═══ 1. Title page ═══

 32 bits Linux ext2 file system driver for OS/2
EXT2-OS2 VERSION 2.40
Copyright (C) Matthieu WILLM 1995, 1996, 1997 (willm@ibm.net)


═══ 2. Introduction ═══

ext2-os2 is a file system driver that allows OS/2 to seamlessly access Linux 
native partitions (ext2fs partitions) in both read and write modes. Once 
installed, Linux partitions appear as standard OS/2 drive letters. 

My initial purpose was to get limited read only access to ext2fs  partitions 
from my OS/2 system, so at the beginning (version 0.1 alpha back in June 1995), 
it was more a quick and dirty rewrite than a clean and full port of the Linux 
ext2fs  file system kernel code. 

I've improved it to reuse as much original ext2fs code as possible in order to 
get an almost full featured ext2fs driver on OS/2.  I wrote a kind of IFS to 
VFS translation layer, and ported some Linux kernel services in order to 
minimize the changes required into the ext2fs sources. Now that ext2-os2 is 32 
bits, these changes are minimal. 

This version is almost full featured : the main restriction is that ext2-os2 
does not support extended attributes (EAS). Here are the main features of 
ext2-os2 : 

     full 32bits ring 0 implementation 
     allows read-write access to Linux ext2 partitions 
     allows OS/2 to swap on Linux ext2 partitions 
     allows OS/2 to boot from a Linux ext2 partition 
     supports huge cache sizes 
     contains a port of most of the Linux ext2 file system utilities from the 
      e2fsprogs-1.10 package : 
        -  mke2fs, to format drives as Linux ext2 
        -  e2fsck, to check and repair Linux ext2 file systems 
        -  tune2fs, to change Linux ext2 file systems parameters 
        -  debugfs, an interactive Linux ext2 file system debugger 
        -  lsattr and chattr, to see or change attributes of files on Linux 
           ext2 file systems 
     can reorder all hard disk drive letters, to avoid partition headaches 
     hard links are supported under OS/2 and can also be created from OS/2 

 DISCLAIMER : 
   THE FACT THAT I WORK FOR IBM HAS NOTHING TO DO WITH THIS DRIVER.  I'M 
 WRITING IT AT HOME, DURING MY SPARE TIME, FOR MY OWN PURPOSES, ON MY OWN 
 MACHINE, WITHOUT ANY IBM MEANS.  I USED ONLY IBM PUBLISHED DOCUMENTS TO WRITE 
 IT :  I AM ABSOLUTELY NOT RELATED TO THE OS/2 DEVELOPMENT TEAM AND THUS I HAVE 
 ABSOLUTELY NO ACCESS TO ANY CONFIDENTIAL OS/2 IFS RELATED INFORMATION. 

 Anyway I hope you'll enjoy trying this driver, and even if I don't have much 
 time, I'll try to keep improving it.  I wait for your feedback on this version 
 :  comments, ideas, bug reports and contributions are welcome. Success stories 
 do also interest me, and disasters too ! 

 Matthieu WILLM 

 willm@ibm.net                 Home 
 mwillm@vnet.ibm.com           Work 

 Note:  English is not my native language (I'm french) ... so this document 
        isn't certainly perfect. 


═══ 3. *** IMPORTANT WARNING - READ THIS FIRST *** ═══

ext2-os2 is a FILE SYSTEM DRIVER. Being a driver it runs at the most privileged 
level of the OS/2 operating system, known as "kernel mode".  IN KERNEL MODE 
THERE IS NO MORE CRASH PROTECTION.  It means that if a bug occurs in the driver 
IT CAN CAUSE A HANG OR A SYSTEM HALT, AND THUS CAN CAUSE LOSS OF DATA. 

     I STRONGLY SUGGEST YOU TO MAKE A BACKUP OF ALL YOUR DATA (EVEN IF THEY 
      ARE NOT LOCATED ON A LINUX PARTITION) PRIOR TO USE THIS DRIVER. 
     DON'T USE THIS DRIVER ON A PRODUCTION MACHINE, OR IF YOU MANIPULATE 
      CRITICAL DATA. 
     IF HUMAN LIFE DEPENDS ON YOUR SYSTEM, DON'T INSTALL THIS DRIVER ON IT. 
     USE IT AT YOUR OWN RISKS, AND ONLY ON A MACHINE YOU ACCEPT TO CRASH AND 
      REINSTALL. 
     BE AWARE THAT THIS DRIVER WILL GIVE FULL UNRESTRICTED (ROOT) ACCESS TO 
      ALL YOUR EXT2FS PARTITIONS FROM OS/2. 

 From experience, ext2-os2 is now stable for normal use, and it is unlikely 
 that a disaster will occur. But as any piece of software, it cannot be 
 bulletproof. Doing regular backups is still the best method to protect against 
 disasters of any kinds ... 


═══ 4. Copyright information ═══

      Copyright notice for the ext2fs IFS driver 
      Copyright notice for the ext2fs partition filter 
      Authors of the original Linux ext2 file system 


═══ 4.1. Copyright notice for the ext2fs IFS driver ═══

    32 bits Linux ext2fs file system driver for OS/2 WARP - Allows OS/2 to
    access your Linux ext2fs partitions as normal drive letters.
    Copyright (C) 1995, 1996, 1997  Matthieu WILLM

    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software
    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.


