We insert padding if the MAC header's size is not a multiple of 4
to ensure that the SNAP header is DWORD aligned. When we do so, we
let the firmware know by setting a bit in Tx command
(TX_CMD_FLG_MH_PAD) which will instruct the firmware to drop those
2 bytes before sending the frame.
However, this is not needed for AMSDU as the sub frame header (14B)
complements the MAC header (26B) so that the SNAP header is DWORD
aligned without adding any pad.

Until 9000, the firmware didn't check the TX_CMD_FLG_MH_PAD bit
but rather checked the length of the MAC header itself and
assumed the entity that enqueued the frame (driver or internal
firmware code) added the pad.
Since the driver inserted the pad even for AMSDU this logic applied.
Note that the padding is a DMA optimization but it's not strictly
needed, so we could pad even if it was not needed.

However, the CSUM hardware introduced for the 9000 devices requires
to not pad AMSDU as it is not needed, and will fail if such a pad
exists.
Due to older FW not checking the padding bit but checking the mac
header size itself - we cannot do this adjustments for older
generations.
Do not align the size if it is an AMSDU and HW checksum is enabled -
which will only happen on 9000 devices and on.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

This makes u-APSD work with more peers.
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

Bjorn pointed out that printing an error value as an
hexadecimal isn't very convenient. Change that.
Reported-by: NBjorn Helgaas <bhelgaas@google.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

Use RX_HANDLER_ASYNC_UNLOCKED instead of unlock and re-lock
the mutex independently.
Signed-off-by: NChaya Rachel Ivgi <chaya.rachel.ivgi@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

We don't use the refcount value anymore, all the refcounting is done
in the runtime PM usage_count value.  Remove it.
Signed-off-by: NLuca Coelho <luciano.coelho@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

When entering suspend the driver calls iwl_disable_interrupts() and
then iwl_pcie_disable_ict().
On resume the driver calls only iwl_pcie_reset_ict() without calling
explicitly to iwl_enable_interrupts().
This mostly works since iwl_pcie_reset_ict is calling to
iwl_enable_interrupts, but it doesn't work when there is no ict_table
in MSIx mode.
The result is that driver tries to resume but fails since it doesn't
get the RX interrupt from FW indicating that d0i3 exit was completed.
Fix it by adding an explicit call to enable interrupts.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

My patch resized the pool size, but neglected to resize
the global table, which is obviously wrong since the global
table maps the pool's rxb to vid one to one. This results
in a panic in 9000 devices.
Add a build bug to avoid such a case in the future.

Fixes: 7b542436 ("iwlwifi: pcie: fine tune number of rxbs")
Reported-by: NHaim Dreyfuss <haim.dreyfuss@intel.com>
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

Add debugfs entry named lqm_send_cmd for kicking a
measurement. This hook takes the duration and the timeout
as parameter.
Signed-off-by: NAviya Erenfeld <aviya.erenfeld@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

LQM stands for Link Quality Measurement. The firmware
will collect a defined set of statitics (see the
notification for details) that allow to know how busy
the medium is. The driver issues a request to the firmware
that includes the duration of the measurement (the firmware
needs to be on channel for that amount of time) and the
timeout (in case the firmware has a lot of offchannel
activities). If the timeout elapses, the firmware will
send partial results which are still valuable.
In case of disassociation / channel switch and alike, the
driver is in charge of stopping the measurements and the
firmware will reply with partial results.

The user space API for now is debugfs only and will be
implmemented in an upcoming patch.
Signed-off-by: NAviya Erenfeld <aviya.erenfeld@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

In case of FW error, support dumping the UMAC internal txfifos.
To do so, support version 2 of shared memory cfg command, which
contains the sizes of the internal txfifos, and move the command
to the system group.
Signed-off-by: NGolan Ben-Ami <golan.ben.ami@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

Paging contains 3 sections in the fw. The first for the paging separator,
The second for the CSS block, the third with the paging data.

Currently if the driver finds the paging separator, and there is only
section left (CSS), once reading the CSS section, the driver will
attempt to read the paging data and will go out of the arrays bounds.

