[Bug 1602755] Re: Tunnel offload indications not stripped from encapsulated packets, causing performance overhead
Launchpad Bug Tracker
1602755 at bugs.launchpad.net
Mon Aug 29 14:10:47 UTC 2016
This bug was fixed in the package linux - 4.4.0-36.55
---------------
linux (4.4.0-36.55) xenial; urgency=low
[ Stefan Bader ]
* Release Tracking Bug
- LP: #1612305
* I2C touchpad does not work on AMD platform (LP: #1612006)
- SAUCE: pinctrl/amd: Remove the default de-bounce time
* CVE-2016-5696
- tcp: make challenge acks less predictable
linux (4.4.0-35.54) xenial; urgency=low
[ Stefan Bader ]
* Release Tracking Bug
- LP: #1611215
* [i915_bpo] Sync with v4.7 (LP: #1609742)
- SAUCE: i915_bpo: Sync with v4.7
* s390/cio: fix reset of channel measurement block (LP: #1609415)
- s390/cio: allow to reset channel measurement block
* in Ubuntu16.10: Hit on Call traces and system goes down when transactional
memory tests are running in 32TB Brazos system (LP: #1606786)
- powerpc/tm: Avoid SLB faults in treclaim/trecheckpoint when RI=0
- powerpc/tm: Fix stack pointer corruption in __tm_recheckpoint()
* Power Menu does not display after press the Power Button (LP: #1609204)
- intel-vbtn: new driver for Intel Virtual Button
- [config] enable CONFIG_INTEL_VBTN=m
* OptiPlex 7450 AIO hangs when rebooting (LP: #1608762)
- x86/reboot: Add Dell Optiplex 7450 AIO reboot quirk
* virtualbox+usb 3.0 breaks boot, -28 kernel works (LP: #1604058)
- SAUCE: xhci: Fix soft lockup in xhci_pci_probe path when XHCI_STATE_HALTED
* linux-kernel: Freeing IRQ from IRQ context (LP: #1597908)
- block: defer timeouts to a workqueue
* Tunnel offload indications not stripped from encapsulated packets, causing
performance overhead (LP: #1602755)
- tunnels: Remove encapsulation offloads on decap.
* lm-sensors is throwing "ERROR: Can't get value of subfeature temp1_input:
I/O error" for be2net driver (LP: #1607387)
- be2net: perform temperature query in adapter regardless of its interface
state
* Dell dock MAC Address pass through doesn't work in Ubuntu (LP: #1579984)
- r8152: Add support for setting pass through MAC address on RTL8153-AD
* vmxnet3 LRO IPv6 performance issues (stalling TCP) (LP: #1605494)
- Driver: Vmxnet3: set CHECKSUM_UNNECESSARY for IPv6 packets
* ISST-LTE:pVM:monklp5:Ubuntu16.04.1:system crashed at
lpfc_sli4_scmd_to_wqidx_distr (LP: #1597974)
- SAUCE: lpfc: fix oops in lpfc_sli4_scmd_to_wqidx_distr() from
lpfc_send_taskmgmt()
* Backport cxlflash shutdown patch to Xenial SRU (LP: #1605405)
- SAUCE: cxlflash: Verify problem state area is mapped before notifying
shutdown
* Xenial update to v4.4.16 stable release (LP: #1607404)
- mac80211: fix fast_tx header alignment
- mac80211: mesh: flush mesh paths unconditionally
- mac80211_hwsim: Add missing check for HWSIM_ATTR_SIGNAL
- mac80211: Fix mesh estab_plinks counting in STA removal case
- EDAC, sb_edac: Fix rank lookup on Broadwell
- IB/cm: Fix a recently introduced locking bug
- IB/mlx4: Properly initialize GRH TClass and FlowLabel in AHs
- powerpc/pseries: Fix IBM_ARCH_VEC_NRCORES_OFFSET since POWER8NVL was added
- powerpc/tm: Always reclaim in start_thread() for exec() class syscalls
- usb: dwc2: fix regression on big-endian PowerPC/ARM systems
- USB: EHCI: declare hostpc register as zero-length array
- usb: common: otg-fsm: add license to usb-otg-fsm
- mnt: fs_fully_visible test the proper mount for MNT_LOCKED
- mnt: Account for MS_RDONLY in fs_fully_visible
- mnt: If fs_fully_visible fails call put_filesystem.