═══ 4.2. Copyright notice for the ext2fs partition filter ═══

/************************************************************************/
/*                       Linux partition filter.                        */
/*          (C) Copyright Deon van der Westhuysen, July 1995.           */
/*                                                                      */
/*  Dedicated to Jesus Christ, my Lord and Saviour.                     */
/*                                                                      */
/* This program is free software; you can redistribute it and/or modify */
/* it under the terms of the GNU General Public License as published by */
/* the Free Software Foundation; either version 2, or (at your option)  */
/* any later version.                                                   */
/*                                                                      */
/* This program is distributed in the hope that it will be useful,      */
/* but WITHOUT ANY WARRANTY; without even the implied warranty of       */
/* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the        */
/* GNU General Public License for more details.                         */
/*                                                                      */
/* You should have received a copy of the GNU General Public License    */
/* along with this program; if not, write to the Free Software          */
/* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.            */
/*                                                                      */
/*  This code is still under development; expect some rough edges.      */
/*                                                                      */
/************************************************************************/


═══ 4.3. Authors of the original Linux ext2 file system ═══

Parts of the ext2-os2 package are based on GNU GPL copyrighted code whose 
owners are : 

     Linus Torvalds, the main author of the Linux kernel. 
     Remy Card, the main author of the Linux ext2fs file system kernel code, 
      as well as part of the ext2fs utilities. 
     Theodore Ts'o, the main author of the Linux ext2fs file system utilities. 
     Stephen Tweedie, who improved the Linux ext2fs allocation algorithms. 
     Werner Almesberger, the author of the Linux FAT file system. 
     Jean Loup Gailly,  the author of gzip 
     ...probably many other contributors to the Linux kernel. 


═══ 5. Installation instructions ═══

 This chapter describes the installation and setup instructions 

      Step 1 - Things to do BEFORE installing ext2-os2 
      Step 2 - Make OS/2 see your Linux ext2fs partitions 
      Step 3 - Copying the files to your hard drive 
      Step 4 - Using the ext2-os2 monitoring tool 
      Step 5 - Specifying command line parameters 


═══ 5.1. Step 1 - Things to do BEFORE installing ext2-os2 ═══

Here are things you must do before installing this version of ext2-os2 : 

   1. Please run e2fsck from Linux on ALL your ext2fs partitions BEFORE 
      installing ext2-os2. (You might have to boot Linux from floppy to do 
      this, especially to run e2fsck on your root partition). 
   2. I STRONGLY suggest you to first backup any data you don't want to loose 
      BEFORE trying this driver. 
   3. Make a backup copy of your CONFIG.SYS 
   4. Have the two OS/2 kicker diskettes handy ,  with a full screen editor, so 
      that you can restore your CONFIG.SYS if the system crashed at boot time. 
      (This is not needed with WARP : Alt+F1 C at boot time will do the trick) 
   5. Prepare a Linux boot diskette, with e2fsck on it. 
   6. If you installed a previous version of ext2-os2, then comment out the 
      following CONFIG.SYS statements and reboot your system to ensure that no 
      ext2-os2 file is locked while installing the new version. 
          IFS = <path>\ext2-os2.ifs <parms> 
          RUN = <path>\ext2_lw.exe 
          BASEDEV = ext2flt.flt <parms> 
          BASEDEV = mwdd32.sys <parms> 


═══ 5.2. Step 2 - Make OS/2 see your Linux ext2fs partitions ═══

 This chapter describes how to make OS/2 see your Linux ext2fs partitions 


═══ 5.2.1. Understanding how does OS/2 assign drive letters ═══

 OS2DASD.DMD is the block device managing fixed partitionnable disks and 
floppies. It is always (as far as I know) the first block device loaded by the 
system. It first scans every .add drivers for physical disks. Then it reads the 
partition table of each disk. It first looks for active primary OS/2 partitions 
(Types 01, 04, 06 and 07) and installs a block device for each of them It 
repeats this for each physical drive in the system. Then it looks for extended 
volumes on the first physical drive (type 05 partition), reads the extended 
boot record, looks for an OS/2 partition and assign it a drive letter, looks 
for another extended volume ... and so on until no more extended volume is 
found on the first disk. It them repeats this for each physical drive. 

Example : 

■ Scans all *.add driver for fixed disks - found 2
■ Reads the MBR of physical disk 1
        ■ Found an OS/2 type 7 partition (active primary OS/2 IFS)  - Install a block device (C:)
        ■ Found an unknown type 0x83 partition                      - skip it
        ■ Found an OS/2 type 5 partition (extended volume)          - skip it
        ■ No more primary partition
■ Reads the MBR of physical disk 2
        ■ Found an OS/2 type 7 partition                            - Install a block device (D:)
        ■ Found an unknown type 17 (inactive primary OS/2 IFS)      - skip it
        ■ Found an OS/2 type 5 partition (extended vol)             - skip it
        ■ No more primary partition
Now done for primary partitions - processing extended volumes
■ Reread the MBR of physical disk 1
        ■ Found an OS/2 type 5 partition (extended volume)
                ■ Reads the extended boot record of the extended volume
                        ■ Found a OS/2 type 7 partition             - Install a block device (E:)
                        ■ Found an OS/2 type 5 partition
                                ■ Reads the extended boot record
                                    ■ Found a OS/2 type 7 partition - Install a block device (F:)
                                    ■ No more partition
■ Reread the MBR of physical disk 2
        ■ Found an OS/2 type 5 partition (extended volume)
                ■ Reads the extended boot record of the extended volume
                        ■ Found a OS/2 type 7 partition             - Install a block device (G:)
                        ■ No more partition

Make OS/2 see your Linux ext2fs partitions 


═══ 5.2.2. Make OS/2 see your Linux ext2fs partitions. ═══

During "normal" operations, OS/2 will not assign a drive letter to the Linux 
partitions,  because their type (0x83) is not one of the types OS/2 normally 
recognize. An installable file system driver does not have any control on how 
block devices are assigned : this is a device driver issue. If we don't do 
anything, the ext2fs IFS won't be very useful, since OS/2 doesn't even call it 
(no block device allocated). 

In order for OS/2 to see your ext2fs partition, you must either change your 
Linux ext2fs partition ID from 0x83 (Linux native) to 0x7 (OS/2 IFS - and not 
OS/2 HPFS as stated in the Linux fdisk documentation), or make OS/2 think type 
0x83 partitions are type 0x7 partitions. 

In the first case we must manually alter the partition table, and in the second 
case we must use a special filter device driver - ext2flt.flt - whose purpose 
is to show a "hacked" partition table to OS2DASD.DMD. 

Here's how ext2flt.flt works : A virtual fixed disk is created for each 
partition under control of the driver. These virtual fixed disks each contain 
one extended partition with a logical drive in it. Since OS2DASD.DMD that 
controls fixed disks first assigns drive letters to primary partitions the 
virtual partitions are tacked on at the end of allocated drive letters. (To 
control the mounting order of all partitions OS2DASD is prevented from directly 
accessing the fixed disks and the filter presents it with an alternate list of 
partitions.) 

