After the commit 9207f9d45b ("net: preserve IP control block
during GSO segmentation"), the GSO CB and the IPoIB CB conflict.
That destroy the IPoIB address information cached there,
causing a severe performance regression, as better described here:
http://marc.info/?l=linux-kernel&m=146787279825501&w=2
This change moves the data cached by the IPoIB driver from the
skb control lock into the IPoIB hard header, as done before
the commit 936d7de3d7 ("IPoIB: Stop lying about hard_header_len
and use skb->cb to stash LL addresses").
In order to avoid GRO issue, on packet reception, the IPoIB driver
stash into the skb a dummy pseudo header, so that the received
packets have actually a hard header matching the declared length.
To avoid changing the connected mode maximum mtu, the allocated
head buffer size is increased by the pseudo header length.
After this commit, IPoIB performances are back to pre-regression
value.
v2 -> v3: rebased
v1 -> v2: avoid changing the max mtu, increasing the head buf size
Fixes: 9207f9d45b ("net: preserve IP control block during GSO segmentation")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a new CM connection is being requested, ipoib driver copies data
from the path pointer in the CM/tx object, the path object might be
invalid at the point and memory corruption will happened later when now
the CM driver will try using that data.
The next scenario demonstrates it:
neigh_add_path --> ipoib_cm_create_tx -->
queue_work (pointer to path is in the cm/tx struct)
#while the work is still in the queue,
#the port goes down and causes the ipoib_flush_paths:
ipoib_flush_paths --> path_free --> kfree(path)
#at this point the work scheduled starts.
ipoib_cm_tx_start --> copy from the (invalid)path pointer:
(memcpy(&pathrec, &p->path->pathrec, sizeof pathrec);)
-> memory corruption.
To fix that the driver now starts the CM/tx connection only if that
specific path exists in the general paths database.
This check is protected with the relevant locks, and uses the gid from
the neigh member in the CM/tx object which is valid according to the ref
count that was taken by the CM/tx.
Fixes: 839fcaba35 ('IPoIB: Connected mode experimental support')
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
In ipoib_remove_one the driver holds the rtnl_lock and tries to do some
operation like dev_change_flags or unregister_netdev, while sysfs
callback like ipoib_vlan_delete holds sysfs mutex and tries to hold the
rtnl_lock via rtnl_trylock() and restart_syscall() if the lock is not
free, meanwhile ipoib_remove_one tries to get the sysfs lock in order to
free its sysfs directory, and we will get a->b, b->a deadlock.
Trace like the following:
schedule+0x37/0x80
schedule_preempt_disabled+0xe/0x10
__mutex_lock_slowpath+0xb5/0x120
mutex_lock+0x23/0x40
rtnl_lock+0x15/0x20
netdev_run_todo+0x17c/0x320
rtnl_unlock+0xe/0x10
ipoib_vlan_delete+0x11b/0x1b0 [ib_ipoib]
delete_child+0x54/0x80 [ib_ipoib]
dev_attr_store+0x18/0x30
sysfs_kf_write+0x37/0x40
mutex_lock+0x16/0x40
SyS_write+0x55/0xc0
entry_SYSCALL_64_fastpath+0x16/0x75
And
schedule+0x37/0x80
__kernfs_remove+0x1a8/0x260
? wake_atomic_t_function+0x60/0x60
kernfs_remove+0x25/0x40
sysfs_remove_dir+0x50/0x80
kobject_del+0x18/0x50
device_del+0x19f/0x260
netdev_unregister_kobject+0x6a/0x80
rollback_registered_many+0x1fd/0x340
rollback_registered+0x3c/0x70
unregister_netdevice_queue+0x55/0xc0
unregister_netdev+0x20/0x30
ipoib_remove_one+0x114/0x1b0 [ib_ipoib]
ib_unregister_client+0x4a/0x170 [ib_core]
? find_module_all+0x71/0xa0
ipoib_cleanup_module+0x10/0x94 [ib_ipoib]
SyS_delete_module+0x1b5/0x210
entry_SYSCALL_64_fastpath+0x16/0x75
The fix is by checking the flag IPOIB_FLAG_INTF_ON_DESTROY in order to
get out from the sysfs function.
Fixes: 862096a8bb ("IB/ipoib: Add more rtnl_link_ops callbacks")
Fixes: 9baa0b0364 ("IB/ipoib: Add rtnl_link_ops support")
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
IPoIB converts skb-fragments to sge adding 1 extra sge when SG is enabled.
Current codepath assumes that the max number of sge a device support
is at least MAX_SKB_FRAGS+1, there is no interaction with upper layers
to limit number of fragments in an skb if a device suports fewer
sges. The assumptions also lead to requesting a fixed number of sge
when IPoIB creates queue-pairs with SG enabled.
A fallback/slowpath is implemented using skb_linearize to
handle cases where the conversion would result in more sges than supported.
Signed-off-by: Hans Westgaard Ry <hans.westgaard.ry@oracle.com>
Reviewed-by: Håkon Bugge <haakon.bugge@oracle.com>
Reviewed-by: Wei Lin Guay <wei.lin.guay@oracle.com>
Reviewed-by: Yuval Shaia <yuval.shaia@oracle.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Instead, use the cached copy of the attributes present on the device.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
This adds an abstraction that allows ULPs to simply pass a completion
object and completion callback with each submitted WR and let the RDMA
core handle the nitty gritty details of how to handle completion
interrupts and poll the CQ.