Make sure that the FW image contains the right amount of sections for
paging.
Signed-off-by: NMatti Gottlieb <matti.gottlieb@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

Currently the driver has 2 buffers for paging:
1. paging db - this contains all of the pages that were in the FW
image, that the driver stores for the FW. This is allocated for each
block separately (not contiguous).
2. download buffer - we need to provide this empty buffer for the
iwl_sdio_load_fw_chunk function to copy the requested pages to the shared
memory. This is one big buffer of contiguous memory whose size is the
size of all the blocks that the fw paging section can contain.

This download buffer size is too big, and causes the allocation to fail
sometimes. Since the driver allocates memory for each block separately,
it is not possible for the FW to request all of the pages in one request
(the FW gives an address and size, so blocks need to be contiguous for
this to happen), therefore the FW is limited to request only one block.

Decrease the size of the paging download buffer to be the size of a
paging block.
Signed-off-by: NMatti Gottlieb <matti.gottlieb@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

Currently when stop flow is performed, there might be transport TX RTPM
references that are not freed in case we unmap a queue that still has
packets not reclaimed. Fix that.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

bond_get_stats() can be called from rtnetlink (with RTNL held)
or from /proc/net/dev seq handler (with RCU held)

The logic added in commit 5f0c5f73 ("bonding: make global bonding
stats more reliable") kind of assumed only one cpu could run there.

If multiple threads are reading /proc/net/dev, stats can be really
messed up after a while.

A second problem is that some fields are 32bit, so we need to properly
handle the wrap around problem.

Given that RTNL is not always held, we need to use
bond_for_each_slave_rcu().

Fixes: 5f0c5f73 ("bonding: make global bonding stats more reliable")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Andy Gospodarek <gospo@cumulusnetworks.com>
Cc: Jay Vosburgh <j.vosburgh@gmail.com>
Cc: Veaceslav Falico <vfalico@gmail.com>
Reviewed-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When un-mapping skb->data in __bcmgenet_tx_reclaim(),
we must use the length that was used in original dma_map_single(),
instead of skb->len that might be bigger (includes the frags)

We simply can store skb_len into tx_cb_ptr->dma_len and use it
at unmap time.

Fixes: 1c1008c7 ("net: bcmgenet: add main driver file")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Acked-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 85743f1e ("net/mlx4_core: Set UAR page size to 4KB regardless
of system page size") introduced dependency where old VF drivers without
this fix fail to load if the PF driver runs with this commit.

To resolve this add a module parameter which disables that functionality
by default.  If both the PF and VFs are running with a driver with that
commit the administrator may set the module param to true.

The module parameter is called enable_4k_uar.

Fixes: 85743f1e ('net/mlx4_core: Set UAR page size to 4KB ...')
Signed-off-by: NEli Cohen <eli@mellanox.com>
Tested-by: NAlexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Drop two extra occurrences of "on" in option title and help text.

Fixes: 379d7ac7 ("phy: mdio-thunder: Add driver for Cavium Thunder SoC MDIO buses.")
Cc: David Daney <david.daney@cavium.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>
Acked-by: NDavid Daney <david.daney@cavium.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add lan78xx_get_stats64 of ndo_get_stats64 to report
all statistics counters including errors from HW statistics.

Read from HW when auto suspend is disabled, use saved counter when
auto suspend is enabled because periodic call to ndo_get_stats64
prevents USB auto suspend.
Signed-off-by: NWoojung Huh <woojung.huh@microchip.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Update to handle statistics counter rollover.
Check statistics counter periodically and compensate it when
counter value rolls over at max (20 or 32bits).

Simple mechanism adjusts monitoring timer to allow USB auto suspend.
Signed-off-by: NWoojung Huh <woojung.huh@microchip.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Convert the dma transfers to be dmaengine based, now pxa has a dmaengine
slave driver. This makes this driver a bit more PXA agnostic.

