[coLinux-devel] [ colinux-Bugs-1745597 ] BSOD on Windows Server 2003 with Adaptec AIC-9580W RAID Ctrl

[SourceForge.net <no...@so...>]

Bugs item #1745597, was opened at 2007-06-29 23:23
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=1745597&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Closed
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: BSOD on Windows Server 2003 with Adaptec AIC-9580W RAID Ctrl

Initial Comment:
I am trying to install coLinux-0.6.4.exe on a system that is running Windows server 2003 and has a Adaptec AIC-9580W RAID controller (aka ServeRAID 8k-l or 8k) installed on it.  The installation gets about 98% complete and then it gives the BSOD (blus screen of death) with the STOP: 0x1111777A (0x0, 0x0, 0x0, 0x0) code at the bottom.

I have been able to install colinux-0.6.4 on systems that are running Windows Server 2003, but do not have the RAID Controller HW.

I was also able to install colinux-0.6.4 on a system that is running Windows 2000 and has a Adaptec AIC-9580W RAID controller (aka ServeRAID 8k-l or 8k) installed on it.

I have tired various configuratoins in the boot.ini file (i.e. /NOEXECUTE, /APE, /NOAPE etc), but the BSOD still shows up.

I have also tried coLinux development version 0.8.0.  This version installs the linux.sys driver (using the colinux-daemon --install-driver commmand) without giving the BSOD, but when coLinux-daemon.exe is executed the BSOD shows up with the following message:
IRQL_NOT_LESS_OR_EQUAL

Here is some system information on the machine that has Windows Server 2003 and the Adaptec AIC-9580W RAID controller (aka ServeRAID 8k-l or 8k) installed on it:
a two socket server utilizing Intel Xeon 5000 series processors
Intel 5000X chipset
Adaptec AIC-9580W RAID controller (aka ServeRAID 8k-l or 8k)
Max memory supported is 32GB.


Anyone have any suggestions as to what the cause of the BSOD is?

Thank you



----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-05-13 15:00

Message:
Does the latest build (0.8.0) able to run under Window 2003 Server R2 SP2
with more than 4GB??

----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2007-09-24 18:46

Message:
Logged In: YES 
user_id=579204
Originator: NO

ok, was an older colinux driver without versions checking. Closed bug now.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-08-21 16:12

Message:
Logged In: NO 


Solved!

one uppon a time it happens that there was a linux.sys which 
did not run with mem=128, running with mem=126 was OK.
That was 0.6.1 or 0.6.2!
And that linux.sys was installed on the box which BSODs.

After deinstalling:
No problems any more!
Thanks a lot!


But i am pretty sure i started the colinux 0.7.1 daemon from the
pen drive. Why didn't it complain about that old driver?
Was the BSOD faster?



Some infos over the old driver:

hardware wizzard shows as details only:
ROOT\LEGACY_COLINUXDRIVER\0000

dir:

29.05.2004  16:59           237.649 coLinux-0.6.1.tar.gz
05.02.2005  10:57           422.441 colinux-0.6.2.tar.gz
29.05.2004  15:20           274.784 colinux-bridged-net-daemon.exe
29.05.2004  15:20           713.682 colinux-console-fltk.exe
29.05.2004  15:20           402.411 colinux-console-nt.exe
29.05.2004  15:20           338.449 colinux-daemon.exe
29.05.2004  15:20           274.032 colinux-net-daemon.exe
14.07.2004  00:54        62.146.666 Debian-20040605-mit.ext3.1610mb.bz2
09.06.2005  15:18        18.329.320 Debian-3.0r0.ext3.1gb.bz2
10.11.2004  17:05     1.073.741.824 Debian-3.0r2.ext3-mit-backports.1gb
27.06.2004  00:36        22.449.829
Debian-3.0r2.ext3-mit-backports.1gb.bz2
25.04.2004  16:35             1.112 default.colinux.xml
10.11.2004  17:03             1.247 default.colinux.xml.$$$
10.11.2004  17:04             1.256 default.colinux.xml.old
29.05.2004  15:20            94.327 linux.sys
09.06.2005  15:10    <DIR>          netdriver
20.04.2004  06:43             5.014 README
09.06.2005  15:10            60.313 Uninstall.exe
29.05.2004  14:32         1.822.230 vmlinux
29.05.2004  15:20           641.763 vmlinux-modules.tar.gz

