kernel-packages team mailing list archive

Thread
Date
[Bug 1547680] Re: bonded interfaces in bridge do not have LRO turned off

To: kernel-packages@xxxxxxxxxxxxxxxxxxx
From: Launchpad Bug Tracker <1547680@xxxxxxxxxxxxxxxxxx>
Date: Mon, 14 Mar 2016 14:09:00 -0000
Reply-to: Bug 1547680 <1547680@xxxxxxxxxxxxxxxxxx>
Sender: bounces@xxxxxxxxxxxxx
This bug was fixed in the package linux - 3.13.0-83.127

---------------
linux (3.13.0-83.127) trusty; urgency=low

  [ Brad Figg ]

  * Release Tracking Bug
    - LP: #1555839

  [ Florian Westphal ]

  * SAUCE: [nf,v2] netfilter: x_tables: don't rely on well-behaving
    userspace
    - LP: #1555338

linux (3.13.0-82.126) trusty; urgency=low

  [ Brad Figg ]

  * Release Tracking Bug
    - LP: #1554732

  [ Upstream Kernel Changes ]

  * Revert "drm/radeon: call hpd_irq_event on resume"
    - LP: #1554608
  * net: generic dev_disable_lro() stacked device handling
    - LP: #1547680

linux (3.13.0-81.125) trusty; urgency=low

  [ Luis Henriques ]

  * Release Tracking Bug
    - LP: #1552316

  [ Upstream Kernel Changes ]

  * Revert "firmware: dmi_scan: Fix UUID endianness for SMBIOS >= 2.6"
    - LP: #1551419
  * bcache: Fix a lockdep splat in an error path
    - LP: #1551327

linux (3.13.0-80.124) trusty; urgency=low

  [ Brad Figg ]

  * Release Tracking Bug
    - LP: #1548519

  [ Andy Whitcroft ]

  * [Debian] hv: hv_set_ifconfig -- convert to python3
    - LP: #1506521
  * [Debian] hv: hv_set_ifconfig -- switch to approved indentation
    - LP: #1540586
  * [Debian] hv: hv_set_ifconfig -- fix numerous parameter handling issues
    - LP: #1540586

  [ Dan Streetman ]

  * SAUCE: nbd: ratelimit error msgs after socket close
    - LP: #1505564

  [ Upstream Kernel Changes ]