The driver was only compile tested. The risk is quite small as no
current PXA platform I'm aware of is using smc911x driver.
Signed-off-by: NRobert Jarzmik <robert.jarzmik@free.fr>
Tested-by: NFabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove unnecessary set of flag IFF_MULTICAST, since ether_setup
already does this.
Signed-off-by: NZhang Shengju <zhangshengju@cmss.chinamobile.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove unnecessary set of flag IFF_MULTICAST, since ether_setup
already does this.
Signed-off-by: NZhang Shengju <zhangshengju@cmss.chinamobile.com>
Reviewed-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NAndy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There were two issues here:
1) dma_mapping_error() return true/false but we want to return -ENOMEM
2) If dmaengine_prep_slave_sg() failed then "err" wasn't set but
   presumably that should be -ENOMEM as well.

I changed the success path to "return 0;" instead of "return ret;" for
clarity.

Fixes: 94fe8c68 ('ks8842: Support DMA when accessed via timberdale')
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

MVNETA_BM has a dependency on MVNETA, so we can only select the former
if the latter is enabled. However, the code dependency is the reverse:
The mvneta module can call into the mvneta_bm module, so mvneta cannot
be a built-in if mvneta_bm is a module, or we get a link error:

drivers/net/built-in.o: In function `mvneta_remove':
drivers/net/ethernet/marvell/mvneta.c:4211: undefined reference to `mvneta_bm_pool_destroy'
drivers/net/built-in.o: In function `mvneta_bm_update_mtu':
drivers/net/ethernet/marvell/mvneta.c:1034: undefined reference to `mvneta_bm_bufs_free'

This avoids the problem by further clarifying the dependency so that
MVNETA_BM is a silent Kconfig option that gets turned on by the
new MVNETA_BM_ENABLE option. This way both the core HWBM module and
the MVNETA_BM code are always built-in when needed.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Fixes: dc35a10f ("net: mvneta: bm: add support for hardware buffer management")
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

  Checkpatch updates for sunvnet.c and sunvnet_common.c.
Signed-off-by: NAaron Young <aaron.young@oracle.com>
Signed-off-by: NRashmi Narasimhan <rashmi.narasimhan@oracle.com>
Reviewed-by: NSowmini Varadhan <sowmini.varadhan@oracle.com>
Reviewed-by: NAlexandre Chartre <Alexandre.Chartre@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

  Add ldmvsw.c driver

  Details:

  The ldmvsw driver very closely follows the sunvnet.c code and makes
  use of the sunvnet_common.c code for core functionality.

  A significant difference between sunvnet and ldmvsw driver is
  sunvnet creates a network interface for each vnet-port *parent*
  node in the MD while the ldmvsw driver creates a network interface
  for every vsw-port node in the Machine Description (MD).
  Therefore the netdev_priv() for sunvnet is a vnet structure while
  the netdev_priv() for ldmvsw is a vnet_port structure.

  Vnet_port structures allocated by ldmvsw have the vsw bit set.
  When finding the net_device associated with a port, the common code keys
  off this bit to use either the net_device found in the vnet_port or the
  net_device in the vnet structure (see the VNET_PORT_TO_NET_DEVICE() macro in
  sunvnet_common.h). This scheme allows the common code to work with
  both drivers with minimal changes.

  Similar to Xen, network interfaces created by the ldmvsw driver will always
  have a HW Addr (i.e. mac address) of FE:FF:FF:FF:FF:FF and each will be
  assigned the devname "vif<cfg_handle>.<port_id>" - where <cfg_handle> and
  <port_id> are a unique handle/port pair assigned to the associated
  vsw-port node in the MD.
Signed-off-by: NAaron Young <aaron.young@oracle.com>
Signed-off-by: NRashmi Narasimhan <rashmi.narasimhan@oracle.com>
Reviewed-by: NSowmini Varadhan <sowmini.varadhan@oracle.com>
Reviewed-by: NAlexandre Chartre <Alexandre.Chartre@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

  Modify sunvnet common code and data structures to be compatible
  with both sunvnet and ldmvsw drivers.

  Details:

  Sunvnet operates on "vnet-port" nodes which appear in the Machine
  Description (MD) in a guest domain. Ldmvsw operates on "vsw-port"
  nodes which appear in the MD of a service domain.