After deinstalling the old colinux and install 0.8.0 i get,
when started from pendrive:

Cooperative Linux Daemon, 0.7.1
Compiled on Sat Jul 14 12:15:18 2007

using 'vmlinux' as kernel image
configuring 128 MB of virtual RAM
mapping cobd4 to \Device\Cdrom0
mapping cobd1 to \??\E:\colinux.disks\swap_file
mapping cobd0 to \??\E:\colinux.disks\Debian-3.0r2.ext3-mit-backports.1gb
mapping hdc to cobd4
mapping hdb to cobd1
mapping hda to cobd0
configured Slirp as eth0
redirections tcp:15022:22/tcp:15080:80/tcp:5901:5901
mapping cofs0 to \??\E:\colinux.disks\exchange
using 'initrd.gz' as initrd image
mapping cofs31 to \??\E:\coLinux
kernel boot parameters: 'root=/dev/cobd0'

colinux: manager open: last error: Das System kann die angegebene Datei
nicht finden.
daemon: exit code 8320a80e
daemon: can't access CoLinuxDriver, please check status driver!
back to "E:\coLinux\colinuxstart.cmd"...


So i it is still a dream to run a "linux-in-a-box" 
without an installed driver or admin rights.
Mark Russinovich (wininterals etc.) one made that trick 
with his "filemon" IIRC.
 


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2007-08-21 08:56

Message:
Logged In: YES 
user_id=579204
Originator: NO

There exist an variable for total ram used by coLinux (your "mem=xx") This
variable exist in a struct on daemon and will transfer to the driver
linux.sys, this calls than the malloc. (has to do with
co_os_physical_memory_pages() and the automatic down rounder for variable
"hostmem_used")

Very old coLinux has no version check (think before 0.6.4, but not
remember).
Current Daemon first checks the version of linux.sys and don't start.

You can start coLinux from a USB stick. But you need to load ones the
driver.
After you have the driver updated, you can run without admin rights.

If you say, this machine has never been seen coLinux before, then the
daemon would not start.

Check the hardware device manager and enable "hidden devices", then you
can see the colinux driver, under "non PnP drivers". This scren shows the
way howto find the driver by mouse clicks:
http://www.henrynestler.com/colinux/screenshoots/driver-details-grayed.png
Remove this driver and restart windows. Or run "colinux-daemon.exe
--remove-driver" (do the same without reboot).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-08-20 22:42

Message:
Logged In: NO 


>>Comment By: Henry N. (henryn)
>Date: 2007-08-20 11:58

>Message:
>Logged In: YES
>user_id=579204
>Originator: NO

>Hello,

>can't resolve the stack to source, because I can't find a matching
>binary linux.sys from there you got the BSOD.

I used the normal linux.sys from the windows install.
I added file sizes and dates.


>Your address "linux+52e6" is inside the
>function co_os_malloc,
>and the condition for this bug is a call "co_os_malloc(0)".

Jepp. i already analyzed the sources but could not see any place 
where co_os_malloc etc. could(!) be called with a "0". At all locations 
(i found) that param was a compile time constant value excpt one. 
So it currently sounds like an ugly pointer error or bad case/table jump.

>Are you shure, that you have called
>"colinux-daemon.exe --remove-driver"
>and "colinux-daemon.exe --install-driver"
>between the change of coLinux versions?

Hm. No. because i assumed that the machine never saw colinux before.
And i was not aware that this could be a problem at all because
I spawned colinux from a 4GB USB-Drive without any install
because in wanted to test it without admin rights...
I was very surprised to BSOD a box (without admin rights, i thought), 
but if an old installed driver was used, a BSOD is possible. 
Would colinux-daemon.exe use any installed linux.sys
without any version check?
That would explain the problem.
I'll have a look if a (very old) linux.sys was installed 
long time ago tomorrow.