BE CAREFUL : IF YOU DECIDE TO CHANGE PARTITION ID MANUALLY BE AWARE THAT IT CAN 
CHANGE YOUR DRIVE LETTERS AND MAKE OS/2 UNBOOTABLE !! 

Examples being better than a long speach, let's look at some different 
situations : 

      Example 1 
      Example 2 
      Example 3 


═══ 5.3. Step 3 - Copying the files to your hard drive ═══

     Installation using the "Device Driver Install" object 
     Manual installation 


═══ 5.3.1. Installation using the "Device Driver Install" object ═══

   1. Open the "OS/2 System" folder 
   2. Open the "System Setup" folder 
   3. Open the "Device Driver Install" object 
        a. Type in the "source directory" entry field the path where you 
           decompacted the files 
        b. Type in the "target directory" your OS/2 boot drive 
        c. Click on the "Install..." push button 
        d. Select the drivers to install (select both of them) : 
               "Linux ext2fs file system driver (IFS) version 2.40 for OS/2" 
               "ext2flt - Filter to make Linux partitions visable to OS/2." 
        e. Click on OK 
   4. Add <boot drive>:\OS2\FS\EXT2 to the PATH and LIBPATH statements in your 
      CONFIG.SYS 
   5. Set and/or change command line parameters for ext2-os2.ifs and 
      ext2flt.flt. 
   6. Reboot your computer 


═══ 5.3.2. Manual installation ═══

ext2flt installation 

   1. copy ext2flt.flt and ext2flt.sym in  <boot drive>\OS2 or in <boot 
      drive>\OS2\BOOT (WARP only) 
   2. add the following line in CONFIG.SYS 

            BASEDEV=EXT2FLT.FLT

   3. if necessary, adjust ext2flt command line parameters (for instance /W if 
      you want write access) 

 ext2-os2 installation 

   1. copy mwdd32.sys and mwdd32.sym in  <boot drive>\OS2 or in <boot 
      drive>\OS2\BOOT (WARP only) 
   2. add the following line in CONFIG.SYS 

            BASEDEV=MWDD32.SYS

   3. make a directory to hold ext2-os2 files somewhere but NOT on a Linux 
      partition. 
   4. add this directory at the end of the PATH and LIBPATH statements in 
      CONFIG.SYS 
   5. copy the following files to that directory : 
          ext2-os2.ifs 
          ext2-os2.exe 
          sync.exe 
          hardlink.exe 
          umount.exe 
          remount.exe 
          ext2_lw.exe 
          ext2-os2.inf 
          ext2-os2.htm 
          ext2-os2.sym 
          ext2-os2.ddp 
          vfsapi.dll 
          vfsapi.lib 
          uext2.dll 
          uuid.dll 
          et.dll 
          ext2fs.dll 
          badblks.exe 
          mke2fs.exe 
          e2fsck.exe 
   6. add the following lines in your CONFIG.SYS 

            IFS=<path>\ext2-os2.ifs -cache:256 -errors=continue
            RUN=<path>\ext2_lw.exe

   7. if necessary, adjust ext2-os2.ifs command line parameters (for instance 
      -rw if you want write access) 

 Note:  In CONFIG.SYS, ext2-os2.ifs MUST be located AFTER the IFS managing the 
        boot drive (HPFS.IFS usually). ext2-os2.ifs can appear at the head of 
        CONFIG.SYS only if you boot OS/2 from a Linux partition. 


═══ 5.4. Step 4 - Using the ext2-os2 monitoring tool ═══

ext2-os2.exe is a monitoring tool to retrieve data from ext2-os2.ifs. It is a 
PM notebook containing : 

     page 1 : disk cache information 
     page 2 : I-node and open file information 
     page 3 : ext2-os2.ifs standard output 
     page 4 : ext2-os2.ifs swapper information (active only when SWAPPER.DAT 
      is located on a Linux ext2fs partition) 


═══ 5.5. Step 5 - Specifying command line parameters ═══

ext2-os2.ifs command line in CONFIG.SYS : 

IFS = <drive>:<path>\ext2-os2.ifs [-q]
                                  [-cache:<max. disk cache size>]
                                  [-rw]
                                  [-no_auto_fsck]
                                  [-case_retensive]
                                  [-<IFS entry point to trace>]
                                  [-no_strat2]
                                  [-errors=[panic|continue]]
                                  [-output=[com1|com2]]
                                  [-tz:<time zone in minutes from UTC>]
                                  [-force_strat2:<removable drive number to be forced to use strat 2 I/O>]

NOTE : case is not significant (-q and -Q are equivalent), and '-' as well as 
'/' can be used as a command line switch on the ext2-os2.ifs command line. 

ext2flt.flt command line in CONFIG.SYS : 

BASEDEV = EXT2FLT.FLT [/Q]
                      [/V]
                      [/W]
                      [/A]
                      [/M <mount list>]


═══ 6. Usage instructions ═══

This chapter describes some imortant usage instructions. 

   1. Lazy writes considerations 
   2. Automatic check disk 
   3. Fatal ext2-os2 errors 
   4. Case sensitivity considerations 
   5. Note to "System Commander" Users 
   6. Dynamic disk cache considerations 
   7. Time zone considerations 
   8. Important removable media considerations 


═══ 6.1. Lazy writes considerations ═══

This file system driver makes heavy use of lazy writes : it means that data are 
not committed to the disk immediately for performance purposes. This means that 
YOU *MUST ALWAYS* SHUTDOWN YOUR SYSTEM PROPERLY so that data can be written to 
disk. IF YOU DON'T FOLLOW THIS RULE, YOUR EXT2FS PARTITION WILL BE PROBABLY 
DAMAGED, AND YOUR DATA LOST. 

If you need to commit data to disk without shutting down the system, you can 
use the sync.exe utility. This utility behaves exactly like its UNIX equivalent 
(but only on Linux ext2fs file systems). 

There is an automatic flush during system shutdown., so running sync.exe prior 
system shutdown is not necessary. 