  A difference between the sunvnet driver and the ldmvsw driver is
  the sunvnet driver creates a network interface (i.e. a struct net_device)
  for every vnet-port *parent* "network" node. Several vnet-ports may appear
  under this common parent network node - each corresponding to a common parent
  network interface.  Conversely, since bridge/vswitch software will need
  to interface with every vsw-port in a system, the ldmvsw driver creates
  a network interface (i.e. a struct net_device) for every vsw-port - not
  every parent node as with sunvnet.  This difference required some special
  handling in the common code as explained below.

  There are 2 key data structures used by the sunvnet and ldmvsw drivers
  (which are now found in sunvnet_common.h):

  1. struct vnet_port
     This structure represents a vnet-port node in sunvnet and a vsw-port
     in the ldmvsw driver.

  2. struct vnet
     This structure represents a parent "network" node in sunvnet and a parent
     "virtual-network-switch" node in ldmvsw.

  Since the sunvnet driver allocates a net_device for every parent "network"
  node, a net_device member appears in the struct vnet. Since the ldmvsw
  driver allocates a net_device for every port, a net_device member was
  added to the vnet_port. The common code distinguishes which structure
  net_device member to use by checking a 'vsw' bit that was added to the
  vnet_port structure. See the VNET_PORT_TO_NET_DEVICE() marco in
  sunvnet_common.h.

  The netdev_priv() in sunvnet is allocated as a vnet. The netdev_priv()
  in ldmvsw is a vnet_port. Therefore, any place in the common code
  where a netdev_priv() call was made, a wrapper function was implemented
  in each driver to first get the vnet and/or vnet_port (in a driver
  specific way) and pass them as newly added parameters to the common
  functions (see wrapper funcs: vnet_set_rx_mode() and vnet_poll_controller()).
  Since these wrapper functions call __tx_port_find(), __tx_port_find() was
  moved from the common code back into sunvnet.c. Note - ldmvsw.c does not
  require this function.

  These changes also required that port_is_up() be made
  into a common function and thus it was given a _common suffix and
  exported like the other common functions.

  A wrapper function was also added for vnet_start_xmit_common() to pass a
  driver-specific function arg to return the port associated with a given
  struct sk_buff and struct net_device. This was required because
  vnet_start_xmit_common() grabs a lock prior to getting the associated
  port. Using a function pointer arg allowed the code to work unchanged
  without risking changes to the non-trivial locking logic in
  vnet_start_xmit_common().
Signed-off-by: NAaron Young <aaron.young@oracle.com>
Signed-off-by: NRashmi Narasimhan <rashmi.narasimhan@oracle.com>
Reviewed-by: NSowmini Varadhan <sowmini.varadhan@oracle.com>
Reviewed-by: NAlexandre Chartre <Alexandre.Chartre@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

  Split sunvnet.c into sunvnet.c and sunvnet_common.c.

  Details:

  Since the sunvnet and ldmvsw drivers will both use common sunvnet code,
  move the functions (and support functions) anticipated to be common code
  from sunvnet.c to sunvnet_common.c. Similarly, sunvnet.h was renamed to
  sunvnet_common.h. The sunvnet_common.c code will be compiled into the
  kernel and act as a library of functions that are linked by either
  (or both) drivers when loaded.

  Function names for external functions in sunvnet_common.c (to be
  called by both the sunvnet and ldmvsw drivers) were tagged with a "_common"
  suffix to clearly designate them as common functions.

  No functional changes as of yet... just moved code verbatim to the new
  sunvnet_common.c/h files.

  Makefile/Kconfig support added to build sunvnet_common.c file. The code
  is included in the kernel if SUN_LDOMS is defined/selected.

  NOTE - per the SubmittingPatches documentation, since the code was just
  moved from one file another, the code was NOT checkpatch'd in this commit
  to aid in review.
Signed-off-by: NAaron Young <aaron.young@oracle.com>
Signed-off-by: NRashmi Narasimhan <rashmi.narasimhan@oracle.com>
Reviewed-by: NSowmini Varadhan <sowmini.varadhan@oracle.com>
Reviewed-by: NAlexandre Chartre <Alexandre.Chartre@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Not all adapters have FC-NPIV configured. If bnx2fc is used with such an
adapter, driver would read irrelevant data from the the nvram and log
"FC-NPIV table with bad length..." In system logs.

