>
> Enhanced switch port 0 can support CC (indicated in CC
> ClassPortInfo:CapabilityMask.EnhancedPort0CC).
>
> portnum 0 is a valid port number and offsets should be calculated based on
> port 0 rather than port 1.
>
> Also, fix some error messages to indicate switch port congestion setting
>
>
> Enhanced switch port 0 can support CC (indicated in CC
> ClassPortInfo:CapabilityMask.EnhancedPort0CC).
>
> portnum 0 is a valid port number and offsets should be calculated based on
> port 0 rather than port 1.
>
> Also, fix some error messages to indicate switch port congestion setting
>
>
> When idebug is >= 1, mad_rpc and mad_rpc_rmpp output only the mad data
> response in debug output. It does not output the request mad data. This is
> confusing since it's not clear if it's request or response output.
>
> When idebug > 1, full mad send and receive buffers are output, which m
When idebug is >= 1, mad_rpc and mad_rpc_rmpp output only
the mad data response in debug output. It does not output
the request mad data. This is confusing since it's not clear
if it's request or response output.
When idebug > 1, full mad send and receive buffers are
output, which makes the mad
Just didn't hit the issue there so didn't think to remove it there a
well. I'll repost
Al
On Fri, 2015-01-09 at 23:27 +, Weiny, Ira wrote:
> Any reason this should not be removed from mad_rpc_rmpp as well?
>
> Ira
>
>
> > -Original Message-
> > From: linux-rdma-ow...@vger.kernel.o
> for consistency with CC Annex which shows meaning of threshold fields in hex
> not decimal
>
> Signed-off-by: Hal Rosenstock
Thanks applied.
Ira
> ---
> Change since v1:
> Change fields to hex rather than SwitchPortCongestionSetting:Threshold to be
> consistent with CC Annex
>
> diff --git
Any reason this should not be removed from mad_rpc_rmpp as well?
Ira
> -Original Message-
> From: linux-rdma-ow...@vger.kernel.org [mailto:linux-rdma-
> ow...@vger.kernel.org] On Behalf Of Albert Chu
> Sent: Friday, December 19, 2014 11:08 AM
> To: linux-rdma@vger.kernel.org
> Subject: [
On Fri, Jan 09, 2015 at 03:40:51PM -0500, Chuck Lever wrote:
>
> On Jan 9, 2015, at 3:39 PM, J. Bruce Fields wrote:
>
> > On Fri, Jan 09, 2015 at 02:21:54PM -0500, Chuck Lever wrote:
> >> In addition to miscellaneous clean up, the following series of
> >> patches for the Linux NFS server introdu
On Jan 9, 2015, at 3:39 PM, J. Bruce Fields wrote:
> On Fri, Jan 09, 2015 at 02:21:54PM -0500, Chuck Lever wrote:
>> In addition to miscellaneous clean up, the following series of
>> patches for the Linux NFS server introduces support in the
>> server’s RPC/RDMA transport implementation for RDMA
On Fri, Jan 09, 2015 at 02:21:54PM -0500, Chuck Lever wrote:
> In addition to miscellaneous clean up, the following series of
> patches for the Linux NFS server introduces support in the
> server’s RPC/RDMA transport implementation for RDMA_NOMSG type
> messages, and fixes a bug that prevents the s
On Jan 9, 2015, at 3:14 PM, J. Bruce Fields wrote:
> On Fri, Jan 09, 2015 at 02:23:11PM -0500, Chuck Lever wrote:
>> Simpler approach, and this needs to be done before the tail is
>> copied to the end of the read list (pre-requisite for a subsequent
>> patch).
>>
>> Fixes: e560e3b510d2 ("svcrdm
On Fri, Jan 09, 2015 at 02:23:11PM -0500, Chuck Lever wrote:
> Simpler approach, and this needs to be done before the tail is
> copied to the end of the read list (pre-requisite for a subsequent
> patch).
>
> Fixes: e560e3b510d2 ("svcrdma: Add zero padding if the client... ")
Wait, does this actu
Simpler approach, and this needs to be done before the tail is
copied to the end of the read list (pre-requisite for a subsequent
patch).
Fixes: e560e3b510d2 ("svcrdma: Add zero padding if the client... ")
Signed-off-by: Chuck Lever
---
net/sunrpc/xprtrdma/svc_rdma_recvfrom.c | 37 ---
Most NFS RPCs place large payload arguments at the end of the RPC
header (eg, NFSv3 WRITE). For NFSv3 WRITE and SYMLINK, RPC/RDMA
sends the complete RPC header inline, and the payload argument in a
read list.
One important case is not like this, however. NFSv4 WRITE compounds
can have an operation
An RPC/RDMA client may send large data payloads via a read chunk
list. This is a list of scatter/gather elements which are RPC call
arguments too large to fit in a small RDMA SEND.