Note:  You will notice that the shutdown process is slightly longer, especially 
       if you used a huge cache size : this is a normal behaviour. 


═══ 6.2. Automatic check disk ═══

When ext2-os2 mounts an ext2fs partition and when read writes are enabled, it 
will set the mount count in the superblock to its maximum value in order to 
force a disk check when Linux mounts the drive. 

This is a safety precaution to ensure that ext2-os2.ifs didn't corrupt the file 
systems. I do NOT recommend to disable this automatic check ; I only included 
the -no_auto_fsck switch for those who have very large drives, where e2fsck 
takes ages to complete. 


═══ 6.3. Fatal ext2-os2 errors ═══

ext2-os2 fatal errors can appear in several forms : 

   1. A system halt with a register dump : this is a fatal exception trapped by 
      the OS/2 kernel, due for example to a NULL pointer assignment. In this 
      case the kernel debugger or a system dump is needed to track down the 
      problem. 
   2. A system halt with a message like ">>>>>>> EXT2-OS2 FATAL ERROR <<<<<<<". 
      This is a system halt initiated by ext2-os2.ifs due to an unexpected 
      error. Normally there should be a few lines with a cryptic message. 
      Please write it down, as it will tell me where the error occured. 
   3. The whole system is frozen : it is probably because an infinite loop was 
      entered. Here also the kernel debugger is the only solution. 
   4. One or more applications are frozen : it is probably a deadlock. 

 In either case, your ext2fs partition is in great danger ; To minimize the 
 risk of data loss, please do the following : 

   1. In case of a system halt, write down the error message. 
   2. In case of a system hang, try to press CTRL+ALT+DEL : this will commit 
      dirty buffers to disk and thus minimise the damages on the partitions 
      (ext2fs as well as FAT or HPFS). If it didn't work, the big red button is 
      the only solution. 
   3. Reboot Linux and run e2fsck ON ALL YOUR LINUX EXT2FS PARTITIONS to check 
      and repair them if needed. DO NOT RESTART OS/2 WITH EXT2-OS2.IFS WITHOUT 
      HAVING RUN E2FSCK ON  ALL YOUR PARTITIONS, THIS COULD SPREAD FILE SYSTEM 
      ERRORS AND MAKE THINGS EVEN WORSE ! 
   4. Please e-mail me the error, as well as a test case if possible, and 
      please join your partition scheme (FDISK /QUERY from OS/2 or fdisk -l 
      from Linux) and your CONFIG.SYS. 


═══ 6.4. Case sensitivity considerations ═══

ext2fs is a case sensitive file system ; ext2-os2.ifs by default does case 
sensitive comparisions, but even if OS/2 supports case sensitive file systems, 
all the apps that uppercase or lowercase file names are broken. Some parts of 
OS/2 are also broken : for instance the OS/2 program loader itself will 
uppercase executable names, and so will fail loading programs or DLL if they 
reside in a path with low case components. 

As a workaround, I introduced a command line switch (-case_retensive) to make 
ext2-os2.ifs behave like HPFS : once set, it will ignore case when comparing 
file names, but it will retain the case within the file system. It means that 
FILE, File, FiLe and file will be treated as the same file if they reside in 
the same directory ; but if you create FiLe, it will keep the case in the file 
system. 

Note:  If you plan to run OS/2 programs from a Linux partition, this option is 
       necessary because the OS/2 program loader uppercase executable names. 

 There's a drawback to this case retensive mode : if 2 files created from Linux 
 in the same directory exist with the same name but different case (FILE and 
 File for instance), and you run ext2-os2.ifs in case retensive mode, RESULTS 
 WILL BE UNPREDICTABLE AND YOU CAN TRASH ONE OR EVEN BOTH FILES ! In fact only 
 the first entry in the "." file of the parent directory will be used by 
 ext2-os2.ifs  ; This "first" entry is not necessarily the first entry 
 displayed by the 'DIR' command, nor is it the first one created. EXT2-OS2.IFS 
 DOES NOTHING TO SOLVE SUCH FILE NAME CONFLICT IN CASE RETENSIVE MODE. 

 If you want to use this case retensive mode, please make sure from Linux that 
 there are never several files with the same name but different case in ANY 
 directory on ANY ext2fs partitions on your system BEFORE trying case retensive 
 mode. If there are some, then rename these files from Linux so that they have 
 have different names. 


═══ 6.5. Note to "System Commander" Users ═══

Note that "System Commander" (from V-Communications) may make changes to the 
partition table at boot time.  This may affect which partitions are visible, or 
what drive letters are assigned by OS/2 when booting from System Commander.  In 
particular, by default System Commander will make primary partitions on the 
second hard drive invisible to OS/2.  If one of these other primary paritions 
contains a Linux partition, it will not be visible to OS/2 or EXT2-FS, (even 
though you are using the /V /A /M options with EXT2FLT.FLT).  To make a primary 
partition on a second hard drive visible when booting OS/2 from System 
Commander, do the following: 

   1. From the main System Commander screen, move to the OS/2 boot selection, 
      and press Alt-S for the setup screen. 
   2. Select "Local special options" from the setup screen. 
   3. Select "YES" for the "Make primary partitions visible on drive 1". 
   4. That's all 

 This has been reported by Steve Enns (ennss@eris.nhrc.sk.doe.ca) 


═══ 6.6. Dynamic disk cache considerations ═══

ext2-os2.ifs now has a dynamic disk cache. The disk cache code is in fact a 
port of the Linux disk cache, but with a few important differences. 

In Linux, the dynamic cache code is coupled to the virtual memory manager code, 
so that all free physical memory can be used as cache : when there's some free 
physical memory, the cache can grow, and when there's not enough free physical 
memory, the virtual memory manager asks the disk cache to free some pages (this 
can range from  "Could you please give me some free pages" ... to "I URGENTLY 
need some memory so make whatever you can to free some pages"). On OS/2 I 
couldn't make things work that way, simply because I can't hack the OS/2 kernel 
code !! On OS/2, you MUST specify an upper limit for the amount of physical 
memory the disk cache can allocate. When needed, ext2-os2.ifs will allocate 
this memory as disk cache, but to simulate VMM calls to shrink the cache, a 
daemon (ext2_lw.exe) periodically tries to shrink the cache. While it is not a 
perfect solution, it seems to be a good compromise, at least if the max cache 
size is reasonable (small enough so that it doesn't cause physical memory 
starvation to the kernel). 