  * Revert "workqueue: make sure delayed work run in local cpu"
    - LP: #1546320
  * [media] gspca: ov534/topro: prevent a division by 0
    - LP: #1542497
  * [media] media: dvb-core: Don't force CAN_INVERSION_AUTO in oneshot mode
    - LP: #1542497
  * tools lib traceevent: Fix output of %llu for 64 bit values read on 32
    bit machines
    - LP: #1542497
  * KVM: x86: correctly print #AC in traces
    - LP: #1542497
  * drm/radeon: call hpd_irq_event on resume
    - LP: #1542497
  * xhci: refuse loading if nousb is used
    - LP: #1542497
  * arm64: Clear out any singlestep state on a ptrace detach operation
    - LP: #1542497
  * time: Avoid signed overflow in timekeeping_get_ns()
    - LP: #1542497
  * rtlwifi: fix memory leak for USB device
    - LP: #1542497
  * wlcore/wl12xx: spi: fix oops on firmware load
    - LP: #1542497
  * EDAC, mc_sysfs: Fix freeing bus' name
    - LP: #1542497
  * EDAC: Don't try to cancel workqueue when it's never setup
    - LP: #1542497
  * EDAC: Robustify workqueues destruction
    - LP: #1542497
  * powerpc: Make value-returning atomics fully ordered
    - LP: #1542497
  * powerpc: Make {cmp}xchg* and their atomic_ versions fully ordered
    - LP: #1542497
  * dm space map metadata: remove unused variable in brb_pop()
    - LP: #1542497
  * dm thin: fix race condition when destroying thin pool workqueue
    - LP: #1542497
  * futex: Drop refcount if requeue_pi() acquired the rtmutex
    - LP: #1542497
  * drm/radeon: clean up fujitsu quirks
    - LP: #1542497
  * mmc: sdio: Fix invalid vdd in voltage switch power cycle
    - LP: #1542497
  * mmc: sdhci: Fix sdhci_runtime_pm_bus_on/off()
    - LP: #1542497
  * udf: limit the maximum number of indirect extents in a row
    - LP: #1542497
  * nfs: Fix race in __update_open_stateid()
    - LP: #1542497
  * USB: cp210x: add ID for ELV Marble Sound Board 1
    - LP: #1542497
  * NFSv4: Don't perform cached access checks before we've OPENed the file
    - LP: #1542497
  * NFS: Fix attribute cache revalidation
    - LP: #1542497
  * posix-clock: Fix return code on the poll method's error path
    - LP: #1542497
  * rtlwifi: rtl8192de: Fix incorrect module parameter descriptions
    - LP: #1542497
  * rtlwifi: rtl8192se: Fix module parameter initialization
    - LP: #1542497
  * rtlwifi: rtl8192ce: Fix handling of module parameters
    - LP: #1542497
  * rtlwifi: rtl8192cu: Add missing parameter setup
    - LP: #1542497
  * bcache: fix a livelock when we cause a huge number of cache misses
    - LP: #1542497
  * bcache: Add a cond_resched() call to gc
    - LP: #1542497
  * bcache: clear BCACHE_DEV_UNLINK_DONE flag when attaching a backing
    device
    - LP: #1542497
  * bcache: fix a leak in bch_cached_dev_run()
    - LP: #1542497
  * bcache: unregister reboot notifier if bcache fails to unregister device
    - LP: #1542497
  * bcache: add mutex lock for bch_is_open
    - LP: #1542497
  * bcache: allows use of register in udev to avoid "device_busy" error.
    - LP: #1542497
  * bcache: Change refill_dirty() to always scan entire disk if necessary
    - LP: #1542497
  * wlcore/wl12xx: spi: fix NULL pointer dereference (Oops)
    - LP: #1542497
  * Input: i8042 - add Fujitsu Lifebook U745 to the nomux list
    - LP: #1542497
  * libxfs: pack the agfl header structure so XFS_AGFL_SIZE is correct
    - LP: #1542497
  * x86/xen: don't reset vcpu_info on a cancelled suspend
    - LP: #1542497
  * udf: Prevent buffer overrun with multi-byte characters
    - LP: #1542497
  * udf: Check output buffer length when converting name to CS0
    - LP: #1542497
  * PCI: host: Mark PCIe/PCI (MSI) IRQ cascade handlers as IRQF_NO_THREAD
    - LP: #1542497
  * iwlwifi: update and fix 7265 series PCI IDs
    - LP: #1542497
  * locks: fix unlock when fcntl_setlk races with a close
    - LP: #1542497
  * ASoC: compress: Fix compress device direction check
    - LP: #1542497
  * dm snapshot: fix hung bios when copy error occurs
    - LP: #1542497
  * uml: fix hostfs mknod()
    - LP: #1542497
  * uml: flush stdout before forking
    - LP: #1542497
  * drm/nouveau/kms: take mode_config mutex in connector hotplug path
    - LP: #1542497
  * x86/boot: Double BOOT_HEAP_SIZE to 64KB
    - LP: #1542497
  * s390: fix normalization bug in exception table sorting
    - LP: #1542497
  * xfs: inode recovery readahead can race with inode buffer creation
    - LP: #1542497
  * clocksource/drivers/vt8500: Increase the minimum delta
