UPDATE: GP in proc_lookupfd with pre-patch-2.0.31-10

Tomasz Motylewski (motyl@stan.chemie.unibas.ch)
Tue, 14 Oct 1997 22:47:50 +0200 (MET DST)


UPDATE: when I have reported this error, it looked like it could be
attributed to the "Linux non executable stack" patch. But now I know that it
was noticed also in upatched 2.0.29 and after installing 2.0.30 with
pre-patch-2.0.31-10 I am still getting it with:

Oct 13 17:42:45 uran kernel: Linux version 2.0.31 (root@crds.chemie.unibas.ch) (gcc version 2.7.2.1.f.1) #1 Sat Oct 11 23:00:31 MET DST 1997
Oct 13 17:42:45 uran kernel: *** Installed patches:
Oct 13 17:42:45 uran kernel: Linux Unofficial Patch Printout (2.0.30)
Oct 13 17:42:45 uran kernel: pre-patch-2.0.31-10 (Linus) http://www.kernel.org/pub/linux/kernel/testing/
Oct 13 17:42:45 uran kernel: *** End of Patches
Oct 13 17:42:45 uran kernel: Serial driver version 4.13 with no serial options enabled
Oct 13 17:42:45 uran kernel: tty00 at 0x03f8 (irq = 4) is a 16550A
Oct 13 17:42:45 uran kernel: tty01 at 0x02f8 (irq = 3) is a 16550A
Oct 13 17:42:45 uran kernel: Real Time Clock Driver v1.07
Oct 13 17:42:45 uran kernel: loop: registered device at major 7
Oct 13 17:42:45 uran kernel: ide: i82371 PIIX (Triton) on PCI bus 0 function 57
Oct 13 17:42:45 uran kernel: ide0: BM-DMA at 0xf000-0xf007
Oct 13 17:42:45 uran kernel: ide1: BM-DMA at 0xf008-0xf00f
Oct 13 17:42:45 uran kernel: hda: WDC AC2850F, 814MB w/64kB Cache, CHS=827/32/63
Oct 13 17:42:45 uran kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Oct 13 17:42:45 uran kernel: Floppy drive(s): fd0 is 1.44M
Oct 13 17:42:45 uran kernel: FDC 0 is a post-1991 82077
Oct 13 17:42:45 uran kernel: md driver 0.35 MAX_MD_DEV=4, MAX_REAL=8
Oct 13 17:42:45 uran kernel: tunnel: version v0.2b2
Oct 13 17:42:45 uran kernel: PPP: version 2.2.0 (dynamic channel allocation)
Oct 13 17:42:45 uran kernel: TCP compression code copyright 1989 Regents of the University of California
Oct 13 17:42:45 uran kernel: PPP Dynamic channel allocation code copyright 1995 Caldera, Inc.
Oct 13 17:42:45 uran kernel: PPP line discipline registered.
Oct 13 17:42:45 uran kernel: SLIP: version 0.8.4-NET3.019-NEWTTY (dynamic channels, max=256) (6 bit encapsulation enabled).
Oct 13 17:42:45 uran kernel: CSLIP: code copyright 1989 Regents of the University of California.
Oct 13 17:42:45 uran kernel: SLIP linefill/keepalive option.
Oct 13 17:42:45 uran kernel: ne.c:v1.10 9/23/94 Donald Becker (becker@cesdis.gsfc.nasa.gov)
Oct 13 17:42:45 uran kernel: NE*000 ethercard probe at 0x6100: 00 40 05 3b 17 11
Oct 13 17:42:45 uran kernel: eth0: NE2000 found at 0x6100, using IRQ 11.
Oct 13 17:42:45 uran kernel: Partition check:
Oct 13 17:42:45 uran kernel: hda: hda1 hda2 < hda5 hda6 > hda3 hda4
Oct 13 17:42:45 uran kernel: VFS: Mounted root (ext2 filesystem) readonly.
Oct 13 17:42:45 uran kernel: Adding Swap: 31244k swap-space (priority -1)