>Replacing the file linux.sys and the exe daemons in the directory
>are not enough.

I thought it would, because i did not install it.
Maybe some else did. My intention was do have "my own linux on a stick"
but if it uses the installed stone aged linux.sys it could always crash.




----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2007-08-20 09:58

Message:
Logged In: YES 
user_id=579204
Originator: NO

Hello,

can't resolve the stack to source, because I can't find a matching binary
linux.sys from there you got the BSOD.

Your address "linux+52e6" is inside the function co_os_malloc, and the
condition for this bug is a call "co_os_malloc(0)".
See line 63:
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/src/colinux/os/winnt/kernel/lowlevel/alloc.c?view=markup

The 1111777A exist only ones in the source. So your address "linux+52e6"
should match an address ones behind the BugCheck-call.

> ef6d92db 687a771111      push    1111777Ah      <-- BugCheck code
> ef6d92e0 ff1518e16def    call    dword ptr [linux+0xa118 (ef6de118)] 
<-- calls KeBugCheck
> ef6d92e6    <--- this is the sddress you see in stack trace

In your disassemble I can see. But, I have no this file, you used.

We must use exactly the same linux.sys for analysis from where you got the
bug. I see different build date and file size for linux.sys. Nothing of my
known files matched with your disassembler output. Are you shure, that you
have called "colinux-daemon.exe --remove-driver" and "colinux-daemon.exe
--install-driver" between the change of coLinux versions? Replacing the
file linux.sys and the exe daemons in the directory are not enough.

I have added more informations to the bugcheck 1111777A to see the caller.
Please do a clean install from
http://www.henrynestler.com/colinux/testing/devel-0.8.0/20070819-Snapshot/

I see, you have WinDbg installed. This is very good. You can send me the
output from "kd> !analyze -v" via Mail. In "debugging.txt" have added the
way howto resolve the symbols from linux.sys by hands. Hope, this helps.

Henry

----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2007-08-08 08:26

Message:
Logged In: YES 
user_id=579204
Originator: NO

BSOD 0x1111777A can have more as one initial errors (more as the known and
fixed call).

Many thanks for the 'STACK_TEXT:', can see the linux calls. Will check
it.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-08-07 15:00

Message:
Logged In: NO 

Sorry, I forgot to emphesis:

I already tried the "developer driver" as the file date "2007-07-14"
indicates.
So:
BSOD 0x111177A is not fixed in 0.71

Have not tried 0.8 because there is no zip where i can take the driver
from. But as 0.8 and 0.71 has the same patches...

Have tried Debug version of 0.71, but the symbol information is not
acepted by kernel debugger and thatverion ist a bit older.
Debug version 0.7.1 (build on 26 March 2007) still show the BSOD.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-08-07 14:51

Message:
Logged In: NO 


linux.sys BSDOD 0x111177A

coLinux-0.7.1
Internal Version: 1.1.0.0
linux.sys 69120 byte  2007-07-14 14:15
Debug: linux.sys  139.730  139.730 2007-03-09 02:42 (But symbols are not
usable)


Starting gives BSOD on an normal Office PC running 
"Windows XP 2002 SP2 Celeron 2,80GHz, 504MB RAM" 
512MiB RAM, 2GiB Swap
Trendmicro Virus scanner
"Slirp" network

So that's no problem of "too much memory", i assume.


