commit 518b1646 ("IPoIB/cm: Fix SRQ WR leak") introduced a severe
performance regression on Mellanox cards, because keeping a QP in the
error state for extended periods of time moves hardware to the slow
path (until the QP is destroyed). For example, MPI latency goes from
~3 usecs to ~7 usecs.
Fix this by posting a send WR on one of the QPs that are being
flushed, instead of using a separate drain QP that is kept in the
error state.
This fixes bug <https://bugs.openfabrics.org/show_bug.cgi?id=636>,
reported and bisected by Scott Weitzenkamp at Cisco and debugged by
Sasha Mikheev at Voltaire.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Since NAPI polling is disabled while ipoib_cm_dev_stop() is running,
ipoib_cm_dev_stop() must poll the CQ itself in order to see the
packets draining.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
time_after() was used backwards, so the timeout occurred immediately.
Signed-off-by: Michael S. Tsirkin <mst@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>
In the presence of some running RX connections, we repeat
queue_delayed_work calls each 4 RX WRs, which is a waste. It's enough
to start stale task when a first passive connection is added, and
rerun it every IPOIB_CM_RX_DELAY as long as there are outstanding
passive connections.
This removes some code from RX data path.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Remove includes of <linux/smp_lock.h> where it is not used/needed.
Suggested by Al Viro.
Builds cleanly on x86_64, i386, alpha, ia64, powerpc, sparc,
sparc64, and arm (all 59 defconfigs).
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
IPoIB: Convert to NAPI
IB: Return "maybe missed event" hint from ib_req_notify_cq()
IB: Add CQ comp_vector support
IB/ipath: Fix a race condition when generating ACKs
IB/ipath: Fix two more spin lock problems
IB/fmr_pool: Add prefix to all printks
IB/srp: Set proc_name
IB/srp: Add orig_dgid sysfs attribute to scsi_host
IPoIB/cm: Don't crash if remote side uses one QP for both directions
RDMA/cxgb3: Support for new abort logic
RDMA/cxgb3: Initialize cpu_idx field in cpl_close_listserv_req message
RDMA/cxgb3: Fail qp creation if the requested max_inline is too large
RDMA/cxgb3: Fix TERM codes
IPoIB/cm: Fix error handling in ipoib_cm_dev_open()
IB/ipath: Don't corrupt pending mmap list when unmapped objects are freed
IB/mthca: Work around kernel QP starvation
IB/ipath: Don't put QP in timeout queue if waiting to send
IB/ipath: Don't call spin_lock_irq() from interrupt context
Convert the IP-over-InfiniBand network device driver over to using
NAPI to handle completions for the main CQ. This covers all receives
as well as datagram mode sends; send completions for connected mode
connections are still handled from interrupt context.
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>
Add an orig_dgid attribute in sysfs for SRP scsi_hosts, so that
userspace can tell what the original dgid value written to the
add_target file was, even if the connection is redirected to a
different port while connecting.
Signed-off-by: Ishai Rabinovitz <ishai@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The IPoIB CM spec allows the use of a single connection in both
active->passive and passive->active directions. The current Linux
code uses one connection for both directions, but if another node only
uses one connection for both directions, we oops when we try to look
up the passive connection. Fix by checking that qp_context is
non-NULL before dereferencing it.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
I noticed that many source files include <linux/pci.h> while they do
not appear to need it. Here is an attempt to clean it all up.
In order to find all possibly affected files, I searched for all
files including <linux/pci.h> but without any other occurence of "pci"
or "PCI". I removed the include statement from all of these, then I
compiled an allmodconfig kernel on both i386 and x86_64 and fixed the
false positives manually.
