Hi all,
I am benchmarking a sockets based application and I want a sanity check on
IPoIB performance expectations when using connected mode (65520 MTU). I am
using the tuning tips in Documentation/infiniband/ipoib.txt. The machines have
Mellanox QDR cards (see below for the verbose ibv_devinfo
Support the ability to run congestion support in an environment
where some IB hardware does not support congestion control.
In addition, by flagging ports that do not support congestion
control, we can reduce the amount of MAD traffic and log messages.
Signed-off-by: Albert Chu
---
include/open
I was able to make ib_write_bw to suffer of the same NUMA issue,
and this is a good news.
The buffer allocation is aligned to 4K, basically the memory allocation
is done in this way:
page_size = sysconf(_SC_PAGESIZE); //this is 4K on my system
buf = memalign(page_size, buffer_size);
with same mod
From: Shlomo Pongratz
Lockdep pointed on a circular locking dependency betwwen the ipoib device
priv spinlock (priv->lock) and the neighbour table rwlock (ntbl->rwlock).
In normal path e.g. neigbour Garbage Collection task, the neigh table
rwlock is taken first and then if the neighbour need to
Hi Roland,
This short series include two more fixes from Shlomo to the newly
introduced IPoIB neighbour table, and a fix for Yishai that completes
the support for memory registration of up to 8TB.
Or.
Shlomo Pongratz (2):
IB/ipoib: Fix memory leak in the neigh table deletion flow
IB/ipoib: F
From: Yishai Hadas
This patch adds on the fixes done in commits 89dd86db78 "Allow large
mlx4_buddy bitmaps" and 3de819e6b "Fix integer overflow issues around
MTT table" such that finally memory registration of up to 8TB
(log_num_mtt=31) works fine.
It handle int overflows in few mlx4_table_yyy r
From: Shlomo Pongratz
If the neighbours hash table is empty when unloading the module,
then ipoib_flush_neighs, the cleanup routine wasn't called, and
the memory used for the hash table itself leaked.
To fix this, ipoib_flush_neighs is allways called, and another
completion object was added to s
On Wed, Aug 29, 2012 at 12:51 AM, Hefty, Sean wrote:
>> I'm not sure if I have to say sorry for the noise or not but it seems
>> that the issue was just an NUMA issue!
>
> that's good news
>
>> May be rsocket has to do a memory affinity on the node with the IB
>> board attached on
>> before to all
