kernel-packages team mailing list archive

Thread
Date
[Bug 1400209] Re: KVM: PPC: BOOK3S: HV: CMA: Reserve cma region only in hypervisor mode

To: kernel-packages@xxxxxxxxxxxxxxxxxxx
From: Launchpad Bug Tracker <1400209@xxxxxxxxxxxxxxxxxx>
Date: Fri, 30 Jan 2015 15:34:11 -0000
Reply-to: Bug 1400209 <1400209@xxxxxxxxxxxxxxxxxx>
Sender: bounces@xxxxxxxxxxxxx
This bug was fixed in the package linux - 3.13.0-45.74

---------------
linux (3.13.0-45.74) trusty; urgency=low

  [ Seth Forshee ]

  * Release Tracking Bug
    - LP: #1410384

  [ Jesse Barnes ]

  * SAUCE: drm/i915/vlv: assert and de-assert sideband reset at boot and
    resume v3
    - LP: #1401963

  [ K. Y. Srinivasan ]

  * SAUCE: storvsc: force SPC-3 compliance on win8 and win8 r2 hosts
    - LP: #1406867

  [ Timo Aaltonen ]

  * SAUCE: Switch VLV/BYT to use i915_bdw.
    - LP: #1401963

  [ Upstream Kernel Changes ]

