Remove unnecessary checks for the IB connection state and for QP
overflow, as conn state changes are reported by iSER to libiscsi and
handled there. QP overflow is theoretically possible only when
unsolicited data-outs are used; anyway it's being checked and handled
by HW drivers.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Two minor flows in iSER's data path still use allocations; move them
to be atomic as a preperation step towards moving to use libiscsi
passthrough mode.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Simplify and shrink the logic/code used for the send descriptors.
Changes include removing struct iser_dto (an unnecessary abstraction),
using struct iser_regd_buf only for handling SCSI commands, using
dma_sync instead of dma_map/unmap, etc.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Use a different CQ for send completions, where send completions are
polled by the interrupt-driven receive completion handler.  Therefore,
interrupts aren't used for the send CQ.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Now that both the posting and reaping of receive buffers is done in
the completion path, the counter of outstanding buffers not be atomic.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Currently, the recv buffer posting logic is based on the transactional
nature of iSER which allows for posting a buffer before sending a PDU.
Change this to post only when the number of outstanding recv buffers
is below a water mark and in a batched manner, thus simplifying and
optimizing the data path.  Use a pre-allocated ring of recv buffers
instead of allocating from kmem cache.  A special treatment is given
to the login response buffer whose size must be 8K unlike the size of
buffers used for any other purpose which is 128 bytes.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

We will make a major change in the recv buffer posting logic, after
which the problem commit bba7ebba "avoid recv buffer exhaustion caused
by unexpected PDUs" comes to solve doesn't exist any more, so revert it.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

IPoIB can miss a change in destination GID under some conditions.  The
problem is caused when ipoib_neigh->dgid contains a stale address.
The fix is to set ipoib_neigh->dgid to zero in ipoib_neigh_alloc().

This can happen when a system using bonding on its IPoIB interfaces
has switched its active interface from interface A to B and back to A.
The system that fails over will not correctly processes the 2nd
address change, as described below.

When an address has changed neighbor->ha is updated with the new
address.  Each neighbor has an associated ipoib_neigh.
ipoib_neigh->dgid also holds a copy of the remote node's hardware
address.  When an address changes neighbor->ha is updated by the
network layer (arp code) with the new address.  IPoIB detects this
change in ipoib_start_xmit() by comparing neighbor->ha with
ipoib_neigh->dgid.  The bug is that ipoib_neigh->dgid may already
contain the new address (A) thus the change from B to A is missed by
ipoib.  Here is the sequence of events:

    ipoib_neigh->dgid = A  and  neighbor->ha = A

The address is switched to B (the first switch)

    neighbor->ha = B

The change is seen in ipoib_start_xmit() -- neighbor->ha !=
ipoib_neigh->dgid so ipoib_neigh is released, and a new one is
allocated.

The allocator may return the same chunk of memory that was just
released, therefore ipoib_neigh->dgid still contains A at this point.

ipoib_neigh->dgid should be updated in neigh_add_path(), but if the
following conditions are true dgid is not updated:

        1) __path_find() returns a path
        2) path->ah is NULL

The remote system now switches from address B to A, neighbor->ha is
updated to A.

Now we have again : ipoib_neigh->dgid = A  and  neighbor->ha = A

Since the addresses are the same ipoib won't process the change in
address.  Fix this by zeroing out the dgid field when allocating a new
struct ipoib_neigh.
Signed-off-by: NDavid Wilder <dwilder@us.ibm.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

When iser enabled lu reset support it did not set the
bit to allow userspace to get/set the timeout. This
sets the tgt and lu reset timeout bits.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

fix some typos and punctuation in comments
Signed-off-by: NThadeu Lima de Souza Cascardo <cascardo@holoscopio.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

After dma-mapping an SG list provided by the SCSI midlayer, iser has
to make sure the mapped SG is "aligned for RDMA" in the sense that its
possible to produce one mapping in the HCA IOMMU which represents the
whole SG. Next, the mapped SG is formatted for registration with the HCA.

This patch re-writes the logic that does the above, to make it clearer
and simpler. It also fixes a bug in the being aligned for RDMA checks,
where a "start" check wasn't done but rather only "end" check.
Signed-off-by: NAlexander Nezhinsky <alexandern@voltaire.com>
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

This patch contains changes that allow iscsi_session_setup
to allocate private space for LLD's
Signed-off-by: NJayamohan Kallickal <jayamohank@serverengines.com>
Acked-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Multicast joins can succeed even if the IB port is down.  This happens
when the SM runs on the same port with the requesting port.  However,
IPoIB calls netif_carrier_on() when the join of the broadcast group
succeeds, without caring about the state of the IB port.  The result
is an IPoIB interface in RUNNING state but without an active IB port
to support it.