In detail there is a new ib_cqe structure which just contains the
completion callback, and which can be used to get at the containing
object using container_of. It is pointed to by the WR and WC as an
alternative to the wr_id field, similar to how many ULPs already use
the field to store a pointer using casts.
A driver using the new completion callbacks allocates it's CQs using
the new ib_create_cq API, which in addition to the number of CQEs and
the completion vectors also takes a mode on how we poll for CQEs.
Three modes are available: direct for drivers that never take CQ
interrupts and just poll for them, softirq to poll from softirq context
using the to be renamed blk-iopoll infrastructure which takes care of
rearming and budgeting, or a workqueue for consumer who want to be
called from user context.
Thanks a lot to Sagi Grimberg who helped reviewing the API, wrote
the current version of the workqueue code because my two previous
attempts sucked too much and converted the iSER initiator to the new
API.
Signed-off-by: Christoph Hellwig <hch@lst.de>
This patch split up struct ib_send_wr so that all non-trivial verbs
use their own structure which embedds struct ib_send_wr. This dramaticly
shrinks the size of a WR for most common operations:
sizeof(struct ib_send_wr) (old): 96
sizeof(struct ib_send_wr): 48
sizeof(struct ib_rdma_wr): 64
sizeof(struct ib_atomic_wr): 96
sizeof(struct ib_ud_wr): 88
sizeof(struct ib_fast_reg_wr): 88
sizeof(struct ib_bind_mw_wr): 96
sizeof(struct ib_sig_handover_wr): 80
And with Sagi's pending MR rework the fast registration WR will also be
down to a reasonable size:
sizeof(struct ib_fastreg_wr): 64
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> [srp, srpt]
Reviewed-by: Chuck Lever <chuck.lever@oracle.com> [sunrpc]
Tested-by: Haggai Eran <haggaie@mellanox.com>
Tested-by: Sagi Grimberg <sagig@mellanox.com>
Tested-by: Steve Wise <swise@opengridcomputing.com>
The pd now has a local_dma_lkey member which completely replaces
ib_get_dma_mr, use it instead.
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Now that there are no ib_cm clients using the compare_data feature for
matching IB CM requests' private data, remove the compare_data parameter of
ib_cm_listen and remove the code implementing the feature.
Signed-off-by: Haggai Eran <haggaie@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
By default, IPoIB-CM driver uses 64k MTU. Larger MTU gives better
performance.
This MTU plus overhead puts the memory allocation for IP based packets at
32 4k pages (order 5), which have to be contiguous.
When the system memory under pressure, it was observed that allocating 128k
contiguous physical memory is difficult and causes serious errors (such as
system becomes unusable).
This enhancement resolve the issue by removing the physically contiguous
memory requirement using Scatter/Gather feature that exists in Linux stack.
With this fix Scatter-Gather will be supported also in connected mode.
This change reverts some of the change made in commit e112373fd6
("IPoIB/cm: Reduce connected mode TX object size").
The ability to use SG in IPoIB CM is possible because the coupling
between NETIF_F_SG and NETIF_F_CSUM was removed in commit
ec5f061564 ("net: Kill link between CSUM and SG features.")
Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>
Acked-by: Christian Marie <christian@ponies.io>
Signed-off-by: Doug Ledford <dledford@redhat.com>
See also patch "IPoIB/cm: Add connected mode support for devices
without SRQs" (commit ID 68e995a295). Detected by smatch.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Pradeep Satyanarayana <pradeeps@linux.vnet.ibm.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
During my recent work on the rtnl lock deadlock in the IPoIB driver, I
saw that even once I fixed the apparent races for a single device, as
soon as that device had any children, new races popped up. It turns
out that this is because no matter how well we protect against races
on a single device, the fact that all devices use the same workqueue,
and flush_workqueue() flushes *everything* from that workqueue means
that we would also have to prevent all races between different devices
(for instance, ipoib_mcast_restart_task on interface ib0 can race with
ipoib_mcast_flush_dev on interface ib0.8002, resulting in a deadlock on
the rtnl_lock).
There are several possible solutions to this problem:
Make carrier_on_task and mcast_restart_task try to take the rtnl for
some set period of time and if they fail, then bail. This runs the
real risk of dropping work on the floor, which can end up being its
own separate kind of deadlock.
Set some global flag in the driver that says some device is in the
middle of going down, letting all tasks know to bail. Again, this can
drop work on the floor.
Or the method this patch attempts to use, which is when we bring an
interface up, create a workqueue specifically for that interface, so
that when we take it back down, we are flushing only those tasks
associated with our interface. In addition, keep the global
workqueue, but now limit it to only flush tasks. In this way, the
flush tasks can always flush the device specific work queues without
having deadlock issues.
Signed-off-by: Doug Ledford <dledford@redhat.com>
This reverts commit 5141861cd5.
The series of IPoIB bug fixes that went into 3.19-rc1 introduce
regressions, and after trying to sort things out, we decided to revert
to 3.18's IPoIB driver and get things right for 3.20.
Signed-off-by: Roland Dreier <roland@purestorage.com>
During my recent work on the rtnl lock deadlock in the IPoIB driver, I
saw that even once I fixed the apparent races for a single device, as
soon as that device had any children, new races popped up. It turns
out that this is because no matter how well we protect against races
on a single device, the fact that all devices use the same workqueue,
and flush_workqueue() flushes *everything* from that workqueue, we can
have one device in the middle of a down and holding the rtnl lock and
another totally unrelated device needing to run mcast_restart_task,
which wants the rtnl lock and will loop trying to take it unless is
sees its own FLAG_ADMIN_UP flag go away. Because the unrelated
interface will never see its own ADMIN_UP flag drop, the interface
going down will deadlock trying to flush the queue. There are several
possible solutions to this problem:
Make carrier_on_task and mcast_restart_task try to take the rtnl for
some set period of time and if they fail, then bail. This runs the
real risk of dropping work on the floor, which can end up being its
own separate kind of deadlock.
