Pull rdma updates from Jason Gunthorpe:
"This has been a smaller cycle with many of the commits being smallish
code fixes and improvements across the drivers.
- Driver updates for bnxt_re, cxgb4, hfi1, hns, mlx5, nes, qedr, and
rxe
- Memory window support in hns
- mlx5 user API 'flow mutate/steering' allows accessing the full
packet mangling and matching machinery from user space
- Support inter-working with verbs API calls in the 'devx' mlx5 user
API, and provide options to use devx with less privilege
- Modernize the use of syfs and the device interface to use attribute
groups and cdev properly for uverbs, and clean up some of the core
code's device list management
- More progress on net namespaces for RDMA devices
- Consolidate driver BAR mmapping support into core code helpers and
rework how RDMA holds poitners to mm_struct for get_user_pages
cases
- First pass to use 'dev_name' instead of ib_device->name
- Device renaming for RDMA devices"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (242 commits)
IB/mlx5: Add support for extended atomic operations
RDMA/core: Fix comment for hw stats init for port == 0
RDMA/core: Refactor ib_register_device() function
RDMA/core: Fix unwinding flow in case of error to register device
ib_srp: Remove WARN_ON in srp_terminate_io()
IB/mlx5: Allow scatter to CQE without global signaled WRs
IB/mlx5: Verify that driver supports user flags
IB/mlx5: Support scatter to CQE for DC transport type
RDMA/drivers: Use core provided API for registering device attributes
RDMA/core: Allow existing drivers to set one sysfs group per device
IB/rxe: Remove unnecessary enum values
RDMA/umad: Use kernel API to allocate umad indexes
RDMA/uverbs: Use kernel API to allocate uverbs indexes
RDMA/core: Increase total number of RDMA ports across all devices
IB/mlx4: Add port and TID to MAD debug print
IB/mlx4: Enable debug print of SMPs
RDMA/core: Rename ports_parent to ports_kobj
RDMA/core: Do not expose unsupported counters
IB/mlx4: Refer to the device kobject instead of ports_parent
RDMA/nldev: Allow IB device rename through RDMA netlink
...
IPCB should be cleared before icmp_send, since it may contain data from
previous layers and the data could be misinterpreted as ip header options,
which later caused the ihl to be set to an invalid value and resulted in
the following stack corruption:
[ 1083.031512] ib0: packet len 57824 (> 2048) too long to send, dropping
[ 1083.031843] ib0: packet len 37904 (> 2048) too long to send, dropping
[ 1083.032004] ib0: packet len 4040 (> 2048) too long to send, dropping
[ 1083.032253] ib0: packet len 63800 (> 2048) too long to send, dropping
[ 1083.032481] ib0: packet len 23960 (> 2048) too long to send, dropping
[ 1083.033149] ib0: packet len 63800 (> 2048) too long to send, dropping
[ 1083.033439] ib0: packet len 63800 (> 2048) too long to send, dropping
[ 1083.033700] ib0: packet len 63800 (> 2048) too long to send, dropping
[ 1083.034124] ib0: packet len 63800 (> 2048) too long to send, dropping
[ 1083.034387] ==================================================================
[ 1083.034602] BUG: KASAN: stack-out-of-bounds in __ip_options_echo+0xf08/0x1310
[ 1083.034798] Write of size 4 at addr ffff880353457c5f by task kworker/u16:0/7
[ 1083.034990]
[ 1083.035104] CPU: 7 PID: 7 Comm: kworker/u16:0 Tainted: G O 4.19.0-rc5+ #1
[ 1083.035316] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu2 04/01/2014
[ 1083.035573] Workqueue: ipoib_wq ipoib_cm_skb_reap [ib_ipoib]
[ 1083.035750] Call Trace:
[ 1083.035888] dump_stack+0x9a/0xeb
[ 1083.036031] print_address_description+0xe3/0x2e0
[ 1083.036213] kasan_report+0x18a/0x2e0
[ 1083.036356] ? __ip_options_echo+0xf08/0x1310
[ 1083.036522] __ip_options_echo+0xf08/0x1310
[ 1083.036688] icmp_send+0x7b9/0x1cd0
[ 1083.036843] ? icmp_route_lookup.constprop.9+0x1070/0x1070
[ 1083.037018] ? netif_schedule_queue+0x5/0x200
[ 1083.037180] ? debug_show_all_locks+0x310/0x310
[ 1083.037341] ? rcu_dynticks_curr_cpu_in_eqs+0x85/0x120
[ 1083.037519] ? debug_locks_off+0x11/0x80
[ 1083.037673] ? debug_check_no_obj_freed+0x207/0x4c6