If a bonding interface uses this IPoIB interface as a slave it might
not detect that this slave is almost useless and failover
functionality will be damaged.  The fix checks the state of the IB
port in the carrier_task before calling netif_carrier_on().

Adresses: https://bugs.openfabrics.org/show_bug.cgi?id=1726Signed-off-by: NMoni Shoua <monis@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Check that the format of multicast link addresses is correct before
taking them from dev->mc_list to priv->multicast_list.  This way we
never try to send a bogus address to the SA, which prevents badness
from erronous 'ip maddr addr add', broken bonding drivers, etc.
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

IPoIB currently must use irqsave locking for priv->lock, since it is
taken from interrupt context in one path.  However, ipoib_send() does
skb_orphan(), and the network stack locking is not IRQ-safe.
Therefore we need to make sure we don't hold priv->lock when calling
ipoib_send() to avoid lockdep warnings (the code was almost certainly
safe in practice, since the only code path that takes priv->lock from
interrupt context would never call into the network stack).

Addresses: http://bugzilla.kernel.org/show_bug.cgi?id=13757Reported-by: NBart Van Assche <bart.vanassche@gmail.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Signed-off-by: NRoland Dreier <rolandd@cisco.com>

The generic packet receive code takes care of setting
netdev->last_rx when necessary, for the sake of the
bonding ARP monitor.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Acked-by: NNeil Horman <nhorman@txudriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Define three accessors to get/set dst attached to a skb

struct dst_entry *skb_dst(const struct sk_buff *skb)

void skb_dst_set(struct sk_buff *skb, struct dst_entry *dst)

void skb_dst_drop(struct sk_buff *skb)
This one should replace occurrences of :
dst_release(skb->dst)
skb->dst = NULL;

Delete skb->dst field
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Last two drivers that need skb->dst in their start_xmit() function

Tell dev_hard_start_xmit() to no release it by unsetting  IFF_XMIT_DST_RELEASE
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If a task did not complete normally due to a TMF, libiscsi will
now complete the task with the state ISCSI_TASK_ABRT_TMF. Drivers
like bnx2i that need to free resources if a command did not complete normally
can then check the task state. If a driver does not need to send
a special command if we have dropped the session then they can check
for ISCSI_TASK_ABRT_SESS_RECOV.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

When we create the tcp/ip connection by calling ep_connect, we currently
just go by the routing table info.

I think there are two problems with this.

1. Some drivers do not have access to a routing table. Some drivers like
qla4xxx do not even know about other ports.

2. If you have two initiator ports on the same subnet, the user may have
set things up so that session1 was supposed to be run through port1. and
session2 was supposed to be run through port2. It looks like we could
end with both sessions going through one of the ports.

Fixes for cxgb3i from Karen Xie.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Network device sysfs files that grab the rtnl_lock unconditionally
will deadlock if accessed when the network device is being
unregistered.  So use trylock and syscall_restart to avoid this
deadlock.
Signed-off-by: NEric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Set target can queue limit to the number of preallocated
session tasks we have.

This along with the cxgb3i can_queue patch will fix a throughput
problem where it could only queue one LU worth of data at a time.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

If NAPI is enabled while IPoIB's CQ is being drained, it creates a
race on priv->ibwc between ipoib_poll() and ipoib_drain_cq(), leading
to memory corruption.

The solution is to enable/disable NAPI in ipoib_ib_dev_{open/stop}()
instead of in ipoib_{open/stop}(), and sync NAPI on the INITIALIZED
flag instead on the ADMIN_UP flag. This way NAPI will be disabled when
ipoib_drain_cq() is called.

This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=1587>.
Signed-off-by: NYossi Etigin <yosefe@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

We have to release the RTNL before calling free_netdev() so that the
device state has a chance to become NETREG_UNREGISTERED.  Otherwise
when removing a child interface, we hit the BUG() that tests the
device state in free_netdev().
Reported-by: NYossi Etigin <yosefe@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Signed-off-by: NStephen Hemminger <shemminger@vyatta.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We do not need to have llds set the host no for the session's
parent, because we know the session's parent is going to be
the host. This removes it from the session creation callback
and converts the drivers.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The qdepth setting was useful when we needed libiscsi to verify
the setting. Now we just need to make sure if older tools
passed in zero then we need to set some default.

