 06:34:27PM +0530, Varun Prakash wrote:
On Mon, Apr 13, 2015 at 10:38:08AM -0600, Jason Gunthorpe wrote:
On Mon, Apr 13, 2015 at 07:34:23PM +0530, Varun Prakash wrote:
define struct ulptx_idata in common header file t4_msg.h
to remove duplicate definitions.
The
On Wed, Apr 15, 2015 at 09:17:14AM +0300, Erez Shitrit wrote:
+ /* parent interface */
+ if (!test_bit(IPOIB_FLAG_SUBINTERFACE, priv-flags))
+ return dev-ifindex;
+
+ /* child/vlan interface */
+ if (!priv-parent)
+ return -1;
Like was
-Original Message-
From: linux-rdma-ow...@vger.kernel.org [mailto:linux-rdma-
ow...@vger.kernel.org] On Behalf Of Or Gerlitz
Sent: Thursday, April 02, 2015 4:09 PM
To: Roland Dreier; Doug Ledford
Cc: linux-rdma@vger.kernel.org; Erez Shitrit; Tal Alon; Amir Vadai; Or Gerlitz
Subject:
On Wed, Apr 15, 2015 at 8:13 PM, Roland Dreier rol...@kernel.org wrote:
On Wed, Apr 15, 2015 at 5:39 AM, Doug Ledford dledf...@redhat.com wrote:
Regardless of what we end up doing long term, Roland is out at the moment,
so 4.1 is happening more or less without him :-/
Sorry. I am back at
On Wed, Apr 15, 2015 at 6:54 AM, Or Gerlitz ogerl...@mellanox.com wrote:
There were two trivial conflicts between the mlx4 changes to another mlx4
series which is present in net-next, I resolved them and you can pull the
fixed series from my git tree,SB.
Not sure how you expect this to work.
Hey Doug/Roland,
Please include this in the 4.1 merge...
Thanks,
Steve.
On 3/19/2015 11:43 AM, Hariprasad Shenai wrote:
Hi,
This patch series changes a macro definition to be consistent with other
register macro defines. Fixes for 32b platform, use BAR2 register for kernel
mode CQ's and
On 4/13/2015 8:22 AM, Michael Wang wrote:
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 65994a1..d54f91e 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -75,10 +75,13 @@ enum rdma_node_type {
};
enum rdma_transport_type {
+ /* legacy
On 4/13/2015 3:50 PM, Jason Gunthorpe wrote:
Less clear is how rocee vs ib work within a device... Can you APM
between those two kinds of ports?
The specs allow this to work but AFAIK it's not implemented.
--
To unsubscribe from this list: send the line unsubscribe linux-rdma in
the body of a
-Original Message-
From: linux-rdma-ow...@vger.kernel.org
[mailto:linux-rdma-ow...@vger.kernel.org] On Behalf Of Doug Ledford
Sent: Monday, April 13, 2015 12:15 PM
To: Jason Gunthorpe
Cc: Yann Droneaud; Hefty, Sean; Roland Dreier; Hal Rosenstock
(h...@dev.mellanox.co.il); David
On Wed, Apr 15, 2015 at 02:36:13PM -0400, Hal Rosenstock wrote:
On 4/13/2015 8:22 AM, Michael Wang wrote:
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 65994a1..d54f91e 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -75,10 +75,13 @@ enum
On Wed, 2015-04-15 at 13:52 -0500, Steve Wise wrote:
Hey Doug/Roland,
Please include this in the 4.1 merge...
These patches no longer apply cleanly.
Thanks,
Steve.
On 3/19/2015 11:43 AM, Hariprasad Shenai wrote:
Hi,
This patch series changes a macro definition to be
On Wed, Apr 15, 2015 at 01:29:31PM -0600, Hefty, Sean wrote:
libibverbs also exposes transport at the device level. Isn't a change to
make transport per port rather than per device needed there as well to
be consistent with these proposed kernel changes ? If so, would the
additional IBoE
Ok, Hari can you please resubmit. Should it be based on your github for-4.1
branch?
-Original Message-
From: Doug Ledford [mailto:dledf...@redhat.com]
Sent: Wednesday, April 15, 2015 3:33 PM
To: Steve Wise
Cc: rol...@purestorage.com; Hariprasad Shenai;
On 04/15/2015 01:42 PM, Andy Lutomirski wrote:
I disagree. We should try to NACK any new code that can't function
without MTRRs.
(Plus, ARM is growing in popularity in the server space, and ARM quite
sensibly doesn't have MTRRs.)