Set some global flag in the driver that says some device is in the
middle of going down, letting all tasks know to bail. Again, this can
drop work on the floor. I suppose if our own ADMIN_UP flag doesn't go
away, then maybe after a few tries on the rtnl lock we can queue our
own task back up as a delayed work and return and avoid dropping work
on the floor that way. But I'm not 100% convinced that we won't cause
other problems.
Or the method this patch attempts to use, which is when we bring an
interface up, create a workqueue specifically for that interface, so
that when we take it back down, we are flushing only those tasks
associated with our interface. In addition, keep the global
workqueue, but now limit it to only flush tasks. In this way, the
flush tasks can always flush the device specific work queues without
having deadlock issues.
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
This addresses a problem where NFS client writes over IPoIB connected
mode may deadlock on memory allocation/writeback.
The problem is not directly memory reclamation. There is an indirect
dependency between network filesystems writing back pages and
ipoib_cm_tx_init() due to how a kworker is used. Page reclaim cannot
make forward progress until ipoib_cm_tx_init() succeeds and it is
stuck in page reclaim itself waiting for network transmission.
Ordinarily this situation may be avoided by having the caller use
GFP_NOFS but ipoib_cm_tx_init() does not have that information.
To address this, take a general approach and add a new QP creation
flag that tells the low-level hardware driver to use GFP_NOIO for the
memory allocations related to the new QP.
Use the new flag in the ipoib connected mode path, and if the driver
doesn't support it, re-issue the QP creation without the flag.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Change CM skb memory allocation to use GFP_KERNEL when possible.
During device init there's no need to use GFP_ATOMIC when allocating
memory for the CM skbs -- use GFP_KERNEL instead.
Signed-off-by: Tal Alon <talal@mellanox.com>
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
In several places, this snippet is used when removing neigh entries:
list_del(&neigh->list);
ipoib_neigh_free(neigh);
The list_del() removes neigh from the associated struct ipoib_path, while
ipoib_neigh_free() removes neigh from the device's neigh entry lookup
table. Both of these operations are protected by the priv->lock
spinlock. The table however is also protected via RCU, and so naturally
the lock is not held when doing reads.
This leads to a race condition, in which a thread may successfully look
up a neigh entry that has already been deleted from neigh->list. Since
the previous deletion will have marked the entry with poison, a second
list_del() on the object will cause a panic:
#5 [ffff8802338c3c70] general_protection at ffffffff815108c5
[exception RIP: list_del+16]
RIP: ffffffff81289020 RSP: ffff8802338c3d20 RFLAGS: 00010082
RAX: dead000000200200 RBX: ffff880433e60c88 RCX: 0000000000009e6c
RDX: 0000000000000246 RSI: ffff8806012ca298 RDI: ffff880433e60c88
RBP: ffff8802338c3d30 R8: ffff8806012ca2e8 R9: 00000000ffffffff
R10: 0000000000000001 R11: 0000000000000000 R12: ffff8804346b2020
R13: ffff88032a3e7540 R14: ffff8804346b26e0 R15: 0000000000000246
ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0000
#6 [ffff8802338c3d38] ipoib_cm_tx_handler at ffffffffa066fe0a [ib_ipoib]
#7 [ffff8802338c3d98] cm_process_work at ffffffffa05149a7 [ib_cm]
#8 [ffff8802338c3de8] cm_work_handler at ffffffffa05161aa [ib_cm]
#9 [ffff8802338c3e38] worker_thread at ffffffff81090e10
#10 [ffff8802338c3ee8] kthread at ffffffff81096c66
#11 [ffff8802338c3f48] kernel_thread at ffffffff8100c0ca
We move the list_del() into ipoib_neigh_free(), so that deletion happens
only once, after the entry has been successfully removed from the lookup
table. This same behavior is already used in ipoib_del_neighs_by_gid()
and __ipoib_reap_neigh().
Signed-off-by: Jim Foraker <foraker1@llnl.gov>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Shlomo Pongratz <shlomop@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Use preferable function name which implies using a pseudo-random number
generator.
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit f0dc117abd ("IPoIB: Fix TX queue lockup with mixed UD/CM
traffic") attempts to solve an issue where unprocessed UD send
completions can deadlock the netdev.
The patch doesn't fully resolve the issue because if more than half
the tx_outstanding's were UD and all of the destinations are RC
reachable, arming the CQ doesn't solve the issue.
This patch uses the IB_CQ_REPORT_MISSED_EVENTS on the
ib_req_notify_cq(). If the rc is above 0, the UD send cq completion
callback is called directly to re-arm the send completion timer.
This issue is seen in very large parallel filesystem deployments
and the patch has been shown to correct the issue.
Cc: <stable@vger.kernel.org>
Reviewed-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
After commit b13912bbb4 ("IPoIB: Call skb_dst_drop() once skb is
enqueued for sending"), using connected mode and running multithreaded
iperf for long time, ie
iperf -c <IP> -P 16 -t 3600
results in a crash.
