kernel-packages team mailing list archive

Thread
Date
[Bug 1270237] Re: prevent the conntrack table from filling up in the kernel

To: kernel-packages@xxxxxxxxxxxxxxxxxxx
From: Launchpad Bug Tracker <1270237@xxxxxxxxxxxxxxxxxx>
Date: Thu, 06 Mar 2014 15:58:46 -0000
Reply-to: Bug 1270237 <1270237@xxxxxxxxxxxxxxxxxx>
Sender: bounces@xxxxxxxxxxxxx
This bug was fixed in the package linux - 3.2.0-60.91

---------------
linux (3.2.0-60.91) precise; urgency=low

  [ Brad Figg ]

  * Release Tracking Bug
    - LP: #1281800

  [ Andy Whitcroft ]

  * [Config] d-i -- add xts.ko to crypto-modules udeb
    - LP: #1276739

  [ Upstream Kernel Changes ]

  * ath9k_htc: properly set MAC address and BSSID mask
    - LP: #1252422
    - CVE-2013-4579
  * SELinux: Fix kernel BUG on empty security contexts.
    - CVE-2014-1874
  * net: do not pretend FRAGLIST support
    - LP: #1281620
  * rds: prevent BUG_ON triggered on congestion update to loopback
    - LP: #1281620
  * ipv6: don't count addrconf generated routes against gc limit
    - LP: #1281620
  * net: drop_monitor: fix the value of maxattr
    - LP: #1281620
  * tg3: Initialize REG_BASE_ADDR at PCI config offset 120 to 0
    - LP: #1281620
  * net: unix: allow bind to fail on mutex lock
    - LP: #1281620
  * net: inet_diag: zero out uninitialized idiag_{src,dst} fields
    - LP: #1281620
  * drivers/net/hamradio: Integer overflow in hdlcdrv_ioctl()
    - LP: #1281620
  * hamradio/yam: fix info leak in ioctl
    - LP: #1281620
  * rds: prevent dereference of a NULL device
    - LP: #1281620
  * net: rose: restore old recvmsg behavior
    - LP: #1281620
  * vlan: Fix header ops passthru when doing TX VLAN offload.
    - LP: #1281620
  * net: llc: fix use after free in llc_ui_recvmsg
    - LP: #1281620
  * bridge: use spin_lock_bh() in br_multicast_set_hash_max
    - LP: #1281620
  * bnx2x: fix DMA unmapping of TSO split BDs
    - LP: #1281620
  * inet_diag: fix inet_diag_dump_icsk() timewait socket state logic
    - LP: #1281620
  * net: avoid reference counter overflows on fib_rules in multicast
    forwarding
    - LP: #1281620
  * xfs: Account log unmount transaction correctly
    - LP: #1281620
  * PCI: Enable ARI if dev and upstream bridge support it; disable
    otherwise
    - LP: #1281620
  * mm/memory-failure.c: recheck PageHuge() after hugetlb page migrate
    successfully
    - LP: #1281620
  * staging: comedi: cb_pcidio: fix for newer PCI-DIO48H
    - LP: #1281620
  * Fix warning from machine_kexec.c
    - LP: #1281620
  * hpfs: fix warnings when the filesystem fills up
    - LP: #1281620
  * KVM: x86: Convert vapic synchronization to _cached functions
    (CVE-2013-6368)
    - LP: #1281620
  * x86, fpu, amd: Clear exceptions in AMD FXSAVE workaround
    - LP: #1281620
  * mm: ensure get_unmapped_area() returns higher address than
    mmap_min_addr
    - LP: #1281620
  * ceph: cleanup aborted requests when re-sending requests.
    - LP: #1281620
  * ceph: wake up 'safe' waiters when unregistering request
    - LP: #1281620
  * sh: always link in helper functions extracted from libgcc
    - LP: #1281620
  * libata: add ATA_HORKAGE_BROKEN_FPDMA_AA quirk for Seagate Momentus
    SpinPoint M8
    - LP: #1281620
  * ext4: call ext4_error_inode() if jbd2_journal_dirty_metadata() fails
    - LP: #1281620
  * ext4: fix use-after-free in ext4_mb_new_blocks
    - LP: #1281620
  * ext4: check for overlapping extents in ext4_valid_extent_entries()
    - LP: #1281620
  * ext2: Fix oops in ext2_get_block() called from ext2_quota_write()
    - LP: #1281620
  * ext4: fix del_timer() misuse for ->s_err_report
    - LP: #1281620
  * xhci: Limit the spurious wakeup fix only to HP machines
    - LP: #1281620
  * iscsi-target: Fix-up all zero data-length CDBs with R/W_BIT set
    - LP: #1281620
  * drm/radeon: Fix sideport problems on certain RS690 boards
    - LP: #1281620
  * ALSA: hda - Add enable_msi=0 workaround for four HP machines
    - LP: #1260225, #1281620
  * gpio: msm: Fix irq mask/unmask by writing bits instead of numbers
    - LP: #1281620
  * radiotap: fix bitmap-end-finding buffer overrun
    - LP: #1281620
  * ftrace: Initialize the ftrace profiler for each possible cpu
    - LP: #1281620
  * libata: disable a disk via libata.force params
    - LP: #1281620
  * drm/ttm: Fix accesses through vmas with only partial coverage
    - LP: #1281620
  * sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities
    - LP: #1281620
  * ALSA: Add SNDRV_PCM_STATE_PAUSED case in wait_for_avail function
    - LP: #1281620
  * drm/i915: Use the correct GMCH_CTRL register for Sandybridge+
    - LP: #1281620
  * rtlwifi: pci: Fix oops on driver unload
    - LP: #1281620
  * ath9k: Fix interrupt handling for the AR9002 family
    - LP: #1281620
  * cpupower: Fix segfault due to incorrect getopt_long arugments
    - LP: #1281620
  * ASoC: wm8904: fix DSP mode B configuration
    - LP: #1281620
  * net_dma: mark broken
    - LP: #1281620
  * dm9601: fix reception of full size ethernet frames on dm9620/dm9621a
    - LP: #1281620
  * dm9601: work around tx fifo sync issue on dm962x
    - LP: #1281620
  * ext4: add explicit casts when masking cluster sizes
    - LP: #1281620
  * drm/radeon: 0x9649 is SUMO2 not SUMO
    - LP: #1281620
  * selinux: fix broken peer recv check
    - LP: #1281620
  * selinux: selinux_setprocattr()->ptrace_parent() needs rcu_read_lock()
    - LP: #1281620
  * ARM: fix footbridge clockevent device
    - LP: #1281620
  * powerpc: Fix bad stack check in exception entry
    - LP: #1281620
  * ahci: Use PCI_VENDOR_ID_MARVELL_EXT for 0x1b4b
    - LP: #1281620
  * ahci: add an observed PCI ID for Marvell 88se9172 SATA controller
    - LP: #1281620
  * pci: Add PCI_DEVICE_SUB() macro
    - LP: #1281620
  * ahci: add PCI ID for Marvell 88SE9170 SATA controller
    - LP: #1281620
  * ARM: fix "bad mode in ... handler" message for undefined instructions
    - LP: #1281620
  * SELinux: Fix possible NULL pointer dereference in
    selinux_inode_permission()
    - LP: #1281620
  * md/raid5: Fix possible confusion when multiple write errors occur.
    - LP: #1281620
  * md/raid10: fix two bugs in handling of known-bad-blocks.
    - LP: #1281620
  * md/raid10: fix bug when raid10 recovery fails to recover a block.
    - LP: #1281620
  * hwmon: (coretemp) Fix truncated name of alarm attributes
    - LP: #1281620
  * nilfs2: fix segctor bug that causes file system corruption
    - LP: #1281620
  * perf/x86/amd/ibs: Fix waking up from S3 for AMD family 10h
    - LP: #1281620
  * mm: fix aio performance regression for database caused by THP
    - LP: #1281620
  * mm: hugetlbfs: fix hugetlbfs optimization
    - LP: #1281620
  * sched/rt: Fix SCHED_RR across cgroups
    - LP: #1281620
  * sched,rt: fix isolated CPUs leaving root_task_group indefinitely
    throttled
    - LP: #1281620
  * sched: Unthrottle rt runqueues in __disable_runtime()
    - LP: #1281620
  * sched/rt: Avoid updating RT entry timeout twice within one tick period
    - LP: #1281620
  * Linux 3.2.55
    - LP: #1281620
  * netfilter: nf_conntrack: avoid large timeout for mid-stream pickup
    - LP: #1270237
 -- Joseph Salisbury <joseph.salisbury@xxxxxxxxxxxxx>   Tue, 18 Feb 2014 16:43:57 -0500