    - LP: #1542497
  * Input: elantech - mark protocols v2 and v3 as semi-mt
    - LP: #1542497
  * x86/reboot/quirks: Add iMac10,1 to pci_reboot_dmi_table[]
    - LP: #1542497
  * ALSA: seq: Fix missing NULL check at remove_events ioctl
    - LP: #1542497
  * ALSA: seq: Fix race at timer setup and close
    - LP: #1542497
  * virtio_balloon: fix race by fill and leak
    - LP: #1542497
  * virtio_balloon: fix race between migration and ballooning
    - LP: #1542497
  * parisc: Fix __ARCH_SI_PREAMBLE_SIZE
    - LP: #1542497
  * scripts/recordmcount.pl: support data in text section on powerpc
    - LP: #1542497
  * powerpc/module: Handle R_PPC64_ENTRY relocations
    - LP: #1542497
  * ALSA: timer: Fix double unlink of active_list
    - LP: #1542497
  * dmaengine: dw: fix cyclic transfer setup
    - LP: #1542497
  * dmaengine: dw: fix cyclic transfer callbacks
    - LP: #1542497
  * mmc: mmci: fix an ages old detection error
    - LP: #1542497
  * ALSA: timer: Fix race among timer ioctls
    - LP: #1542497
  * sparc64: fix incorrect sign extension in sys_sparc64_personality
    - LP: #1542497
  * cifs: Ratelimit kernel log messages
    - LP: #1542497
  * cifs: fix race between call_async() and reconnect()
    - LP: #1542497
  * cifs_dbg() outputs an uninitialized buffer in cifs_readdir()
    - LP: #1542497
  * m32r: fix m32104ut_defconfig build fail
    - LP: #1542497
  * dma-debug: switch check from _text to _stext
    - LP: #1542497
  * scripts/bloat-o-meter: fix python3 syntax error
    - LP: #1542497
  * ocfs2/dlm: ignore cleaning the migration mle that is inuse
    - LP: #1542497
  * ALSA: timer: Harden slave timer list handling
    - LP: #1542497
  * mm: soft-offline: check return value in second __get_any_page() call
    - LP: #1542497
  * memcg: only free spare array when readers are done
    - LP: #1542497
  * panic: release stale console lock to always get the logbuf printed out
    - LP: #1542497
  * kernel/panic.c: turn off locks debug before releasing console lock
    - LP: #1542497
  * printk: do cond_resched() between lines while outputting to consoles
    - LP: #1542497
  * ALSA: hda - Fix bass pin fixup for ASUS N550JX
    - LP: #1542497
  * crypto: af_alg - Disallow bind/setkey/... after accept(2)
    - LP: #1542497
  * crypto: af_alg - Fix socket double-free when accept fails
    - LP: #1542497
  * crypto: af_alg - Add nokey compatibility path
    - LP: #1542497
  * crypto: hash - Add crypto_ahash_has_setkey
    - LP: #1542497
  * crypto: af_alg - Allow af_af_alg_release_parent to be called on nokey
    path
    - LP: #1542497
  * crypto: af_alg - Forbid bind(2) when nokey child sockets are present
    - LP: #1542497
  * ALSA: hrtimer: Fix stall by hrtimer_cancel()
    - LP: #1542497
  * ALSA: pcm: Fix snd_pcm_hw_params struct copy in compat mode
    - LP: #1542497
  * ALSA: seq: Fix snd_seq_call_port_info_ioctl in compat mode
    - LP: #1542497
  * ALSA: control: Avoid kernel warnings from tlv ioctl with numid 0
    - LP: #1542497
  * crypto: algif_skcipher - Load TX SG list after waiting
    - LP: #1542497
  * crypto: crc32c - Fix crc32c soft dependency
    - LP: #1542497
  * IB/qib: fix mcast detach when qp not attached
    - LP: #1542497
  * iscsi-target: Fix potential dead-lock during node acl delete
    - LP: #1542497
  * ocfs2: NFS hangs in __ocfs2_cluster_lock due to race with
    ocfs2_unblock_lock
    - LP: #1542497
  * [media] rc: allow rc modules to be loaded if rc-main is not a module
    - LP: #1542497
  * SCSI: initio: remove duplicate module device table
    - LP: #1542497
  * clk: xgene: Fix divider with non-zero shift value
    - LP: #1542497
  * ath9k_htc: check for underflow in ath9k_htc_rx_msg()
    - LP: #1542497
  * mtd: nand: fix ONFI parameter page layout
    - LP: #1542497
  * ALSA: fm801: propagate TUNER_ONLY bit when autodetected
    - LP: #1542497
  * pinctrl: bcm2835: Fix memory leak in error path
    - LP: #1542497
  * kconfig: return 'false' instead of 'no' in bool function
    - LP: #1542497
  * perf/x86: Fix filter_events() bug with event mappings
    - LP: #1542497
  * power: test_power: correctly handle empty writes
    - LP: #1542497
  * firmware: actually return NULL on failed request_firmware_nowait()
    - LP: #1542497
  * mmc: sd: limit SD card power limit according to cards capabilities
    - LP: #1542497
  * Btrfs: clean up an error code in btrfs_init_space_info()
    - LP: #1542497
