[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: System becomes very slow after update to squeeze 6.0.1

Hi Andu,

sorry for reply so later.

於 日，2011-04-03 於 22:33 +0300，Alexandru IANCU 提到：
Hi Jackie,

the buggy package is: linux-image-2.6.32-5-amd64, Version 2.6.32-31, with description:
  [ Ian Campbell ]
  * xen: blkback: fix potential leak of kernel thread. (CVE-2010-3699)

  [ Moritz Muehlenhoff ]
  * rds: Fix rds_iovec page count overflow (CVE-2010-3865)

  [ Ben Hutchings ]
  * tty: Fix information leaks from SIOCGICOUNT handlers (CVE-2010-4075,
    CVE-2010-4076, CVE-2010-4077)
  * bonding: Ensure that we unshare skbs prior to calling pskb_may_pull
    (Closes: #610838)
  * r8169: Keep firmware in memory (Closes: #609538)
  * linux-base: Convert LILO entries for /boot/vmlinuz, /boot/vmlinuz.old
    (Closes: #613200)
  * aufs: Fix VM race leading to kernel panic (Closes: #607879)
  * rt2500usb: Fall back to SW encryption for TKIP+AES (Closes: #611390)
  * Add longterm 2.6.32.29:
    - SCSI: Fix medium error problems with some arrays which can cause
      data corruption
    - ptrace: Use safer wake up on ptrace_detach()
    - [x86] mm: Avoid possible bogus TLB entries by clearing prev
      mm_cpumask after switching mm
    - sched: Fix softirq time accounting
    - sched: Use group weight, idle cpu metrics to fix imbalances during
      idle
    - [openvz,vserver] Revert sched changes since they conflict
  * Revert "USB: Prevent buggy hubs from crashing the USB stack", included
    in longterm 2.6.32.29 and reported to cause a regression
  * virtio_net: Further fixes for out-of-memory conditions (Closes: #603835)
    - Fix OOM handling on TX
    - Add schedule check to napi_enable call
  * af_unix: Limit recursion level of passing sockets through sockets
    (variant of CVE-2010-4249)
  * iowarrior: Don't trust report_size for buffer size (CVE-2010-4656)
  * drm: Fix unsigned vs signed comparison issue in modeset ctl ioctl
    (CVE-2011-1013)
  * brcm80211: Fix suspend/resume in brcmsmac (Closes: #600769, #604802)
  * brcm80211: Fix race between scanning and calibration on SMP
    (Closes: 602444)
  * drm/i915: Overlay on gen2 can't address above 1G
  * drm/i915: Fix memory corruption with GM965 and >4GB RAM
  * ipv6: Silence privacy extensions initialization (Closes: #590653)
  * [x86] Enable VT6656, loading firmware from a separate file (requires
    firmware-linux-nonfree 0.28+squeeze1) (Closes: #568454)
  * usbfs: Show correct speed for SuperSpeed USB devices (Closes: #613531)
  * drm/i915: Add pipe A force quirk for some laptops (Closes: #608148)
  * psmouse/elantech: Fix detection and decoding for newer Elantech
    touchpads (Closes: #613335)

  [ dann frazier ]
  * xfs: Fix information leak using stale NFS handle (CVE-2010-2943)
  * CAN: Use inode instead of kernel address for /proc file (CVE-2010-4565)

