This patch fixes a critical reset issue that resulting to the server
reboot when an Admin changes VF configuration on the host, for example
changing VF to Trusted/non_Trusted mode, the PF driver send reset
notification to AVF driver while also continue with reset flow. However,
AVF driver schedule another reset due to notification, which causes two
concurrent reset going on, and trigger lock up in the FW, with AQ call to
delete VSI.
Signed-off-by: NAkeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The total number of queues available on the device is divided between
multiple physical functions (PF) in the firmware and provided to the
driver when it gets function capabilities from the firmware. Thus
each PF knows how many Tx/Rx queues it has. These queues are then
doled out to different VSIs (for LAN traffic, SR-IOV VF traffic, etc.)

To track usage of these queues at the PF level, the driver uses two
bitmaps avail_txqs and avail_rxqs. At the VSI level (i.e. struct ice_vsi
instances) the driver uses two arrays txq_map and rxq_map, to track
ownership of VSIs' queues in avail_txqs and avail_rxqs respectively.

The aforementioned bitmaps and arrays should be allocated dynamically,
because the number of queues supported by a PF is only available once
function capabilities have been queried. The current static allocation
consumes way more memory than required.

This patch removes the DECLARE_BITMAP for avail_txqs and avail_rxqs
and instead uses bitmap_zalloc to allocate the bitmaps during init.
Similarly txq_map and rxq_map are now allocated in ice_vsi_alloc_arrays.
As a result ICE_MAX_TXQS and ICE_MAX_RXQS defines are no longer needed.
Also as txq_map and rxq_map are now allocated and freed, some code
reordering was required in ice_vsi_rebuild for correct functioning.
Signed-off-by: NAnirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The VF driver can call VIRTCHNL_OP_[ENABLE|DISABLE]_QUEUES separately
for each queue. Add support for virtchnl_queue_select.[tx|rx]_queues
bitmap which is used to indicate which queues to enable and disable.

Add tracing of VF Tx/Rx per queue enable state to avoid enabling enabled
queues and disabling disabled queues. Add total queues enabled count and
clear ICE_VF_STATE_QS_ENA when count is zero.
Signed-off-by: NPaul Greenwalt <paul.greenwalt@intel.com>
Signed-off-by: NPeng Huang <peng.huang@intel.com>
Signed-off-by: NTony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Refactor the queue handling functions that are going through queue
arrays in a way that the logic done for a single queue is pulled out and
it will be called for each ring when traversing ring array. This implies
that when disabling Tx rings we won't fill up q_ids, q_teids and
q_handles arrays.  Drop also 'offset' parameter; the value from vsi's
txq_map is stored in ring->reg_idx and that drops the need for mentioned
parameter. Introduce the ice_vsi_cfg_txq, ice_vsi_stop_tx_ring and
ice_vsi_ctrl_rx_ring that are the functions with pulled out logic.

There's several Tx queue meta data (q_id, q_handle, q_teid and other)
that need to be set up during Tx queue disablement, so let's as well add
a helper structure that wraps it up and a function that will be filling
it up.
Signed-off-by: NMaciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: NTony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The loop counter of a for-loop is a u8 however this is being compared
to an int upper bound and this can lead to an infinite loop if the
upper bound is greater than 255 since the loop counter will wrap back
to zero. Fix this potential issue by making the loop counter an int.

Addresses-Coverity: ("Infinite loop")
Fixes: c7aeb4d1 ("ice: Disable VFs until reset is completed")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

ice_is_tc_ena is used to check whether a given traffic class is
enabled. Because there are only 8 traffic classes, the function took
a u8 bitmap. This causes problems because it is cast to an unsigned
long causing a static analysis warning regarding Out-of-bounds read.

Fix this by simply updating ice_is_tc_ena to take an unsigned long.
Passing a u8 to this function should implicitly convert the value.
Signed-off-by: NJacob Keller <jacob.e.keller@intel.com>
Signed-off-by: NTony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Check num_queue_pairs to avoid access to unallocated field of
vsi->tx_rings/vsi->rx_rings. Without this validation we can set
vsi->alloc_txq/vsi->alloc_rxq to value smaller than ICE_MAX_BASE_QS_PER_VF
and send this command with num_queue_pairs greater than
vsi->alloc_txq/vsi->alloc_rxq. This lead to access to unallocated memory.

In VF vsi alloc_txq and alloc_rxq should be the same. Get minimum
because looks more readable.

