This patch enables IPoIB to use 4K UD messages (when the underlying
device and fabrics support a 4K MTU) by using two scatter buffers when
PAGE_SIZE is less than or equal to thhe HCA IB MTU size. The first
buffer is for IPoIB header + GRH header, and the second buffer is the
IPoIB payload, which is 4K-4.
Signed-off-by: Shirley Ma <xma@us.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
For HCAs that support TCP segmentation offload (IB_DEVICE_UD_TSO), set
NETIF_F_TSO and use HW LSO to offload TCP segmentation.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
This patch acts as a preparation for using checksum offload for IB
devices capable of inserting/verifying checksum in IP packets. The
patch does not actaully turn on NETIF_F_SG - we defer that to the
patches adding checksum offload capabilities.
We only add support for send gathers for datagram mode, since existing
HW does not support checksum offload on connected QPs.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Some IB adapters (notably IBM's eHCA) do not implement SRQs (shared
receive queues). The current IPoIB connected mode support only works
on devices that support SRQs.
Fix this by adding support for using the receive queue of each
connected mode receive QP. The disadvantage of this compared to using
an SRQ is that it means a full queue of receives must be posted for
each remote connected mode peer, which means that total memory usage
is potentially much higher than when using SRQs. To manage this, add
a new module parameter "max_nonsrq_conn_qp" that limits the number of
connections allowed per interface.
The rest of the changes are fairly straightforward: we use a table of
struct ipoib_cm_rx to hold all the active connections, and put the
table index of the connection in the high bits of receive WR IDs.
This is needed because we cannot rely on the struct ib_wc.qp field for
non-SRQ receive completions. Most of the rest of the changes just
test whether or not an SRQ is available, and post receives or find
received packets in the right place depending on the answer.
Cleaning up dead connections actually becomes simpler, because we do
not have to do the "last WQE reached" dance that is required to
destroy QPs attached to an SRQ. We just move the QP to the error
state and wait for all pending receives to be flushed.
Signed-off-by: Pradeep Satyanarayana <pradeeps@linux.vnet.ibm.com>
[ Completely rewritten and split up, based on Pradeep's work. Several
bugs fixed and no doubt several bugs introduced. - Roland ]
Signed-off-by: Roland Dreier <rolandd@cisco.com>
ipoib_transport_dev_init() calls ipoib_cm_dev_init(), so it needs to
call ipoib_cm_dev_cleanup() to unwind that on the error path.
Found by Dotan Barak of Mellanox.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
SRQ WR leakage has been observed with IPoIB/CM: e.g. flipping ports on
and off will, with time, leak out all WRs and then all connections
will start getting RNR NAKs. Fix this in the way suggested by spec:
move the QP being destroyed to the error state, wait for "Last WQE
Reached" event and then post WR on a "drain QP" connected to the same
CQ. Once we observe a completion on the drain QP, it's safe to call
ib_destroy_qp.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
SM reconfiguration or failover possibly causes a shuffling of the values
in the P_Key table. Right now, IPoIB only queries for the P_Key index
once when it creates the device QP, and hence there are problems if the
index of a P_Key value changes. Fix this by using the PKEY_CHANGE event
to trigger a recheck of the P_Key index.
Signed-off-by: Yosef Etigin <yosefe@voltaire.com>
Acked-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Add a num_comp_vectors member to struct ib_device and extend
ib_create_cq() to pass in a comp_vector parameter -- this parallels
the userspace libibverbs API. Update all hardware drivers to set
num_comp_vectors to 1 and have all ULPs pass 0 for the comp_vector
value. Pass the value of num_comp_vectors to userspace rather than
hard-coding a value of 1.
