Total e-switch vports are already stored in mlx5_eswitch total_vports.
Avoid copy of it in nvports and reuse existing total_vports calculation.
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

When eswitch is enabled, VFs might not be enabled. Hence, consider
maximum number of VFs.
This further closes the gap between handling VF vports between ECPF and
PF.

Fixes: ea2128fd ("net/mlx5: E-switch, Reduce dependency on num_vfs during mode set")
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Add function ID to reclaim pages debug log for better user visibility.
Signed-off-by: Avihu Hagag <avihuh@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Per page request event, FW request to allocated or release pages for a
single function. Driver maintains FW pages object per function, so there
is no need to hold one global page data-base. Instead, have a page
data-base per function, which will improve performance release flow in all
cases, especially for "release all pages".

As the range of function IDs is large and not sequential, use xarray to
store a per function ID page data-base, where the function ID is the key.

Upon first allocation of a page to a function ID, create the page
data-base per function. This data-base will be released only at pagealloc
mechanism cleanup.

NIC: ConnectX-4 Lx
CPU: Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz
Test case: 32 VFs, measure release pages on one VF as part of FLR
Before: 0.021 Sec
After:  0.014 Sec

The improvement depends on amount of VFs and memory utilization
by them. Time measurements above were taken from idle system.
Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Replace the existing /* fall through */ comments and its variants with
the new pseudo-keyword macro fallthrough[1].

[1] https://www.kernel.org/doc/html/v5.7/process/deprecated.html?highlight=fallthrough#implicit-switch-case-fall-throughSigned-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Tony Nguyen says:

====================
1GbE Intel Wired LAN Driver Updates 2020-07-27

This series contains updates to igc driver only.

Sasha cleans up double definitions, unneeded and non applicable
registers, and removes unused fields in structs. Ensures the Receive
Descriptor Minimum Threshold Count is cleared and fixes a static checker
error.

v2: Remove fields from hw_stats in patches that removed their uses.
Reworded patch descriptions for patches 1, 2, and 4.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently n_rq_elems is being assigned to params.elem_size instead of the
field params.num_elems.  Coverity is detecting this as a double assingment
to params.elem_size and reporting this as an usused value on the first
assignment.  Fix this.

Addresses-Coverity: ("Unused value")
Fixes: b6db3f71 ("qed: simplify chain allocation with init params struct")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Edward Cree says:

====================
sfc: driver for EF100 family NICs, part 1

EF100 is a new NIC architecture under development at Xilinx, based
 partly on existing Solarflare technology.  As many of the hardware
 interfaces resemble EF10, support is implemented within the 'sfc'
 driver, which previous patch series "commonised" for this purpose.

In order to maintain bisectability while splitting into patches of a
 reasonable size, I had to do a certain amount of back-and-forth with
 stubs for things that the common code may try to call, mainly because
 we can't do them until we've set up MCDI, but we can't set up MCDI
 without probing the event queues, at which point a lot of the common
 machinery becomes reachable from event handlers.
Consequently, this first series doesn't get as far as actually sending
 and receiving packets.  I have a second series ready to follow it
 which implements the datapath (and a few other things like ethtool).

Changes from v4:
 * Fix build on CONFIG_RETPOLINE=n by using plain prototypes instead
   of INDIRECT_CALLABLE_DECLARE.

Changes from v3:
 * combine both drivers (sfc_ef100 and sfc) into a single module, to
   make non-modular builds work.  Patch #4 now adds a few indirections
   to support this; the ones in the RX and TX path use indirect-call-
   wrappers to minimise the performance impact.

Changes from v2:
 * remove MODULE_VERSION.
 * call efx_destroy_reset_workqueue() from ef100_exit_module().
 * correct uint32_ts to u32s.  While I was at it, I fixed a bunch of
   other style issues in the function-control-window code.
All in patch #4.

Changes from v1:
 * kernel test robot spotted a link error when sfc_ef100 was built
   without mdio.  It turns out the thing we were trying to link to
   was a bogus thing to do on anything but Falcon, so new patch #1
   removes it from this driver.
 * fix undeclared symbols in patch #4 by shuffling around prototypes
   and #includes and adding 'static' where appropriate.
 * fix uninitialised variable 'rc2' in patch #7.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In ef100_reset(), make the MCDI call to do the reset.
Also, do a reset at start-of-day during probe, to put the function in
 a clean state.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

MC_CMD_GET_CAPABILITIES now has a third word of flags; extend the
 efx_has_cap() machinery to cover it.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently RX and TX-completion events are unhandled, as neither the RX
 nor the TX path has been implemented yet.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Channels are probed, but actual event handling is still stubbed out.