Some important points : 

     Disk cache memory is allocated/freed by chunks of 4 Kb (a page). 
     The disk cache is allocaled in PHYSICAL MEMORY, it means that this memory 
      CANNOT BE SWAPPED OUT to make room for other needs by the kernel (Having 
      a swappable disk cache would be stupid IMHO !). It means that the maximun 
      cache size passed to ext2-os2.ifs command line in CONFIG.SYS (-cache:xxx 
      option) should be reasonable. 
     You MUST ALWAYS HAVE EXT2_LW.EXE RUNNING, even in read only mode. If you 
      kill it or if you suppress the starting of ext2_lw.exe in CONFIG.SYS, the 
      disk cache size WILL NEVER SHRINK. 
     You can monitor the disk cache size using ext2-os2.exe. 


═══ 6.7. Time zone considerations ═══

ext2-os2 can now take your local time zone into account. File times should 
match Linux file times (local time) instead of being relative to UTC . The time 
zone can be set in CONFIG.SYS on the ext2-os2.ifs command line through the -tz 
switch . It should be set to the  minute shift from UTC (-1440 to +1440). 0 
means GMT, 60 means GMT+1, 120 means GMT+2, -60 means GMT-1 and so on. 

This option is rather experimental and can't be perfect because OS/2 does not 
natively support time zones. Please tell me if you find it useful and if it 
works correctly, otherwise simply remove the -tz switch and tell me the 
problems you encountered. 

Note:  The OS/2 system time is not affected by this variable, so when you use 
       -tz, you make ext2-os2.ifs consistent with Linux, but you introduce a 
       shift between OS/2 system time, and file times. In some cases this can 
       be a problem. Please tell me if you encountered some problems due to 
       this (makefiles ?). 


═══ 6.8. Important removable media considerations ═══

ext2-os2 does support removable media formatted with the Linux ext2fs file 
system. To safely use ext2-os2 on removable media, you MUST FOLLOW THE THE 
PROCEDURE DESCRIBED BELOW !!! If you don't follow this procedure, you can face 
data loss or even system halts. 

When you insert a new media in its drive, OS/2 will automatically attach the 
correct file system on it. No special step is required to make OS/2 recognize 
this new media. If you cannot access the drive, then you can try to run the 
remount.exe program to make OS/2 redetermine the file system to be attached to 
the drive. If it still fails, then check that it is actually formatted with the 
ext2 file system, and if it is the case, try to run e2fsck on that drive. 

When you plan to remove a media from its device, YOU MUST RUN THE UMOUNT.EXE 
PROGRAM ON THAT DRIVE TO COMMIT ALL DATA TO DISK AND MAKE EXT2-OS2 FREE ANY 
REFERENCES TO THAT PARTICULAR VOLUME. After that, the media can be safely 
removed from its device. umount.exe actually issues an unmount operation, like 
on UNIX. 

Note:  The old method of using sync.exe before removing a media could cause 
       problem in some cases. That's why I strongly recommend to use umount.exe 
       instead each time you need to remove a volume from its device 
       (umount.exe does a lot more work than sync.exe, as it does exactly what 
       the UNIX "umount" call does). 


═══ 7. Features, limitations and known bugs for this release. ═══

   1. Features and limitations 
   2. Known bugs 


═══ 7.1. Features and limitations ═══

     ext2-os2.ifs is a full 32 bits ring 0 FLAT memory model file system 
      driver (like WARP Server's HPFS386). 
     All flavours of OS/2 WARP are supported (V3 and V4) except SMP versions. 
     I used the Linux 1.2.1 kernel sources, corresponding to ext2fs 0.5b. It 
      should work even with ext2fs from latest linux kernels (up to 2.1.13 
      which is the latest kernel as of today). 
     Pseudo dynamic disk cache size (limited by the physical memory 
      available), ported from Linux. 
     Global I-node cache, identical to Linux one. 
     Global directory entry cache, identical to Linux one. 
     Uses almost only original Linux ext2fs routines. 
     Can use a blocksize of 1024, 2048 or 4096 bytes (1024 is the default). 
     Doesn't support fragments (neither do the original Linux ext2fs 0.5b 
      code) 
     Linux special files (fifo, symbolic links, sockets, block and character 
      devices) are mapped as SYSTEM files : you can see them with DIR /A but 
      you normally can't read them 
     OS/2 mounts ext2fs partitions as root : it means that from OS/2 you have 
      full unrestricted access to your Linux partitions. 
     Unlike Linux, file names beginning with '.' are not hidden 
     Hard links are properly handled and can be created from OS/2. 
     Symbolic links are not followed (treated as SYSTEM regular files) 
     Each ext2 partition has its own drive letter (provided you made them 
      visible from OS/2) 
     OS/2 extended attributes are not supported, because Linux ext2fs does not 
      support them for the moment. 
     Swapping on a Linux ext2fs partition is supported : it means you can put 
      SWAPPER.DAT on a Linux partition. 
     Booting OS/2 from a Linux ext2fs partition is now supported, but you must 
      still have a FAT or HPFS partition to store the WPS desktop tree. 
     Some OS/2 API are not or partially supported, for instance : 
        1. DosSetFileInfo : level 2 (used to store/change EAS) is ignored 
           because EAS are not supported by ext2fs. 
        2. DosSetPathInfo : level 2 (used to store/change EAS) is ignored 
           because EAS are not supported by ext2fs. 
        3. DosSetVerify  : completely ignored on ext2fs partitions ! 
     ext2fs specific behaviour of some OS/2 API : 
        1. DosQueryFSInfo with FSIL_ALLOC (level 1) information returns the 
           free space minus the reserved space. This should be consistent with 
           Minix style df output (no fs overhead taken into account). 
        2. DosQueryFSInfo with FSIL_VOLSER (level 2) information returns the 
           Linux ext2 volume name (if any) truncated to 11 characters, and a 32 
           bits CRC checksum of the Linux ext2 UUID as the volume serial 
           number. 
     A file is considered read only if neither UNIX user, group and other 
      'write' bits are set. 
     When a file is set read-write, the UNIX user 'write' bit is set. 
     When the block device driver supports it, strategy 2 I/O are used on non 
      removable media. They provide the following advantages : 
        -  Uses asynchronous I/O request lists. 
        -  Priorities can be assigned to I/O. 
        -  Multiple non contiguous I/O requests can be sent to the device 
           driver in one shot, this permits the the IFS and the device driver 
           to sort requests for better performance. OS2DASD.DMD, the standard 
      fixed disk block device driver does support strategy 2 requests. 