Also add validation for ring_len param. It should be greater than 32 and
be multiple of 32. Incorrect value leads to hang traffic on PF.
Signed-off-by: NMichal Swiatkowski <michal.swiatkowski@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

In case of MDD events on VF, don't clog kernel log with unlimited VF MDD
events message "VF 0 has had 1018 MDD events since last boot" - limit
events log message to 30, based on the observation in some experimentation
with sending malicious packet once, and number of events reported before
device stopped observing MDD events.

Also removed defunct macro "ICE_DFLT_NUM_MDD_EVENTS_ALLOWED" for tracking
number of MDD events allowed before disabling the interface...
Signed-off-by: NAkeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

When a VSI is accessed inside the ice_for_each_vsi macro in the rebuild
path (ice_vsi_rebuild_all() and ice_vsi_replay_all()), it is referred to
as pf->vsi[i]. Introduce local variables to improve readability.
Signed-off-by: NKrzysztof Kazimierczak <krzysztof.kazimierczak@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Add some verbose debugging for dyndbg to help us when
we are having issues with link and/or PHY.

While there, shorten some strings used by locals that
were causing long line wrapping.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NTony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

1. ndo_open and ndo_stop are implemented by ice_open and ice_stop
   respectively. When enabling/disabling VSIs, just call
   ice_open/ice_stop instead of ndo_open/ndo_stop.

2. Rework logic around rtnl_lock/rtnl_unlock

3. In ice_ena_vsi, remove an unnecessary stack variable and return
   0 instead of err when __ICE_NEEDS_RESTART is not set.
Signed-off-by: NAnirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

There was a bug in the previous code which never traverses all the
children to get the first node of the requested layer. Add a sibling
head pointer to point the first node of each layer per TC. This helps
traverse easier and quicker and also removes the recursion.
Signed-off-by: NVictor Raj <victor.raj@intel.com>
Signed-off-by: NTony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch fixes the issue where port and PFC statistics counters are
incrementing at the wrong port with 4x25G cards.
Read the GLPRT port registers using lport parameter instead of pf_id to
update the statistics otherwise the pf_ids are flipped for ports 2 and 3
when read from the HW register PF_FUNC_RID and this is expected as per
hardware specification.
Signed-off-by: NUsha Ketineni <usha.k.ketineni@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Add MODULE_FIRMWARE entries for AMDA0058 boards.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: NDirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Move the call to dma_sync_single_for_cpu after calling napi_alloc_skb.
This avoids calling dma_sync_single_for_cpu w/o handing control back
to device if the memory allocation should fail.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Vlad Buslov says:

====================
Refactor cls hardware offload API to support rtnl-independent drivers

Currently, all cls API hardware offloads driver callbacks require caller
to hold rtnl lock when calling them. This patch set introduces new API
that allows drivers to register callbacks that are not dependent on rtnl
lock and unlocked classifiers to offload filters without obtaining rtnl
lock first, which is intended to allow offloading tc rules in parallel.

Recently, new rtnl registration flag RTNL_FLAG_DOIT_UNLOCKED was added.
TC rule update handlers (RTM_NEWTFILTER, RTM_DELTFILTER, etc.) are
already registered with this flag and only take rtnl lock when qdisc or
classifier requires it. Classifiers can indicate that their ops
callbacks don't require caller to hold rtnl lock by setting the
TCF_PROTO_OPS_DOIT_UNLOCKED flag. Unlocked implementation of flower
classifier is now upstreamed. However, this implementation still obtains
rtnl lock before calling hardware offloads API.

Implement following cls API changes:

- Introduce new "unlocked_driver_cb" flag to struct flow_block_offload
  to allow registering and unregistering block hardware offload
  callbacks that do not require caller to hold rtnl lock. Drivers that
  doesn't require users of its tc offload callbacks to hold rtnl lock
  sets the flag to true on block bind/unbind. Internally tcf_block is
  extended with additional lockeddevcnt counter that is used to count
  number of devices that require rtnl lock that block is bound to. When
  this counter is zero, tc_setup_cb_*() functions execute callbacks
  without obtaining rtnl lock.