Simply accept that reading '0' as the feature offset in nvram indicates
the feature isn't there and return.
Reported-by: NAndrew Patterson <andrew.patterson@hpe.com>
Signed-off-by: NYuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The result value is overwritten by a return value of
ravb_ptp_interrupt().
Signed-off-by: NYoshihiro Kaneko <ykaneko0929@gmail.com>
Acked-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When running small packets [length < 256 bytes] traffic, packets were
being dropped due to invalid data in those packets which were
delivered by the driver upto the stack. Using pci_dma_sync_single_for_cpu
ensures copying latest and updated data into skb from the receive buffer.
Signed-off-by: NSony Chacko <sony.chacko@qlogic.com>
Signed-off-by: NManish Chopra <manish.chopra@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If a dt mdio entry has been added least assume that we wont
search for phys attached. The DT and of_mdiobus_register already do
this. This stops DSA phys being found and phys created for them, as
this is handled by the DSA driver.
Signed-off-by: NPhil Reid <preid@electromag.com.au>
Acked-by: NGiuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There was a missing unlock on the error path.

Fixes: 656e7052 ('net-next: mediatek: add support for MT7623 ethernet')
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Acked-by: NJohn Crispin <blogic@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

of_phy_connect() returns NULL on error, it never returns error pointers.

Fixes: 656e7052 ('net-next: mediatek: add support for MT7623 ethernet')
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Acked-by: NJohn Crispin <blogic@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

No code changes.  Since OCTEON is a Cavium product, move the driver to
the vendor directory to unclutter things a bit.
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Added support for Gemalto's Cinterion PHxx WWAN interfaces
by adding QMI_FIXED_INTF with Cinterion's VID and PID.

PHxx can have:
2 RmNet Interfaces (PID 0x0082) or
1 RmNet + 1 USB Audio interface (PID 0x0083).
Signed-off-by: NHans-Christoph Schemmel <hans-christoph.schemmel@gemalto.com>
Acked-by: NBjørn Mork <bjorn@mork.no>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

vlan drivers lack proper propagation of gso_max_segs from
lower device.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Some callers of strtobool() were passing a pointer to unterminated
strings.  In preparation of adding multi-character processing to
kstrtobool(), update the callers to not pass single-character pointers,
and switch to using the new kstrtobool_from_user() helper where
possible.
Signed-off-by: NKees Cook <keescook@chromium.org>
Cc: Amitkumar Karwar <akarwar@marvell.com>
Cc: Nishant Sarmukadam <nishants@marvell.com>
Cc: Kalle Valo <kvalo@codeaurora.org>
Cc: Steve French <sfrench@samba.org>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Joe Perches <joe@perches.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The success of CMA allocation largely depends on the success of
migration and key factor of it is page reference count.  Until now, page
reference is manipulated by direct calling atomic functions so we cannot
follow up who and where manipulate it.  Then, it is hard to find actual
reason of CMA allocation failure.  CMA allocation should be guaranteed
to succeed so finding offending place is really important.

In this patch, call sites where page reference is manipulated are
converted to introduced wrapper function.  This is preparation step to
add tracepoint to each page reference manipulation function.  With this
facility, we can easily find reason of CMA allocation failure.  There is
no functional change in this patch.

In addition, this patch also converts reference read sites.  It will
help a second step that renames page._count to something else and
prevents later attempt to direct access to it (Suggested by Andrew).
Signed-off-by: NJoonsoo Kim <iamjoonsoo.kim@lge.com>
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Acked-by: NVlastimil Babka <vbabka@suse.cz>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It is possible, although unlikely, that probing will find the
phy_device for the first LMAC of a thunder BGX device, but then need
to fail with -EPROBE_DEFER on a subsequent LMAC.  In this case, we
need to call put_device() on each of the phy_devices that were
obtained, but will be unused due to returning -EPROBE_DEFER.

Also, since we can break out of the probing loop early, we need to
explicitly call of_node_put() outside of the loop.
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>