[ 1083.037841] ? check_flags.part.27+0x450/0x450
[ 1083.037995] ? debug_check_no_obj_freed+0xc3/0x4c6
[ 1083.038169] ? debug_locks_off+0x11/0x80
[ 1083.038318] ? skb_dequeue+0x10e/0x1a0
[ 1083.038476] ? ipoib_cm_skb_reap+0x2b5/0x650 [ib_ipoib]
[ 1083.038642] ? netif_schedule_queue+0xa8/0x200
[ 1083.038820] ? ipoib_cm_skb_reap+0x544/0x650 [ib_ipoib]
[ 1083.038996] ipoib_cm_skb_reap+0x544/0x650 [ib_ipoib]
[ 1083.039174] process_one_work+0x912/0x1830
[ 1083.039336] ? wq_pool_ids_show+0x310/0x310
[ 1083.039491] ? lock_acquire+0x145/0x3a0
[ 1083.042312] worker_thread+0x87/0xbb0
[ 1083.045099] ? process_one_work+0x1830/0x1830
[ 1083.047865] kthread+0x322/0x3e0
[ 1083.050624] ? kthread_create_worker_on_cpu+0xc0/0xc0
[ 1083.053354] ret_from_fork+0x3a/0x50
For instance __ip_options_echo is failing to proceed with invalid srr and
optlen passed from another layer via IPCB
[ 762.139568] IPv4: __ip_options_echo rr=0 ts=0 srr=43 cipso=0
[ 762.139720] IPv4: ip_options_build: IPCB 00000000f3cd969e opt 000000002ccb3533
[ 762.139838] IPv4: __ip_options_echo in srr: optlen 197 soffset 84
[ 762.139852] IPv4: ip_options_build srr=0 is_frag=0 rr_needaddr=0 ts_needaddr=0 ts_needtime=0 rr=0 ts=0
[ 762.140269] ==================================================================
[ 762.140713] IPv4: __ip_options_echo rr=0 ts=0 srr=0 cipso=0
[ 762.141078] BUG: KASAN: stack-out-of-bounds in __ip_options_echo+0x12ec/0x1680
[ 762.141087] Write of size 4 at addr ffff880353457c7f by task kworker/u16:0/7
Signed-off-by: Denis Drozdov <denisd@mellanox.com>
Reviewed-by: Erez Shitrit <erezsh@mellanox.com>
Reviewed-by: Feras Daoud <ferasda@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
IPoIB netlink support was broken by the below commit since integrating
the rdma_netdev support relies on an allocation flow for netdevs that
was controlled by the ipoib driver while netdev's rtnl_newlink
implementation assumes that the netdev will be allocated by netlink.
Such situation leads to crash in __ipoib_device_add, once trying to
reuse netlink device.
This patch fixes the kernel oops for both mlx4 and mlx5
devices triggered by the following command:
Fixes: cd565b4b51 ("IB/IPoIB: Support acceleration options callbacks")
Signed-off-by: Denis Drozdov <denisd@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Feras Daoud <ferasda@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
netdev has several interfaces that expect to call alloc_netdev_mqs from
the core code, with the driver only providing the arguments. This is
incompatible with the rdma_netdev interface that returns the netdev
directly.
Thus re-organize the API used by ipoib so that the verbs core code calls
alloc_netdev_mqs for the driver. This is done by allowing the drivers to
provide the allocation parameters via a 'get_params' callback and then
initializing an allocated netdev as a second step.
Fixes: cd565b4b51 ("IB/IPoIB: Support acceleration options callbacks")
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Denis Drozdov <denisd@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Some tools may currently be using only the deprecated attribute;
let's print an elaborate and clear deprecation notice to kmsg.
To do that, we have to replace the whole sysfs file, since we inherit
the original one from netdev.
Signed-off-by: Arseny Maslennikov <ar@cs.msu.ru>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Some InfiniBand network devices have multiple ports on the same PCI
function. This initializes the `dev_port' sysfs field of those
network interfaces with their port number.
Prior to this the kernel erroneously used the `dev_id' sysfs
field of those network interfaces to convey the port number to userspace.
The use of `dev_id' was considered correct until Linux 3.15,
when another field, `dev_port', was defined for this particular
purpose and `dev_id' was reserved for distinguishing stacked ifaces
(e.g: VLANs) with the same hardware address as their parent device.
Similar fixes to net/mlx4_en and many other drivers, which started
exporting this information through `dev_id' before 3.15, were accepted
into the kernel 4 years ago.