- Extend cls API single hardware rule update tc_setup_cb_call() function
  with tc_setup_cb_add(), tc_setup_cb_replace(), tc_setup_cb_destroy()
  and tc_setup_cb_reoffload() functions. These new APIs are needed to
  move management of block offload counter, filter in hardware counter
  and flag from classifier implementations to cls API, which is now
  responsible for managing them in concurrency-safe manner. Access to
  cb_list from callback execution code is synchronized by obtaining new
  'cb_lock' rw_semaphore in read mode, which allows executing callbacks
  in parallel, but excludes any modifications of data from
  register/unregister code. tcf_block offloads counter type is changed
  to atomic integer to allow updating the counter concurrently.

- Extend classifier ops with new ops->hw_add() and ops->hw_del()
  callbacks which are used to notify unlocked classifiers when filter is
  successfully added or deleted to hardware without releasing cb_lock.
  This is necessary to update classifier state atomically with callback
  list traversal and updating of all relevant counters and allows
  unlocked classifiers to synchronize with concurrent reoffload without
  requiring any changes to driver callback API implementations.

New tc flow_action infrastructure is also modified to allow its user to
execute without rtnl lock protection. Function tc_setup_flow_action() is
modified to conditionally obtain rtnl lock before accessing action
state. Action data that is accessed by reference is either copied or
reference counted to prevent concurrent action overwrite from
deallocating it. New function tc_cleanup_flow_action() is introduced to
cleanup/release all such data obtained by tc_setup_flow_action().

Flower classifier (only unlocked classifier at the moment) is modified
to use new cls hardware offloads API and no longer obtains rtnl lock
before calling it.
====================
Acked-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Don't manually take rtnl lock in flower classifier before calling cls
hardware offloads API. Instead, pass rtnl lock status via 'rtnl_held'
parameter.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to remove dependency on rtnl lock, modify tc_setup_flow_action()
to copy tunnel info, instead of just saving pointer to tunnel_key action
tunnel info. This is necessary to prevent concurrent action overwrite from
releasing tunnel info while it is being used by rtnl-unlocked driver.

Implement helper tcf_tunnel_info_copy() that is used to copy tunnel info
with all its options to dynamically allocated memory block. Modify
tc_cleanup_flow_action() to free dynamically allocated tunnel info.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to remove dependency on rtnl lock when calling hardware offload
API, take reference to action mirred dev when initializing flow_action
structure in tc_setup_flow_action(). Implement function
tc_cleanup_flow_action(), use it to release the device after hardware
offload API is done using it.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to allow using new flow_action infrastructure from unlocked
classifiers, modify tc_setup_flow_action() to accept new 'rtnl_held'
argument. Take rtnl lock before accessing tc_action data. This is necessary
to protect from concurrent action replace.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to remove dependency on rtnl lock from offloads code of
classifiers, take rtnl lock conditionally before executing driver
callbacks. Only obtain rtnl lock if block is bound to devices that require
it.

Block bind/unbind code is rtnl-locked and obtains block->cb_lock while
holding rtnl lock. Obtain locks in same order in tc_setup_cb_*() functions
to prevent deadlock.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Extend struct flow_block_offload with "unlocked_driver_cb" flag to allow
registering and unregistering block hardware offload callbacks that do not
require caller to hold rtnl lock. Extend tcf_block with additional
lockeddevcnt counter that is incremented for each non-unlocked driver
callback attached to device. This counter is necessary to conditionally
obtain rtnl lock before calling hardware callbacks in following patches.

Register mlx5 tc block offload callbacks as "unlocked".
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

To remove dependency on rtnl lock, extend classifier ops with new
ops->hw_add() and ops->hw_del() callbacks. Call them from cls API while
holding cb_lock every time filter if successfully added to or deleted from
hardware.

Implement the new API in flower classifier. Use it to manage hw_filters
list under cb_lock protection, instead of relying on rtnl lock to
synchronize with concurrent fl_reoffload() call.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Without rtnl lock protection filters can no longer safely manage block
offloads counter themselves. Refactor cls API to protect block offloadcnt
with tcf_block->cb_lock that is already used to protect driver callback
list and nooffloaddevcnt counter. The counter can be modified by concurrent
tasks by new functions that execute block callbacks (which is safe with
previous patch that changed its type to atomic_t), however, block
bind/unbind code that checks the counter value takes cb_lock in write mode
to exclude any concurrent modifications. This approach prevents race
conditions between bind/unbind and callback execution code but allows for
concurrency for tc rule update path.

