linux-azure 4.10.0-1006.6 source package in Ubuntu

Changelog

linux-azure (4.10.0-1006.6) xenial; urgency=low

  [ Marcelo Cerri ]

  * linux-azure: 4.10.0-1006.6 -proposed tracker (LP: #1694731)

  * [Hyper-V][SAUCE] hv_sock for 4.10 (LP: #1690196)
    - SAUCE: azure: hv_sock: introduce Hyper-V Sockets
    - [Config] azure: CONFIG_HYPERV_SOCK=m
    - SAUCE: azure: hv_sock: 4.11 update
    - SAUCE: 9p: net/9p: add hv_sock transport

  * Azure kernel fails in ADT tests (LP: #1693494)
    - Revert "UBUNTU: [config] azure: disable VIRTIO drivers"

  * linux-azure should be able to boot in a local Hyper-V (LP: #1692920)
    - Revert "UBUNTU: [config] azure: disable CONFIG_EFI"

  * [Hyper-V] Rebase Hyper-V to upstream 4.11 kernel plus some cherry-picks from
    linux-next (LP: #1690177)
    - Input: serio - drop unnecessary calls to device_init_wakeup
    - scsi: mvumi: remove fake transport template
    - scsi: libsas: remove sas_scsi_timed_out
    - scsi: remove eh_timed_out methods in the transport template
    - netvsc: negotiate checksum and segmentation parameters
    - netvsc: report number of rx queues in ethtool
    - netvsc: add ethtool ops to get/set RSS key
    - netvsc: allow more flexible setting of number of channels
    - netvsc: remove unused variables
    - netvsc: remove no longer needed receive staging buffers
    - netvsc: enhance transmit select_queue
    - netvsc: group all per-channel state together
    - netvsc: optimize receive path
    - netvsc: don't pass void * to internal device_add
    - netvsc: simplify rndis_filter_remove
    - netvsc: eliminate per-device outstanding send counter
    - netvsc: account for packets/bytes transmitted after completion
    - netvsc: call netif_receive_skb
    - vmbus: use kernel bitops for traversing interrupt mask
    - vmbus: drop no longer used kick_q argument
    - vmbus: remove no longer used signal_policy
    - vmbus: remove unused kickq argument to sendpacket
    - vmbus: put related per-cpu variable together
    - vmbus: callback is in softirq not workqueue
    - vmbus: change to per channel tasklet
    - vmbus: add direct isr callback mode
    - vmbus: expose hv_begin/end_read
    - vmbus: remove conditional locking of vmbus_write
    - vmbus: constify parameters where possible
    - net-next: treewide use is_vlan_dev() helper function.
    - scripts/spelling.txt: add "intialization" pattern and fix typo instances
    - netvsc: handle select_queue when device is being removed
    - netvsc: simplify get next send section
    - netvsc: fix calculation of available send sections
    - PCI: hv: Specify CPU_AFFINITY_ALL for MSI affinity when >= 32 CPUs
    - refcount_t: Introduce a special purpose refcount type
    - locking/refcounts: Out-of-line everything
    - locking/refcounts: Add missing kernel.h header to have UINT_MAX defined
    - locking/refcount: Add refcount_t API kernel-doc comments
    - netvsc: don't overload variable in same function
    - vmbus: introduce in-place packet iterator
    - netvsc: implement NAPI
    - netvsc: replace netdev_alloc_skb_ip_align with napi_alloc_skb
    - net: hyperv: use new api ethtool_{get|set}_link_ksettings
    - vmbus: only reschedule tasklet if time limit exceeded
    - vmbus: remove useless return's
    - vmbus: remove unnecessary initialization
    - vmbus: fix spelling errors
    - vmbus: make channel_message table constant
    - vmbus: cleanup header file style
    - vmbus: expose debug info for drivers
    - Drivers: hv: Fix a typo
    - Drivers: hv: Base autoeoi enablement based on hypervisor hints
    - scsi: storvsc: Prefer kcalloc over kzalloc with multiply
    - scsi: storvsc: remove return at end of void function
    - scsi: scsi_transport_fc: Add dummy initiator role to rport
    - scsi: storvsc: Add support for FC rport.
    - netvsc: remove unused #define
    - netvsc: need napi scheduled during removal
    - netvsc: avoid race with callback
    - netvsc: add comments about callback's and NAPI
    - netvsc: fix NAPI performance regression
    - netvsc: allow get/set of RSS indirection table
    - netvsc: handle offline mtu and channel change
    - netvsc: change max channel calculation
    - qed*: Update to dual-license
    - qede: Break datapath logic into its own file
    - qede: Split filtering logic to its own file
    - qed*: Change maximal number of queues
    - qede: Postpone reallocation until NAPI end
    - qed*: RSS indirection based on queue-handles
    - qed*: Add support for ndo_set_vf_trust
    - net: make ndo_get_stats64 a void function
    - netvsc: report per-channel stats in ethtool statistics
    - netvsc: enable GRO
    - netvsc: use RCU to protect inner device structure
    - netvsc: fix use-after-free in netvsc_change_mtu()
    - netvsc: uses RCU instead of removal flag
    - SAUCE: include refcount.h in hyperv_net.h
    - netvsc: use refcount_t for keeping track of sub channels
    - netvsc: remove unnecessary lock on shutdown
    - netvsc: eliminate unnecessary skb == NULL checks
    - netvsc: fix and cleanup rndis_filter_set_packet_filter
    - Drivers: hv: Issue explicit EOI when autoeoi is not enabled
    - hv_netvsc: Fix the queue index computation in forwarding case
    - hv_netvsc: change netvsc device default duplex to FULL
    - netvsc: Deal with rescinded channels correctly

 -- Marcelo Henrique Cerri <email address hidden>  Wed, 31 May 2017 11:59:08 -0300