Jump to content











Photo
- - - - -

Illegal Opcode


  • Please log in to reply
17 replies to this topic

#1 artreader

artreader

    Newbie

  • Members
  • 12 posts
  •  
    Germany

Posted 08 November 2012 - 01:01 PM

While trying to install syslinux on small (700 MB) FAT16 Partition on HP Server got Red Screen of Death with a register dump.

Illegal OpCode

EAX=00001F01 EBX=000001A4 ECX=00000000 EDX=00000019F
EBP=0000FBFA ESI=000007BE EDI=00000800
DS=0000 ES=0000 FS=0000 GS=0000
CS:EIP=0000:000083A6 SS:ESP=1000:0000AFF6
EFLAGS=00000206

Tested many versions of syslinux: from 3.63 to 4.06: unfortunately the same
It works on "normal" PC, e.g. without HP RAID controller

Anyone can interpret the register dump?
Any other idea?

#2 Sha0

Sha0

    WinVBlock Dev

  • Developer
  • 1682 posts
  • Location:reboot.pro Forums
  • Interests:Booting
  •  
    Canada

Posted 08 November 2012 - 02:23 PM

While trying to install syslinux on small (700 MB) FAT16 Partition on HP Server got Red Screen of Death with a register dump.

Illegal OpCode

EAX=00001F01 EBX=000001A4 ECX=00000000 EDX=00000019F
EBP=0000FBFA ESI=000007BE EDI=00000800
DS=0000 ES=0000 FS=0000 GS=0000
CS:EIP=0000:000083A6 SS:ESP=1000:0000AFF6
EFLAGS=00000206

Tested many versions of syslinux: from 3.63 to 4.06: unfortunately the same
It works on "normal" PC, e.g. without HP RAID controller

Anyone can interpret the register dump?
Any other idea?

If you can tie this output to an exact version and specify where you got it (the Syslinux version), it might be possible to know what has happened by examining what probably would've been at CS:EIP.

#3 MedEvil

MedEvil

    Platinum Member

  • .script developer
  • 7771 posts

Posted 08 November 2012 - 02:24 PM

Wow, didn't knew that illegal OpCodes still exist.
I though those died out with the C64.

:cheers:

#4 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 16066 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 08 November 2012 - 05:06 PM

It is not an "Illegal Opcode" in the sense we old timers may remember it, it is just a deceiving message from the HP Bios (or Raid bios extension or whatever).

The good HP guys manage to create a RSOD (Red Screen Of Death) whenever *anything* isn't of the exact liking of their Bios (or Raid bios extension or whatever).
Example:
Posted Image

@artreader
See if anything here (though related to GRUB, seemingly GRUB legacy):
http://www.voipone.c...-40lenny41.html
can be of help/have points in common.

:cheers:
Wonko

#5 artreader

artreader

    Newbie

  • Members
  • 12 posts
  •  
    Germany

Posted 09 November 2012 - 12:09 PM

This dump came from syslinux 4.06 release compiled with gcc 4.2.4 on Linux LFS 6.3

#6 artreader

artreader

    Newbie

  • Members
  • 12 posts
  •  
    Germany

Posted 30 November 2012 - 01:39 PM

Just tested syslinux 5.00 pre 12 with same result - same register dump as before

#7 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 16066 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 30 November 2012 - 03:10 PM

Just tested syslinux 5.00 pre 12 with same result - same register dump as before

The link I posted suggested that the RSOD could be caused by an invalid or missing MBR; ad that case was solved checking the MBR of the disk that is booted by BIOS.
  • Did you do that?
  • Is the MBR code "correct"?
  • Is it the Syslinux one?
  • Is it anyone else?
  • Is the MBR data "correct"?
If you don't know how to check, make a copy of it (first sector of .PhysicalDrive) with *any* dd-like tool or use hdhacker:
http://dimio.altervista.org/eng/
compress it to a .zip archive, upload the archive to *any* file hosting service and post a link to it, and I'll have a look at it.

:cheers:
Wonko

#8 Sha0

Sha0

    WinVBlock Dev

  • Developer
  • 1682 posts
  • Location:reboot.pro Forums
  • Interests:Booting
  •  
    Canada

Posted 30 November 2012 - 04:39 PM

Just tested syslinux 5.00 pre 12 with same result - same register dump as before

It is not clear to me whether you are getting this red screen when you install Syslinux (using syslinux.com, syslinux.exe, etc.) or when you boot Syslinux. Your first post says "install," but please confirm.