Stub implementation of check_caps is needed because ptp.c will call into
 it from efx_ptp_use_mac_tx_timestamps() to decide if it wants TXQs.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We handle everything ourselves in ef100_reset(), rather than relying on
 the generic down/up routines.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We can't actually do the MCDI to probe it fully until we have working
 MCDI, which comes later, but we need efx->phy_data to be allocated so
 that when we get MCDI events the link-state change handler doesn't
 NULL-dereference.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We don't actually do the efx_mcdi_reset() because we don't have MCDI yet.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

No TX or RX path, no MCDI, not even an ifup/down handler.
Besides stubs, the bulk of the patch deals with reading the Xilinx
 extended PCIe capability, which tells us where to find our BAR.

Though in the same module, EF100 has its own struct pci_driver,
 which is named sfc_ef100.

A small number of additional nic_type methods are added; those in the
 TX (tx_enqueue) and RX (rx_packet) paths are called through indirect
 call wrappers to minimise the performance impact.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

EF100 adds a few new valid addresses for efx_writed_page(), as well as
 a Function Control Window in the BAR whose location is variable.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

An MDIO-based n-way restart does not make sense for any of the NICs
 supported by this driver, nor for the coming EF100.
Unlike on Falcon (which was already split off into a separate driver),
 the PHY on all of Siena, EF10 and EF100 is managed by MC firmware.
While Siena can talk to the PHY over MDIO, doing so for anything other
 than debugging purposes (mdio_mii_ioctl) is likely to confuse the
 firmware.
(According to the SFC firmware team, this support was originally added
 to the Siena driver early in the development of that product, before
 it was decided to have firmware manage the PHY.)
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Murali Karicheri says:

====================
Add PRP driver

This series is dependent on the following patches sent out to
netdev list. All (1-3) are already merged to net/master as of
sending this, but not on the net-next master branch. So need
to apply them to net-next before applying this series. v3 of
the iproute2 patches can be merged to work with this series
as there are no updates since then.

[1] https://marc.info/?l=linux-netdev&m=159526378131542&w=2
[2] https://marc.info/?l=linux-netdev&m=159499772225350&w=2
[3] https://marc.info/?l=linux-netdev&m=159499772425352&w=2

This series adds support for Parallel Redundancy Protocol (PRP)
in the Linux HSR driver as defined in IEC-62439-3. PRP Uses a
Redundancy Control Trailer (RCT) the format of which is
similar to HSR Tag. This is used for implementing redundancy.
RCT consists of 6 bytes similar to HSR tag and contain following
fields:-

- 16-bit sequence number (SeqNr);
- 4-bit LAN identifier (LanId);
- 12 bit frame size (LSDUsize);
- 16-bit suffix (PRPsuffix).

The PRPsuffix identifies PRP frames and distinguishes PRP frames
from other protocols that also append a trailer to their useful
data. The LSDUsize field allows the receiver to distinguish PRP
frames from random, nonredundant frames as an additional check.
LSDUsize is the size of the Ethernet payload inclusive of the
RCT. Sequence number along with LanId is used for duplicate
detection and discard.

PRP node is also known as Dual Attached Node (DAN-P) since it
is typically attached to two different LAN for redundancy.
DAN-P duplicates each of L2 frames and send it over the two
Ethernet links. Each outgoing frame is appended with RCT.
Unlike HSR, these are added to the end of L2 frame and will be
treated as pad by bridges and therefore would be work with
traditional bridges or switches, where as HSR wouldn't as Tag
is prefixed to the Ethenet frame. At the remote end, these are
received and the duplicate frame is discarded before the stripped
frame is send up the networking stack. Like HSR, PRP also sends
periodic Supervision frames to the network. These frames are
received and MAC address from the SV frames are populated in a
database called Node Table. The above functions are grouped into
a block called Link Redundancy Entity (LRE) in the IEC spec.

As there are many similarities between HSR and PRP protocols,
this patch re-uses the code from HSR driver to implement PRP
driver. As per feedback from the RFC series, the implementation
uses the existing HSR Netlink socket interface to create the
PRP interface by adding a new proto parameter to the ip link
command to identify the PRP protocol. iproute2 is enhanced to
implement this new parameter. The hsr_netlink.c is enhanced
to handle the new proto parameter. As suggested during the RFC
review, the driver introduced a proto_ops structure to hold
protocol specfic functions to handle HSR and PRP specific
function pointers and use them in the code based on the
protocol to handle protocol specific part differently in the
driver.

Please review this and provide me feedback so that I can work to
incorporate them and spin the next version if needed.

The patch was tested using two TI AM57x IDK boards for PRP which
are connected back to back over two CPSW Ethernet ports.

PRP Test setup
---------------

--------eth0             eth0 --------
|AM572x|----------------------|AM572x|
|      |----------------------|      |
--------eth1             eth1 --------

To build, enable CONFIG_HSR=y or m
make omap2plus_defconfig
make zImage; make modules; make dtbs
Copy the zImage and dtb files to the file system on SD card
and power on the AM572x boards.
This can be tested on any platforms with 2 Ethernet interfaces.
So will appreciate if you can give it a try and provide your
Tested-by.

Command to create PRP interface
-------------------------------
ifconfig eth0 0.0.0.0 down
ifconfig eth1 0.0.0.0 down
ifconfig eth0 hw ether 70:FF:76:1C:0E:8C
ifconfig eth1 hw ether 70:FF:76:1C:0E:8C
ifconfig eth0 up
ifconfig eth1 up
ip link add name prp0 type hsr slave1 eth0 slave2 eth1 supervision 45 proto 1
ifconfig prp0 192.168.2.10

ifconfig eth0 0.0.0.0 down
ifconfig eth1 0.0.0.0 down
ifconfig eth0 hw ether 70:FF:76:1C:0E:8D
ifconfig eth1 hw ether 70:FF:76:1C:0E:8D
ifconfig eth0 up
ifconfig eth1 up
ip link add name prp0 type hsr slave1 eth0 slave2 eth1 supervision 45 proto 1
ifconfig prp0 192.168.2.20

command to show node table
----------------------------
Ping the peer board after the prp0 interface is up.

The remote node (DAN-P) will be shown in the node table as below.

root@am57xx-evm:~# cat /sys/kernel/debug/hsr/prp0/node_table
Node Table entries for (PRP) device
MAC-Address-A,    MAC-Address-B,    time_in[A], time_in[B], Address-B port, SAN-A, SAN-B, DAN-P
70:ff:76:1c:0e:8c 00:00:00:00:00:00   ffffe83f,   ffffe83f,              0,     0,     0,     1

Try to capture the raw PRP frames at the eth0 interface as
tcpdump -i eth0 -xxx

Sample Supervision frames and ARP frames shown below.

==================================================================================
Successive Supervision frames captured with tcpdump (with RCT at the end):

03:43:29.500999 70:ff:76:1c:0e:8d (oui Unknown) > 01:15:4e:00:01:2d (oui Unknown), ethertype Unknown (0x88f
        0x0000:  0115 4e00 012d 70ff 761c 0e8d 88fb 0001
        0x0010:  7e0a 1406 70ff 761c 0e8d 0000 0000 0000
        0x0020:  0000 0000 0000 0000 0000 0000 0000 0000
        0x0030:  0000 0000 0000 0000 0000 0000 fc2b a034
        0x0040:  88fb

03:43:31.581025 70:ff:76:1c:0e:8d (oui Unknown) > 01:15:4e:00:01:2d (oui Unknown), ethertype Unknown (0x88f
        0x0000:  0115 4e00 012d 70ff 761c 0e8d 88fb 0001
        0x0010:  7e0b 1406 70ff 761c 0e8d 0000 0000 0000
        0x0020:  0000 0000 0000 0000 0000 0000 0000 0000
        0x0030:  0000 0000 0000 0000 0000 0000 fc2c a034
        0x0040:  88fb

ICMP Echo request frame with RCT
03:43:33.805354 IP 192.168.2.20 > 192.168.2.10: ICMP echo request, id 63748, seq 1, length 64
        0x0000:  70ff 761c 0e8c 70ff 761c 0e8d 0800 4500
        0x0010:  0054 26a4 4000 4001 8e96 c0a8 0214 c0a8
        0x0020:  020a 0800 c28e f904 0001 202e 1c3d 0000
        0x0030:  0000 0000 0000 0000 0000 0000 0000 0000
        0x0040:  0000 0000 0000 0000 0000 0000 0000 0000
        0x0050:  0000 0000 0000 0000 0000 0000 0000 0000
        0x0060:  0000 fc31 a05a 88fb
==================================================================================
The iperf3 traffic test logs can be accessed at the links below.
DUT-1: https://pastebin.ubuntu.com/p/8SkQzWJMn8/
DUT-2: https://pastebin.ubuntu.com/p/j2BZvvs7p4/

Other tests done.
 - Connect a SAN (eth0 and eth1 without prp interface) and
   do ping test from eth0 (192.168.2.40) to prp0 (192.168.2.10)
   verify the SAN node shows at the correct link A and B as shown
   in the node table dump
 - Regress HSR interface using 3 nodes connected in a ring topology.
   create hsr link version 0. Do iperf3 test between all nodes
   create hsr link version 1. Do iperf3 test between all nodes.

         --------eth0             eth1 --------eth0      eth1-------|
         |AM572x|----------------------|AM572x|--------------|AM572x|
         |      |                      |      |        ------|      |
         --------eth1---|               -------        | eth0 -------
                        |-------------------------------

   command used for HSR interface

   HSR V0

   ifconfig eth0 0.0.0.0 down
   ifconfig eth1 0.0.0.0 down
   ifconfig eth0 hw ether 70:FF:76:1C:0E:8C
   ifconfig eth1 hw ether 70:FF:76:1C:0E:8C
   ifconfig eth0 up
   ifconfig eth1 up
   ip link add name hsr0 type hsr slave1 eth0 slave2 eth1 supervision 45 version 0
   ifconfig hsr0 192.168.2.10

   HSR V1

   ifconfig eth0 0.0.0.0 down
   ifconfig eth1 0.0.0.0 down
   ifconfig eth0 hw ether 70:FF:76:1C:0E:8C
   ifconfig eth1 hw ether 70:FF:76:1C:0E:8C
   ifconfig eth0 up
   ifconfig eth1 up
   ip link add name hsr0 type hsr slave1 eth0 slave2 eth1 supervision 45 version 1
   ifconfig hsr0 192.168.2.10

   Logs at
   DUT-1 : https://pastebin.ubuntu.com/p/6PSJbZwQ6y/
   DUT-2 : https://pastebin.ubuntu.com/p/T8TqJsPRHc/
   DUT-3 : https://pastebin.ubuntu.com/p/VNzpv6HzKj/
 - Build tests :-
   Build with CONFIG_HSR=m
   allmodconfig build
   build with CONFIG_HSR=y and rebuild with sparse checker
   make C=1 zImage; make modules

Version history:
  v5 : Fixed comments about Kconfig changes on Patch 1/7 against v4
       Rebased to netnext/master branch.

  v4 : fixed following vs v3
       reverse xmas tree for local variables
       check for return type in call to skb_put_padto()

  v3 : Separated bug fixes from this series and send them for immediate merge
       But for that this is same as v2.

  v2 : updated comments on RFC. Following are the main changes:-
       - Removed the hsr_prp prefix
       - Added PRP information in header files to indicate
         the support for PRP explicitely
       - Re-use netlink socket interface with an added
         parameter proto for identifying PRP.
       - Use function pointers using a proto_ops struct
         to do things differently for PRP vs HSR.

   RFC: initial version posted and discussed at
       https://www.spinics.net/lists/netdev/msg656229.html
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Print PRP specific information from node table as part of debugfs
node table display. Also display the node as DAN-H or DAN-P depending
on the info from node table.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

DAN-P (Dual Attached Nodes PRP) nodes are expected to receive
traditional IP packets as well as PRP (Parallel Redundancy
Protocol) tagged (trailer) packets. PRP trailer is 6 bytes
of PRP protocol unit called RCT, Redundancy Control Trailer
(RCT) similar to HSR tag. PRP network can have traditional
devices such as bridges/switches or PC attached to it and
should be able to communicate. Regular Ethernet devices treat
the RCT as pads.  This patch adds logic to format L2 frames
from network stack to add a trailer (RCT) and send it as
duplicates over the slave interfaces when the protocol is
PRP as per IEC 62439-3. At the ingress, it strips the trailer,
do duplicate detection and rejection and forward a stripped
frame up the network stack. PRP device should accept frames
from Singly Attached Nodes (SAN) and thus the driver mark
the link where the frame came from in the node table.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

As a preparatory patch to introduce PRP, refactor the code specific to
handling HSR frames into separate functions and call them through
proto_ops function pointers.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support for generation of PRP supervision frames. For PRP,
supervision frame format is similar to HSR version 0, but have
a PRP Redundancy Control Trailer (RCT) added and uses a different
message type, PRP_TLV_LIFE_CHECK_DD. Also update
is_supervision_frame() to include the new message type used for
PRP supervision frame.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

As a preparatory patch to introduce support for PRP protocol, add a
protocol ops ptr in the private hsr structure to hold function
pointers as some of the functions at protocol level packet
handling is different for HSR vs PRP. It is expected that PRP will
add its of set of functions for protocol handling. Modify existing
hsr_announce() function to call proto_ops->send_sv_frame() to send
supervision frame for HSR. This is expected to be different for PRP.
So introduce a ops function ptr, send_sv_frame() for the same and
initialize it to send_hsr_supervsion_frame(). Modify hsr_announce()
to call proto_ops->send_sv_frame().
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

As a preparatory patch to introduce PRP protocol support in the
driver, refactor the skb init code to a separate function.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Parallel Redundancy Protocol (PRP) is another redundancy protocol
introduced by IEC 63439 standard. It is similar to HSR in many
aspects:-

 - Use a pair of Ethernet interfaces to created the PRP device
 - Use a 6 byte redundancy protocol part (RCT, Redundancy Check
   Trailer) similar to HSR Tag.
 - Has Link Redundancy Entity (LRE) that works with RCT to implement
   redundancy.