After the above-mentioned patch, the driver is calling skb_orphan() and
skb_dst_drop() after calling post_send() in ipoib_cm.c::ipoib_cm_send()
(also in ipoib_ib.c::ipoib_send())
The problem with this is, as is written in a comment in both routines,
"it's entirely possible that the completion handler will run before we
execute anything after the post_send()." This leads to running the
skb cleanup routines simultaneously in two different contexts.
The solution is to always perform the skb_orphan() and skb_dst_drop()
before queueing the send work request. If an error occurs, then it
will be no different than the regular case where dev_free_skb_any() in
the completion path, which is assumed to be after these two routines.
Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Currently, IPoIB delays collecting send completions for TX packets in
order to batch work more efficiently. It does skb_orphan() right after
queuing the packets so that destructors run early, to avoid problems
like holding socket send buffers for too long (since we might not
collect a send completion until a long time after the packet is
actually sent).
However, IPoIB clears IFF_XMIT_DST_RELEASE because it actually looks
at skb_dst() to update the PMTU when it gets a too-long packet. This
means that the packets sitting in the TX ring with uncollected send
completions are holding a reference on the dst. We've seen this lead
to pathological behavior with respect to route and neighbour GC. The
easy fix for this is to call skb_dst_drop() when we call skb_orphan().
Also, give packets sent via connected mode (CM) the same skb_orphan()
/ skb_dst_drop() treatment that packets sent via datagram mode get.
Signed-off-by: Roland Dreier <roland@purestorage.com>
With the new netlink support in commit 862096a8bb ("IB/ipoib: Add more
rtnl_link_ops callbacks") we need ipoib_set_mode() to be available even
if connected mode isn't built. Move the function from ipoib_cm.c to
ipoib_main.c (and make a few CM-related macros available unconditonally).
This fixes the build error
drivers/built-in.o: In function 'ipoib_changelink':
ipoib_netlink.c:(.text+0x6a5fc9): undefined reference to 'ipoib_set_mode'
ipoib_netlink.c:(.text+0x6a5fe3): undefined reference to 'ipoib_set_mode'
when CONFIG_INFINIBAND_IPOIB_CM isn't set.
Reported-by: Randy Dunlap <rdunlap@xenotime.net>
Reported-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Add the rtnl_link_ops changelink and fill_info callbacks, through
which the admin can now set/get the driver mode, etc policies.
Maintain the proprietary sysfs entries only for legacy childs.
For child devices, set dev->iflink to point to the parent
device ifindex, such that user space tools can now correctly
show the uplink relation as done for vlan, macvlan, etc
devices. Pointed out by Patrick McHardy <kaber@trash.net>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit b63b70d877 ("IPoIB: Use a private hash table for path lookup
in xmit path") introduced a bug where in ipoib_cm_destroy_tx() a CM
object is moved between lists without any supported locking. Under a
stress test, this eventually leads to list corruption and a crash.
Previously when this routine was called, callers were taking the
device priv lock. Currently this function is called from the RCU
callback associated with neighbour deletion. Fix the race by taking
the same lock we used to before.
Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Dave Miller <davem@davemloft.net> provided a detailed description of
why the way IPoIB is using neighbours for its own ipoib_neigh struct
is buggy:
Any time an ipoib_neigh is changed, a sequence like the following is made:
spin_lock_irqsave(&priv->lock, flags);
/*
* It's safe to call ipoib_put_ah() inside
* priv->lock here, because we know that
* path->ah will always hold one more reference,
* so ipoib_put_ah() will never do more than
* decrement the ref count.
*/
if (neigh->ah)
ipoib_put_ah(neigh->ah);
list_del(&neigh->list);
ipoib_neigh_free(dev, neigh);
spin_unlock_irqrestore(&priv->lock, flags);
ipoib_path_lookup(skb, n, dev);
This doesn't work, because you're leaving a stale pointer to the freed up
ipoib_neigh in the special neigh->ha pointer cookie. Yes, it even fails
with all the locking done to protect _changes_ to *ipoib_neigh(n), and
with the code in ipoib_neigh_free() that NULLs out the pointer.
The core issue is that read side calls to *to_ipoib_neigh(n) are not
being synchronized at all, they are performed without any locking. So
whether we hold the lock or not when making changes to *ipoib_neigh(n)
you still can have threads see references to freed up ipoib_neigh
objects.
cpu 1 cpu 2
n = *ipoib_neigh()
*ipoib_neigh() = NULL
kfree(n)
n->foo == OOPS
[..]
Perhaps the ipoib code can have a private path database it manages
entirely itself, which holds all the necessary information and is
looked up by some generic key which is available easily at transmit
time and does not involve generic neighbour entries.
See <http://marc.info/?l=linux-rdma&m=132812793105624&w=2> and
<http://marc.info/?l=linux-rdma&w=2&r=1&s=allows+references+to+freed+memory&q=b>
for the full discussion.
This patch aims to solve the race conditions found in the IPoIB driver.
The patch removes the connection between the core networking neighbour
structure and the ipoib_neigh structure. In addition to avoiding the
race described above, it allows us to handle SKBs carrying IP packets
that don't have any associated neighbour.
We add an ipoib_neigh hash table with N buckets where the key is the
destination hardware address. The ipoib_neigh is fetched from the
hash table and instead of the stashed location in the neighbour
structure. The hash table uses both RCU and reference counting to
guarantee that no ipoib_neigh instance is ever deleted while in use.
Fetching the ipoib_neigh structure instance from the hash also makes
the special code in ipoib_start_xmit that handles remote and local
bonding failover redundant.