#9 artreader

artreader

    Newbie

  • Members
  • 12 posts
  •  
    Germany

Posted 30 November 2012 - 05:17 PM

Hi to all,

@Wonko the Sane: the MBR is correct and it is the syslinux one. With non HP threre is no problem.
@Shao: Syslinux is installed with linux sysinstall sucessfully and it is the boot with syslinux, which makes a problem

#10 Sha0

Sha0

    WinVBlock Dev

  • Developer
  • 1682 posts
  • Location:reboot.pro Forums
  • Interests:Booting
  •  
    Canada

Posted 30 November 2012 - 05:43 PM

@Wonko the Sane: the MBR is correct and it is the syslinux one. With non HP threre is no problem.
@Shao: Syslinux is installed with linux sysinstall sucessfully and it is the boot with syslinux, which makes a problem

Thanks. I replied to this on the Syslinux mailing-list, as there are more Syslinux experts, there.

#11 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 16066 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 30 November 2012 - 06:16 PM

@Wonko the Sane: the MBR is correct and it is the syslinux one.

Try a non-syslinux "Plain" MBR, like the 2K/XP one or the Vista :ph34r: /7 one. (or any other "Normal" MBR CODE that just chainloads the Active partition bootsector).
See if this changes anything.
If it doesn't, try booting *somehow* (boot CD, USB stick, etc.) to grub4dos and try chainloading from it the LDLINUX (provided that Sha0 confirms that this is now possible :unsure:)
http://reboot.pro/to...ading-syslinux/

This way only the BPB (and NO code) of the bootsector is used....

:cheers:
Wonko

#12 artreader

artreader

    Newbie

  • Members
  • 12 posts
  •  
    Germany

Posted 30 November 2012 - 06:45 PM

To make sure there is no "foreing" code in partition table and MBR,
MBR was nulled with mbr-sys and partition table produced new with fdisk.
Cause it is a machine with HP Raid, the old-fashioned cciss driver aund the new one hpsa were tested. The FAT Partition created with parted and than syslinux install.

To make sure it is not a machine-specific problem, another HP RAID was borrowed - the same results. Additionally I visited some friends and tested the syslinux versions on some desktops and servers, partially with RAID controllers (non HP) - no single problem.

So, from the logical part, it should be syslinux, that makes the problem

@Wonko the Sane,thanks. This idea to use boot chaining is interesting, but unfortunately i've already tried it out without sucess.

#13 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 16066 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 30 November 2012 - 07:18 PM

There is something that doesn't "fit" :unsure:. (or I am missing something :dubbio:).
You are talking of Linux fdisk, right?
AFAICR it does not put MBR CODE in the MBR, only the partition table and the magic bytes 55AA.
Then you need to install to it (and you seem like having done this at the same time with Syslinux install) the MBR CODE (and this obviously results in SYSLINUX MBR CODE).
What I suggested trying was ANOTHER MBR CODE, what I seem to get is that you made sure (or you think you made sure) that the Syslinux MBR code was "correctly written" to the MBR.

I will repeat myself, the Syslinux MBR code is an almost "standard" MBR code, that is aimed to chainload the bootsector of the Active partition in the MBR partition table, but since it is possible that it causes the issue, to verify it's not that part of the booting that goes wrong, the idea is to test an alternative MBR code.

If you don't have the mentioned 2K/XP or Vista :ph34r: /7 MBR code to do this test, try using MBLDR:
http://mbldr.sourceforge.net/

So, from the logical part, it should be syslinux, that makes the problem

Yes :thumbsup:, but what I was suggesting is that Syslinux occupies THREE "stages" of the booting which goes like this:
BIOS->1. MBR code->2. Active partition PBR or bootsector code ->3. ldlinux.sys
Right now we don't have any evidence that the issue is caused by any particular one among the THREE parts of Syslinux involved in the booting process:
1. MBR code <-part of syslinux install
2. Active partition PBR or bootsector code <-part of syslinux install
3. ldlinux.sys <-part of syslinux install

By trying an alternate MBR code, you make sure that one of the TWO later stages are involved.

