On Wed, 2015-12-16 at 13:58 -0600, Faisal Latif wrote:
> From: Anjali Singhai Jain
>
> This patch adds a Client interface for i40iw driver
> support. Also expands the Virtchannel to support messages
> from i40evf driver on behalf of i40iwvf driver.
[]
> diff --git
On Wed, 16 Dec 2015, Christoph Lameter wrote:
> DRAFT: This is missing the check if this device supports
> extended counters.
Found some time and here is the patch with the detection of the extended
attribute through sending a mad request. Untested. Got the info on how
to do the proper mad
On Tue, Dec 08, 2015 at 07:31:56AM -0600, Steve Wise wrote:
>
>
> > -Original Message-
> > From: Chuck Lever [mailto:chuck.le...@oracle.com]
> > Sent: Monday, December 07, 2015 9:45 AM
> > To: Steve Wise
> > Cc: linux-rdma@vger.kernel.org; Veeresh U. Kokatnur; Linux NFS Mailing List
> >
> --- a/include/uapi/rdma/rdma_netlink.h
> +++ b/include/uapi/rdma/rdma_netlink.h
> @@ -5,6 +5,7 @@
>
> enum {
> RDMA_NL_RDMA_CM = 1,
> + RDMA_NL_I40IW,
> RDMA_NL_NES,
> RDMA_NL_C4IW,
> RDMA_NL_LS, /* RDMA Local Services */
This changes the values for the
i40iw_main.c contains routines for i40e <=> i40iw interface and setup.
i40iw.h is header file for main device data structures.
i40iw_status.h is for return status codes.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by:
i40iw_puda.[ch] are files to handle iwarp connection packets as
well as exception packets over multiple privilege mode uda queues.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
From: Anjali Singhai Jain
This patch adds a Client interface for i40iw driver
support. Also expands the Virtchannel to support messages
from i40evf driver on behalf of i40iwvf driver.
This client API is used by the i40iw and i40iwvf driver
to access the core driver
i40iw_ctrl.c provides for hardware wqe supporti and cqp.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_ctrl.c | 4774
i40iw_verbs.[ch] are to handle iwarp interface.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_ucontext.h | 110 ++
IAINTAINERS< Kconfig, Makefile and rdma_netlink.h to include
i40iw
Signed-off-by: Faisal Latif
---
MAINTAINERS | 10 ++
drivers/infiniband/Kconfig | 1 +
drivers/infiniband/hw/Makefile | 1 +
include/uapi/rdma/rdma_netlink.h | 1 +
i40iw_vf.[ch] and i40iw_virtchnl[ch] are used for virtual
channel support for iWARP VF module.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
Kconfig and Kbuild needed to build iwarp module.
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/Kbuild | 43 +
drivers/infiniband/hw/i40iw/Kconfig | 7 ++
2 files changed, 50 insertions(+)
create mode 100644
i40iw_pble.[ch] to manage pble resource for iwarp clients.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_pble.c | 618
i40iw_user.h and i40iw_uk.c are used by both user library as well as
kernel requests.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_uk.c |
i40iw_hmc.[ch] are to manage hmc for the device.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_hmc.c | 823
X722 Hardware registers defines for iWARP component.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_register.h | 1027
i40iw_hw.c, i40iw_utils.c and i40iw_osdep.h are files to handle
interrupts and processing.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_hw.c
i40iw_cm.c i40iw_cm.h are used for connection management.
Acked-by: Anjali Singhai Jain
Acked-by: Shannon Nelson
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_cm.c | 4447
header files for hardware accesses
Signed-off-by: Faisal Latif
---
drivers/infiniband/hw/i40iw/i40iw_d.h| 1713 ++
drivers/infiniband/hw/i40iw/i40iw_p.h| 106 ++
drivers/infiniband/hw/i40iw/i40iw_type.h | 1308 +++
Hi Faisal,
[auto build test WARNING on net/master]
[also build test WARNING on v4.4-rc5 next-20151216]
[cannot apply to net-next/master]
url:
https://github.com/0day-ci/linux/commits/Faisal-Latif/add-Intel-R-X722-iWARP-driver/20151217-040340
config: arm-allyesconfig (attached as .config
This series contains the addition of the i40iw.ko driver.