Aged ipoib_neigh instances are deleted by a garbage collection task
that runs every M seconds and deletes every ipoib_neigh instance that
was idle for at least 2*M seconds. The deletion is safe since the
ipoib_neigh instances are protected using RCU and reference count
mechanisms.
The number of buckets (N) and frequency of running the GC thread (M),
are taken from the exported arb_tbl.
Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
- Updates to the qib low-level driver
- First chunk of changes for SR-IOV support for mlx4 IB
- RDMA CM support for IPv6-only binding
- Other misc cleanups and fixes
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=f09j
-----END PGP SIGNATURE-----
Merge tag 'rdma-for-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband
Pull InfiniBand/RDMA changes from Roland Dreier:
- Updates to the qib low-level driver
- First chunk of changes for SR-IOV support for mlx4 IB
- RDMA CM support for IPv6-only binding
- Other misc cleanups and fixes
Fix up some add-add conflicts in include/linux/mlx4/device.h and
drivers/net/ethernet/mellanox/mlx4/main.c
* tag 'rdma-for-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (30 commits)
IB/qib: checkpatch fixes
IB/qib: Add congestion control agent implementation
IB/qib: Reduce sdma_lock contention
IB/qib: Fix an incorrect log message
IB/qib: Fix QP RCU sparse warnings
mlx4: Put physical GID and P_Key table sizes in mlx4_phys_caps struct and paravirtualize them
mlx4_core: Allow guests to have IB ports
mlx4_core: Implement mechanism for reserved Q_Keys
net/mlx4_core: Free ICM table in case of error
IB/cm: Destroy idr as part of the module init error flow
mlx4_core: Remove double function declarations
IB/mlx4: Fill the masked_atomic_cap attribute in query device
IB/mthca: Fill in sq_sig_type in query QP
IB/mthca: Warning about event for non-existent QPs should show event type
IB/qib: Fix sparse RCU warnings in qib_keys.c
net/mlx4_core: Initialize IB port capabilities for all slaves
mlx4: Use port management change event instead of smp_snoop
IB/qib: RCU locking for MR validation
IB/qib: Avoid returning EBUSY from MR deregister
IB/qib: Fix UC MR refs for immediate operations
...
This will be used so that we can compose a full flow key.
Even though we have a route in this context, we need more. In the
future the routes will be without destination address, source address,
etc. keying. One ipv4 route will cover entire subnets, etc.
In this environment we have to have a way to possess persistent storage
for redirects and PMTU information. This persistent storage will exist
in the FIB tables, and that's why we'll need to be able to rebuild a
full lookup flow key here. Using that flow key will do a fib_lookup()
and create/update the persistent entry.
Signed-off-by: David S. Miller <davem@davemloft.net>
* 'modsplit-Oct31_2011' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux: (230 commits)
Revert "tracing: Include module.h in define_trace.h"
irq: don't put module.h into irq.h for tracking irqgen modules.
bluetooth: macroize two small inlines to avoid module.h
ip_vs.h: fix implicit use of module_get/module_put from module.h
nf_conntrack.h: fix up fallout from implicit moduleparam.h presence
include: replace linux/module.h with "struct module" wherever possible
include: convert various register fcns to macros to avoid include chaining
crypto.h: remove unused crypto_tfm_alg_modname() inline
uwb.h: fix implicit use of asm/page.h for PAGE_SIZE
pm_runtime.h: explicitly requires notifier.h
linux/dmaengine.h: fix implicit use of bitmap.h and asm/page.h
miscdevice.h: fix up implicit use of lists and types
stop_machine.h: fix implicit use of smp.h for smp_processor_id
of: fix implicit use of errno.h in include/linux/of.h
of_platform.h: delete needless include <linux/module.h>
acpi: remove module.h include from platform/aclinux.h
miscdevice.h: delete unnecessary inclusion of module.h
device_cgroup.h: delete needless include <linux/module.h>
net: sch_generic remove redundant use of <linux/module.h>
net: inet_timewait_sock doesnt need <linux/module.h>
...
Fix up trivial conflicts (other header files, and removal of the ab3550 mfd driver) in
- drivers/media/dvb/frontends/dibx000_common.c
- drivers/media/video/{mt9m111.c,ov6650.c}
- drivers/mfd/ab3550-core.c
- include/linux/dmaengine.h
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (62 commits)
mlx4_core: Deprecate log_num_vlan module param
IB/mlx4: Don't set VLAN in IBoE WQEs' control segment
IB/mlx4: Enable 4K mtu for IBoE
RDMA/cxgb4: Mark QP in error before disabling the queue in firmware
RDMA/cxgb4: Serialize calls to CQ's comp_handler
RDMA/cxgb3: Serialize calls to CQ's comp_handler
IB/qib: Fix issue with link states and QSFP cables
IB/mlx4: Configure extended active speeds
mlx4_core: Add extended port capabilities support
IB/qib: Hold links until tuning data is available
IB/qib: Clean up checkpatch issue
IB/qib: Remove s_lock around header validation
IB/qib: Precompute timeout jiffies to optimize latency
IB/qib: Use RCU for qpn lookup
IB/qib: Eliminate divide/mod in converting idx to egr buf pointer
IB/qib: Decode path MTU optimization
IB/qib: Optimize RC/UC code by IB operation
IPoIB: Use the right function to do DMA unmap pages
RDMA/cxgb4: Use correct QID in insert_recv_cqe()
RDMA/cxgb4: Make sure flush CQ entries are collected on connection close
...