:cheers:
Wonko

#14 Sha0

Sha0

    WinVBlock Dev

  • Developer
  • 1682 posts
  • Location:reboot.pro Forums
  • Interests:Booting
  •  
    Canada

Posted 01 December 2012 - 04:11 AM

If it doesn't, try booting *somehow* (boot CD, USB stick, etc.) to grub4dos and try chainloading from it the LDLINUX (provided that Sha0 confirms that this is now possible :unsure:)

Unfortunately, it's not possible at this time (but in a mental to-do list).

...Stuff about MBR versus VBR versus LDLINUX.SYS...

That was very a detailed response. :) Hopefully, combined with the mailing-list, that means that the same requests for information/suggestions for tests are coming "from all sides." ;)

#15 artreader

artreader

    Newbie

  • Members
  • 12 posts
  •  
    Germany

Posted 06 December 2012 - 03:23 PM

Hi to all,

It was the trivial MBR from parted on HP RAID Controller, which made a problem.

With syslinux MBR works syslinux (3.86, 4.00, 5.00pre12 tested), independend on hardware

The functioning sequence:
fdisk /dev/sda < "wnon"
ms-sys -z /dev/sda
parted -s /dev/sda mklabel msdos
ms-sys -s /dev/sda
parted -s /dev/sda mkpart primary fat16 0 700
parted -s /dev/sda mkfs 1 fat16
parted -s /dev/sda set 1 boot on
partprobe
syslinux -f -U /dev/sda1 or syslinux /dev/sda1 (3.86)

Very many thanks to all



#16 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 16066 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 06 December 2012 - 04:25 PM

Which translates as:
 
 
Had I followed exactly the suggestions from the beginning I could have solved the issue much earlier.
 
Additionally, when asked to check the MBR CODE, you explicitely stated (falsely) how you did that and that the MBR CODE was the Syslinux one, and now you are saying that it wasn't it.
 
The good thing is that you solved the problem :thumbup: though not through the "normal" procedure, but rather through this more complex and lengthy one:
  • asking for help
  • NOT following the given suggestions (aggravated by saying that you were following them)
  • finally solve the problem by means of doing EXACTLY what was initially suggested to you
 http://verydemotivat...inted-eagle.jpg
 
 

:cheers:
Wonko

#17 Sha0

Sha0

    WinVBlock Dev

  • Developer
  • 1682 posts
  • Location:reboot.pro Forums
  • Interests:Booting
  •  
    Canada

Posted 07 December 2012 - 03:26 PM

The good thing is that you solved the problem :thumbup: though not through the "normal" procedure, but rather through this more complex and lengthy one:


    [*]asking for help
    [*]NOT following the given suggestions (aggravated by saying that you were following them)
    [*]finally solve the problem by means of doing EXACTLY what was initially suggested to you
    [/list]


To be fair, working out where the MBR was coming from eventually happened via correspondence on a mailing-list, where the whole sequence of commands was eventually shared.

#18 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 16066 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 07 December 2012 - 04:48 PM

To be fair, working out where the MBR was coming from eventually happened via correspondence on a mailing-list, where the whole sequence of commands was eventually shared.

Well, to be fair, an exact question was asked and an exactly corresponding (though false - evidently by mistake) answer to it was given.

 

If OP had asked (here, where he was given the suggestion to check the MBR CODE) "How do I check the MBR CODE?" or he had done what was suggested him in the case he was not confident in doing the check (post a copy of his MBR, INCLUDING EXACT instructions on how to do the latter) he would have solved the issue 4 or 5 days before.

 

A tentative hint (that resulted afterwards as the actual issue) was actually given to the OP some three weeks earlier.

 

I stand by the statement that had the OP:

  • listened to given advice
  • asked for support about what he was not initially capable of doing properly 

he would have solved his issue earlier.

 

This not-so-casually leads us back to the known "how to ask questions the smart way":

http://catb.org./~es...-questions.html

and to the old (and made void of any relevance) amendment #f. to Rules:

http://reboot.pro/to...82-board-rules/

 

As said however, the main thing is that the issue, one way or the other, is solved. :)

 

:cheers:

Wonko






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users