This driver provides iWARP RDMA functionality for the Intel(R) X722
Ethernet controller for PCI Physical Functions. It also has support
for Virtual Function driver (i40iwvf.ko) that will be part of seperate
patch series.
It cooperates
On Wed, 2015-12-16 at 13:58 -0600, Faisal Latif wrote:
> i40iw_main.c contains routines for i40e <=> i40iw interface and setup.
> i40iw.h is header file for main device data structures.
> i40iw_status.h is for return status codes.
[]
> diff --git a/drivers/infiniband/hw/i40iw/i40iw.h
>
On 12/16/2015 01:56 AM, Moni Shoua wrote:
>> The part that bothers me about this is that this statement makes sense
>> when just thinking about the spec, as you say. However, once you
>> consider namespaces, security implications make this statement spec
>> compliant, but still unacceptable. The
On Wed, Dec 16, 2015 at 04:11:04PM +0100, Christoph Hellwig wrote:
> We now alwasy have a per-PD local_dma_lkey available. Make use of that
> fact in svc_rdma and stop registering our own MR.
>
> Signed-off-by: Christoph Hellwig
Reviewed-by: Jason Gunthorpe
On Wed, Dec 16, 2015 at 08:56:01AM +0200, Moni Shoua wrote:
> I can't object to that but I really would like to get an example of a
> security risk.
How can anyone give you an example when nobody knows exactly how mlx
hardware works in this area?
>From an kapi prespective, the security design
On Wed, Dec 16, 2015 at 11:26:39AM +0100, Michael Wang wrote:
>
> On 12/15/2015 06:30 PM, Jason Gunthorpe wrote:
> > On Tue, Dec 15, 2015 at 05:38:34PM +0100, Michael Wang wrote:
> >> The hop_limit is only suggest that the package allowed to be
> >> routed, not have to, correct?
> >
> > If the
On Wed, Dec 16, 2015 at 09:57:02AM +, Liran Liss wrote:
> Currently, namespaces are not supported for RoCE.
IMHO, we should not be accepting rocev2 without at least basic
namespace support too, since it is fairly trivial to do based on the
work that is already done for verbs. An obvious
There is a window between the time the RPC reply handler wakes the
waiting RPC task and when xprt_release() invokes ops->buf_free.
During this time, memory regions containing the data payload may
still be accessed by a broken or malicious server, but the RPC
application has already been allowed
FRWR's ro_unmap is asynchronous. The new ro_unmap_sync posts
LOCAL_INV Work Requests and waits for them to complete before
returning.
Note also, DMA unmapping is now done _after_ invalidation.
Signed-off-by: Chuck Lever
Tested-by: Devesh Sharma
The root of the problem was that sends (especially unsignalled
FASTREG and LOCAL_INV Work Requests) were not properly flow-
controlled, which allowed a send queue overrun.
Now that the RPC/RDMA reply handler waits for invalidation to
complete, the send queue is properly flow-controlled. Thus this
For FRWR FASTREG and LOCAL_INV, move the ib_*_wr structure off
the stack. This allows frwr_op_map and frwr_op_unmap to chain
WRs together without limit to register or invalidate a set of MRs
with a single ib_post_send().
(This will be for chaining LOCAL_INV requests).
Signed-off-by: Chuck Lever
Preserve any rpcrdma_req that is attached to rpc_rqst's allocated
for the backchannel. Otherwise, after all the pre-allocated
backchannel req's are consumed, incoming backward calls start
writing on freed memory.
Somehow this hunk got lost.
Fixes: f531a5dbc451 ('xprtrdma: Pre-allocate backward
Clean up.