  [ maximilian attems]
  * Update openvz patch to feoktistov (ipv6, checkpointing, stability,
    ipsec, nfs, ppp, tc, ve). (closes: #607041, #613501, #613170)
  * HID: add support for Acan FG-8100 barcode reader. (closes: #615888)
  * Add longterm 2.6.32.30:
    - nfsd: Memory corruption due to writing beyond the stat array
    - av7110: check for negative array offset (CVE-2011-0521)
    - cred: Fix get_task_cred() and task_state() to not resurrect dead
      credentials
    - cred: Fix kernel panic upon security_file_alloc() failure
    - cred: Fix BUG() upon security_cred_alloc_blank() failure
    - cred: Fix memory and refcount leaks upon security_prepare_creds()
      failure
    - dm/raid1: Fail writes if errors are not handled and log fails
    - GFS2: Fix bmap allocation corner-case bug
    - [s390] remove task_show_regs (CVE-2011-0710)
    - PM/hibernate: Return error code when alloc_image_page() fails
    - fs/partitions: Validate map_count in Mac partition tables
      (CVE-2011-1010)
    - ALSA: caiaq - Fix possible string-buffer overflow (CVE-2011-0712)
    - acer-wmi, asus_acpi, tc1100-wmi: Restrict write permissions on files
      in procfs/sysfs
    - [x86] usbip/vhci: Update reference count for usb_device
    - [x86] usbip/vhci: Give back URBs from in-flight unlink requests
    - [x86] usbip/vhci: Refuse to enqueue for dead connections
    - epoll: Prevent creating circular epoll structures
    - fs/partitions/ldm: Corrupted partition table can cause kernel oops
    - xhci: Avoid BUG() in interrupt context
    - xhci: Fix errors in the running total calculations in the TRB math
    - xhci: Fix an error in count_sg_trbs_needed()
    - x25: Do not reference freed memory
  * Add longterm 2.6.32.31.
  * Add longterm 2.6.32.32-rc1:
    - netxen: fix set mac addr. (closes: #616058)
    - [xen] do not release any memory under 1M in domain 0. (closes: #613823)
    - virtio: set pci bus master enable bit. (closes: #610360)
    - sctp: Fix oops when sending queued ASCONF chunks (CVE-2010-1173).
    - drm/ttm: Fix two race conditions + fix busy codepaths (closes: #591061)
  * Add Slovak translation by Slavko. (closes: #608684)
  * Add drm changes from 2.6.32.28+drm33.13:
    - drm/i915: Add dependency on CONFIG_TMPFS.
    - drm/i915/lvds: Add AOpen i915GMm-HFS to the list of false-positive LVDS.
    - drm/radeon/kms: add pll debugging output.
    - drm/radeon/kms: add quirk for Mac Radeon HD 2600 card.
    - drm/radeon/kms: fix s/r issues with bios scratch regs.
    - drm/radeon/kms: make the mac rv630 quirk generic.
    - drm/radeon: remove 0x4243 pci id.

  [ Aurelien Jarno ]
  * init: fix race between init and kthreadd, fixes a kernel panic on
    mips/5kc-malta.
  * mips/swarm: enable PATA drivers that have been lost during IDE -> PATA
    conversion.

  [ Martin Michlmayr ]
  * Orion: add support for Buffalo LS-CHL (Closes: #590105).
  * Kirkwood: initialize PCIE1 for QNAP TS-419P+ (Closes: #613499).

  [ Jurij Smakov ]
  * sparc: add sparc-console-handover.patch to address problems with
    console handover on sparc causing kernel to hang during boot on
    systems using atyfb driver.
    Thanks to Fabio M. Di Nitto <fabbione@fabbione.net> for the patch.
    (Closes: #602853)

  [ Bastian Blank ]
  * Add supportt for AMD Family 10h/11h CPU internal temperatur sensor.
    (closes: #614555)

Squeeze 6.0.0 comes with 2.6.32-30 of this package.

So to conclude:
1. If you installed Squeeze 6.0.1(a) you can fix your system by restoring /boot folder. Can you make a copy of those files from a 6.0.0 install ? If not, I can give you a couple of binaries but it's not wise running binaries from strangers!
i'm afraid that i can't restore from 6.0.0, i do a fresh install from 6.0.1a.
2. I don't know how to roll-back to a previous package version. Maybe you should start a new thread here.
me too. i've downloaded a testing DVD, i'll try later.
3. We need to speak with linux-image's package maintainer and tell him/her this package ruins our grannie's laptop. Do you know where to write an email ?
sorry, i dont know, maybe the development maillist can help?


Best,
Andu.

On Fri, Apr 1, 2011 at 10:00 AM, Jackie Wang <jackie2004.wang@gmail.com> wrote:
於 四，2011-03-31 於 22:21 +0300，Alexandru IANCU 提到：

Hi Jackie,

so you can reproduce this by reinstalling and upgrading.
More useful than lsmod, lspci and w commands would be to upgrade package by package and restart after each install to see what package slows down the computer.

Cheers,
Andu.
Hi Andu,

Yes, I guess it's related Driver or FW, because i use back to the old kernel, is seems normal. maybe the new kernel get rid of non-free or contrib driver/fw. i dont know. some days later, i'll try to reinstall the newest testing debian.




Reply to: