MFC r317518, r318647, r319197, r319198, r319199, r319200, r320625, r320626,

r320628, r319201, r320629, r320630, r320631, r320632.

r317518:
Import Amazon Elastic Network Adapter (ENA) HAL to sys/contrib/

Import from vendor-sys/ena-com/1.1.4.1
SVN rev.: 317516
Version: 1.1.4.1

r318647:
Add support for Amazon Elastic Network Adapter (ENA) NIC

ENA is a networking interface designed to make good use of modern CPU
features and system architectures.

The ENA device exposes a lightweight management interface with a
minimal set of memory mapped registers and extendable command set
through an Admin Queue.

The driver supports a range of ENA devices, is link-speed independent
(i.e., the same driver is used for 10GbE, 25GbE, 40GbE, etc.), and has
a negotiated and extendable feature set.

Some ENA devices support SR-IOV. This driver is used for both the
SR-IOV Physical Function (PF) and Virtual Function (VF) devices.

ENA devices enable high speed and low overhead network traffic
processing by providing multiple Tx/Rx queue pairs (the maximum number
is advertised by the device via the Admin Queue), a dedicated MSI-X
interrupt vector per Tx/Rx queue pair, and CPU cacheline optimized
data placement.

The ENA driver supports industry standard TCP/IP offload features such
as checksum offload and TCP transmit segmentation offload (TSO).
Receive-side scaling (RSS) is supported for multi-core scaling.

The ENA driver and its corresponding devices implement health
monitoring mechanisms such as watchdog, enabling the device and driver
to recover in a manner transparent to the application, as well as
debug logs.

Some of the ENA devices support a working mode called Low-latency
Queue (LLQ), which saves several more microseconds. This feature will
be implemented for driver in future releases.

r319197:
Add mbuf defragmentation to the ENA driver

When mbuf chain is too long and device cannot handle that number
of segments in DMA transaction, mbuf chain will be defragmented.
Initially, driver was dropping all mbuf chains that were exceeding
supported number of segments.

r319198:
Add locks before each ena_up and ena_down

Lock only ena_up and ena_down calls in ioctl handler, instead of whole
ioctl. Locking ioctl with sx lock that is sleepable, is not allowed in
some cases, e.g. when multicast options are being changed.
Additional locking was added in deatch function to prevent race condition
with ioctl function.

r319199:
Add error handling to the ENA driver if init of the reset task fails

Also, to simplify cleaning routine, reset task is initialized before
allocating statistics and other resources.

r319200:
Move ENA's hw stats updating routine to separate task

Initially, stats were being updated each time OS was requesting for
the first statistic.
To read statistics from hw, condvar was used. cv_timedwait cannot be
called when unsleepable lock is held, and this happens when FreeBSD
is requesting statistic.
Seperate task is reading statistics from NIC each 1 second.

r319201:
Introduce additional locks when releasing TX resources and buffers in ENA

There could be race condition with TX cleaning routine when cleaning mbufs,
when it was called directly from main sending thread (ena_mq_start).

r320625:
Add missing lock upon initialization of the interface

Lack of this lock was causing crash if down was called in
parallel with the initialization routine.

r320626:
Acquire locks before calling drbr_flush()

It is required to hold lock that is associated with buffer ring before
flushing drbr.

r320628:
Unmask all IO irqs after driver state is set as running

If driver left MSI-x handlling routine because interface was put down,
it is not unmasking IRQs, so any requesting interrupt will be awaiting
for unmasking.

On ena_up() routine all interrupts are being unmasked and any awaiting
interrupt will be handled right away.

If handler was executed before driver state was set as running, handling
routine is being ended immediately, leaving IO irqs for given queue
masked.

r320629:
Call drbr_advance() before leaving TX routine

If drbr_advance() is not called before doing cleanup and packet is
already enqueued for sending (tx_info is holding pointer to mbuf), then
mbuf is cleaned both in drbr_flush() and in cleanup routine, when all
mbufs hold by tx_buffer_info are being released.

This causes panic, because mbuf is released twice.

r320630:
Remove RX mtx from ENA driver

RX lock is no longer required. There can only be one RX cleanup task
running at a time, RX cleanup cannot be executed if interface is not
yet initialized and ena_down() will not free any RX resources if any io
interrupt is being handled - RX cleanup task is only called from an
interrupt handler.

r320631:
Fix creation of dma tags and TSO settings

TSO settings were not reflecting real HW capabilities.

DMA tags were created with wrong window - high address was the same as
low, so excluding window was not working.

Capabilities of TX dma transaction were not set properly - TSO max size
had been increased and size of one segment had been adjusted.

r320632:
Replace mbuf defragmentation with collapse

Collapse should be more effective than defragmentation.
Added missing declaration of ena_check_and_collapse_mbuf().

Submitted by:	Michal Krawczyk <mk@semihalf.com>
		Jakub Palider <jpa@semihalf.com>
		Jan Medala <jan@semihalf.com>
Obtained from: Semihalf, Amazon.com Inc.
Sponsored by: Amazon.com Inc.

1 files changed, 255 insertions, 0 deletions

diff --git a/share/man/man4/ena.4 b/share/man/man4/ena.4
new file mode 100644
index 0000000..90e5839
--- /dev/null
+++ b/share/man/man4/ena.4
@@ -0,0 +1,255 @@
+.\" Copyright (c) 2015-2017 Amazon.com, Inc. or its affiliates.
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\"
+.\" 1. Redistributions of source code must retain the above copyright
+.\"    notice, this list of conditions and the following disclaimer.
+.\"
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\"    notice, this list of conditions and the following disclaimer in
+.\"    the documentation and/or other materials provided with the
+.\"    distribution.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+.\" "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+.\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+.\" A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+.\" OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+.\" SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+.\" LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+.\" OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+.\"
+.\" $FreeBSD$
+.\"
+.Dd May 04, 2017
+.Dt ENA 4
+.Os
+.Sh NAME
+.Nm ena
+.Nd "FreeBSD kernel driver for Elastic Network Adapter (ENA) family"
+.Sh SYNOPSIS
+To compile this driver into the kernel,
+place the following line in your
+kernel configuration file:
+.Bd -ragged -offset indent
+.Cd "device ena"
+.Ed
+.Pp
+Alternatively, to load the driver as a
+module at boot time, place the following line in
+.Xr loader.conf 5 :
+.Bd -literal -offset indent
+if_ena_load="YES"
+.Ed
+.Sh DESCRIPTION
+The ENA is a networking interface designed to make good use of modern CPU
+features and system architectures.
+.Pp
+The ENA device exposes a lightweight management interface with a
+minimal set of memory mapped registers and extendable command set
+through an Admin Queue.
+.Pp
+The driver supports a range of ENA devices, is link-speed independent
+(i.e., the same driver is used for 10GbE, 25GbE, 40GbE, etc.), and has
+a negotiated and extendable feature set.
+.Pp
+Some ENA devices support SR-IOV. This driver is used for both the
+SR-IOV Physical Function (PF) and Virtual Function (VF) devices.
+.Pp
+The ENA devices enable high speed and low overhead network traffic
+processing by providing multiple Tx/Rx queue pairs (the maximum number
+is advertised by the device via the Admin Queue), a dedicated MSI-X
+interrupt vector per Tx/Rx queue pair, and CPU cacheline optimized
+data placement.
+.Pp
+The
+.Nm
+driver supports industry standard TCP/IP offload features such
+as checksum offload and TCP transmit segmentation offload (TSO).
+Receive-side scaling (RSS) is supported for multi-core scaling.
+.Pp
+The
+.Nm
+driver and its corresponding devices implement health
+monitoring mechanisms such as watchdog, enabling the device and driver
+to recover in a manner transparent to the application, as well as
+debug logs.
+.Pp
+Some of the ENA devices support a working mode called Low-latency
+Queue (LLQ), which saves several more microseconds. This feature will
+be implemented for driver in future releases.
+.Sh HARDWARE
+Supported PCI vendor ID/device IDs:
+.Pp
+.Bl -bullet -compact
+.It
+1d0f:0ec2 - ENA PF
+.It
+1d0f:1ec2 - ENA PF with LLQ support
+.It
+1d0f:ec20 - ENA VF
+.It
+1d0f:ec21 - ENA VF with LLQ support
+.El
+.Sh DIAGNOSTICS
+.Ss Device initialization phase:
+.Bl -diag
+.It ena%d: failed to init mmio read less
+.Pp
+Error occured during initialization of the mmio register read request.
+.It ena%d: Can not reset device
+.Pp
+Device could not be reset; device may not be responding or is already
+during reset.
+.It ena%d: device version is too low
+.Pp
+Version of the controller is too low and it is not supported by the driver.
+.It ena%d: Invalid dma width value %d
+.Pp
+The controller is able to request dma transcation width. Device stopped
+responding or it demanded invalid value.
+.It ena%d: Can not initialize ena admin queue with device
+.Pp
+Initialization of the Admin Queue failed; device may not be responding or there
+was a problem with initialization of the resources.
+.It ena%d: Cannot get attribute for ena device rc: %d
+.Pp
+Failed to get attributes of the device from the controller.
+.It ena%d: Cannot configure aenq groups rc: %d
+.Pp
+Errors occured when trying to configure AENQ groups.
+.El
+.Ss Driver initialisation/shutdown phase:
+.Bl -diag
+.It ena%d: PCI resource allocation failed!
+.It ena%d: allocating ena_dev failed
+.It ena%d: failed to pmap registers bar
+.It ena%d: Error while setting up bufring
+.It ena%d: Error with initialization of IO rings
+.It ena%d: can not allocate ifnet structure
+.It ena%d: Error with network interface setup
+.It ena%d: Failed to enable and set the admin interrupts
+.It ena%d: Failed to allocate %d, vectors %d
+.It ena%d: Failed to enable MSIX, vectors %d rc %d
+.It ena%d: Error with MSI-X enablement
+.It ena%d: could not allocate irq vector: %d
+.It ena%d: Unable to allocate bus resource: registers
+.Pp
+Resource allocation failed when initializing the device; driver will not
+be attached.
+.It ena%d: ENA device init failed (err: %d)
+.Pp
+Device initialization failed; driver will not be attached.
+.It ena%d: could not activate irq vector: %d
+.Pp
+Error occured when trying to activate interrupt vectors for Admin Queue.
+.It ena%d: failed to register interrupt handler for irq %ju: %d
+.Pp
+Error occured when trying to register Admin Queue interrupt handler.
+.It ena%d: Cannot setup mgmnt queue intr
+.Pp
+Error occured during configuration of the Admin Queue interrupts.
+.It ena%d: Enable MSI-X failed
+.Pp
+Configuration of the MSI-X for Admin Queue failed; there could be lack
+of resources or interrupts could not have been configured; driver will
+not be attached.
+.It ena%d: VLAN is in use, detach first
+.Pp
+VLANs are being used when trying to detach the driver; VLANs should be detached
+first and then detach routine should be called again.
+.It ena%d: Unmapped RX DMA tag associations
+.It ena%d: Unmapped TX DMA tag associations
+.Pp
+Error occured when trying to destroy RX/TX DMA tag.
+.It ena%d: Cannot init RSS
+.It ena%d: Cannot fill indirect table
+.It ena%d: Cannot fill indirect table
+.It ena%d: Cannot fill hash function
+.It ena%d: Cannot fill hash control
+.It ena%d: WARNING: RSS was not properly initialized, it will affect bandwidth
+.Pp
+Error occured during initialization of one of RSS resources; device is still
+going to work but it will affect performance because all RX packets will be
+passed to queue 0 and there will be no hash information.
+.It ena%d: failed to tear down irq: %d
+.It ena%d: dev has no parent while releasing res for irq: %d
+Release of the interrupts failed.
+.El
+.Ss Additional diagnostic:
+.Bl -diag
+.It ena%d: Cannot get attribute for ena device
+.Pp
+This message appears when trying to change MTU and driver is unable to get
+attributes from the device.
+.It ena%d: Invalid MTU setting. new_mtu: %d
+.Pp
+Requested MTU value is not supported and will not be set.
+.It ena%d: keep alive watchdog timeout
+.Pp
+Device stopped responding and will be reset.
+.It ena%d: Found a Tx that wasn't completed on time, qid %d, index %d.
+.Pp
+Packet was pushed to the NIC but not sent within given time limit; it may
+be caused by hang of the IO queue.
+.It ena%d: The number of lost tx completion is aboce the threshold (%d > %d). Reset the device
+.Pp
+If too many Tx wasn't completed on time the device is going to be reset; it may
+be caused by hanged queue or device.
+.It ena%d: trigger reset is on
+.Pp
+Device will be reset; reset is triggered either by watchdog or if too many TX
+packets were not completed on time.
+.It ena%d: invalid value recvd
+.Pp
+Link status received from the device in the AENQ handler is invalid.
+.It ena%d: Allocation for Tx Queue %u failed
+.It ena%d: Allocation for Rx Queue %u failed
+.It ena%d: Unable to create Rx DMA map for buffer %d
+.It ena%d: Failed to create io TX queue #%d rc: %d
+.It ena%d: Failed to get TX queue handlers. TX queue num %d rc: %d
+.It ena%d: Failed to create io RX queue[%d] rc: %d
+.It ena%d: Failed to get RX queue handlers. RX queue num %d rc: %d
+.It ena%d: failed to request irq
+.It ena%d: could not allocate irq vector: %d
+.It ena%d: failed to register interrupt handler for irq %ju: %d
+.Pp
+IO resources initialization failed. Interface will not be brought up.
+.It ena%d: LRO[%d] Initialization failed!
+.Pp
+Initialization of the LRO for the RX ring failed.
+.It ena%d: failed to alloc buffer for rx queue
+.It ena%d: failed to add buffer for rx queue %d
+.It ena%d: refilled rx queue %d with %d pages only
+.Pp
+Allocation of resources used on RX path failed; if happened during
+initialization of the IO queue, the interface will not be brought up.
+.It ena%d: ioctl promisc/allmulti
+.Pp
+IOCTL request for the device to work in promiscuous/allmulti mode; see
+.Xr ifconfig 8
+for more details.
+.It ena%d: too many fragments. Last fragment: %d!
+.Pp
+Packet with unsupported number of segments was queued for sending to the
+device; packet will be dropped.
+.Sh SUPPORT
+If an issue is identified with the released source code with a supported adapter
+email the specific information related to the issue to
+.Aq Mt mk@semihalf.com
+and
+.Aq Mt mw@semihalf.com .
+.Sh SEE ALSO
+.Xr vlan 4 ,
+.Xr ifconfig 8
+.Sh AUTHORS
+The
+.Nm
+driver was written by
+.An Semihalf.