Fixes: 63cae47005af ('xprtrdma: Handle incoming backward direction')
Signed-off-by: Chuck Lever
Tested-by: Devesh Sharma
---
net/sunrpc/xprtrdma/backchannel.c | 16 +---
1 file changed, 9 insertions(+), 7 deletions(-)
In the current xprtrdma implementation, some memreg strategies
implement ro_unmap synchronously (the MR is knocked down before the
method returns) and some asynchonously (the MR will be knocked down
and returned to the pool in the background).
To guarantee the MR is truly invalid before the RPC
physical's ro_unmap is synchronous already. The new ro_unmap_sync
method just has to DMA unmap all MRs associated with the RPC
request.
Signed-off-by: Chuck Lever
Tested-by: Devesh Sharma
---
net/sunrpc/xprtrdma/physical_ops.c | 13
FMR's ro_unmap method is already synchronous because ib_unmap_fmr()
is a synchronous verb. However, some improvements can be made here.
1. Gather all the MRs for the RPC request onto a list, and invoke
ib_unmap_fmr() once with that list. This reduces the number of
doorbells when there is
On Wed, 2015-12-16 at 13:58 -0600, Faisal Latif wrote:
> This series contains the addition of the i40iw.ko driver.
This series should probably be respun against -next
instead of linus' tree.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to
On Wed, Dec 16, 2015 at 03:39:16PM -0500, Doug Ledford wrote:
> These patches add the concept of duplicate GIDs that are differentiated
> by their RoCE version (also called network type).
and by vlan, and smac, and ... Basically everything network unique about
a namespace has to be encapsulted
Clean up.
rb_lock critical sections added in rpcrdma_ep_post_extra_recv()
should have first been converted to use normal spin_lock now that
the reply handler is a work queue.
The backchannel set up code should use the appropriate helper
instead of open-coding a rb_recv_bufs list add.
Problem
For 4.5, I'd like to address the send queue accounting and
invalidation/unmap ordering issues Jason brought up a couple of
months ago.
Also available in the "nfs-rdma-for-4.5" topic branch of this git repo:
git://git.linux-nfs.org/projects/cel/cel-2.6.git
Or for browsing:
From: Mitko Haralanov
This commit "flips the switch" on the TID caching feature
implemented in this patch series.
As well as enabling the new feature by tying the new function
with the PSM API, it also cleans up the old unneeded code,
data structure members, and
From: Mitko Haralanov
The previous patch in the series added the free/invalidate
function bodies. Now, it's time for the programming side.
This large function takes the user's buffer, breaks it up
into manageable chunks, allocates enough RcvArray groups
and programs
From: Mitko Haralanov
TID caching will rely on the MMU notifier to be told
when memory is being invalidated. When the callback
is called, the driver will find all RcvArray entries
that span the invalidated buffer and "schedule" them
to be freed by the PSM library.
From: Mitko Haralanov
The upcoming TID caching feature requires different data
structures and, by extension, different initialization for each
of the MPI processes.
The two new functions (currently unused) perform the required
initialization and freeing of required
From: Mitko Haralanov
Convert bit definitions to use BIT() macro as per checkpatch.pl
requirements.
Reviewed-by: Ira Weiny
Signed-off-by: Mitko Haralanov
---
include/uapi/rdma/hfi/hfi1_user.h | 56
From: Mitko Haralanov
Definitions and functions use to manage sets of TID/RcvArray groups.
These will be used by the TID cacheline functionality coming with
later patches.
TID groups (or RcvArray groups) are groups of TID/RcvArray entries
organized in sets of 8 and
From: Mitko Haralanov
The exp_lock lock does not need to be a spinlock as
all its uses are in process context and allowing the
process to sleep when the mutex is contended might
be benefitial.
Reviewed-by: Ira Weiny
Signed-off-by: Mitko Haralanov
From: Mitko Haralanov
Functions added by this patch are building blocks for the upcoming
TID caching functionality. The functions added are currently unsed
(and marked as such.)