We want multiple CQ event vector support (via MSI-X or similar for
adapters that can generate multiple interrupts), but it's not clear
how many vectors we want, or how we want to deal with policy issues
such as how to decide which vector to use or how to set up interrupt
affinity. This patch is useful for experimenting, since no core
changes will be necessary when updating a driver to support multiple
vectors, and we know that we want to make at least these changes
anyway.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
An asynchronous event carries the port number that the event occurred
on, so there's no reason for an IPoIB interface to process an event
associated with a different local HCA port.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The following patch adds experimental support for IPoIB connected
mode, as defined by the draft from the IETF ipoib working group. The
idea is to increase performance by increasing the MTU from the maximum
of 2K (theoretically 4K) supported by IPoIB on top of UD. With this
code, I'm able to get 800MByte/sec or more with netperf without
options on a Mellanox 4x back-to-back DDR system.
Some notes on code:
1. SRQ is used for scalability to large cluster sizes
2. Only RC connections are used (UC does not support SRQ now)
3. Retry count is set to 0 since spec draft warns against retries
4. Each connection is used for data transfers in only 1 direction, so
each connection is either active(TX) or passive (RX). 2 sides that
want to communicate create 2 connections.
5. Each active (TX) connection has a separate CQ for send completions -
this keeps the code simple without CQ resize and other tricks
6. To detect stale passive side connections (where the remote side is
down), we keep an LRU list of passive connections (updated once per
second per connection) and destroy a connection after it has been
unused for several seconds. The LRU rule makes it possible to avoid
scanning connections that have recently been active.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Handle client reregister events by treating them just like LID or
SM changes -- flush all cached paths and rejoin multicast groups.
Signed-off-by: Leonid Arsh <leonida@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Make IPoIB's send and receive queue sizes tunable via module
parameters ("send_queue_size" and "recv_queue_size"). This allows the
queue sizes to be enlarged to fix disastrously bad performance on some
platforms and workloads, without bloating memory usage when large
queues aren't needed.
Signed-off-by: Shirley Ma <xma@us.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
This patch causes the network interface to respond to P_Key change
events correctly. As a result, you'll see a child interface in the
"RUNNING" state (netif_carrier_on()) only when the corresponding P_Key
is configured by the SM. When SM removes a P_Key, the "RUNNING" state
will be disabled for the corresponding network interface. To
implement this, I added IB_EVENT_PKEY_CHANGE event handling. To
prevent flushing the device before the device is open by the "delay
open" mechanism, I added an additional device flag called
IPOIB_FLAG_INITIALIZED.
This also prevents the child network interface from trying to join to
multicast groups until the PKEY is configured. We used to get error
messages like:
ib0.f2f2: couldn't attach QP to multicast group ff12:401b:f2f2:0:0:0:ffff:ffff
in this case. To fix this, I just check IPOIB_FLAG_OPER_UP flag in
ipoib_set_mcast_list().
Signed-off-by: Leonid Arsh <leonida@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
With the current IPoIB driver, the status of network interfaces stays
"RUNNING" even if the link goes down (for example because a cable is
unplugged). Fix this by flushing the IPoIB interface when the link
goes down.
Signed-off-by: Leonid Arsh <leonida@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Move ipoib_ib_dev_flush() to ipoib's workqueue. This keeps it ordered
with respect to other work scheduled by the ipoib driver. This fixes
problems with races, for example:
- ipoib_ib_dev_flush() has started running because of an IB event
- user does ifconfig ib0 down
- ipoib_mcast_stop_thread() gets called twice and waits for the same
completion twice
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
semaphore to mutex conversion by Ingo and Arjan's script.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
[ Sanity-checked on real IB hardware ]
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Minor cleanups: fix a misleading comment, and get rid of attr_mask
variables that are only used to hold constants (just use the constants
directly).
Signed-off-by: Roland Dreier <rolandd@cisco.com>
ipoib_create_qp() no longer creates IPoIB's QP, so it shouldn't
destroy the QP on failure -- that unwinding happens elsewhere, so the
current code can cause a double free. While we're at it, the
function's name should match what it actually does, so rename it to
ipoib_init_qp().
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Move the InfiniBand headers from drivers/infiniband/include to include/rdma.
This allows InfiniBand-using code to live elsewhere, and lets us remove the
ugly EXTRA_CFLAGS include path from the InfiniBand Makefiles.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Make some lawyers happy and add copyright notices for people who
forgot to include them when they actually touched the code.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
Shirley Ma