These files were getting the moduleparam infrastructure from the
implicit presence of module.h being everywhere, but that is going
away soon.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
To ease skb->truesize sanitization, its better to be able to localize
all references to skb frags size.
Define accessors : skb_frag_size() to fetch frag size, and
skb_frag_size_{set|add|sub}() to manipulate it.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pages that were mapped using ib_dma_map_page() should be unmapped
using ib_dma_unmap_page().
Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Reviewed-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Currently, there is only a single ("basic") type of SRQ, but with XRC
support we will add a second. Prepare for this by defining an SRQ type
and setting all current users to IB_SRQT_BASIC.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Roland Dreier <roland@kernel.org>
Cc: Sean Hefty <sean.hefty@intel.com>
Cc: Hal Rosenstock <hal.rosenstock@gmail.com>
Cc: linux-rdma@vger.kernel.org
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Print the return code of ib_post_send() if it fails to make these
debugging messages more useful.
Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The IPoIB UD QP reports send completions to priv->send_cq, which is
usually left unarmed; it only gets armed when the number of
outstanding send requests reaches the size of the TX queue. This
arming is done only in the send path for the UD QP. However, when
sending CM packets, the net queue may be stopped for the same reasons
but no measures are taken to recover the UD path from a lockup.
Consider this scenario: a host sends high rate of both CM and UD
packets, with a TX queue length of N. If at some time the number of
outstanding UD packets is more than N/2 and the overall outstanding
packets is N-1, and CM sends a packet (making the number of
outstanding sends equal N), the TX queue will be stopped. When all
the CM packets complete, the number of outstanding packets will still
be higher than N/2 so the TX queue will not be restarted.
Fix this by calling ib_req_notify_cq() when the queue is stopped in
the CM path.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Dunno, what was the idea, it wasn't used for a long time.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next-2.6: (1623 commits)
netxen: update copyright
netxen: fix tx timeout recovery
netxen: fix file firmware leak
netxen: improve pci memory access
netxen: change firmware write size
tg3: Fix return ring size breakage
netxen: build fix for INET=n
cdc-phonet: autoconfigure Phonet address
Phonet: back-end for autoconfigured addresses
Phonet: fix netlink address dump error handling
ipv6: Add IFA_F_DADFAILED flag
net: Add DEVTYPE support for Ethernet based devices
mv643xx_eth.c: remove unused txq_set_wrr()
ucc_geth: Fix hangs after switching from full to half duplex
ucc_geth: Rearrange some code to avoid forward declarations
phy/marvell: Make non-aneg speed/duplex forcing work for 88E1111 PHYs
drivers/net/phy: introduce missing kfree
drivers/net/wan: introduce missing kfree
net: force bridge module(s) to be GPL
Subject: [PATCH] appletalk: Fix skb leak when ipddp interface is not loaded
...
Fixed up trivial conflicts:
- arch/x86/include/asm/socket.h
converted to <asm-generic/socket.h> in the x86 tree. The generic
header has the same new #define's, so that works out fine.
- drivers/net/tun.c
fix conflict between 89f56d1e9 ("tun: reuse struct sock fields") that
switched over to using 'tun->socket.sk' instead of the redundantly
available (and thus removed) 'tun->sk', and 2b980dbd ("lsm: Add hooks
to the TUN driver") which added a new 'tun->sk' use.
Noted in 'next' by Stephen Rothwell.
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: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Neil Horman <nhorman@txudriver.com>
Signed-off-by: David 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: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
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: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace all uses of IPOIB_GID_FMT, IPOIB_GID_RAW_ARG() and IPOIB_GID_ARG()
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, IPoIB is an LLTX driver that uses its own IRQ-disabling
tx_lock. Not only do we want to get rid of LLTX, this actually causes
problems because of the skb_orphan() done with this tx_lock held: some
skb destructors expect to be run with interrupts enabled.
The simplest fix for this is to get rid of the driver-private tx_lock
and stop using LLTX. We kill off priv->tx_lock and use
netif_tx_lock[_bh]() instead; the patch to do this is a tiny bit
tricky because we need to update places that take priv->lock inside
the tx_lock to disable IRQs, rather than relying on tx_lock having
already disabled IRQs.
Also, there are a couple of places where we need to disable BHs to
make sure we have a consistent context to call netif_tx_lock() (since
we no longer can use _irqsave() variants), and we also have to change
ipoib_send_comp_handler() to call drain_tx_cq() through a timer rather
than directly, because ipoib_send_comp_handler() runs in interrupt
context and drain_tx_cq() must run in BH context so it can call
netif_tx_lock().
Signed-off-by: Roland Dreier <rolandd@cisco.com>
There are users that are running UDP applications that require a large
receive queue size in order to get good performance. To prevent
allocation failures for rx_rings when using non-SRQ mode and large
recv_queue_size (1K or larger), use vmalloc() instead of kcalloc() to
alocate rx_rings.
Signed-off-by: David Wilder <dwilder@us.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
wr->sg_list should be set to the sge pointer passed in, not
priv->cm.rx_sge.
Reported-by: Hoang-Nam Nguyen <HNGUYEN@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Since IPoIB connected mode does not NETIF_F_SG, we only have one DMA
mapping per send, so we don't need a mapping[] array. Define a new
struct with a single u64 mapping member and use it for the CM tx_ring.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
When the driver sets the MTU of the net device outside of its
change_mtu method, it should make use of dev_set_mtu() instead of
directly setting the mtu field of struct netdevice. Otherwise
functions registered to be called upon MTU change will not get called
(this is done through call_netdevice_notifiers() in dev_set_mtu()).