My tests covered 66% of the affected files, so there could be false
positives remaining. Untested files are:
arch/alpha/kernel/err_common.c
arch/alpha/kernel/err_ev6.c
arch/alpha/kernel/err_ev7.c
arch/ia64/sn/kernel/huberror.c
arch/ia64/sn/kernel/xpnet.c
arch/m68knommu/kernel/dma.c
arch/mips/lib/iomap.c
arch/powerpc/platforms/pseries/ras.c
arch/ppc/8260_io/enet.c
arch/ppc/8260_io/fcc_enet.c
arch/ppc/8xx_io/enet.c
arch/ppc/syslib/ppc4xx_sgdma.c
arch/sh64/mach-cayman/iomap.c
arch/xtensa/kernel/xtensa_ksyms.c
arch/xtensa/platform-iss/setup.c
drivers/i2c/busses/i2c-at91.c
drivers/i2c/busses/i2c-mpc.c
drivers/media/video/saa711x.c
drivers/misc/hdpuftrs/hdpu_cpustate.c
drivers/misc/hdpuftrs/hdpu_nexus.c
drivers/net/au1000_eth.c
drivers/net/fec_8xx/fec_main.c
drivers/net/fec_8xx/fec_mii.c
drivers/net/fs_enet/fs_enet-main.c
drivers/net/fs_enet/mac-fcc.c
drivers/net/fs_enet/mac-fec.c
drivers/net/fs_enet/mac-scc.c
drivers/net/fs_enet/mii-bitbang.c
drivers/net/fs_enet/mii-fec.c
drivers/net/ibm_emac/ibm_emac_core.c
drivers/net/lasi_82596.c
drivers/parisc/hppb.c
drivers/sbus/sbus.c
drivers/video/g364fb.c
drivers/video/platinumfb.c
drivers/video/stifb.c
drivers/video/valkyriefb.c
include/asm-arm/arch-ixp4xx/dma.h
sound/oss/au1550_ac97.c
I would welcome test reports for these files. I am fine with removing
the untested files from the patch if the general opinion is that these
changes aren't safe. The tested part would still be nice to have.
Note that this patch depends on another header fixup patch I submitted
to LKML yesterday:
[PATCH] scatterlist.h needs types.h
http://lkml.org/lkml/2007/3/01/141
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
If skb allocation fails when we start the device, we call
ipoib_cm_dev_stop() even though ipoib_cm_dev_open() did not run to
completion, so we pass an invalid pointer to ib_destroy_cm_id and get
an oops.
Fix by clearing cm.id on error, and testing it in cm_dev_stop().
This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=561>
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband: (49 commits)
IB: Set class_dev->dev in core for nice device symlink
IB/ehca: Implement modify_port
IB/umad: Clarify documentation of transaction ID
IPoIB/cm: spin_lock_irqsave() -> spin_lock_irq() replacements
IB/mad: Change SMI to use enums rather than magic return codes
IB/umad: Implement GRH handling for sent/received MADs
IB/ipoib: Use ib_init_ah_from_path to initialize ah_attr
IB/sa: Set src_path_bits correctly in ib_init_ah_from_path()
IB/ucm: Simplify ib_ucm_event()
RDMA/ucma: Simplify ucma_get_event()
IB/mthca: Simplify CQ cleaning in mthca_free_qp()
IB/mthca: Fix mthca_write_mtt() on HCAs with hidden memory
IB/mthca: Update HCA firmware revisions
IB/ipath: Fix WC format drift between user and kernel space
IB/ipath: Check that a UD work request's address handle is valid
IB/ipath: Remove duplicate stuff from ipath_verbs.h
IB/ipath: Check reserved memory keys
IB/ipath: Fix unit selection when all CPU affinity bits set
IB/ipath: Don't allow QPs 0 and 1 to be opened multiple times
IB/ipath: Disable IB link earlier in shutdown sequence
...
For the common, open coded 'skb->mac.raw = skb->data' operation, so that we can
later turn skb->mac.raw into a offset, reducing the size of struct sk_buff in
64bit land while possibly keeping it as a pointer on 32bit.
This one touches just the most simple case, next will handle the slightly more
"complex" cases.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are quite a few places in ipoib_cm.c where we know IRQs are
enabled because we do something that sleeps in the same function, so
we can convert several occurrences of spin_lock_irqsave() to a plain
spin_lock_irq(). This cleans up the source a little and makes the
code smaller too:
add/remove: 0/0 grow/shrink: 1/5 up/down: 3/-51 (-48)
function old new delta
ipoib_cm_tx_reap 403 406 +3
ipoib_cm_stale_task 146 145 -1
ipoib_cm_dev_stop 173 172 -1
ipoib_cm_tx_handler 964 956 -8
ipoib_cm_rx_handler 956 937 -19
ipoib_cm_skb_reap 212 190 -22
Signed-off-by: Roland Dreier <rolandd@cisco.com>
To support destinations that are not on the local IB subnet, IPoIB
should include the GRH information when constructing an address
handle. Using the existing ib_init_ah_from_path() call will do this
for us.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
There's no point in printing the opcode field in the completion
handling debugging output, since the type of completion is already
printed at the beginning of the line. In fact the opcode field is not
even defined for completions with a status other than success.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Receive buffers need to be mapped with DMA_FROM_DEVICE. Incorrectly
mapping with DMA_TO_DEVICE causes a hard lock on ppc64 machines with
an IOMMU.