Invoked by from USB flash drive by:
set ETH=eth0=slirp,,tcp:5022:22
set WINSHARE=cofs0=\colinux.disks\exchange
set MEM=128
set CDROM0=cobd4=\Device\Cdrom0 hdc=:cobd4

set BOOTDEV=cobd0
set ROOTFS=%BOOTDEV%=\colinux.disks\Debian-3.0r2.ext3-mit-backports.imi 
set ROOTFS=%BOOTDEV%=\colinux.disks\Debian-3.0r2.ext3-mit-backports.1gb
set ROOTFS=%ROOTFS% root=/dev/%BOOTDEV% hda=:%BOOTDEV%
set SWAPDEV=cobd1
set SWAPFS=%SWAPDEV%=\colinux.disks\swap_file hdb=:%SWAPDEV%
set OPTIONS=-k
set OPTIONS=%OPTIONS% -t NT
\coLinux\colinux-daemon.exe %OPTIONS% kernel=vmlinux initrd=initrd.gz
mem=%MEM% %CDROM0% %SWAPFS% %ROOTFS% %WINSHARE% %ETH% 

Reducing paramters to minimum does not change behavior.
---


Analysis of core dump:


kd -z c:\Dump\MEMORY.DMP -i c:\coLinux\ -y c:\Symbols\

*** ERROR: Symbol file could not be found.  Defaulted to export symbols
for NDIS.sys -
Probably caused by : linux.sys ( linux+52e6 )


Windows XP Kernel Version 2600 (Service Pack 2) UP Free x86 compatibl
Product: WinNt, suite: TerminalServer SingleUserTS
Built by: 2600.xpsp_sp2_gdr
Kernel base = 0x804d7000 PsLoadedModuleList = 0x8055a620
Debug session time: Tue Aug  7 11:27:53.602 2007 (GMT+2)
System Uptime: 0 days 0:25:19.109


PEB is paged out (Peb.Ldr = 7ffdc00c).  Type ".hh dbgerr001" for details


BugCheck 1111777A, {0, 0, 0, 0}
Probably caused by : linux.sys ( linux+52e6 )


MODULE_NAME: linux
FAULTING_MODULE: 804d7000 nt
DEBUG_FLR_IMAGE_TIMESTAMP:  40b88e1b
DEFAULT_BUCKET_ID:  WRONG_SYMBOLS
BUGCHECK_STR:  0x111777A
LAST_CONTROL_TRANSFER:  from ef6d92e6 to 80533523

STACK_TEXT:
WARNING: Stack unwind information not available. Following frames may be
wrong.
ee441b14 ef6d92e6 1111777a 818fc65c ee441b74 nt!KeBugCheck+0x14
ee441b34 ef6d7742 00000000 00000000 ee441bc4 linux+0x52e6
ee441b74 ef6d7aa6 81beb3a8 00000000 818fc638 linux+0x3742
ee441b94 ef6d7c5b 81beb3a8 81705000 80561b40 linux+0x3aa6
ee441bc4 ef6d86c4 81beb3a8 00000011 81705000 linux+0x3c5b
ee441c14 ef6d86f3 81beb2f0 81a85418 81705000 linux+0x46c4
ee441c34 804e37f7 81beb2f0 81a85418 806ee2d0 linux+0x46f3
ee441c58 8057ad03 81beb2f0 81a85418 818fc650 nt!IofCallDriver+0x32
ee441d00 8057d281 000007c8 00000000 00000000
nt!SeQueryAuthenticationIdToken+0x6c6
ee441d34 804de7ec 000007c8 00000000 00000000
nt!NtDeviceIoControlFile+0x2a
ee441ddc 804f827a f8332b85 82187f48 00000000 nt!ZwYieldExecution+0xb78
ee441de0 f8332b85 82187f48 00000000 0000037f nt!KeInitializeTimer+0x107
ee441de4 82187f48 00000000 0000037f 00000000
NDIS!NdisFreeToBlockPool+0x15e1
ee441de8 00000000 0000037f 00000000 00000000 0x82187f48


STACK_COMMAND:  kb
FOLLOWUP_IP:
linux+52e6
ef6d92e6 83c40c          add     esp,0Ch