Move block offload counter, filter in hardware counter and filter flags
management from classifiers into cls hardware offloads API. Make functions
tcf_block_offload_{inc|dec}() and tc_cls_offload_cnt_update() to be cls API
private. Implement following new cls API to be used instead:

  tc_setup_cb_add() - non-destructive filter add. If filter that wasn't
  already in hardware is successfully offloaded, increment block offloads
  counter, set filter in hardware counter and flag. On failure, previously
  offloaded filter is considered to be intact and offloads counter is not
  decremented.

  tc_setup_cb_replace() - destructive filter replace. Release existing
  filter block offload counter and reset its in hardware counter and flag.
  Set new filter in hardware counter and flag. On failure, previously
  offloaded filter is considered to be destroyed and offload counter is
  decremented.

  tc_setup_cb_destroy() - filter destroy. Unconditionally decrement block
  offloads counter.

  tc_setup_cb_reoffload() - reoffload filter to single cb. Execute cb() and
  call tc_cls_offload_cnt_update() if cb() didn't return an error.

Refactor all offload-capable classifiers to atomically offload filters to
hardware, change block offload counter, and set filter in hardware counter
and flag by means of the new cls API functions.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As a preparation for running proto ops functions without rtnl lock, change
offload counter type to atomic. This is necessary to allow updating the
counter by multiple concurrent users when offloading filters to hardware
from unlocked classifiers.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to remove dependency on rtnl lock, extend tcf_block with 'cb_lock'
rwsem and use it to protect flow_block->cb_list and related counters from
concurrent modification. The lock is taken in read mode for read-only
traversal of cb_list in tc_setup_cb_call() and write mode in all other
cases. This approach ensures that:

- cb_list is not changed concurrently while filters is being offloaded on
  block.

- block->nooffloaddevcnt is checked while holding the lock in read mode,
  but is only changed by bind/unbind code when holding the cb_lock in write
  mode to prevent concurrent modification.
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/net/ethernet/cirrus/cs89x0.c: In function 'cs89x0_platform_probe':
drivers/net/ethernet/cirrus/cs89x0.c:1847:20: warning:
 variable 'lp' set but not used [-Wunused-but-set-variable]
Reported-by: NHulk Robot <hulkci@huawei.com>
Fixes: 6751edeb ("cirrus: cs89x0: Use managed interfaces")
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This reverts commit ee641b0c.

Actually it is not clear whether this register read is not
needed for it's HW side effects or not.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fixes gcc '-Wunused-but-set-variable' warning:
drivers/net/ethernet/mediatek/mtk_eth_soc.c: In function mtk_handle_irq:
drivers/net/ethernet/mediatek/mtk_eth_soc.c:1951:6: warning: variable status set but not used [-Wunused-but-set-variable]

Fixes: 296c9120 ("net: ethernet: mediatek: Add MT7628/88 SoC support")
Signed-off-by: NMao Wenan <maowenan@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

SFPs can report two different power values, the transmit power and the
receive power. Add labels to make it clear which is which. Also add
labels to the other sensors, VCC power supply, bias and module
temperature.

sensors(1) now shows:

sff2-isa-0000
Adapter: ISA adapter
VCC:          +3.23 V
temperature:  +33.4 C
TX_power:    276.00 uW
RX_power:     20.00 uW
bias:         +0.01 A
Signed-off-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
100GbE Intel Wired LAN Driver Updates 2019-08-23

This series contains updates to ice driver only.

Dave adds logic for the necessary bits to be set in the VSI context for
the PF_VSI and the TX_descriptors for control packets egressing the
PF_VSI.  Updated the logic to detect both DCBx and LLDP states in the
firmware engine to account for situations where DCBx is enabled and LLDP
is disabled.  Fixed the driver to treat the DCBx state of "NOT_STARTED"
as a valid state and should not assume "is_fw_lldp" true automatically.
Since "enable-fw-lldp" flag was confusing and cumbersome, change the
flag to "fw-lldp-agent" with a value of on or off to help clarify
whether the LLDP agent is running or not.

Brett fixes an issue where synchronize_irq() was being called from the
host of VF's, which should not be done.

Michal fixed an issue when rebuilding the DCBx configuration while in
IEEE mode versus CEE mode, so add a check before copying the
configuration value to ensure we are only in CEE mode.

Jake fixes the PF to reject any VF request to setup head writeback since
the support has been deprecated.

Mitch adds an additional check to ensure the VF is active before sending
out an error message that a message was unable to be sent to a
particular VF.

Chinh updates the driver to use "topology" mode when checking the PHY
for status, since this mode provides us the current module type that is
available.  Fixes the driver from clearing the auto_fec_enable bit which
was blocking a user from forcing non-spec compliant FEC configurations.