This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=431>
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
When a connection is terminated asynchronously from the iSCSI layer's
perspective, iSER needs to notify the iSCSI layer that the connection
has failed. This is done using a workqueue (switched to from the iSER
tasklet context). Meanwhile, the connection object (that holds the
work struct) is released. If the workqueue function wasn't called
yet, it will be called later with a NULL pointer, which will crash the
kernel.
The context switch (tasklet to workqueue) is not required, and
everything can be done from the iSER tasklet. This eliminates the NULL
work struct bug (and simplifies the code).
Signed-off-by: Erez Zilber <erezz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
IB/iser: Handle aborting a command after it is sent
IB/mthca: Fix thinko in init_mr_table()
RDMA/cxgb3: Fix resource leak in cxio_hal_init_ctrl_qp()
The SCSI midlayer may abort a command that was already sent. If the
initiator is still trying to send the command (or data-out PDUs for
that command), the QP may time out after the midlayer times
out. Therefore, when aborting the command, iSER may still have
references for the command's buffers. When sending these PDUs, the
sends will complete with an error and their resources will be released
then.
Signed-off-by: Erez Zilber <erezz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
->neigh_destructor() is killed (not used), replaced with
->neigh_cleanup(), which is called when neighbor entry goes to dead
state. At this point everything is still valid: neigh->dev,
neigh->parms etc.
The device should guarantee that dead neighbor entries (neigh->dead !=
0) do not get private part initialized, otherwise nobody will cleanup
it.
I think this is enough for ipoib which is the only user of this thing.
Initialization private part of neighbor entries happens in ipib
start_xmit routine, which is not reached when device is down. But it
would be better to add explicit test for neigh->dead in any case.
Signed-off-by: David S. Miller <davem@davemloft.net>
The packet length checks in ipoib are broken: we add 4 bytes (IPoIB
encapsulation header) when sending a packet, not 20 bytes (hardware
address length) to each packet. Therefore, if connected mode is
enabled so that the interface MTU is larger than the multicast MTU,
IPoIB may end up trying to send too-long multicast packets. For
example, multicast is broken if a message of size 2048 bytes is sent
on an interface with UD MTU 2048, because 2048 is bigger than the real
limit of 2044 but the code tests against the wrong limit of 2060.
This patch fixes <https://bugs.openfabrics.org/show_bug.cgi?id=418>,
submitted by Scott Weitzenkamp <sweitzen@cisco.com>.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The connected mode code added the possibility that an neigh struct
gets freed in the list_for_each_entry() loop in path_rec_completion(),
which causes a use-after-free. Fix this by changing to the _safe
variant of the list walking macro.
This was spotted by the Coverity checker (CID 1567).
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
There's a race between ipoib_mcast_leave() and ipoib_mcast_join_finish()
where we can try to detach from a multicast group before we've
attached to it. Fix this by reordering the code in ipoib_mcast_leave
to free the multicast group first, which waits for the multicast
callback thread (which calls ipoib_mcast_join_finish()) to complete
before detaching from the group.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The sense of the time_after_eq() test in ipoib_cm_stale_task() is
reversed so that only non-stale connections are reaped. Fix this by
changing to time_before_eq().
Noticed by Pradeep Satyanarayana <pradeep@us.ibm.com>.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
This patch renames DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH to avoid
confusion with the drivers default values (DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH
is the iscsi RFC specific default).
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Do netif_carrier_on() right after the IPv4 broadcast multicast group
is joined, rather than waiting for all of the initial set of multicast
group joins to finish. This allows at least IPv4 traffic to limp
along on broken fabrics where not all multicast groups can be joined.