═══ 7.2. Known bugs ═══

The following list includes some common problems encountered with ext2-os2. 
There are of several kinds : permanent restrictions, problems due to bugs in 
the OS/2 operating system itself, or problems I couldn't reproduce yet (for 
instance because I don't have the related hardware). This list is not complete 
but I'll try to keep it as much up to date as possible , if you give me your 
feedback... 

     The mount process will mount the ext2fs partition even if it has not been 
      unmounted cleanly  : I STRONLY SUGGEST YOU TO UNMOUNT YOUR FILE SYSTEMS 
      CLEANLY FROM LINUX - RUNNING THIS DRIVER ON A CORRUPTED EXT2FS PARTITION 
      IS LIKELY TO CRASH YOU SYSTEM. 
     DosFindNext() called from a 32 bits app (aka DosFindFromName) doesn't 
      restart from the specified name, but from the last directory entry found. 
     As ext2fs doesn't support EAS, there are many cosmetic bugs with the 
      workplace shell (some icons stay hatched....) 
     Error codes returned from the IFS to OS/2 are sometimes not translated 
      from Linux to OS/2. This will result in messages like "Cannot find 
      message 56347" instead of "Directory not empty" for instance. 
     Some people may encounter a hang or a system halt at boot time, if they 
      have removable devices like MO drive, Syquest and such. There's 
      apparently a problem between EXT2FLT.FLT and some filter device drivers 
      that attempt to make these devices appear like fixed disks (these drivers 
      are LOCKDRV.FLT, SYQLOCK.FLT ...). To solve the problem, edit CONFIG.SYS 
      so that EXT2FLT.FLT appears as the first .FLT driver. 


═══ 8. How to build ext2-os2.ifs ═══

This chapter describes how to build the IFS. 

   1. Tools needed 
   2. The build tree - How to build 


═══ 8.1. Tools needed ═══

Table or required compiler, assembler and linker : 

┌────────────────────┬──────────┬──────────┬──────────┬──────────┐
│                    │32 bits C │16 bits C │80x86     │Linker    │
│                    │compiler  │compiler  │assembler │          │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│ext2-os2.ifs        │(1)       │(none)    │(2)       │(7)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│mwdd32.sys          │(1)       │(none)    │(2)       │(7)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│ext2flt.flt         │(none)    │(3) or (4)│(5) or (6)│(8) or (9)│
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│vfsapi.dll          │(1)       │(none)    │(none)    │(7)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│ext2-os2.exe        │(1)       │(none)    │(none)    │(7)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│ext2_lw.exe         │(1)       │(none)    │(none)    │(7)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│sync.exe            │(1)       │(none)    │(none)    │(7)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│uext2.dll           │(1)       │(none)    │(none)    │(7)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│microfsd.fsd        │(none)    │(3)       │(5)       │(8)       │
├────────────────────┼──────────┼──────────┼──────────┼──────────┤
│minifsd.fsd         │(none)    │(3)       │(5)       │(8)       │
└────────────────────┴──────────┴──────────┴──────────┴──────────┘

List of required toolkits : 

     The IBM OS/2 WARP developer's toolkit (included in IBM Visualage C++ 
      package) 
     The IBM OS/2 Developer Connection Device Driver Kit version 2 or later 
     The IBM 'IFS toolkit'. I don't know the exact status of this package ; it 
      seems to be publically available (I've seen it on Hobbes - a file named 
      ifsinf.zip). If it's not publically available, ask your local IBM rep for 
      a copy. 


═══ 8.2. The build tree - How to build ═══

   1. make a directory somewhere on your disk. 
   2. in that directory, unzip the file ext2_src.zip 
   3. in that directory, unzip the file mwdd_src.zip (included in 32drv130.zip 
      included in this package). 
   4. update the file <your directory>\32bits\ext2-os2\makefile.inc, according 
      to the table below. 
   5. open a DOS session, and in that session : 
          change the current directory to <your 
           directory>\32bits\ext2-os2\ext2flt 
          type 'nmake -f makefile.msc' to build ext2flt.flt. 
   6. exit from the DOS session 
   7. change the current directory to <your directory>\32bits\ext2-os2 
   8. type "nmake" to build the whole ext2-os2 package 

  ┌──────────────────┬─────────────────────────────────────────────┐
  │EXT2OS2_BASE      │The root path of ext2-os2 (usually \ext2-os2)│
  ├──────────────────┼─────────────────────────────────────────────┤
  │DDKPATH           │The path where is installed your DDK (usually│
  │                  │\DDKX86 or \DDK)                             │
  ├──────────────────┼─────────────────────────────────────────────┤
  │IFSTKTPATH        │The path where you put the "IFS toolkit"     │
  │                  │(fsd.h, fsh.h and fshelper.lib)              │
  ├──────────────────┼─────────────────────────────────────────────┤
  │MSVCPATH          │The path where MS Visual C++ is installed    │
  │                  │(usually \MSVC)                              │
  ├──────────────────┼─────────────────────────────────────────────┤
  │VACPATH           │The path where IBM Visualage C++ is installed│
  │                  │(usually \IBMCPP)                            │
  └──────────────────┴─────────────────────────────────────────────┘

 The build tree looks like : 