SYMBOL_STACK_INDEX:  1
FOLLOWUP_NAME:  MachineOwner
IMAGE_NAME:  linux.sys
SYMBOL_NAME:  linux+52e6
BUCKET_ID:  WRONG_SYMBOLS

kd> r
eax=ffdff13c ebx=00000000 ecx=00000000 edx=00000032 esi=81beb3a8
edi=818fc65c
eip=80533523 esp=ee441afc ebp=ee441b14 iopl=0         nv up ei ng nz na pe
nc
cs=0008  ss=0010  ds=0023  es=0023  fs=0030  gs=0000            
efl=00000286
nt!KeBugCheck+0x14:
80533523 5d              pop     ebp


kd> u
nt!KeBugCheck+0x14:
80533523 5d              pop     ebp
80533524 c20400          ret     4
80533527 90              nop

kd> u ef6d92d8
linux+0x52d8:
ef6d92d8 83ec0c          sub     esp,0Ch
ef6d92db 687a771111      push    1111777Ah
ef6d92e0 ff1518e16def    call    dword ptr [linux+0xa118 (ef6de118)]
ef6d92e6 83c40c          add     esp,0Ch
ef6d92e9 ebd5            jmp     linux+0x52c0 (ef6d92c0)
ef6d92eb 90              nop
ef6d92ec 8d742600        lea     esi,[esi]
ef6d92f0 55              push    ebp

linux+0x52c0:
ef6d92c0 83ec04          sub     esp,4
ef6d92c3 68636f6c78      push    786C6F63h
ef6d92c8 53              push    ebx
ef6d92c9 6a00            push    0
ef6d92cb ff15fce06def    call    dword ptr [linux+0xa0fc (ef6de0fc)]
ef6d92d1 8b5dfc          mov     ebx,dword ptr [ebp-4]
ef6d92d4 89ec            mov     esp,ebp
ef6d92d6 5d              pop     ebp



C:\Programme\Debugging Tools for Windows>dir d:\coLinux\linux.sys
19.03.2007  02:42           139.730 linux.sys

C:\Programme\Debugging Tools for Windows>symchk.exe d:\coLinux\linux.sys
SYMCHK: linux.sys            FAILED  - Built without debugging
information.
SYMCHK: FAILED files = 1
SYMCHK: PASSED + IGNORED files = 0



Source: System Error
Cat: (102)
ID: 1003

Error code 1111777a, 1. Parameter 00000000, 2. Parameter 00000000, 3.
Parameter 00000000, 4. Parameter 00000000.


0000: 53 79 73 74 65 6d 20 45   System E
0008: 72 72 6f 72 20 20 45 72   rror  Er
0010: 72 6f 72 20 63 6f 64 65   ror code
0018: 20 31 31 31 31 37 37 37    1111777
0020: 61 20 20 50 61 72 61 6d   a  Param
0028: 65 74 65 72 73 20 30 30   eters 00
0030: 30 30 30 30 30 30 2c 20   000000, 
0038: 30 30 30 30 30 30 30 30   00000000
0040: 2c 20 30 30 30 30 30 30   , 000000
0048: 30 30 2c 20 30 30 30 30   00, 0000
0050: 30 30 30 30               0000    


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2007-07-02 18:18

Message:
Logged In: YES 
user_id=579204
Originator: NO

Yes, the BSOD (0x1111777A) is fixed at 2007-03-24 in Version 0.8.0
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/src/colinux/os/winnt/kernel/lowlevel/misc.c?view=log
BSOD also fixed in pre version 0.7.1 (build on 26 March 2007).
Version 0.6.4 would no fix, is to old.

The changes on linux.sys installer prevents the install of this driver, if
you have more than 4GB RAM, but should no crash with
IRQL_NOT_LESS_OR_EQUAL, if the daemon tries to load it. How many memory you
have? 32GB? What says the screen by pressing keys Win+Pause?

The main problem is: You have to many memory.
CoLinux would run, if you limit your host memory to 4GB totaly RAM
   /MAXMEM=4096 in your C:\BOOT.INI
Sorry more is not usable by current design ;-(
CoLinux can not run on machines with more as 4GB RAM.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=1745597&group_id=98788