Swap Corruption in 2.4.3 ?

Swap Corruption in 2.4.3 ?

[Richard Russon]

Hi all,

When I unmount the swapfile I get:

VM: Undead swap entry 000bb300
VM: Undead swap entry 00abb300
VM: Undead swap entry 016fb300

I can repeat this fairly reliably.  First fill up the conventional
memory, and a few 10s of Mb of swap, then put the machine under a bit
of a load.  It may not be the loading that's causing this, just the
movement in swap.

First things stop working, e.g. shared libraries appear to be missing
or corrupt, every program segfaults on startup, occasionally the whole
machines hangs.  I have seen Oopses but if memory is being corrupted I
doubt they'll be any use.

If I reboot, everything is fine again.

The machine is a 800Mhz PIII, all IDE, all ext2, not overclocked, with a
vanilla 2.4.3 kernel (compiled with gcc-2.91.66).

I can reproduce this with DMA turned off.
I cannot reproduce this with the swap file off, however hard I try.
If the swapfile isn't touched the machine seems stable under load.

Below is my .config (if it's useful)

Can anyone think of anything else I can test to give you any pointers?

Cheers,
  FlatCap (Richard Russon)
  kernel@flatcap.org

CONFIG_X86=y
CONFIG_ISA=y
CONFIG_UID16=y
CONFIG_EXPERIMENTAL=y
CONFIG_MODULES=y
CONFIG_MODVERSIONS=y
CONFIG_KMOD=y
CONFIG_MPENTIUMIII=y
CONFIG_X86_WP_WORKS_OK=y
CONFIG_X86_INVLPG=y
CONFIG_X86_CMPXCHG=y
CONFIG_X86_BSWAP=y
CONFIG_X86_POPAD_OK=y
CONFIG_X86_L1_CACHE_SHIFT=5
CONFIG_X86_TSC=y
CONFIG_X86_GOOD_APIC=y
CONFIG_X86_PGE=y
CONFIG_X86_USE_PPRO_CHECKSUM=y
CONFIG_NOHIGHMEM=y
CONFIG_MTRR=y
CONFIG_X86_UP_IOAPIC=y
CONFIG_X86_IO_APIC=y
CONFIG_X86_LOCAL_APIC=y
CONFIG_NET=y
CONFIG_PCI=y
CONFIG_PCI_GOANY=y
CONFIG_PCI_BIOS=y
CONFIG_PCI_DIRECT=y
CONFIG_PCI_NAMES=y
CONFIG_SYSVIPC=y
CONFIG_SYSCTL=y
CONFIG_KCORE_ELF=y
CONFIG_BINFMT_ELF=y
CONFIG_PARPORT=y
CONFIG_PARPORT_PC=y
CONFIG_PNP=y
CONFIG_ISAPNP=y
CONFIG_BLK_DEV_FD=y
CONFIG_BLK_DEV_RAM_SIZE=4096
CONFIG_PACKET=y
CONFIG_NETLINK=y
CONFIG_UNIX=y
CONFIG_INET=y
CONFIG_IDE=y
CONFIG_BLK_DEV_IDE=y
CONFIG_BLK_DEV_IDEDISK=y
CONFIG_BLK_DEV_CMD640=y
CONFIG_BLK_DEV_RZ1000=y
CONFIG_BLK_DEV_IDEPCI=y
CONFIG_IDEPCI_SHARE_IRQ=y
CONFIG_BLK_DEV_IDEDMA_PCI=y
CONFIG_IDEDMA_PCI_AUTO=y
CONFIG_BLK_DEV_IDEDMA=y
CONFIG_IDEDMA_AUTO=y
CONFIG_BLK_DEV_IDE_MODES=y
CONFIG_SD_EXTRA_DEVS=40
CONFIG_SR_EXTRA_DEVS=2
CONFIG_SCSI_DEBUG_QUEUES=y
CONFIG_SCSI_MULTI_LUN=y
CONFIG_NETDEVICES=y
CONFIG_NET_ETHERNET=y
CONFIG_NET_PCI=y
CONFIG_VT=y
CONFIG_VT_CONSOLE=y
CONFIG_SERIAL=y
CONFIG_UNIX98_PTYS=y
CONFIG_UNIX98_PTY_COUNT=256
CONFIG_MOUSE=y
CONFIG_PSMOUSE=y
CONFIG_AGP_SIS=y
CONFIG_AUTOFS4_FS=y
CONFIG_JOLIET=y
CONFIG_PROC_FS=y
CONFIG_DEVPTS_FS=y
CONFIG_EXT2_FS=y
CONFIG_NFS_V3=y
CONFIG_NFSD_V3=y
CONFIG_LOCKD_V4=y
CONFIG_SMB_NLS_DEFAULT=y
CONFIG_SMB_NLS_REMOTE="cp437"
CONFIG_MSDOS_PARTITION=y
CONFIG_SMB_NLS=y
CONFIG_NLS=y
CONFIG_NLS_DEFAULT="iso8859-1"
CONFIG_NLS_CODEPAGE_437=y
CONFIG_NLS_CODEPAGE_850=y
CONFIG_NLS_ISO8859_1=y
CONFIG_NLS_ISO8859_15=y
CONFIG_VGA_CONSOLE=y
CONFIG_VIDEO_SELECT=y
CONFIG_SOUND=y
CONFIG_BLK_DEV_IDECD=m
CONFIG_8139TOO=m
CONFIG_NE2K_PCI=m
CONFIG_BLK_DEV_IDECD=m
CONFIG_FAT_FS=m
CONFIG_VFAT_FS=m
CONFIG_SOUND_ES1371=m

Re: Swap Corruption in 2.4.3 ?

[Rik van Riel]

On 10 Apr 2001, Richard Russon wrote:

> VM: Undead swap entry 000bb300
> VM: Undead swap entry 00abb300
> VM: Undead swap entry 016fb300

Known bug ... unknown cause ;(

http://www.linux-mm.org/bugzilla.shtml has it already listed

regards,

Rik
--
Virtual memory is like a game you can't win;
However, without VM there's truly nothing to lose...

		http://www.surriel.com/
http://www.conectiva.com/	http://distro.conectiva.com.br/

Re: Swap Corruption in 2.4.3 ?

[Ishikawa]

On 10 Apr 2001, Richard Russon wrote:

> VM: Undead swap entry 000bb300
> VM: Undead swap entry 00abb300
> VM: Undead swap entry 016fb300

I have seen similar mysterious crashes of X server
when I began accessing large web page using Netscape
navigator. It was reproducible the first few times I noticed.

Today, a similar problem occurred and luckily I noticed
console messages printed on a virtual console.
There ware many lines like the following.

swap_free: Trying to freee non-existent swap page
Bad Swap Entry: 00000008
Bad Swap Entry: 0000008
    ...
VM: Kiling process jserver
Swap_free: Trying to free non-existent swap  page
   ...

It seems that there was some sort of kernel data corruption.
(In my case, the kernel could not find the swap page and
failed to swap out memory and thus the
VM's selective process killer was invoked to free up memory?)

Then eventually I got
Unable to handle kernel NULL pointer defreence at ...
message and the system crashed.
(I was also typing Alt+SysReq+keystroke to see if I could sync disk.)
.

Anyway, this probem seems to be in 2.4.2, too from what I recall about
the
first mysterious X server crash.

On Apri 20 Rik van Riel wrote:
>Known bug ... unknown cause ;(
>
>http://www.linux-mm.org/bugzilla.shtml has it already listed

The symptom there didn't match mine, but I do suspect
a (new) problem in VM of 2.4.3 (2.4.2, too. 2.4.1, I am not sure.).

Happy Hacking