Hi Devesh.
I wish that there was a header that maps the return errors to errno,
to make things much more canonical. But currently we don't have it :(
AFAIK, the guidelines is that kernel code returns negative values
and userspace code should return positive values.
Thanks
Dotan
On 08/12/2014
On 29/11/2013 22:50, Anuj Kalia wrote:
Dotan,
Thanks for your reply again.
I just wanted to confirm about the order in which messages are sent
and received. I'm not worried about the memory policy.
If you are talking about InfiniBand then this is the case ...
Dotan
--
To unsubscribe from
On 29/11/2013 16:53, Anuj Kalia wrote:
Dotan,
In the example that I asked about, the 5 packets do not form one
message - they are each individual messages. I assume this in the
following.
I now understand that packet reordering cannot happen in IB - thanks!
Only packet loss can happen. From
Hi Anuj.
On 28/11/2013 14:24, Anuj Kalia wrote:
Hi.
I had 2 questions related to RDMA operations (reads and writes) over
unreliable connections:
1. Are WQEs issued over an unreliable connection supposed to generate
CQEs? The Infiniband spec says that such WQEs may not generate an
'ACK'. I'm
at 1:19 AM, Dotan Barak dota...@gmail.com wrote:
Hi Anuj.
On 28/11/2013 14:24, Anuj Kalia wrote:
Hi.
I had 2 questions related to RDMA operations (reads and writes) over
unreliable connections:
1. Are WQEs issued over an unreliable connection supposed to generate
CQEs? The Infiniband spec says
Hi Xavier.
I'm trying to understand exactly what you did...
I'm guessing here the scenario, please correct me if I'm wrong:
You've tried to work with RoCE over ConnectX-3 using RC QP.
You were able to send messages up to 1024 (including this value) and it failed.
You've to execute the same test
Hi Xavier.
On Tue, Oct 16, 2012 at 8:13 PM, Xavier Guérin guer...@gmail.com wrote:
Hello list,
I am having a puzzling error with IBV_WR_SEND_WITH_IMM. Polling fails
with IBV_WC_RETRY_EXC_ERR when I use buffer bigger that 1KB.
Send/Receive buffers are of a larger size and properly registered.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Yishai Hadas yish...@mellanox.com
---
src/rsocket.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/src/rsocket.c b/src/rsocket.c
index 278c8ec..4665ce7 100644
--- a/src/rsocket.c
+++ b/src/rsocket.c
This will make rping to exit with return value other than zero in case of an
error.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Yishai Hadas yish...@mellanox.com
---
examples/rping.c | 21 ++---
1 files changed, 18 insertions(+), 3 deletions(-)
diff --git
Here is a small batch of fixes for librdmacm.
Dotan Barak (5):
rsocket: add missing va_end() after calling va_end()
rsocket: added missing break
rstream: initialize return value in server_connect()
rstream: added missing return is accept() failed
rping: added checks to the return values
Not doing so, may lead to resource leak.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Yishai Hadas yish...@mellanox.com
---
src/rsocket.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/src/rsocket.c b/src/rsocket.c
index 4c4821b..278c8ec 100644
If use_async == 0 and rs_accept() passes (i.e. non negative value), then
the return value from the function was uninitialized.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Yishai Hadas yish...@mellanox.com
---
examples/rstream.c |2 +-
1 files changed, 1 insertions(+), 1
Replace sprintf with snprintf to protects from buffer overflow.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Jack Morgenstein ja...@dev.mellanox.co.il
---
linux/acme_linux.c |4 ++--
src/acm.c |6 +++---
src/parse.c|2 +-
3 files changed, 6
After allocation of dynamic memory blocks, check that the allocation succeeded,
if it failed handle it gracefully.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Erez Shitrit ere...@mellanox.co.il
---
src/parse.c | 22 --
1 files changed, 16 insertions
After allocation of dynamic memory blocks, check that the allocation succeeded,
if it failed handle it gracefully.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Erez Shitrit ere...@mellanox.co.il
---
src/parse.c | 22 --
1 files changed, 16 insertions
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
---
src/cmd.c|4 +++-
src/compat-1_0.c |5 -
src/init.c |3 ++-
3 files changed, 9 insertions(+), 3 deletions(-)
diff --git a/src/cmd.c b/src/cmd.c
index 39af833..1a04228 100644
--- a/src/cmd.c
+++ b/src/cmd.c
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
---
examples/rc_pingpong.c | 43 ---
examples/srq_pingpong.c | 51 ++
examples/uc_pingpong.c | 43 ---
examples
Added the man page verbs.7 which is an introduction to libibverbs man pages.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Signed-off-by: Or Gerlitz ogerl...@voltaire.com
---
Makefile.am|2 +-
libibverbs.spec.in |1 +
man/verbs.7| 201
Clean the idr as part of the cleanup flow since it is a resource too.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Jack Morgenstein ja...@dev.mellanox.co.il
Reviewed-by: Sean Hefty sean.he...@intel.com
---
V2 - don't check if idr_pre_get() failed, , we'll just try again
The function idr_pre_get may fail, so there is a need to check the status of
it. Since this function allocate resources, we need to clean them during the
resource cleaning in case of error.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Jack Morgenstein ja...@dev.mellanox.co.il
In this case, using strdup will provide a cleaner code
(and maybe a little bit faster too).