- of: fix autoloading due to broken modalias with no 'compatible'
- of: irq: fix of_irq_get[_byname]() kernel-doc
- locking/ww_mutex: Report recursive ww_mutex locking early
- locking/qspinlock: Fix spin_unlock_wait() some more
- locking/static_key: Fix concurrent static_key_slow_inc()
- x86, build: copy ldlinux.c32 to image.iso
- kprobes/x86: Clear TF bit in fault on single-stepping
- x86/amd_nb: Fix boot crash on non-AMD systems
- Revert "gpiolib: Split GPIO flags parsing and GPIO configuration"
- uvc: Forward compat ioctls to their handlers directly
- thermal: cpu_cooling: fix improper order during initialization
- writeback: use higher precision calculation in domain_dirty_limits()
- nfsd4/rpc: move backchannel create logic into rpc code
- nfsd: Always lock state exclusively.
- nfsd: Extend the mutex holding region around in nfsd4_process_open2()
- posix_acl: Add set_posix_acl
- nfsd: check permissions when setting ACLs
- make nfs_atomic_open() call d_drop() on all ->open_context() errors.
- NFS: Fix another OPEN_DOWNGRADE bug
- ARM: imx6ul: Fix Micrel PHY mask
- ARM: 8578/1: mm: ensure pmd_present only checks the valid bit
- ARM: 8579/1: mm: Fix definition of pmd_mknotpresent
- MIPS: KVM: Fix modular KVM under QEMU
- mm: Export migrate_page_move_mapping and migrate_page_copy
- UBIFS: Implement ->migratepage()
- sched/fair: Fix cfs_rq avg tracking underflow
- packet: Use symmetric hash for PACKET_FANOUT_HASH.
- net_sched: fix mirrored packets checksum
- cdc_ncm: workaround for EM7455 "silent" data interface
- ipv6: Fix mem leak in rt6i_pcpu
- ARCv2: Check for LL-SC livelock only if LLSC is enabled
- ARCv2: LLSC: software backoff is NOT needed starting HS2.1c
- kvm: Fix irq route entries exceeding KVM_MAX_IRQ_ROUTES
- KVM: nVMX: VMX instructions: fix segment checks when L1 is in long mode.
- HID: elo: kill not flush the work
- HID: hiddev: validate num_values for HIDIOCGUSAGES, HIDIOCSUSAGES commands
- tracing: Handle NULL formats in hold_module_trace_bprintk_format()
- base: make module_create_drivers_dir race-free
- iommu/arm-smmu: Wire up map_sg for arm-smmu-v3
- iommu/vt-d: Enable QI on all IOMMUs before setting root entry
- iommu/amd: Fix unity mapping initialization race
- drm/mgag200: Black screen fix for G200e rev 4
- ipmi: Remove smi_msg from waiting_rcv_msgs list before handle_one_recv_msg()
- arm64: Rework valid_user_regs
- vfs: add d_real_inode() helper
- af_unix: fix hard linked sockets on overlay
- btrfs: account for non-CoW'd blocks in btrfs_abort_transaction
- drm/radeon: fix asic initialization for virtualized environments
- drm/amdgpu/gfx7: fix broken condition check
- ubi: Make recover_peb power cut aware
- drm/amdkfd: unbind only existing processes
- drm/amdkfd: destroy dbgmgr in notifier release
- drm/dp/mst: Always clear proposed vcpi table for port.