here goes "/usr/sbin/hdparm -c1 -m16 -d1 -X34 /dev/hda" in rc.local

Oct 13 17:42:56 uran kernel: hda: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
Oct 13 17:42:56 uran kernel: hda: drive_cmd: error=0x04 { DriveStatusError }
Oct 13 17:47:30 uran kernel: registered device ppp0
Oct 13 18:06:34 uran kernel: registered device ppp1
Oct 13 21:22:11 uran kernel: registered device ppp2
Oct 14 01:50:03 uran kernel: general protection: 0000
Oct 14 01:50:03 uran kernel: CPU: 0
Oct 14 01:50:03 uran kernel: EIP: 0010:[proc_lookupfd+360/436]
Oct 14 01:50:03 uran kernel: EFLAGS: 00010286
Oct 14 01:50:03 uran kernel: eax: f000ef6f ebx: ffffffff ecx: 00000002 edx: 00b5e810
Oct 14 01:50:03 uran kernel: esi: 0066e004 edi: 016dc9d8 ebp: 00001e16 esp: 01b71f14
Oct 14 01:50:03 uran kernel: ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
Oct 14 01:50:03 uran kernel: Process fuser (pid: 7700, process nr: 48, stackpage=01b71000)
Oct 14 01:50:03 uran kernel: Stack: 016dc9d8 01b71f70 01b71f70 00000001 001e3724 0012a766 016dc9d8 0066e004
Oct 14 01:50:03 uran kernel: 00000001 01b71f70 00000000 01b71f70 01b71fb4 bfffed38 00000000 0012a97f
Oct 14 01:50:03 uran kernel: 016dc9d8 0066e003 00000001 01b71f70 00000000 01b71fb4 bfffed60 00000000
Oct 14 01:50:03 uran kernel: Call Trace: [lookup+222/248] [_namei+91/228] [lnamei+48/72] [sys_readlink+63/180] [system_call+85/124]
Oct 14 01:50:03 uran kernel: Code: 83 78 34 00 0f 84 b8 fe ff ff 89 e8 c1 e0 10 8d 84 08 00 01
Oct 14 02:10:02 uran kernel: general protection: 0000
Oct 14 02:10:02 uran kernel: CPU: 0
Oct 14 02:10:02 uran kernel: EIP: 0010:[proc_lookupfd+360/436]
Oct 14 02:10:02 uran kernel: EFLAGS: 00010286
Oct 14 02:10:02 uran kernel: eax: f000ef6f ebx: ffffffff ecx: 00000002 edx: 00ddd810
Oct 14 02:10:02 uran kernel: esi: 01add004 edi: 01fbcbd0 ebp: 00001f06 esp: 01711f14
Oct 14 02:10:02 uran kernel: ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
Oct 14 02:10:02 uran kernel: Process fuser (pid: 7946, process nr: 44, stackpage=01711000)
Oct 14 02:10:02 uran kernel: Stack: 01fbcbd0 01711f70 01711f70 00000001 001e3724 0012a766 01fbcbd0 01add004
Oct 14 02:10:02 uran kernel: 00000001 01711f70 00000000 01711f70 01711fb4 bfffed38 00000000 0012a97f
Oct 14 02:10:02 uran kernel: 01fbcbd0 01add003 00000001 01711f70 00000000 01711fb4 bfffed60 00000000
Oct 14 02:10:02 uran kernel: Call Trace: [lookup+222/248] [_namei+91/228] [lnamei+48/72] [sys_readlink+63/180] [system_call+85/124]
Oct 14 02:10:02 uran kernel: Code: 83 78 34 00 0f 84 b8 fe ff ff 89 e8 c1 e0 10 8d 84 08 00 01

This is a dial-in PPP, mail, HTTP, DNS server using 2 "fourport" serial
cards.

uran:~# cat /proc/cpuinfo
processor : 0
cpu : 586
model : 0
vendor_id : AuthenticAMD
stepping : 1
fdiv_bug : no
hlt_bug : no
fpu : yes
fpu_exception : yes
cpuid : yes
wp : yes
flags : fpu vme de pse tsc msr mce cx8 apic
bogomips : 179.40