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

** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2013-4579

** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2013-6368

** CVE added: http://www.cve.mitre.org/cgi-
bin/cvename.cgi?name=2014-1874

-- 
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/1270237

Title:
  prevent the conntrack table from filling up in the kernel

Status in “linux” package in Ubuntu:
  Fix Released
Status in “linux-lts-raring” package in Ubuntu:
  Invalid
Status in “linux” source package in Precise:
  Fix Released
Status in “linux-lts-raring” source package in Precise:
  Fix Committed
Status in “linux” source package in Quantal:
  Fix Committed
Status in “linux-lts-raring” source package in Quantal:
  Invalid
Status in “linux” source package in Raring:
  Invalid
Status in “linux-lts-raring” source package in Raring:
  Invalid

Bug description:
  [Impact]
  When running a server for an extended amount of time the conntrack table can fill up.
  Here is the netfilter discussion: http://www.spinics.net/lists/netfilter-devel/msg26759.html

  [Fix]
  https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=6547a221871f139cc56328a38105d47c14874cbe

  Present in 3.11 >

  [Test Case]
  From the patch:
  When loose tracking is enabled (default), non-syn packets cause
  creation of new conntracks in established state with default timeout for
  established state (5 days).  This causes the table to fill up with UNREPLIED
  when the 'new ack' packet happened to be the last-ack of a previous,
  already timed-out connection.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1270237/+subscriptions
References

[Bug 1270237] [NEW] prevent the conntrack table from filling up in the kernel
From: Chris J Arges, 2014-01-17