See 76a066f2a2 (`net/mlx4_en: Expose port number through sysfs').
Signed-off-by: Arseny Maslennikov <ar@cs.msu.ru>
Signed-off-by: Doug Ledford <dledford@redhat.com>
It is illegal to change MTU to a value lower than the minimum MTU
stated in ethernet spec. In addition to that we need to add 4 bytes
for encapsulation header (IPOIB_ENCAP_LEN).
Before "ifconfig ib0 mtu 0" command, succeeds while it obviously shouldn't.
Signed-off-by: Muhammad Sammar <muhammads@mellanox.com>
Reviewed-by: Feras Daoud <ferasda@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Inside of start_xmit() the call to check if the connection is up and the
queueing of the packets for later transmission is not atomic which leaves
a window where cm_rep_handler can run, set the connection up, dequeue
pending packets and leave the subsequently queued packets by start_xmit()
sitting on neigh->queue until they're dropped when the connection is torn
down. This only applies to connected mode. These dropped packets can
really upset TCP, for example, and cause multi-minute delays in
transmission for open connections.
Here's the code in start_xmit where we check to see if the connection is
up:
if (ipoib_cm_get(neigh)) {
if (ipoib_cm_up(neigh)) {
ipoib_cm_send(dev, skb, ipoib_cm_get(neigh));
goto unref;
}
}
The race occurs if cm_rep_handler execution occurs after the above
connection check (specifically if it gets to the point where it acquires
priv->lock to dequeue pending skb's) but before the below code snippet in
start_xmit where packets are queued.
if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) {
push_pseudo_header(skb, phdr->hwaddr);
spin_lock_irqsave(&priv->lock, flags);
__skb_queue_tail(&neigh->queue, skb);
spin_unlock_irqrestore(&priv->lock, flags);
} else {
++dev->stats.tx_dropped;
dev_kfree_skb_any(skb);
}
The patch acquires the netif tx lock in cm_rep_handler for the section
where it sets the connection up and dequeues and retransmits deferred
skb's.
Fixes: 839fcaba35 ("IPoIB: Connected mode experimental support")
Cc: stable@vger.kernel.org
Signed-off-by: Aaron Knister <aaron.s.knister@nasa.gov>
Tested-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Move all the checking for pkey and other validity to the __ipoib_vlan_add
function. This removes the last difference from the control flow
of the __ipoib_vlan_add to make the overall design simpler to
understand.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
This fixes a bug in the netlink path where the vlan_rwsem was not
held around __ipoib_vlan_add causing the child_intfs to be manipulated
unsafely.
In the process this greatly simplifies the vlan_rwsem write side locking
to only cover a single non-sleeping statement.
This also further increases the safety of the removal ordering by holding
the netdev of the parent while the child is active to ensure most bugs
become either an oops on a NULL priv or a deadlock on the netdev refcount.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Switching to priv_destructor and needs_free_netdev created a subtle
ordering problem in ipoib_remove_one.
Now that unregister_netdev frees the netdev and priv we must ensure that
the children are unregistered before trying to unregister the parent,
or child unregister will use after free.
The solution is to unregister the children, then parent, in the same batch
all while holding the rtnl_lock. This closes all the races where a new
child could have been added and ensures proper ordering.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
This mutex was introduced to deal with the deadlock formed by calling
unregister_netdev from within the sysfs callback of a netdev.
Now that we have priv_destructor and needs_free_netdev we can switch
to the more targeted solution of running the unregister from a
work queue. This avoids the deadlock and gets rid of the mutex.
The next patch in the series needs this mutex eliminated to create
atomicity of unregisteration.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Now that the unregister_netdev flow for IPoIB no longer relies on external
code we can now introduce the use of priv_destructor and
needs_free_netdev.
The rdma_netdev flow is switched to use the netdev common priv_destructor
instead of the special free_rdma_netdev and the IPOIB ULP adjusted:
- priv_destructor needs to switch to point to the ULP's destructor
which will then call the rdma_ndev's in the right order
- We need to be careful around the error unwind of register_netdev
as it sometimes calls priv_destructor on failure
- ULPs need to use ndo_init/uninit to ensure proper ordering
of failures around register_netdev
Switching to priv_destructor is a necessary pre-requisite to using
the rtnl new_link mechanism.
The VNIC user for rdma_netdev should also be revised, but that is left for
another patch.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Denis Drozdov <denisd@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Now that we have a proper ndo_uninit, move code that naturally pairs
with the ndo_uninit into ndo_init. This allows the netdev core to natually
handle ordering.