On Mon, 22 Sep 1997, Tomasz Motylewski wrote:

> I am getting the GP with aprox 10% probability when executing the
> following command from cron:
>
> */10 * * * * if /usr/bin/fuser -s /tmp/.pop/* 2>/dev/null ; then :; else /bin/rm -rf /tmp/.pop/* 2>/dev/null ; fi
>
>
> Sep 22 07:30:02 uran kernel: general protection: 0000
> Sep 22 07:30:02 uran kernel: CPU: 0
> Sep 22 07:30:02 uran kernel: EIP: 0010:[proc_lookupfd+360/436]
> Sep 22 07:30:02 uran kernel: EFLAGS: 00010286
> Sep 22 07:30:02 uran kernel: eax: f000ef6f ebx: ffffffff ecx: 00000002 edx: 01864414
> Sep 22 07:30:02 uran kernel: esi: 00ff7004 edi: 014d75e8 ebp: 0000590c esp: 01093f14
> Sep 22 07:30:02 uran kernel: ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
> Sep 22 07:30:02 uran kernel: Process fuser (pid: 22798, process nr: 52, stackpage=01093000)
> Sep 22 07:30:02 uran kernel: Stack: 014d75e8 01093f70 01093f70 00000001 001e39d8 0012a902 014d75e8 00ff7004
> Sep 22 07:30:02 uran kernel: 00000001 01093f70 00000000 01093f70 01093fb4 bfffed24 00000000 0012ab1b
> Sep 22 07:30:02 uran kernel: 014d75e8 00ff7003 00000001 01093f70 00000000 01093fb4 bfffed4c 00000000
> Sep 22 07:30:02 uran kernel: Call Trace: [lookup+222/248] [_namei+91/228] [lnamei+48/72] [sys_readlink+63/180] [system_call+85/124] [pipe_write+216/516]
> Sep 22 07:30:02 uran kernel: Code: 83 78 30 00 0f 84 b8 fe ff ff 89 e8 c1 e0 10 8d 84 08 00 01
>
>
> I get this with both 2.0.29 with applied linux-2.0.30-stack-symlink.diff
>
> and 2.0.pre31-9 with:
>
> Sep 17 20:52:11 uran kernel: Linux version 2.0.31 (root@crds.chemie.unibas.ch) (gcc version 2.7.2.1.f.1) #6 Wed Sep 17 15:38:35 MET DST 1997
> Sep 17 20:52:11 uran kernel: *** Installed patches:
> Sep 17 20:52:11 uran kernel: aic7xxx-abort-Sep05.patch (Doug Ledford <dledford@dialnet.net>)
> Sep 17 20:52:11 uran kernel: awedrv-0.4.1d
> Sep 17 20:52:11 uran kernel: commands_sent.ledford.Sep09.advice if (p->device_status[TARGET_INDEX(cmd)].commands_sent < 50)
> Sep 17 20:52:11 uran kernel: interrupt.timeout.zubkoff.Sep09.patch (Leonard Zubkoff)
> Sep 17 20:52:11 uran kernel: linux-2.0.30-stack-symlink.diff (Solar+)
> Sep 17 20:52:11 uran kernel: Linux Unofficial Patch Printout (2.0.30)
> Sep 17 20:52:11 uran kernel: pre-patch-2.0.31-9 (Linus)
> Sep 17 20:52:11 uran kernel: test.1.werner.patch (dr. Werner Fink), Sep 09 1997
> Sep 17 20:52:11 uran kernel: *** End of Patches
>
> I have tried the second kernel with stack-nonexec on and (symlink-patch on
> and off). But the problem may be with hardlink protection which is not
> configurable in .config . Trampoline detection was disabled. The system
> works properly with plain 2.0.29.

The last sentence was not true.

> After the GP the system continues to run (end generate GPs) for days.
>

--
Tomasz Motylewski