The functions' purposes are to find physically contigous pages in
the user's virtual
From: Mitko Haralanov
The HFI1_CAP_TID_UNMAP comment was incorrectly implying the
opposite of what capability actually did. Correct this error.
Reviewed-by: Ira Weiny
Signed-off-by: Mitko Haralanov
---
From: Mitko Haralanov
There is no need to use a separate variable for a
return value and a label when returning right away
would do just as well.
Reviewed-by: Ira Weiny
Signed-off-by: Mitko Haralanov
---
From: Mitko Haralanov
In preparation for adding the TID caching support, there is a set
of headers, structures, and variables which will be needed. This
commit adds them to the hfi.h header file.
Reviewed-by: Ira Weiny
Signed-off-by: Mitko
From: Mitko Haralanov
Up to now, the functions which cleared the programmed
TID entries and gave PSM the list of invalidated TID entries
were just stubs. With this commit, the bodies of these
functions are added.
This commit is a bit asymmetric as it only contains the
Hi Faisal,
[auto build test WARNING on net/master]
[also build test WARNING on v4.4-rc5 next-20151216]
[cannot apply to net-next/master]
url:
https://github.com/0day-ci/linux/commits/Faisal-Latif/add-Intel-R-X722-iWARP-driver/20151217-040340
config: sparc-allyesconfig (attached as .config
From: Ira Weiny
Expected receives work by user-space libraries (PSM) calling into the driver
with information about the user's receive buffer and have the driver DMA-map
that buffer and program the HFI to receive data directly into it.
This is an expensive operation as it
From: Mitko Haralanov
Add mmu notify helper functions and TID caching function
stubs in preparation for the TID caching implementation.
TID caching makes use of the MMU notifier to allow the driver
to respond to the user freeing memory which is allocated to
the HFI.
From: Mitko Haralanov
TID caching will use a new event to signal userland that cache
invalidation has occurred and needs a matching command code that
will be used to read the invalidated TIDs.
Add the event bit and the new command to the exported header file.
The
Hi Matt,
Yes, you are right.
Since the patch is already merged in, I am going to make a separated
patch for that.
thanks,
wengang
在 2015年12月12日 04:28, Matthew Finlay 写道:
Hi Wengang,
I was going through your patch set here, and it seems that you missed changing
kfree to kvfree in
Commit 0ef2f05c7e02ff99c0b5b583d7dee2cd12b053f2 uses vmalloc for WR buffers
when needed and uses kvfree to free the buffers. It missed changing kfree
to kvfree in mlx4_ib_destroy_srq().
Reported-by: Matthew Finaly
Signed-off-by: Wengang Wang
---
From: Leon Romanovsky
Convert kmalloc to be kmalloc_array to fix warnings below:
WARNING: Prefer kmalloc_array over kmalloc with multiply
+ qp->sq.wrid = kmalloc(qp->sq.wqe_cnt * sizeof(u64),
WARNING: Prefer kmalloc_array over kmalloc with multiply
+
From: Leon Romanovsky
Failure in kmalloc memory allocations will throw a warning about it.
Such warnings are not needed anymore, since in commit 0ef2f05c7e02
("IB/mlx4: Use vmalloc for WR buffers when needed"), fallback mechanism
from kmalloc() to __vmalloc() was added.
+static void
+__frwr_dma_unmap(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
+int rc)
+{
+ struct ib_device *device = r_xprt->rx_ia.ri_device;
+ struct rpcrdma_mw *mw = seg->rl_mw;
+ int nsegs = seg->mr_nsegs;
+
+ seg->rl_mw = NULL;
+
+
Hi Doug,
Lets stop beating, both horses and people.
I do understand that
1. you don't link the removal of the attr
2. you do like the removal of all the query calls
I am proposing to take the path of a patch that
does exactly #2 while avoiding #1.
I really don't have a strong preference
Hi Chuck,
Sorry for the last minute comment.
On 12/14/2015 04:19 PM, Chuck Lever wrote:
> I'm about to add code in the RPC/RDMA reply handler between the
> xprt_lookup_rqst() and xprt_complete_rqst() call site that needs
> to execute outside of spinlock critical sections.