Each read chunk list entry has a "position" which is the offset in
the RPC header where the chunk list payload should
Currently the Linux server can not decode RDMA_NOMSG type requests.
The client sends the RPC/RDMA header via RDMA SEND. The header has
a read chunk list that conveys the RPC and NFS headers and the NFS
payload (if any).
Large operations like NFSv4 CREATE(NF4LNK) will by conveyed via
RDMA_NOMSG, so
The byte_count argument is not used, and the function is called
only from one place.
Signed-off-by: Chuck Lever
---
include/linux/sunrpc/svc_rdma.h |2 --
net/sunrpc/xprtrdma/svc_rdma_marshal.c | 16
net/sunrpc/xprtrdma/svc_rdma_recvfrom.c | 15 ---
In addition to miscellaneous clean up, the following series of
patches for the Linux NFS server introduces support in the
server’s RPC/RDMA transport implementation for RDMA_NOMSG type
messages, and fixes a bug that prevents the server from handling
RPC/RDMA messages with inline content following a
The RDMA reader function doesn't change once an svcxprt is
instantiated. Instead of checking sc_devcap during every incoming
RPC, set the reader function once when the connection is accepted.
Signed-off-by: Chuck Lever
---
include/linux/sunrpc/svc_rdma.h | 10
net/sunrpc/xprtrdm
Nit: Fix inconsistent white space in dprintk messages.
Signed-off-by: Chuck Lever
---
net/sunrpc/xprtrdma/svc_rdma_recvfrom.c |8
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
index
Nit: remove an unused variable to squelch a compiler warning.
Signed-off-by: Chuck Lever
---
net/sunrpc/xprtrdma/svc_rdma_transport.c |2 --
1 files changed, 0 insertions(+), 2 deletions(-)
diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c
b/net/sunrpc/xprtrdma/svc_rdma_transport.c
in
Current convention is to avoid using BUG_ON() in places where an
oops could cause complete system failure.
Replace BUG_ON() call sites in svcrdma with an assertion error
message and allow execution to continue safely.
Some BUG_ON() calls are removed because they have never fired in
production (th
xdr_start() can return the wrong rmsgp address if an assumption
about how the xdr_buf was constructed changes. When it gets it
wrong, the client receives a reply that has gibberish in the
RPC/RDMA header, preventing it from matching a waiting RPC request.
Instead, make (and document) just one ass
On 1/9/2015 10:46 AM, Weiny, Ira wrote:
> [off-list]
>
> This replaces the first one?
This is ibccconfig. Other was for ibccquery. Similar subjects but
separated by tool rather than one patch for both.
>
>> -Original Message-
>> From: Hal Rosenstock [mailto:h...@dev.mellanox.co.il]
>> S
[off-list]
This replaces the first one?
> -Original Message-
> From: Hal Rosenstock [mailto:h...@dev.mellanox.co.il]
> Sent: Friday, January 09, 2015 7:18 AM
> To: Weiny, Ira
> Cc: linux-rdma (linux-rdma@vger.kernel.org); Chu, Al
> Subject: [PATCH infiniband-diags] ibccconfig.c: Account f
Enhanced switch port 0 can support CC (indicated in CC
ClassPortInfo:CapabilityMask.EnhancedPort0CC).
portnum 0 is a valid port number and offsets should be calculated based on port 0
rather than port 1.
Also, fix some error messages to indicate switch port congestion setting
rather then switch
Enhanced switch port 0 can support CC (indicated in CC
ClassPortInfo:CapabilityMask.EnhancedPort0CC).
portnum 0 is a valid port number and offsets should be calculated based on port 0
rather than port 1.
Also, fix some error messages to indicate switch port congestion setting
rather then switch
RDMA can use the same kind of weak security as TCP by checking the
client can bind to a privileged port, which is better than nothing
if TAUTH isn't implemented.
Signed-off-by: Dominique Martinet
---
Please note that this does NOT checkpatch because option tokens
"Opt_foo" are considered CamelCa
for consistency with CC Annex which shows meaning of threshold fields in hex
not decimal
Signed-off-by: Hal Rosenstock
---
Change since v1:
Change fields to hex rather than SwitchPortCongestionSetting:Threshold to be
consistent with CC Annex
diff --git a/src/fields.c b/src/fields.c
index 33a63
On Wed, Jan 7, 2015 at 5:04 PM, Or Gerlitz wrote:
> From: Erez Shitrit
>
> Following commit 016d9fb25cd9 "IPoIB: fix MCAST_FLAG_BUSY usage"
> both IPv6 traffic and for the most cases all IPv4 multicast traffic
> aren't working.
Doug, can you ack the breakage introduced by your commit and the fix
30 matches
Mail list logo