NOT SPEAKING FOR INTEL HERE
Yes. People need to
libibverbs also exposes transport at the device level. Isn't a change to
make transport per port rather than per device needed there as well to
be consistent with these proposed kernel changes ? If so, would the
additional IBoE transport be exposed ? We also need to worry about
backward
On Wed, Apr 15, 2015 at 3:50 PM, Andy Walls awa...@md.metrocast.net wrote:
On Wed, 2015-04-15 at 13:42 -0700, Andy Lutomirski wrote:
On Mon, Apr 13, 2015 at 10:49 AM, Luis R. Rodriguez mcg...@suse.com wrote:
c) ivtv: the driver does not have the PCI space mapped out separately, and
in fact
On Wed, Apr 15, 2015 at 01:42:47PM -0700, Andy Lutomirski wrote:
On Mon, Apr 13, 2015 at 10:49 AM, Luis R. Rodriguez mcg...@suse.com wrote:
c) ivtv: the driver does not have the PCI space mapped out separately, and
in fact it actually does not do the math for the framebuffer, instead it
On Wed, 2015-04-15 at 13:42 -0700, Andy Lutomirski wrote:
On Mon, Apr 13, 2015 at 10:49 AM, Luis R. Rodriguez mcg...@suse.com wrote:
c) ivtv: the driver does not have the PCI space mapped out separately, and
in fact it actually does not do the math for the framebuffer, instead it
lets
On Wed, 2015-04-15 at 16:52 -0700, Andy Lutomirski wrote:
On Wed, Apr 15, 2015 at 3:50 PM, Andy Walls awa...@md.metrocast.net wrote:
On Wed, 2015-04-15 at 13:42 -0700, Andy Lutomirski wrote:
On Mon, Apr 13, 2015 at 10:49 AM, Luis R. Rodriguez mcg...@suse.com
wrote:
c) ivtv: the driver
On Wed, 2015-04-15 at 15:55 -0500, Steve Wise wrote:
Ok, Hari can you please resubmit. Should it be based on your github for-4.1
branch?
Yes please.
-Original Message-
From: Doug Ledford [mailto:dledf...@redhat.com]
Sent: Wednesday, April 15, 2015 3:33 PM
To: Steve
There was network accident in the office. I can't find this email in
public mailing list. So, I reset it. If you had recived this, please
ignore it.
thanks
On Wed, Apr 15, 2015 at 01:16:40PM +0800, Honggang LI wrote:
On Tue, Apr 14, 2015 at 07:30:03PM +0300, Erez Shitrit wrote:
@@ -846,7
On 4/15/2015 9:21 AM, Honggang Li wrote:
Fixes: bf0bf77f6519 ('mlx5: Support communicating arbitrary host page size to
firmware')
Signed-off-by: Honggang Liho...@redhat.com
ain't over... avoid the blank line between the Fixes tag to your SOB
---
On 4/14/2015 11:41 PM, Jason Gunthorpe wrote:
On Tue, Apr 14, 2015 at 07:30:03PM +0300, Erez Shitrit wrote:
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c
b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 657b89b..11ea6e2 100644
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@
On Wed, Apr 15, 2015 at 08:27:23AM +0300, Or Gerlitz wrote:
On 4/15/2015 6:19 AM, Honggang Li wrote:
Fix(bf0bf77 mlx5: Support communicating arbitrary host page size to firmware)
This isn't the way to write the fix note, do it that way
Fixes: bf0bf77f6519 ('mlx5: Support communicating
If CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for x86 systems and physical
memory is more than 4GB, dma_map_page may return a valid memory
address which greater than 0x. As a result, the mlx5 device page
allocator RB tree will be initialized with valid addresses greater than
0xfff.
However,
On Tue, Apr 14, 2015 at 07:30:03PM +0300, Erez Shitrit wrote:
@@ -846,7 +846,10 @@ static int ipoib_get_iflink(const struct net_device
*dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
- return priv-parent-ifindex;
+ if (priv priv-parent)
+
Loading OpenIB kernel modules:
BUG: unable to handle kernel NULL pointer dereference at
0120
IP: [a06b9060] ipoib_get_iflink+0x10/0x20 [ib_ipoib]
PGD 475540067 PUD 473541067 PMD 0
Oops: [#1] SMP
Modules linked in: ib_ipoib(+) rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm
ib_cm
On 04/14/2015 08:21 PM, Jason Gunthorpe wrote:
On Tue, Apr 14, 2015 at 06:02:47PM +, Hefty, Sean wrote:
Yes, that is the only reasonable thing that could happen.
init failure should only be possible under exceptional cases (OOM).
The only system response is to call ib_umad_add_one
If CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for x86 systems and physical
memory is more than 4GB, dma_map_page may return a valid memory
address which greater than 0x. As a result, the mlx5 device page
allocator RB tree will be initialized with valid addresses greater than
0xfff.
However,
On Wed, Apr 15, 2015 at 10:11:51AM +0300, Or Gerlitz wrote:
On 4/15/2015 9:21 AM, Honggang Li wrote:
Fixes: bf0bf77f6519 ('mlx5: Support communicating arbitrary host page size
to firmware')
Signed-off-by: Honggang Liho...@redhat.com
ain't over... avoid the blank line between the Fixes tag
Fix a wrong setting for a single DMA entry. This
is menifested only in DIX mode where the protection
information buffer reg info overrides the data buffer
reg info.
Fixes: 6e19ed0dba03 (IB/iser: Pass struct iser_mem_reg to [...])
Signed-off-by: Sagi Grimberg sa...@mellanox.com
---
Roland,
This
46 matches
Mail list logo