So this patch just has us use the sht->cmd_per_lun or if
for LLD does a host per session then we can set it on per
host basis.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

We were using the shost work queue which ended up being
a little akward since all iscsi hosts need a thread for
scanning, but only drivers hooked into libiscsi need
a workqueue for transmitting. So this patch moves the
xmit workqueue to the lib.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

There is no need to cap the queue depth in the modules. We set
this in userspace and can do that there. For performance testing
with ram based targets, this is helpful since we can have very
high queue depths.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

iser has its own logging inrfastrucutre. Convert it to use
it instead of libiscsi.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Remove hard setting of the IB MTU used by iSER's RC queue-pair to 1K,
as this was done due to inter-op issues with an old iser target which
is not used any more.
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

If path_rec_start() returns error, call path_free() only if the path
was newly-created.  If we free an existing path whose valid flag was zero,
(but do not detach it from the list) we cause corruption of the
path list (of which it is a member), and get a kernel crash.

The simplest solution is to not free an existing path -- just leave it
in the list as-is (i.e., with its valid flag cleared).

Thanks to Yossi Etigin of Voltaire for identifying the problem flow
which caused the kernel crash.
Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: NMoni Shua <monis@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Following the removal of the unused struct net_device * parameter from
the NAPI functions named *netif_rx_* in commit 908a7a16, they are
exactly equivalent to the corresponding *napi_* functions and are
therefore redundant.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Acked-by: NNeil Horman <nhorman@tuxdriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When IPoIB tries to join a multicast group, and the SA module's SM
address handle is NULL (because of an SM change, etc), the join
returns with -EAGAIN status.  In that case, don't print an error
message unless multicast debugging is enabled.
Signed-off-by: NYossi Etigin <yosefe@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Fix a deadlock between child interface creation/deletion and ipoib
start/stop.  The former takes vlan_mutex, and then might take RTNL via
register_netdev()/unregister_netdev().  The latter is executed with
RTNL held, and tries to take vlan_mutex, which can lead to an AB-BA
deadlock.

Fix this by having the child interface creation/deletion code take the
RTNL first so vlan_mutex always nests inside RTNL.  We can use
register_netdevice() for child interfaces because we form the
interface name from the parent interface and hence don't need the '%'
expansion of register_netdev().
Reported-by: NYossi Etigin <yosefe@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

After commit fe25c561 ("IPoIB: Don't enable NAPI when it's already
enabled"), if an interface is brought up but the corresponding P_Key
never appears, then ipoib_stop() will hang in napi_disable(), because
ipoib_open() returns before it does napi_enable().

Fix this by changing ipoib_open() to call napi_enable() even if the
P_Key isn't present.
Reported-by: NYossi Etigin <yosefe@Voltaire.COM>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Fix ib_iser build to depend on INFINIBAND_ADDR_TRANS; if INET=y but
IPV6=n, then the RDMA CM is not built but INFINIBAND_ISER can be
enabled, leading to:

    ERROR: "rdma_destroy_id" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
    ERROR: "rdma_connect" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
    ERROR: "rdma_destroy_qp" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
    ERROR: "rdma_create_id" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
    ERROR: "rdma_create_qp" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
    ERROR: "rdma_resolve_route" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
    ERROR: "rdma_disconnect" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
    ERROR: "rdma_resolve_addr" [drivers/infiniband/ulp/iser/ib_iser.ko] undefined!
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>

Because the ipoib_workqueue is not flushed when ipoib interface is
brought down, ipoib_mcast_join() may trigger a join to the broadcast
group after priv->broadcast was set to NULL (during cleanup).  This
will cause the system to be a member of the broadcast group when
interface is down.  As a side effect, this breaks the optimization of
setting the Q_key only when joining the broadcast group.
Signed-off-by: NYossi Etigin <yosefe@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Fix bonding failover in the case both peers failover and the
gratuitous ARP is lost.  In that case, the sender side will create an
ipoib_neigh and issue a path request with the old GID first.  When
skb->dst->neighbour->ha changes due to ARP refresh, this ipoib_neigh
will not be added to the path->list of the path of the new GID,
because the ipoib_neigh already exists.  It will not have an AH
either, because of sender-side failover.  Therefore, it will not get
an AH when the path is resolved.

The solution here is to compare GIDs in ipoib_start_xmit() even if
neigh->ah is invalid.  Comparing with an uninitialized value of
neigh->dgid should be fine, since a spurious match is harmless (and
astronomically unlikely too).
Signed-off-by: NMoni Shoua <monis@voltaire.com>
Signed-off-by: NYossi Etigin <yosefe@voltaire.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>