  * Revert "xhci: clear root port wake on bits if controller isn't wake-up
    capable"
    - LP: #1408779
  * KVM: PPC: BOOK3S: HV: CMA: Reserve cma region only in hypervisor mode
    - LP: #1400209
  * e1000e: Fix no connectivity when driver loaded with cable out
    - LP: #1400365
  * net/mlx4_core: Enable CQE/EQE stride support
    - LP: #1400127
  * net/mlx4_core: Cache line EQE size support
    - LP: #1400127
  * net/mlx4_en: Add mlx4_en_get_cqe helper
    - LP: #1400127
  * net/mlx4_core: Introduce mlx4_get_module_info for cable module info
    reading
    - LP: #1400127
  * ethtool, net/mlx4_en: Cable info, get_module_info/eeprom ethtool
    support
    - LP: #1400127
  * net/mlx4_core: Introduce ACCESS_REG CMD and eth_prot_ctrl dev cap
    - LP: #1400127
  * net/mlx4_core: Add ethernet backplane autoneg device capability
    - LP: #1400127
  * ethtool, net/mlx4_en: Add 100M, 20G, 56G speeds ethtool reporting
    support
    - LP: #1400127
  * net/mlx4_en: Use PTYS register to query ethtool settings
    - LP: #1400127
  * net/mlx4_en: Use PTYS register to set ethtool settings (Speed)
    - LP: #1400127
  * net/mlx4_en: Add support for setting rxvlan offload OFF/ON
    - LP: #1400127
  * net/mlx4_en: Add ethtool support for [rx|tx]vlan offload set to OFF/ON
    - LP: #1400127
  * net/mlx4_core: Prevent VF from changing port configuration
    - LP: #1400127
  * net/mlx4_en: mlx4_en_set_settings() always fails when autoneg is set
    - LP: #1400127
  * ipv4: fix nexthop attlen check in fib_nh_match
    - LP: #1408779
  * vxlan: fix a use after free in vxlan_encap_bypass
    - LP: #1408779
  * vxlan: using pskb_may_pull as early as possible
    - LP: #1408779
  * vxlan: fix a free after use
    - LP: #1408779
  * ipv4: fix a potential use after free in ip_tunnel_core.c
    - LP: #1408779
  * ax88179_178a: fix bonding failure
    - LP: #1408779
  * tcp: md5: do not use alloc_percpu()
    - LP: #1408779
  * ipv4: dst_entry leak in ip_send_unicast_reply()
    - LP: #1408779
  * drivers/net, ipv6: Select IPv6 fragment idents for virtio UFO packets
    - LP: #1408779
  * drivers/net: macvtap and tun depend on INET
    - LP: #1408779
  * ip6_tunnel: Use ip6_tnl_dev_init as the ndo_init function.
    - LP: #1408779
  * vti6: Use vti6_dev_init as the ndo_init function.
    - LP: #1408779
  * sit: Use ipip6_tunnel_init as the ndo_init function.
    - LP: #1408779
  * gre6: Move the setting of dev->iflink into the ndo_init functions.
    - LP: #1408779
  * vxlan: Do not reuse sockets for a different address family
    - LP: #1408779
  * net: sctp: fix memory leak in auth key management
    - LP: #1408779
  * smsc911x: power-up phydev before doing a software reset.
    - LP: #1408779
  * sunvdc: add cdrom and v1.1 protocol support
    - LP: #1408779
  * sunvdc: compute vdisk geometry from capacity
    - LP: #1408779
  * sunvdc: limit each sg segment to a page
    - LP: #1408779
  * vio: fix reuse of vio_dring slot
    - LP: #1408779
  * sunvdc: don't call VD_OP_GET_VTOC
    - LP: #1408779
  * sparc64: Fix crashes in schizo_pcierr_intr_other().
    - LP: #1408779
  * sparc64: Do irq_{enter,exit}() around generic_smp_call_function*().
    - LP: #1408779
  * sparc32: Implement xchg and atomic_xchg using ATOMIC_HASH locks
    - LP: #1408779
  * sparc64: Fix constraints on swab helpers.
    - LP: #1408779
  * inetdevice: fixed signed integer overflow
    - LP: #1408779
  * ipv4: Fix incorrect error code when adding an unreachable route
    - LP: #1408779
  * ieee802154: fix error handling in ieee802154fake_probe()
    - LP: #1408779
  * qmi_wwan: Add support for HP lt4112 LTE/HSPA+ Gobi 4G Modem
    - LP: #1408779
  * pptp: fix stack info leak in pptp_getname()
    - LP: #1408779
  * ipx: fix locking regression in ipx_sendmsg and ipx_recvmsg
    - LP: #1408779
  * aio: fix uncorrent dirty pages accouting when truncating AIO ring
    buffer
    - LP: #1408779
  * spi: dw: Fix dynamic speed change.
    - LP: #1408779
  * USB: serial: cp210x: add IDs for CEL MeshConnect USB Stick
    - LP: #1408779
  * iio: Fix IIO_EVENT_CODE_EXTRACT_DIR bit mask
    - LP: #1408779
  * usb: serial: ftdi_sio: add PIDs for Matrix Orbital products
    - LP: #1408779
  * USB: keyspan: fix tty line-status reporting
    - LP: #1408779
  * USB: keyspan: fix overrun-error reporting
    - LP: #1408779
  * USB: ssu100: fix overrun-error reporting
    - LP: #1408779
  * nfsd: correctly define v4.2 support attributes
    - LP: #1408779
  * SUNRPC: Fix locking around callback channel reply receive
    - LP: #1408779
  * nfsd: Fix slot wake up race in the nfsv4.1 callback code
    - LP: #1408779
  * bnx2fc: do not add shared skbs to the fcoe_rx_list
    - LP: #1408779
  * scsi: add Intel Multi-Flex to scsi scan blacklist
    - LP: #1408779
  * ARM: 8216/1: xscale: correct auxiliary register in suspend/resume
    - LP: #1408779
  * USB: xhci: don't start a halted endpoint before its new dequeue is set
    - LP: #1408779
  * USB: xhci: Reset a halted endpoint immediately when we encounter a
    stall.
    - LP: #1408779
  * usb: xhci: rework root port wake bits if controller isn't allowed to
    wakeup
    - LP: #1408779
  * ALSA: hda - Limit 40bit DMA for AMD HDMI controllers
    - LP: #1408779
  * PCI/MSI: Add device flag indicating that 64-bit MSIs don't work
    - LP: #1408779
  * gpu/radeon: Set flag to indicate broken 64-bit MSI
    - LP: #1408779
  * sound/radeon: Move 64-bit MSI quirk from arch to driver
    - LP: #1408779
  * powerpc/powernv: Honor the generic "no_64bit_msi" flag
    - LP: #1408779
  * powerpc/pseries: Honor the generic "no_64bit_msi" flag
    - LP: #1408779
  * MIPS: Loongson: Make platform serial setup always built-in.
    - LP: #1408779
  * net/ping: handle protocol mismatching scenario
    - LP: #1408779
  * usb-quirks: Add reset-resume quirk for MS Wireless Laser Mouse 6000
    - LP: #1408779
  * Input: xpad - use proper endpoint type
    - LP: #1408779
  * powerpc: 32 bit getcpu VDSO function uses 64 bit instructions
    - LP: #1408779
  * ARM: 8222/1: mvebu: enable strex backoff delay
    - LP: #1408779
  * ARM: 8226/1: cacheflush: get rid of restarting block
    - LP: #1408779
  * staging: r8188eu: Add new device ID for DLink GO-USB-N150
    - LP: #1408779
  * btrfs: zero out left over bytes after processing compression streams
    - LP: #1408779
  * smiapp: Only some selection targets are settable
    - LP: #1408779
  * i2c: omap: fix NACK and Arbitration Lost irq handling
    - LP: #1408779
  * drm/nouveau/gf116: remove copy1 engine
    - LP: #1408779
  * drm/i915: More cautious with pch fifo underruns
    - LP: #1408779
  * drm/i915: Unlock panel even when LVDS is disabled
    - LP: #1408779
  * AHCI: Add DeviceIDs for Sunrise Point-LP SATA controller
    - LP: #1408779
  * sata_fsl: fix error handling of irq_of_parse_and_map
    - LP: #1408779
  * drm/radeon: kernel panic in drm_calc_vbltimestamp_from_scanoutpos with
    3.18.0-rc6
    - LP: #1408779
  * mm: frontswap: invalidate expired data on a dup-store failure
    - LP: #1408779
  * mm/vmpressure.c: fix race in vmpressure_work_fn()
    - LP: #1408779
  * drivers/input/evdev.c: don't kfree() a vmalloc address
    - LP: #1408779
  * mm: fix swapoff hang after page migration and fork
    - LP: #1408779
  * mm: fix anon_vma_clone() error treatment
    - LP: #1408779
  * slab: fix nodeid bounds check for non-contiguous node IDs
    - LP: #1408779
  * ahci: disable MSI on SAMSUNG 0xa800 SSD
    - LP: #1408779
  * i2c: davinci: generate STP always when NACK is received
    - LP: #1408779
  * ip_tunnel: the lack of vti_link_ops' dellink() cause kernel panic
    - LP: #1408779
  * ipv6: gre: fix wrong skb->protocol in WCCP
    - LP: #1408779
  * Fix race condition between vxlan_sock_add and vxlan_sock_release
    - LP: #1408779
  * tg3: fix ring init when there are more TX than RX channels
    - LP: #1408779
  * net/mlx4_core: Limit count field to 24 bits in qp_alloc_res
    - LP: #1408779
  * rtnetlink: release net refcnt on error in do_setlink()
    - LP: #1408779
  * net: mvneta: fix Tx interrupt delay
    - LP: #1408779
  * net: mvneta: fix race condition in mvneta_tx()
    - LP: #1408779
  * net: sctp: use MAX_HEADER for headroom reserve in output path
    - LP: #1408779
  * Linux 3.13.11-ckt13
    - LP: #1408779
  * ipv6: fix swapped ipv4/ipv6 mtu_reduced callbacks
    - LP: #1404558
  * arm64: Fix machine_shutdown() definition
    - LP: #1404335
  * arm64: Fix deadlock scenario with smp_send_stop()
    - LP: #1404335
  * iwlwifi: mvm: a few more SKUs for 7260 and 3160
  * iwlwifi: fix and add 7265 series HW IDs
    - LP: #1408222
 -- Seth Forshee <seth.forshee@xxxxxxxxxxxxx>   Tue, 13 Jan 2015 11:37:56 -0600