  * batman-adv: Avoid recursive call_rcu for batadv_bla_claim
    - LP: #1542497
  * batman-adv: Avoid recursive call_rcu for batadv_nc_node
    - LP: #1542497
  * batman-adv: Drop immediate orig_node free function
    - LP: #1542497
  * printk: help pr_debug and pr_devel to optimize out arguments
    - LP: #1542497
  * mmc: debugfs: correct wrong voltage value
    - LP: #1542497
  * IB/mlx4: Initialize hop_limit when creating address handle
    - LP: #1542497
  * veth: don’t modify ip_summed; doing so treats packets with bad
    checksums as good.
    - LP: #1542497
  * sctp: sctp should release assoc when sctp_make_abort_user return NULL
    in sctp_close
    - LP: #1542497
  * connector: bump skb->users before callback invocation
    - LP: #1542497
  * unix: properly account for FDs passed over unix sockets
    - LP: #1542497
  * bridge: Only call /sbin/bridge-stp for the initial network namespace
    - LP: #1542497
  * net: sctp: prevent writes to cookie_hmac_alg from accessing invalid
    memory
    - LP: #1542497
  * tcp_yeah: don't set ssthresh below 2
    - LP: #1542497
  * bonding: Prevent IPv6 link local address on enslaved devices
    - LP: #1542497
  * phonet: properly unshare skbs in phonet_rcv()
    - LP: #1542497
  * ipv6: update skb->csum when CE mark is propagated
    - LP: #1542497
  * team: Replace rcu_read_lock with a mutex in team_vlan_rx_kill_vid
    - LP: #1542497
  * Linux 3.13.11-ckt34
    - LP: #1542497
  * qeth: initialize net_device with carrier off
    - LP: #1541907
  * umount: Do not allow unmounting rootfs.
    - LP: #1541313
  * [media] usbvision fix overflow of interfaces array
    - LP: #1546273
  * [media] usbvision: fix leak of usb_dev on failure paths in
    usbvision_probe()
    - LP: #1546273
  * [media] usbvision: fix crash on detecting device with invalid
    configuration
    - LP: #1546273
  * tty: Fix unsafe ldisc reference via ioctl(TIOCGETD)
    - LP: #1546273
  * USB: serial: visor: fix crash on detecting device without write_urbs
    - LP: #1546273
  * ASN.1: Fix non-match detection failure on data overrun
    - LP: #1546273
  * iio: adis_buffer: Fix out-of-bounds memory access
    - LP: #1546273
  * x86/irq: Call chip->irq_set_affinity in proper context
    - LP: #1546273
  * usb: cdc-acm: handle unlinked urb in acm read callback
    - LP: #1546273
  * usb: cdc-acm: send zero packet for intel 7260 modem
    - LP: #1546273
  * cdc-acm:exclude Samsung phone 04e8:685d
    - LP: #1546273
  * usb: hub: do not clear BOS field during reset device
    - LP: #1546273
  * USB: cp210x: add ID for IAI USB to RS485 adaptor
    - LP: #1546273
  * USB: visor: fix null-deref at probe
    - LP: #1546273
  * USB: serial: option: Adding support for Telit LE922
    - LP: #1546273
  * ALSA: seq: Fix incorrect sanity check at snd_seq_oss_synth_cleanup()
    - LP: #1546273
  * ALSA: seq: Degrade the error message for too many opens
    - LP: #1546273
  * USB: serial: ftdi_sio: add support for Yaesu SCU-18 cable
    - LP: #1546273
  * USB: option: fix Cinterion AHxx enumeration
    - LP: #1546273
  * ALSA: compress: Disable GET_CODEC_CAPS ioctl for some architectures
    - LP: #1546273
  * ALSA: usb-audio: Fix TEAC UD-501/UD-503/NT-503 usb delay
    - LP: #1546273
  * arm64: errata: Add -mpc-relative-literal-loads to build flags
    - LP: #1533009, #1546273
  * SCSI: fix crashes in sd and sr runtime PM
    - LP: #1546273
  * n_tty: Fix unsafe reference to "other" ldisc
    - LP: #1546273
  * ALSA: dummy: Disable switching timer backend via sysfs
    - LP: #1546273
  * drm/vmwgfx: respect 'nomodeset'
    - LP: #1546273
  * x86/mm/pat: Avoid truncation when converting cpa->numpages to address
    - LP: #1546273
  * perf annotate browser: Fix behaviour of Shift-Tab with nothing focussed
    - LP: #1546273
  * powerpc/perf: Remove PPMU_HAS_SSLOT flag for Power8
    - LP: #1546273
  * Linux 3.13.11-ckt35
    - LP: #1546273
  * netfilter: bridge: don't use nf_bridge_info data to store mac header
    - LP: #1463911
  * netfilter: bridge: restore vlan tag when refragmenting
    - LP: #1463911
  * netfilter: bridge: forward IPv6 fragmented packets
    - LP: #1463911
  * netfilter: bridge: Use __in6_dev_get rather than in6_dev_get in
    br_validate_ipv6
    - LP: #1463911
  * ALSA: usb-audio: avoid freeing umidi object twice
    - LP: #1546177
    - CVE-2016-2384
  * vmstat: explicitly schedule per-cpu work on the CPU we need it to run
    on
    - LP: #1546320