Key difference is that the protocol unit is a trailer instead of a
prefix as in HSR. That makes it inter-operable with tradition network
components such as bridges/switches which treat it as pad bytes,
whereas HSR nodes requires some kind of translators (Called redbox) to
talk to regular network devices. This features allows regular linux box
to be converted to a DAN-P box. DAN-P stands for Dual Attached Node - PRP
similar to DAN-H (Dual Attached Node - HSR).

Add a comment at the header/source code to explicitly state that the
driver files also handles PRP protocol as well.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Dan reports static checker warning:

"The patch 9b6ee3cf: "qed: sanitize PBL chains allocation" from Jul
23, 2020, leads to the following static checker warning:

	drivers/net/ethernet/qlogic/qed/qed_chain.c:299 qed_chain_alloc_pbl()
	error: uninitialized symbol 'pbl_virt'.

drivers/net/ethernet/qlogic/qed/qed_chain.c
   249  static int qed_chain_alloc_pbl(struct qed_dev *cdev, struct qed_chain *chain)
   250  {
   251          struct device *dev = &cdev->pdev->dev;
   252          struct addr_tbl_entry *addr_tbl;
   253          dma_addr_t phys, pbl_phys;
   254          __le64 *pbl_virt;
                ^^^^^^^^^^^^^^^^
[...]
   271          if (chain->b_external_pbl)
   272                  goto alloc_pages;
                        ^^^^^^^^^^^^^^^^ uninitialized
[...]
   298                  /* Fill the PBL table with the physical address of the page */
   299                  pbl_virt[i] = cpu_to_le64(phys);
                        ^^^^^^^^^^^
[...]
"

This issue was introduced with commit c3a321b0 ("qed: simplify
initialization of the chains with an external PBL"), when
chain->pbl_sp.table_virt initialization was moved up to
qed_chain_init_params().
Fix it by initializing pbl_virt with an already filled chain struct field.

Fixes: c3a321b0 ("qed: simplify initialization of the chains with an external PBL")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Michael Chan says:

====================
bnxt_en update.

This patchset removes the PCIe histogram and other debug register
data from ethtool -S. The removed data are not counters and they have
very large and constantly fluctuating values that are not suitable for
the ethtool -S decimal counter display.

The rest of the patches implement counter rollover for all hardware
counters that are not 64-bit counters.  Different generations of
hardware have different counter widths.  The driver will now query
the counter widths of all counters from firmware and implement
rollover support on all non-64-bit counters.

The last patch adds the PCIe histogram and other PCIe register data back
using the ethtool -d interface.

v2: Fix bnxt_re RDMA driver compile issue.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support to dump PXP registers and PCIe statistics.
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Now we can report all the full 64-bit CPU endian software accumulated
counters instead of the hw counters, some of which may be less than
64-bit wide.  Define the necessary macros to access the software
counters.
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Now that we have the infrastructure in place, add the new function
bnxt_accumulate_all_stats() to periodically accumulate and check for
counter rollover of all ring stats and port stats.

A chip bug was also discovered that could cause some ring counters to
become 0 during DMA.  Workaround by ignoring zeros on the affected
chips.

Some older frimware will reset port counters during ifdown.  We need
to check for that and free the accumulated port counters during ifdown
to prevent bogus counter overflow detection during ifup.
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If supported by newer firmware, make the firmware call to query all
the port counter masks.  If not supported, assume 40-bit port
counter masks.
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Newer firmware has a new call HWRM_FUNC_QSTATS_EXT to retrieve the
masks of all ring counters.  Make this call when supported to
initialize the hardware masks of all ring counters.  If the call
is not available, assume 48-bit ring counter masks on P5 chips.
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some of these DMAed hardware counters are not full 64-bit counters and
so we need to accumulate them as they overflow.  Allocate copies of these
DMA statistics memory blocks with the same size for accumulation.  The
hardware counter widths are also counter specific so we allocate
memory for masks that correspond to each counter.
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>