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

-- 
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/1400209

Title:
  KVM: PPC: BOOK3S: HV: CMA: Reserve cma region only in hypervisor mode

Status in linux package in Ubuntu:
  Fix Released
Status in linux source package in Trusty:
  Fix Released
Status in linux source package in Utopic:
  Fix Released

Bug description:
  SRU Justification:
  [Impact]
  KVM guests on PowerNV allocate memory for CMA in kvm_cma_reserve unnecessarily leading to additional memory consumption.

  [Test Case]
  Boot a KVM guest on PowerNV and measure the differences in memory consumption with and without this patch.

  [Fix]
  commit cec26bc3c125b5dd12a02f04133cd91eae3f1622 upstream
  in v3.18-rc1

  --

  Problem Description
  ==================
  This bugzilla is opened to track the inclusion of the upstream commit mentioned below:

  commit cec26bc3c125b5dd12a02f04133cd91eae3f1622
  Author: Aneesh Kumar K.V <aneesh.kumar@xxxxxxxxxxxxxxxxxx>
  Date:   Mon Sep 29 13:32:38 2014 +0530

      KVM: PPC: BOOK3S: HV: CMA: Reserve cma region only in hypervisor
  mode

      We use cma reserved area for creating guest hash page table.
      Don't do the reservation in non-hypervisor mode. This avoids unnecessary
      CMA reservation when booting with limited memory configs like
      fadump and kdump.

  ---uname output---
  ubuntu14.10

  Machine Type = ppc64

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1400209/+subscriptions