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Use of this lock is required to synchronize changes to the netdvice's
data structs. Also move the call to ipoib_flush_paths() after the
modification of the netdevice flags in set_mode().
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
For devices that don't support SRQs, ipoib_cm_post_receive_nonsrq() is
called from both ipoib_cm_handle_rx_wc() and ipoib_cm_nonsrq_init_rx(),
and these two callers are not synchronized against each other.
However, ipoib_cm_post_receive_nonsrq() always reuses the same receive
work request and scatter list structures, so multiple callers can end
up stepping on each other, which leads to posting garbled work
requests.
Fix this by having the caller pass in the ib_recv_wr and ib_sge
structures to use, and allocating new local structures in
ipoib_cm_nonsrq_init_rx().
Based on a patch by Pradeep Satyanarayana <pradeep@us.ibm.com> and
David Wilder <dwilder@us.ibm.com>, with debugging help from Hoang-Nam
Nguyen <hnguyen@de.ibm.com>.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The connected mode implementation in the IPoIB driver has a large
overhead in the way SKBs are handled in the receive flow. It usually
allocates an SKB with as big as was used in the currently received SKB
and moves unused fragments from the old SKB to the new one. This
involves a loop on all the remaining fragments and incurs overhead on
the CPU. This patch, for small SKBs, allocates an SKB just large
enough to contain the received data and copies to it the data from the
received SKB. The newly allocated SKB is passed to the stack and the
old SKB is reposted.
When running netperf, UDP small messages, without this pach I get:
UDP UNIDIRECTIONAL SEND TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to
14.4.3.178 (14.4.3.178) port 0 AF_INET
Socket Message Elapsed Messages
Size Size Time Okay Errors Throughput
bytes bytes secs # # 10^6bits/sec
114688 128 10.00 5142034 0 526.31
114688 10.00 1130489 115.71
With this patch I get both send and receive at ~315 mbps.
The reason that send performance actually slows down is as follows:
When using this patch, the overhead of the CPU for handling RX packets
is dramatically reduced. As a result, we do not experience RNR NAK
messages from the receiver which cause the connection to be closed and
reopened again; when the patch is not used, the receiver cannot handle
the packets fast enough so there is less time to post new buffers and
hence the mentioned RNR NACKs. So what happens is that the
application *thinks* it posted a certain number of packets for
transmission but these packets are flushed and do not really get
transmitted. Since the connection gets opened and closed many times,
each time netperf gets the CPU time that otherwise would have been
given to IPoIB to actually transmit the packets. This can be verified
when looking at the port counters -- the output of ifconfig and the
oputput of netperf (this is for the case without the patch):
tx packets
==========
port counter: 1,543,996
ifconfig: 1,581,426
netperf: 5,142,034
rx packets
==========
netperf 1,1304,089
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Use a dedicated CQ for UD send completions. Also, do not arm the UD
send CQ, which reduces the number of interrupts generated. This patch
farther reduces overhead by not calling poll CQ for every posted send
WR -- it does polls only when there 16 or more outstanding work requests.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
If a P_Key is deleted and then re-added at the same index, then IPoIB
gets confused because __ipoib_ib_dev_flush() only checks whether the
index is the same without checking whether the P_Key was present, so
the interface is stopped when the P_Key is deleted, but the event when
the P_Key is re-added gets ignored and the interface never gets
restarted.
Also, switch to using ib_find_pkey() instead of ib_find_cached_pkey()
everywhere in IPoIB, since none of the places that look for P_Keys are
in a fast path or in non-sleeping context, and in general we want to
kill off the whole caching infrastructure eventually. This also fixes
consistency problems caused because some IPoIB queries were cached and
some were uncached during the window where the cache was not updated.
Thanks to Venkata Subramonyam <vsubramo@cisco.com> for debugging this
problem and testing this fix.
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>
For HCAs that support checksum offload (ie that set IB_DEVICE_UD_IP_CSUM
in the device capabilities flags), have IPoIB set NETIF_F_IP_CSUM and
use the HCA to generate and verify IP checksums.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Commit 7143740d ("IPoIB: Add send gather support") made struct
ipoib_tx_buf significantly larger, since the mapping member changed
from a single u64 to an array with MAX_SKB_FRAGS + 1 entries. This
means that allocating tx_rings with kzalloc() may fail because there
is not enough contiguous memory for the new, much bigger size. Fix
this regression by allocating the rings with vmalloc() instead.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Commit 7143740d ("IPoIB: Add send gather support") made it possible
for tx_wr.num_sge to be != 1 -- this happens if send gather support is
enabled. However, the code in the connected mode post_send() function
assumes the old invariant, namely that tx_wr.num_sge is always 1. Fix
this by explicitly setting tx_wr.num_sge to 1 in the CM post_send().
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Commit efcd9971 ("IPoIB/cm: Factor out ipoib_cm_free_rx_reap_list()")
introduced a bug in ipoib_cm_dev_stop() when the receive drain times
out. In that case, the function moves all the pending rx stuff into a
private list but then calls ipoib_cm_free_rx_reap_list(), which
handles a different list.
Fix this by moving everything to the rx_reap_list that will actually
get freed up.
This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=906>.
Signed-off-by: Pradeep Satyanarayana <pradeeps@linux.vnet.ibm.com>
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 HCAs (such as ehca2) support SRQ, but only support fewer than 16 SG
entries for SRQs. Currently IPoIB/CM implicitly assumes all HCAs will
support 16 SG entries for SRQs (to handle a 64K MTU with 4K pages). This
patch removes that restriction by limiting the maximum MTU in connected
mode to what the maximum number of SRQ SG entries allows.