- drm/nouveau/disp/sor/gf119: both links use the same training register
- drm/nouveau/gr/gf100-: update sm error decoding from gk20a nvgpu headers
- drm/nouveau/fbcon: fix out-of-bounds memory accesses
- drm/nouveau: fix for disabled fbdev emulation
- drm/nouveau/disp/sor/gf119: select correct sor when poking training pattern
- drm/i915/ilk: Don't disable SSC source if it's in use
- drm/i915: Refresh cached DP port register value on resume
- drm/i915: Update ifdeffery for mutex->owner
- drm/i915: Update CDCLK_FREQ register on BDW after changing cdclk frequency
- drm: add missing drm_mode_set_crtcinfo call
- drm: make drm_atomic_set_mode_prop_for_crtc() more reliable
- drm: atmel-hlcdc: actually disable scaling when no scaling is required
- drm/ttm: Make ttm_bo_mem_compat available
- drm/vmwgfx: Add an option to change assumed FB bpp
- drm/vmwgfx: Work around mode set failure in 2D VMs
- drm/vmwgfx: Check pin count before attempting to move a buffer
- drm/vmwgfx: Delay pinning fbdev framebuffer until after mode set
- drm/vmwgfx: Fix error paths when mapping framebuffer
- memory: omap-gpmc: Fix omap gpmc EXTRADELAY timing
- perf/x86: Fix undefined shift on 32-bit kernels
- xen/balloon: Fix declared-but-not-defined warning
- iio: Fix error handling in iio_trigger_attach_poll_func
- iio:st_pressure: fix sampling gains (bring inline with ABI)
- iio: light apds9960: Add the missing dev.parent
- iio: proximity: as3935: correct IIO_CHAN_INFO_RAW output
- iio: proximity: as3935: remove triggered buffer processing
- iio: proximity: as3935: fix buffer stack trashing
- iio: humidity: hdc100x: correct humidity integration time mask
- iio: humidity: hdc100x: fix IIO_TEMP channel reporting
- iio: hudmidity: hdc100x: fix incorrect shifting and scaling
- staging: iio: accel: fix error check
- iio: accel: kxsd9: fix the usage of spi_w8r8()
- iio:ad7266: Fix broken regulator error handling
- iio:ad7266: Fix support for optional regulators
- iio:ad7266: Fix probe deferral for vref
- tty/vt/keyboard: fix OOB access in do_compute_shiftstate()
- hwmon: (dell-smm) Restrict fan control and serial number to CAP_SYS_ADMIN by
default
- hwmon: (dell-smm) Disallow fan_type() calls on broken machines
- hwmon: (dell-smm) Cache fan_type() calls and change fan detection
- ALSA: dummy: Fix a use-after-free at closing
- ALSA: hda - Fix the headset mic jack detection on Dell machine
- ALSA: hda / realtek - add two more Thinkpad IDs (5050,5053) for tpt460 fixup
- ALSA: au88x0: Fix calculation in vortex_wtdma_bufshift()
- ALSA: echoaudio: Fix memory allocation
- ALSA: timer: Fix negative queue usage by racy accesses
- ALSA: hda/realtek: Add Lenovo L460 to docking unit fixup
- ALSA: hda - Add PCI ID for Kabylake-H
- ALSA: hda - fix read before array start
- ALSA: hda/realtek - add new pin definition in alc225 pin quirk table
- ALSA: pcm: Free chmap at PCM free callback, too
- ALSA: ctl: Stop notification after disconnection
- ALSA: hda - fix use-after-free after module unload
- ALSA: hda: add AMD Stoney PCI ID with proper driver caps
- ARM: sunxi/dt: make the CHIP inherit from allwinner,sun5i-a13
- ARM: dts: armada-38x: fix MBUS_ID for crypto SRAM on Armada 385 Linksys
- ARM: mvebu: fix HW I/O coherency related deadlocks
- ovl: Copy up underlying inode's ->i_mode to overlay inode
- ovl: verify upper dentry in ovl_remove_and_whiteout()
- scsi: fix race between simultaneous decrements of ->host_failed
- 53c700: fix BUG on untagged commands
- Fix reconnect to not defer smb3 session reconnect long after socket
reconnect
- cifs: dynamic allocation of ntlmssp blob
- File names with trailing period or space need special case conversion
- xen/acpi: allow xen-acpi-processor driver to load on Xen 4.7
- crypto: qat - make qat_asym_algs.o depend on asn1 headers
- tmpfs: don't undo fallocate past its last page
- tmpfs: fix regression hang in fallocate undo
- drm/i915: Revert DisplayPort fast link training feature
- ovl: verify upper dentry before unlink and rename
- Linux 4.4.16
* Regression caused by `fuse: Add support for pid namespaces` in 4.4.0-6.21
(LP: #1605344)
- SAUCE: (namespace) fuse: Permit requests from other pid namespaces
* CVE-2016-5400
- media: fix airspy usb probe error path
* Cannot mount proc in unprivileged containers if /proc/xen is mounted
(LP: #1607374)
- SAUCE: xenbus: Use proc_create_mount_point() to create /proc/xen
* Mic mute key does not work for Ideapad laptops (LP: #1607153)
- ideapad_laptop: Add an event for mic mute hotkey
* NVMe stress test fails after 12 hours on Ubuntu 16.04 (LP: #1604995)
- block: atari: Return early for unsupported sector size
* Console extremely slow with 4.4 kernels for servers with Matrox G200er2 or
similar (LP: #1605662)
- SAUCE: vesafb: Set mtrr:3 (write-combining) as default
* Ubuntu 16.04 - Full EEH Recovery Support for NVMe devices (LP: #1602724)
- nvme: use a work item to submit async event requests
- nvme: don't poll the CQ from the kthread
- nvme: replace the kthread with a per-device watchdog timer
- NVMe: Fix reset/remove race
- nvme: Avoid reset work on watchdog timer function during error recovery
- NVMe: Always use MSI/MSI-x interrupts
* [LTC-Test] - NMI watchdog Bug and call traces when trinity is executed.