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
---
src/addrinfo.c |4 +---
1 files changed, 1 insertions(+), 3 deletions(-)
diff --git a/src/addrinfo.c b/src/addrinfo.c
index 695430b..34c7fb4
Fixed a resource leak in case of an error level in the udaddy example.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
---
examples/udaddy.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/examples/udaddy.c b/examples/udaddy.c
index 637306a..1534df5 100644
Pages that were mapped using ib_dma_map_page() should be unmapped using the
right function (i.e. ib_dma_unmap_page()).
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Jack Morgenstein ja...@dev.mellanox.co.il
---
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
b/drivers
Roland: did you have a change to review this patch?
Thanks
Dotan
-- Forwarded Message --
Subject: [PATCH] mthca: Fix the attribute sq_sig_type value in query qp
Date: Monday 05 September 2011 16:44
From: Dotan Barak dot...@sw.voltaire.com
To: rol...@kernel.org
Cc: linux-rdma
Fix the attribute sq_sig_type value in query qp.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Jack Morgenstein ja...@dev.mellanox.co.il
---
diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c
b/drivers/infiniband/hw/mthca/mthca_qp.c
index 9601049..57b4325 100644
When the user queries for device capabilities, fill the attribute
masked_atomic_cap
according to the support level of the regular atomic capabilities and not using
a
hard coded value.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Eli Cohen e...@mellanox.co.il
---
drivers
When calculating the required size of an RC QP send queue, leave enough space
for masked atomic operation (which requires more space than regular atomic
operation).
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Eli Cohen e...@mellanox.co.il
---
drivers/infiniband/hw/mlx4/qp.c
)
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Eli Cohen e...@mellanox.co.il
---
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index 3a91d9d..d91acbf 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -283,7
Prevent resource leak by destroying the event channel before returning from
function in an error flow.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
---
diff --git a/src/cma.c b/src/cma.c
index bad1ba3..1367779 100755
--- a/src/cma.c
+++ b/src/cma.c
@@ -2122,8 +2122,11 @@ int
If the macro CMA_CREATE_MSG_CMD_RESP is being called and there is a failure,
the macro should release the allocated resources before returning from the
called
function
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
---
diff --git a/src/cma.c b/src/cma.c
index bad1ba3..0512187 100755
The function handle_outgoing_dr_smp() is only called if the
MAD to be sent is a directed route SMP. Thus, the check for
IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE is redundant.
Signed-off-by: Ralph Campbell ralph.campb...@qlogic.com
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Acked-by: Sean
Fix the memory release function to be consistent with the memory allocation one
to prevent problems where the implementation of pci_free_consistent and
dma_free_coherent are different.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Eli Cohen e...@mellanox.co.il
---
diff --git
Many features were added to this driver, and the driver version should change
too.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
---
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index dd7d745..ea529b4 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
Use BlueFlame for RDMA Write with immediate without any data (no s/g) too.
This improves latency.
Signed-off-by: Dotan Barak dot...@dev.mellanox.co.il
Reviewed-by: Jack Morgenstein ja...@dev.mellanox.co.il
---
src/qp.c |5 -
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git
Hi.
I noticed minor issue:
@@ -564,6 +597,16 @@ struct ibv_srq {
pthread_mutex_t mutex;
pthread_cond_t cond;
uint32_tevents_completed;
+
+ /* Following fields only available if device supports extensions */
+ enum ibv_srq_type
I call rdma_create_id to create an ib id, then do resolve remote addr,
resolve route work, then
setup qp and call rdma_connect to setup connection, before ack or
error replies, the thread will
wait on a wait queue. The listening ib id of remote node will catch
the connect request,
setup qp,
Hi.
On 10/05/2010 08:42, lihaidong wrote:
Hi,
I have a question about atomic operations.
According to IB specification o10-48, all atomic operation request made to
the same HCA, referencing the same physical memory are serialized with
respect to each other. I know this should be
Hi.
On 11/06/2010 10:51, Ding Dinghua wrote:
Hi all:
I'm using RDMA to do fs-metadata mirror between nodes. I
encountered a strange problem when the program was running:
Complete queue handler reported that the RDMA-Write operation failed,
the status of corresponding struct ib_wc
38 matches
Mail list logo