>
> Add a hook to
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index 4197191..e60d817 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -206,6 +206,8 @@ struct rpcrdma_frmr {
enum rpcrdma_frmr_state fr_state;
struct
On 12/16/2015 3:40 PM, Sagi Grimberg wrote:
I really don't have a strong preference on either of the approaches. I
just want to see this included one way or the other.
sure, agree, I will send my patches tomorrow
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the
On Wed, Dec 16, 2015 at 03:21:02PM +0200, Sagi Grimberg wrote:
This question is not directly related to this patch, but given that
this is a copy-paste from the qib driver I'll go ahead and take it
anyway. How does qib (and rvt now) do memory key invalidation? I didn't
see any reference to
> From: linux-rdma-ow...@vger.kernel.org [mailto:linux-rdma-
> ow...@vger.kernel.org] On Behalf Of Doug Ledford
> In particular, Liran piped up with this comment:
>
> "Also, I don't want to do any route resolution on the Rx path. A UD QP
> completion just reports the details of the packet it
Hi Anna-
> On Dec 16, 2015, at 8:48 AM, Anna Schumaker wrote:
>
> Hi Chuck,
>
> Sorry for the last minute comment.
>
> On 12/14/2015 04:19 PM, Chuck Lever wrote:
>> I'm about to add code in the RPC/RDMA reply handler between the
>> xprt_lookup_rqst() and
On 15/12/2015 20:30, Matan Barak wrote:
> Extending core and vendor verb commands require us to check that the
> unknown part of the user's given command is all zeros.
> Adding ib_is_udata_cleared in order to do so.
>
> Signed-off-by: Matan Barak
Reviewed-by: Haggai Eran
Reporting the hca_core_clock (in kHZ) and the timestamp_mask in
query_device extended verb. timestamp_mask is used by users in order
to know what is the valid range of the raw timestamps, while
hca_core_clock reports the clock frequency that is used for
timestamps.
Hi Matan,
Shouldn't this
enum mlx5_ib_mmap_cmd {
MLX5_IB_MMAP_REGULAR_PAGE = 0,
- MLX5_IB_MMAP_GET_CONTIGUOUS_PAGES = 1, /* always last */
+ MLX5_IB_MMAP_GET_CONTIGUOUS_PAGES = 1,
+ /* 5 is chosen in order to be compatible with old versions of libmlx5 */
+
We now alwasy have a per-PD local_dma_lkey available. Make use of that
fact in svc_rdma and stop registering our own MR.
Signed-off-by: Christoph Hellwig
---
include/linux/sunrpc/svc_rdma.h| 2 --
net/sunrpc/xprtrdma/svc_rdma_backchannel.c | 2 +-
This makes use of the now always available local_dma_lkey, and goes on top
of Chuck's "[PATCH v4 00/11] NFS/RDMA server patches for v4.5" series.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at
This patch exists to provide parity for what is in qib. Should we not
have it? If not, why do we have:
commit 38071a461f0a ("IB/qib: Support the new memory registration API")
That was done by me because I saw this in qib and assumed that it was
supported. Now that I found out that it isn't,
> On Dec 16, 2015, at 9:00 AM, Sagi Grimberg wrote:
>
>
>> diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h
>> b/net/sunrpc/xprtrdma/xprt_rdma.h
>> index 4197191..e60d817 100644
>> --- a/net/sunrpc/xprtrdma/xprt_rdma.h
>> +++ b/net/sunrpc/xprtrdma/xprt_rdma.h
>> @@
On Wed, Dec 16, 2015 at 10:06:33AM -0500, Chuck Lever wrote:
> > Would it make sense to unionize these as they are guaranteed not to
> > execute together? Some people don't like this sort of savings.
>
> I dislike unions because they make the code that uses
> them less readable. I can define
> On Dec 16, 2015, at 10:11 AM, Christoph Hellwig wrote:
>
> On Wed, Dec 16, 2015 at 10:06:33AM -0500, Chuck Lever wrote:
>>> Would it make sense to unionize these as they are guaranteed not to
>>> execute together? Some people don't like this sort of savings.