This patch addresses <https://bugs.openfabrics.org/show_bug.cgi?id=728>
Signed-off-by: Pradeep Satyanarayana <pradeeps@linux.vnet.ibm.com>
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>
Factor out the code for going through the rx_reap list of struct
ipoib_cm_rx and freeing each one. This consolidates the code
duplicated between ipoib_cm_dev_stop() and ipoib_cm_rx_reap() and
reduces the risk of error when adding additional accounting.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Factor out the code to create an SRQ and allocate the receive ring in
ipoib_cm_dev_init() into a new function ipoib_cm_create_srq(). This
will make the code neater when support for devices that don't implement
SRQs is added.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Factor out the code to unmap/free skbs and free the receive ring in
ipoib_cm_dev_cleanup() into a new function ipoib_cm_free_rx_ring().
This function will be called from a couple of other places when
support for devices that don't implement SRQs is added.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Commit 1b524963 ("IPoIB/cm: Use common CQ for CM send completions")
changed how the high-order bits of work request IDs were used, which
had the effect that IPOIB_CM_RX_DRAIN_WRID was no longer handled as a
connected mode receive completion. This leads to the messages
ib1: cm send completion event with wrid 1073741823 (> 64)
ib1: RX drain timing out
when an interface with connected mode QPs is brought down. Fix this
by making sure that both IPOIB_OP_CM and IPOIB_OP_RECV are set in
IPOIB_CM_RX_DRAIN_WRID.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Use the same CQ for CM send completions as for all other IPoIB
completions. This means all completions are processed via the same
NAPI polling routine. This should help reduce the number of
interrupts for bi-directional traffic (such as TCP) and fixes "driver
is hogging interrupts" errors reported for IPoIB send side, e.g.
<https://bugs.openfabrics.org/show_bug.cgi?id=508>
To do this, keep a per-interface counter of outstanding send WRs, and
stop the interface when this counter reaches the send queue size to
avoid CQ overruns.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
It's too hard to figure out what "!likely(...)" really means, and who
knows how compilers interpret the hint.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (87 commits)
mlx4_core: Fix section mismatches
IPoIB: Allow setting policy to ignore multicast groups
IB/mthca: Mark error paths as unlikely() in post_srq_recv functions
IB/ipath: Minor fix to ordering of freeing and zeroing of tid pages.
IB/ipath: Remove redundant link state checks
IB/ipath: Fix IB_EVENT_PORT_ERR event
IB/ipath: Better handling of unexpected GPIO interrupts
IB/ipath: Maintain active time on all chips
IB/ipath: Fix QHT7040 serial number check
IB/ipath: Indicate a couple of chip bugs to userspace
IB/ipath: iba6110 rev4 no longer needs recv header overrun workaround
IB/ipath: Use counters in ipath_poll and cleanup interrupts in ipath_close
IB/ipath: Remove duplicate copy of LMC
IB/ipath: Add ability to set the LMC via the sysfs debugging interface
IB/ipath: Optimize completion queue entry insertion and polling
IB/ipath: Implement IB_EVENT_QP_LAST_WQE_REACHED
IB/ipath: Generate flush CQE when QP is in error state
IB/ipath: Remove redundant code
IB/ipath: Future proof eeprom checksum code (contents reading)
IB/ipath: UC RDMA WRITE with IMMEDIATE doesn't send the immediate
...
Use the stats member of struct netdevice in IPoIB, so we can save
memory by deleting the stats member of struct ipoib_dev_priv, and save
code by deleting ipoib_get_stats().
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make the way QP is being created in ipoib_cm_create_tx_qp()
consistent with ipoib_cm_create_rx_qp().
Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The IB CM should include the HCA ACK delay when calculating the local
ACK timeout value to use for RC QPs. If the HCA ACK delay is large
enough relative to the packet life time, then if it is not taken into
account, the calculated timeout value ends up being too small, which
can result in "retry exceeded" errors.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Fix
drivers/infiniband/ulp/ipoib/ipoib_cm.c:1151: warning: unused variable 'dev'
by getting rid of the variable dev, which is only used if CONFIG_IPV6
is enabled, and replacing the one use of it with the value it is
assigned, namely priv->dev.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
If a page can't be allocated for the frag list of a skb, the code to
unmap the partially allocated list is off by one. For exaple, if
'frags' equals one, i == 0, and the alloc_page() fails, then the old
loop would have unmapped mapping[1] which is uninitialized. The same
would happen if the call to ib_dma_map_page() failed.
Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Acked-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
IPoIB connected mode currently rejects a connection request unless the
supported MTU is >= the local netdevice MTU. This breaks
interoperability with implementations that might have tweaked
IPOIB_CM_MTU, and there's real no longer a reason to do so: this test
is just a leftover from when we did not tweak MTU per-connection. Fix
this by making the test as permissive as possible.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Fix a crasher bug in IPoIB CM: once a QP is in the RTR state, a
receive completion (or even an asynchronous error) might be observed
on this QP, so we have to initialize all of our receive data
structures before moving to the RTR state.
As an optimization (since modify_qp might take a long time), the
jiffies update done when moving RX to the passive_ids list is also
left in place to reduce the chance of the RX being misdetected as
stale.
This fixes bug <https://bugs.openfabrics.org/show_bug.cgi?id=662>.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
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>
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>
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>
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>
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>
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>