  ┌────────────────────┬─────────────────────────────────────────────┐
  │.\                  │contains the main makefile an the files I    │
  │                    │couldn't put elsewhere.                      │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\vfs               │Linux kernel services and VFS interfaces.    │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\ext2              │Linux ext2fs sources (from                   │
  │                    │/usr/src/linux-1.2.1/fs/ext2)                │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\include\os2       │OS/2 specific include files                  │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\include\linux     │Linux include files                          │
  │                    │(/usr/src/linux-1.2.1/include/linux)         │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\include\asm       │Linux include files                          │
  │                    │(/usr/src/linux-1.2.1/include/asm)           │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\console           │ext2-os2.exe monitoring tool sources         │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\doc               │The ext2-os2 documentation                   │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\distrib           │The directory in which is built the ZIP file.│
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\vfsapi            │The VFSAPI library sources.                  │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\ext2flt           │contains Deon van der Westhuysen's           │
  │                    │ext2flt.flt sources, patched to compile with │
  │                    │MS Visual C++ as well as Borland C++.        │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\microfsd          │The ext2-os2 micro FSD (used to boot OS/2    │
  │                    │from a Linux partition)                      │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\minifsd           │The ext2-os2 mini FSD (used to boot OS/2 from│
  │                    │a Linux partition)                           │
  ├────────────────────┼─────────────────────────────────────────────┤
  │.\uext2             │The dummy uext2.dll library sources          │
  │                    │(necessary to make BOOTOS2 happy)            │
  └────────────────────┴─────────────────────────────────────────────┘


═══ 9. The VFSAPI library ═══

The purpose of the VFSAPI library is to provide an extension of the standard 
OS/2 standard file system oriented calls to access Linux VFS specific features 
like I-nodes, UNIX file modes, file ownership and so on. For the moment the 
library is quite limited, but with the time I'll include some more calls ! 

To be able to do what they are designed for, these routines talk directly to 
the ext2-os2.ifs file system driver, through the DosFsCtl call. DosFsCtl 
(almost) directly calls the IFS's FS_FSCTL entry point. Most of the work is 
done in kernel mode inside ext2-os2.ifs. DosFsCtl is the IFS equivalent of 
DosDevIOCtl : an extended interface to the IFS. 

The library is compiled using IBM Visualage C++ 3.0, but it should also compile 
without problem with emx/gcc. The precompiled vfsapi.dll should work with 
emx/gcc (and other compilers) without recompilation. 

The entry points follow the standard OS/2 32 bits API linkage convention 
(_System for IBM Visualage C++, _syscall for Borland C++). 

Note:  The funtion names, as well as data structure names, have been prefixed 
       with "vfs_" in order to prevent name collisions with existing C/C++ 
       runtime libraries. 

     vfs_fstat 
     vfs_stat 
     vfs_sync 
     vfs_link 


═══ 9.1. vfs_fstat ═══

int vfs_fstat(int fd, struct new_stat *s);

This is the ext2-os2 implementation of the Unix fstat() system call. It returns 
a "true" stat structure containing the "true" Unix file mode, the "true" file 
ownership... This stat structure is the same as the Linux stat structure 
defined in /usr/include/sys/stat.h. All the macros and definitions for file 
modes and file ownership are also the same as in Linux. 

┌────────────────────┬──────────┬─────────────────────────────────────────────┐
│Parameter type      │Input     │Description                                  │
│                    │Output    │                                             │
├────────────────────┼──────────┼─────────────────────────────────────────────┤
│int fd              │input     │must be a valid OS/2 file handle (for example│
│                    │          │returned by DosOpen, or DosDupHandle).       │
├────────────────────┼──────────┼─────────────────────────────────────────────┤
│struct vfs_stat *s  │in/out    │On input : must point to a vfs_stat          │
│                    │          │structure. On output : contains the I-node   │
│                    │          │information.                                 │
└────────────────────┴──────────┴─────────────────────────────────────────────┘


═══ 9.2. vfs_stat ═══

int vfs_stat(const char *pathname, struct new_stat *s);

This is the ext2-os2 implementation of the Unix stat() system call. It returns 
a "true" stat structure containing the "true" Unix file mode, the "true" file 
ownership... This stat structure is the same as the Linux stat structure 
defined in /usr/include/sys/stat.h. All the macros and definitions for file 
modes and file ownership are also the same as in Linux. 

┌────────────────────┬──────────┬─────────────────────────────────────────────┐
│Parameter type      │Input     │Description                                  │
│                    │Output    │                                             │
├────────────────────┼──────────┼─────────────────────────────────────────────┤
│const char *pathname│input     │Must be a valid OS/2 file name (a file name  │
│                    │          │acceptable by DosOpen).                      │
├────────────────────┼──────────┼─────────────────────────────────────────────┤
│struct vfs_stat *s  │in/out    │On input : must point to a vfs_stat          │
│                    │          │structure. On output : contains the I-node   │
│                    │          │information.                                 │
└────────────────────┴──────────┴─────────────────────────────────────────────┘


═══ 9.3. vfs_sync ═══

int vfs_sync(void);

This is the ext2-os2 implementation of the Unix sync() system call. It commits 
all data to the disk. 

Note:  vfs_sync will ONLY work on ext2fs partitions, not on other partitions 
       (FAT, HPFS). 


═══ 9.4. vfs_link ═══

int vfs_link(const char *from, const char *to);

This call creates a hard link to an existing file. 

┌────────────────────┬──────────┬─────────────────────────────────────────────┐
│Parameter type      │Input     │Description                                  │
│                    │Output    │                                             │
├────────────────────┼──────────┼─────────────────────────────────────────────┤
│const char *from    │input     │File name to be linked. Must be a valid OS/2 │
│                    │          │file name (a file name acceptable by         │
│                    │          │DosOpen).                                    │
├────────────────────┼──────────┼─────────────────────────────────────────────┤
│const char *to      │input     │New file name. Must be a valid OS/2 file name│
│                    │          │(a file name acceptable by DosOpen).         │
└────────────────────┴──────────┴─────────────────────────────────────────────┘


═══ 10. Booting OS/2 from a Linux ext2fs partition ═══

Ext2-os2 allows OS/2 to boot from a Linux ext2fs partition. It can even makes 
OS/2 reside alongside Linux on the same partition (I mean, having vmlinuz and 
OS2KRNL on the same file system !). 