 -- Brad Figg <brad.figg@xxxxxxxxxxxxx>  Thu, 10 Mar 2016 14:41:56 -0800

** Changed in: linux (Ubuntu Trusty)
       Status: Fix Committed => Fix Released

** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2016-2384

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1547680

Title:
  bonded interfaces in bridge do not have LRO turned off

Status in linux package in Ubuntu:
  Fix Released
Status in linux-lts-utopic package in Ubuntu:
  New
Status in linux source package in Trusty:
  Fix Released
Status in linux-lts-utopic source package in Trusty:
  Fix Committed

Bug description:
  The "large receive offload" functionality of some NICs is not
  compatible with packet forwarding, because it modifies incoming
  packets in an irreversible way so the packets can't be correctly
  forwarded across the bridge.  When an interface is added to a bridge,
  the kernel automatically disables LRO via dev_disable_lro(), however
  in trusty this only works for normal interfaces (or vlan/macvlan)
  added to a bridge.  When interfaces belong to a bond, and the bond is
  added to a bridge, the interfaces' LRO is not disabled, and (some)
  incoming packets are modified by LRO and then lost when passing
  through the bridge.

  Commit fbe168ba91f7c327856f205699404284c2f09e36 updates
  dev_disable_lro() to work for all types of interfaces, including
  bonds.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1547680/+subscriptions
References

[Bug 1547680] [NEW] bonded interfaces in bridge do not have LRO turned off
From: Dan Streetman, 2016-02-19