Amruth does a refactor on the code to first check, then assign in the
virtual channel space.

Bruce updates the driver to actually update the stats when a user runs
the ethtool command 'ethtool -S <iface>' instead of providing a snapshot
of the stats that maybe from a second ago.

Akeem fixes up the adding/removing of VSI MAC filters for VFs, so that
VFs cannot add/remove a filter from another VSI.  We now track the
number of filters added right from when the VF resources get allocated
and won't get into MAC filter mis-match issue in the switch.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Julian Wiedmann says:

====================
s390/qeth: updates 2019-08-23

please apply one more round of qeth patches. These implement support for
a bunch of TX-related features - namely TX NAPI, BQL and xmit_more.

Note that this includes two qdio patches which lay the necessary
groundwork, and have been acked by Vasily.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

IQD devices offer limited support for bulking: all frames in a TX buffer
need to have the same target. qeth_iqd_may_bulk() implements this
constraint, and allows us to defer the TX doorbell until
(a) the buffer is full (since each buffer needs its own doorbell), or
(b) the entire TX queue is full, or
(b) we reached the BQL limit.
Signed-off-by: NJulian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Each TX buffer may contain multiple skbs. So just accumulate the sent
byte count in the buffer struct, and later use the same count when
completing the buffer.
Signed-off-by: NJulian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This allows the stack to bulk-free our TX-completed skbs.
Signed-off-by: NJulian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Due to their large MTU and potentially low utilization of TX buffers,
IQD devices in particular require fast TX recycling. This makes them
a prime candidate for a TX NAPI path in qeth.

qeth_tx_poll() uses the recently introduced qdio_inspect_queue() helper
to poll the TX queue for completed buffers. To avoid hogging the CPU for
too long, we yield to the stack after completing an entire queue's worth
of buffers.
While IQD is expected to transfer its buffers synchronously (and thus
doesn't support TX interrupts), a timer covers for the odd case where a
TX buffer doesn't complete synchronously. Currently this timer should
only ever fire for
(1) the mcast queue,
(2) the occasional race, where the NAPI poll code observes an update to
    queue->used_buffers while the TX doorbell hasn't been issued yet.
Signed-off-by: NJulian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This consolidates the SW statistics code, and improves it to
(1) account for the header overhead of each segment on a TSO skb,
(2) count dangling packets as in-error (during eg. shutdown), and
(3) only count offloads when the skb was successfully transmitted.

We also count each segment of an TSO skb as one packet - except for
tx_dropped, to be consistent with dev->tx_dropped.
Signed-off-by: NJulian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If a driver wants to use the new Output Queue poll code, then the qdio
layer must disable its internal Queue scanning. Let the driver select
this mode by passing a special scan_threshold of 0.

As the scan_threshold is the same for all Output Queues, also move it
into the main qdio_irq struct. This allows for fast opt-out checking, a
driver is expected to operate either _all_ or none of its Output Queues
in polling mode.
Signed-off-by: NJulian Wiedmann <jwi@linux.ibm.com>
Acked-by: NVasily Gorbik <gor@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

While commit d36deae7 ("qdio: extend API to allow polling") enhanced
the qdio layer so that drivers can poll their Input Queues, we don't
have the corresponding infrastructure for Output Queues yet.

Factor out a helper that scans a single QDIO Queue, so that qeth can
implement TX NAPI on top of it.
While doing so, remove the duplicated tracking of the next-to-scan index
(q->first_to_check vs q->first_to_kick) in this code path.

qdio_handle_aobs() needs to move slightly upwards in the code hierarchy,
so that it's still called from the polling path.
Signed-off-by: NJulian Wiedmann <jwi@linux.ibm.com>
Acked-by: NVasily Gorbik <gor@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Recent commit added logic to determine the appropriate statistics block
size to allocate and the size is stored in bp->hw_ring_stats_size.  But
if the firmware spec is older than 1.6.0, it is 0 and not initialized.
This causes the allocation to fail with size 0 and bnxt_open() to
abort.  Fix it by always initializing bp->hw_ring_stats_size to the
legacy default size value.

Fixes: 4e748506 ("bnxt_en: Allocate the larger per-ring statistics block for 57500 chips.")
Reported-by: NJonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: NMichael Chan <michael.chan@broadcom.com>
Tested-by: NJonathan Lemon <jonathan.lemon@gmail.com>
Acked-by: NJonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>