>>
>> I
On Wed, Dec 16, 2015 at 4:43 PM, Sagi Grimberg wrote:
>
>> Reporting the hca_core_clock (in kHZ) and the timestamp_mask in
>> query_device extended verb. timestamp_mask is used by users in order
>> to know what is the valid range of the raw timestamps, while
>>
> On Dec 16, 2015, at 8:57 AM, Sagi Grimberg wrote:
>
>
>> +static void
>> +__frwr_dma_unmap(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
>> + int rc)
>> +{
>> +struct ib_device *device = r_xprt->rx_ia.ri_device;
>> +struct rpcrdma_mw
On Tue, 15 Dec 2015, Doug Ledford wrote:
> On 12/15/2015 04:42 PM, Hal Rosenstock wrote:
> > On 12/15/2015 4:20 PM, Jason Gunthorpe wrote:
> >>> The unicast/multicast extended counters are not always supported -
> depends on setting of PerfMgt ClassPortInfo
>
On Wed, Dec 16, 2015 at 10:13:31AM -0500, Chuck Lever wrote:
> > Shouldn't be an issue with transparent unions these days:
> >
> > union {
> > struct ib_reg_wrfr_regwr;
> > struct ib_send_wr fr_invwr;
> > };
>
> Right, but isn't that
> On Dec 16, 2015, at 10:11 AM, Christoph Hellwig wrote:
>
> We now alwasy have a per-PD local_dma_lkey available. Make use of that
> fact in svc_rdma and stop registering our own MR.
>
> Signed-off-by: Christoph Hellwig
Reviewed-by: Chuck Lever
Looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi Chuck,
iozone passed on ocrdma device. Link bounce fails to recover iozone
traffic, however failure is not related to this patch series. I am in
processes of finding out the patch which broke it.
Tested-By: Devesh Sharma
On Wed, Dec 16, 2015 at 1:07 AM, Anna
iozone passed on ocrdma device. Link bounce fails to recover iozone
traffic, however failure is not related to this patch series. I am in
processes of finding out the patch which broke it.
Tested-By: Devesh Sharma
On Tue, Dec 15, 2015 at 3:00 AM, Chuck Lever
This question is not directly related to this patch, but given that
this is a copy-paste from the qib driver I'll go ahead and take it
anyway. How does qib (and rvt now) do memory key invalidation? I didn't
see any reference to IB_WR_LOCAL_INV anywhere in the qib driver...
What am I missing?
On Tue, 15 Dec 2015, Jason Gunthorpe wrote:
> > The unicast/multicast extended counters are not always supported -
> > depends on setting of PerfMgt ClassPortInfo
> > CapabilityMask.IsExtendedWidthSupportedNoIETF (bit 10).
>
> Yes.. certainly this proposed patch needs to account for that and
>
On 12/15/2015 04:46 PM, Doug Ledford wrote:
> On 12/15/2015 04:42 PM, Hal Rosenstock wrote:
>> On 12/15/2015 4:20 PM, Jason Gunthorpe wrote:
The unicast/multicast extended counters are not always supported -
> depends on setting of PerfMgt ClassPortInfo
>
On 12/15/2015 06:30 PM, Jason Gunthorpe wrote:
> On Tue, Dec 15, 2015 at 05:38:34PM +0100, Michael Wang wrote:
>> The hop_limit is only suggest that the package allowed to be
>> routed, not have to, correct?
>
> If the hop limit is >= 2 (?) then the GRH is mandatory. The
> SM will return this
> From: linux-rdma-ow...@vger.kernel.org [mailto:linux-rdma-
> > Since you and Jason did not reach a consensus, I have to dig in and
> > see if these patches make it possible to break namespace confinement,
> > either accidentally or with intentionally tricky behavior. That's
> > going to take
90 matches
Mail list logo