(LP: #1602524)
- ext4: factor out determining of hole size
- ext4: return hole from ext4_map_blocks()
- ext4: more efficient SEEK_DATA implementation
* changelog: add CVEs as first class citizens (LP: #1604344)
- avoid duplicate CVE numbers in changelog
* [LTCTest][Opal][OP820] Machine crashed with Oops: Kernel access of bad area,
sig: 11 [#1] while executing Froze PE Error injection (LP: #1603449)
- powerpc/eeh: Fix invalid cached PE primary bus
* Hotplug remove and re-add adds PCI adapter to next PCI domain (PCI)
(LP: #1603574)
- powerpc/pci: Assign fixed PHB number based on device-tree properties
* nvme - reset_controller is not working after adapter's firmware upgrade
(adapter quirk is needed) (LP: #1602726)
- NVMe: Create discard zero quirk white list
- nvme/quirk: Add a delay before checking for adapter readiness
* ovs nat: conntrack netlink event are missing (LP: #1603468)
- openvswitch: fix conntrack netlink event delivery
* FlashGT - In Tuleta 8284-22A with card in card slot P1-C9, system Fails to
boot operating system (LP: #1602785)
- cxl: Ignore CAPI adapters misplaced in switched slots
* CVE-2016-5728
- misc: mic: Fix for double fetch security bug in VOP driver
* CVE-2016-5244 (LP: #1589041)
- rds: fix an infoleak in rds_inc_info_copy
* Miscellaneous Ubuntu changes
- Added Snapcraft files
- SAUCE: snapcraft: cleanup and remove unnecessary elements
-- Stefan Bader <stefan.bader at canonical.com> Thu, 11 Aug 2016 17:34:14
+0200
** Changed in: linux (Ubuntu Xenial)
Status: Fix Committed => Fix Released
** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2016-5244
** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2016-5400
** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2016-5696
** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2016-5728
--
You received this bug notification because you are a member of Ubuntu
OpenStack, which is subscribed to the bug report.
https://bugs.launchpad.net/bugs/1602755
Title:
Tunnel offload indications not stripped from encapsulated packets,
causing performance overhead
Status in linux package in Ubuntu:
Fix Committed
Status in linux source package in Xenial:
Fix Released
Bug description:
Commit:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=a09a4c8dd1ec7f830e1fb9e59eb72bddc965d168
Description:
tunnels: Remove encapsulation offloads on decap.
If a packet is either locally encapsulated or processed through GRO
it is marked with the offloads that it requires. However, when it is
decapsulated these tunnel offload indications are not removed. This
means that if we receive an encapsulated TCP packet, aggregate it with
GRO, decapsulate, and retransmit the resulting frame on a NIC that does
not support encapsulation, we won't be able to take advantage of hardware
offloads even though it is just a simple TCP packet at this point.
This fixes the problem by stripping off encapsulation offload indications
when packets are decapsulated.
The performance impacts of this bug are significant. In a test where a
Geneve encapsulated TCP stream is sent to a hypervisor, GRO'ed, decapsulated,
and bridged to a VM performance is improved by 60% (5Gbps->8Gbps) as a
result of avoiding unnecessary segmentation at the VM tap interface.
Reported-by: Ramu Ramamurthy <sramamur at linux.vnet.ibm.com>
Fixes: 68c33163 ("v4 GRE: Add TCP segmentation offload for GRE")
Signed-off-by: Jesse Gross <jesse at kernel.org>
Signed-off-by: David S. Miller <davem at davemloft.net>
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1602755/+subscriptions
More information about the Ubuntu-openstack-bugs
mailing list