This fixes the situation where register_netdev can fail before calling
ndo_init, in which case it wouldn't call ndo_uninit either.
Also move a bunch of duplicated init code that is shared between child
and parent for clarity. Now the child and parent register functions look
very similar.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Currently uninit is sometimes done twice in error flows, and is sprinkled
a bit all over the place.
Improve the clarity of the design by moving all uninit only into
ndo_uinit.
Some duplication is removed:
- Sometimes IPOIB_STOP_NEIGH_GC was done before unregister, but
this duplicates the process in ipoib_neigh_hash_init
- Flushing priv->wq was sometimes done before unregister,
but that duplicates what has been done in ndo_uninit
Uniniting the IB event queue must remain before unregister_netdev as it
requires the RTNL lock to be dropped, this is moved to a helper to make
that flow really clear and remove some duplication in error flows.
If register_netdev fails (and ndo_init is NULL) then it almost always
calls ndo_uninit, which lets us remove all the extra code from the error
unwinds. The next patch in the series will close the 'almost always' hole
by pairing a proper ndo_init with ndo_uninit.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
The neigh_reap_task is self restarting, but so long as we call
cancel_delayed_work_sync() it will be guaranteed to not be running and
never start again. Thus we don't need to have the racy
IPOIB_STOP_NEIGH_GC bit, or the confusing mismatch of places sometimes
calling flush_workqueue after the cancel.
This fixes a situation where the GC work could have been left running
in some rare situations.
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
This essentially duplicates the netdev's reg_state, so just use that
directly. The reg_state is updated under the rntl_lock, and all places
using GOING_DOWN already acquire the rtnl_lock so checking is safe.
Since the only place we use GOING_DOWN is for the parent device this
does not fix any bugs, but it is a step to tidy up the unregister flow
so that after later patches the flow is uniform and sane.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
The change of ipoib_ah data structure with adding "valid" flag and
checks of ah->valid in ipoib_start_xmit affected multicast packet flow.
Since the multicast flow doesn't invoke path_rec_start, "ah->valid" flag
remains unset, so that ipoib_start_xmit end up with neigh_refresh_path
instead of sending the packet using neigh.
"ah->valid" has to be set in multicast send flow. As a result IPoIB
starts sending packets via neigh immediately and eliminates 60sec delay
of neigh keep alive interval.
The typical example of this issue are two sequential arpings:
arping 11.134.208.9 -> got response (mcast_send)
arping 11.134.208.9 -> no response (ah->valid = 0)
Fixes: fa9391dbad ("RDMA/ipoib: Update paths on CLIENT_REREG/SM_CHANGE events")
Signed-off-by: Denis Drozdov <denisd@mellanox.com>
Reviewed-by: Erez Shitrit <erezsh@mellanox.com>
Reviewed-by: Feras Daoud <ferasda@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
The proper return code is "-EOPNOTSUPP" when the modify_cq() callback is
not supported, all drivers should generate this and all users should check
for it when detecting not supported functionality.
Signed-off-by: Kamal Heib <kamalheib1@gmail.com>
Acked-by: Leon Romanovsky <leonro@mellanox.com> (for mlx5)
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Constify several pointers such as path_rec, ib_cm_event and listen_id
pointers in several functions.
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Make sure to check for "-EOPNOTSUPP" instead of "-ENOSYS" which is the
return code from ib_create_srq() in case that it not supported.
Signed-off-by: Kamal Heib <kamalheib1@gmail.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Instead of declaring and passing a dummy 'bad_wr' pointer, pass NULL
as third argument to ib_post_(send|recv|srq_recv)().
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Fix to return a negative error code from the ipoib_neigh_hash_init()
error handling case instead of 0, as done elsewhere in this function.
Fixes: 515ed4f3aa ("IB/IPoIB: Separate control and data related initializations")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Yuval Shaia <yuval.shaia@oracle.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
This table by default takes 32KiB which is 3rd memory order. Meanwhile,
this memory is not aimed for DMA operation and could be safely allocated
by vmalloc.
Signed-off-by: Jan Dakinevich <jan.dakinevich@virtuozzo.com>
Reviewed-by: Håkon Bugge <haakon.bugge@oracle.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Linus Torvalds
Denis Drozdov
Jason Gunthorpe
Arseny Maslennikov
Muhammad Sammar
Aaron Knister
Erez Shitrit
Kamal Heib
Parav Pandit
Bart Van Assche
Wei Yongjun
Jan Dakinevich