Signed-off-by: Shirley Ma <xma@us.ibm.com>
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>
If path_rec_completion() is passed a non-NULL path record pointer
along with an unsuccessful status value, the tracing code incorrectly
prints the (invalid) DLID from the path record rather than the more
interesting status code. The actual logic of the function correctly
uses the path record only if the status indicates a successful lookup.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Now that low-level drivers handle the conversion from an absolute rate
to a relative rate, there's no need for the IPoIB driver to keep track
of the local port's data rate.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Avoid the overhead of freeing/reallocating and mapping/unmapping for
DMA pages that have not been written to by hardware.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The IB SA tracks multicast join/leave requests on a per port basis and
does not do any reference counting: if two users of the same port join
the same group, and one leaves that group, then the SA will remove the
port from the group even though there is one user who wants to stay a
member left. Therefore, in order to support multiple users of the
same multicast group from the same port, we need to perform reference
counting locally.
To do this, add an multicast submodule to ib_sa to perform reference
counting of multicast join/leave operations. Modify ib_ipoib (the
only in-kernel user of multicast) to use the new interface.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
ipoib_cm_alloc_rx_skb() might be called from IRQ context, so it must
use dev_kfree_skb_any(), not kfree_skb().
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
IB/mthca: Always fill MTTs from CPU
IB/mthca: Merge MR and FMR space on 64-bit systems
IB/mthca: Fix access to MTT and MPT tables on non-cache-coherent CPUs
IB/mthca: Give reserved MTTs a separate cache line
IB/mthca: Fix reserved MTTs calculation on mem-free HCAs
RDMA/cxgb3: Add driver for Chelsio T3 RNIC
IB: Remove redundant "_wq" from workqueue names
RDMA/cma: Increment port number after close to avoid re-use
IB/ehca: Fix memleak on module unloading
IB/mthca: Work around gcc bug on sparc64
IPoIB: Connected mode experimental support
IB/core: Use ARRAY_SIZE macro for mandatory_table
IB/mthca: Use correct structure size in call to memset()
Many struct file_operations in the kernel can be "const". Marking them const
moves these to the .rodata section, which avoids false sharing with potential
dirty data. In addition it'll catch accidental writes at compile time to
these shared resources.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
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>
This lets the network core have the ability to handle suspend/resume
issues, if it wants to.
Thanks to Frederik Deweerdt <frederik.deweerdt@gmail.com> for the arm
driver fixes.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
When there is a call to send_tsk_mgmt SRP posts a send and waits for 5
seconds to get a response.
When the QP is in the error state it is obvious that there will be no
response so it is quite useless to wait. In fact, the timeout causes
SRP to wait a long time to reconnect when a QP error occurs. (Each
abort and each reset_device calls send_tsk_mgmt, which waits for the
timeout). The following patch solves this problem by identifying the
failure and returning an immediate error code.
Signed-off-by: Ishai Rabinovitz <ishai@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Checks if the kmalloc in match_strdup() was successful, and bail out
on looking at the token if it failed.
Signed-off-by: Ishai Rabinovitz <ishai@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
iSER limits the number of outstanding PDUs to send. When this threshold
is reached, it should return an error code (-ENOBUFS) instead of setting
the suspend_tx bit (which should be used only by libiscsi).
Signed-off-by: Erez Zilber <erezz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
struct srp_device.fmr_page_mask was unsigned long, which means that
the top part of addresses above 4G was being chopped off on 32-bit
architectures. Of course nothing good happens when data from SRP
targets is DMAed to the wrong place.
Fix this by changing fmr_page_mask to u64, to match the addresses
actually used by IB devices.
Thanks to Brian Cain <Brian.Cain@ge.com> and David McMillen
<davem@systemfabricworks.com> for help diagnosing the bug and testing
the fix.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Move the initialization of ipoib_neigh's skb_queue into
ipoib_neigh_alloc(), since commit 2745b5b7 ("IPoIB: Fix skb leak when
freeing neighbour") will make iterate over the skb_queue to free any
packets left over when freeing the ipoib_neigh structure.
This fixes a crash when freeing ipoib_neigh structures allocated in
ipoib_mcast_send(), which otherwise don't have their skb_queue
initialized.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