Note:  There is only one restriction : you must still have a small FAT or HPFS 
       partition to store the WPS desktop tree (usually <boot 
       drive>:\DESKTOP\*), because the Linux ext2 file system does not support 
       EAS (extended attributes), which are necessary to the WPS. 

 Anyway, it is not completely useless though : with this feature you can for 
 instance install an OS/2 maintenance partition on a Linux ext2fs drive, and 
 thus save a partition. You can also install a full OS/2 system, if you store 
 the desktop and the swap file on a FAT or HPFS partition. 

 The following chapters will describe how to setup an OS/2 system on a Linux 
 ext2fs partition. It's not an easy thing to do, because as you will see it's 
 like if we gathered on a single partition the usual problems encountered when 
 installing a Linux+OS/2 system !! 

 For the moment booting OS/2 from a Linux partition is still quite experimental 
 and is still at beta level stabiliy ! To try this feature, you'll have to be 
 very familiar with OS/2 installation, Linux installation, and LILO + OS/2 Boot 
 Manager configuration. Please make a full backup of your system first ! 

 I'm waiting for your feedback on this new feature, both failures and success 
 stories ! 

   1. Understanding the normal OS/2 HPFS boot sequence 
   2. Understanding the normal Linux ext2fs boot sequence 
   3. Understanding the ext2-os2 OS/2 ext2fs boot sequence 
   4. Installing and configuring OS/2 on a Linux ext2fs partition 


═══ 10.1. Understanding the normal OS/2 HPFS boot sequence ═══

 Assume you have an OS/2 system with Boot Manager installed. Here's a 
simplified HPFS boot sequence : 

   1. Ctrl+Alt+Del or Power on gives control to the power-on-self-test (POST) 
      code. the POST code after hardware initializations issues an INT 19h. 
   2. INT 19h load the first sector of the first drive (the MBR), and executes 
      it. This code will load and execute the boot record of the active primary 
      partition. 
   3. This active partition is Boot Manager, it will present you a menu of 
      bootable systems. 
   4. When you select the OS/2 system, boot manager reads and executes the code 
      in the boot record of the selected partition. 
   5. This code is called a "black box". Its role is to load a micro IFS in 
      memory. This micro IFS is a real mode IFS with just enough brain to read 
      a few files. 
   6. When the micro IFS receives control, it loads the OS/2 loader (stored in 
      OS2LDR). It also loads a mini IFS : a mini IFS is a protect mode IFS with 
      limited read only capabilities, just a little bit more clever than a 
      micro IFS ! (The HPFS mini IFS is stored in OS2BOOT) It then passes 
      control to OS2LDR, along with the current state of memory. 
   7. OS2LDR then loads OS2KRNL from disk, applies required fixups, and gives 
      control to the kernel. 
   8. The kernel then loads the mini IFS from its memory image. The mini IFS is 
      used to load the base device drivers, then the normal device drivers and 
      IFSes, then the VDD drivers. 
   9. When the normal HPFS IFS is loaded, it replaces the mini IFS. 

 As you can see, the OS/2 kernel doesn't know anything about HPFS. It even 
 makes no assumption of booting off of disk. 


═══ 10.2. Understanding the normal Linux ext2fs boot sequence ═══

 Assume you have a Linux system with OS/2 Boot Manager installed as the main 
boot loader, and LILO installed in the superblock of the Linux boot partition. 

   1. Ctrl+Alt+Del or Power on gives control to the power-on-self-test (POST) 
      code. the POST code after hardware initializations issues an INT 19h. 
   2. INT 19h load the first sector of the first drive (the MBR), and executes 
      it. This code will load and execute the boot record of the active primary 
      partition. 
   3. This active partition is Boot Manager, it will present you a menu of 
      bootable systems. 
   4. When you select the Linux system, boot manager reads and executes the 
      code in the boot record of the selected partition. 
   5. This code is actually the LILO first stage loader. its role is to load 
      the LILO second stage loader. 
   6. The LILO second stage loader loads the disk map of the boot image, then 
      using this map it loads the Linux kernel, as well as command line 
      parameters. It then gives control to the setup code stored at the 
      beginning of the Linux boot image. 
   7. The setup code then explores the hardware configuration, switches to 
      protect mode and passes control to the Linux kernel. 
   8. When the Linux kernel receives control, it executes its initialization 
      routines, and initializes the device drivers present in the boot image. 
   9. It then executes the init process. 

 The main difference here is that the Linux kernel does not load dynamically 
 the device drivers, and especially the ext2fs driver. They are all present in 
 the boot image. While it greatly simplifies the boot sequence (no mini IFS, no 
 micro IFS ...) it requires a reconmpilation whenever you want to add or remove 
 a driver. 

 When you run LILO to install a new kernel, it installs a map of the kernel 
 boot image, and a boot loader in the boot record. LILO's main advantage is to 
 be independant of the file system it uses, and even more : it knows absolutely 
 nothing about ext2fs, FAT and so on. It uses the disk map to load the boot 
 image. This is the reason why you must re-run LILO whenever you touch your 
 kernel file. LILO makes no assumption of booting a Linux kernel image : it can 
 load other systems as well : it loads the specified file in a well known place 
 in memory regardless of its type, and gives it control. 


═══ 10.3. Understanding the ext2-os2 OS/2 ext2fs boot sequence ═══

ext2-os2 uses a mix of these two different sequences to boot OS/2 from a Linux 
ext2fs partition. In this package there are two new files : 

     MICROFSD.FSD : the Linux ext2fs micro IFS 
     MINIFSD.FSD : the Linux ext2fs mini FSD 

 The main difference between the normal OS/2 boot sequence and the ext2-os2 
 boot sequence is that ext2-os2 uses LILO as the "black box" described above. 
 LILO is actually configured to load MICROFSD.FSD instead of a standard Linux 
 kernel boot image. 

 Note:  LILO is required to boot OS/2 from a Linux partition. I tested it with 
        LILO 0.15 but it should also work with later versions. 

 The boot sequence will be (assuming you have OS/2 boot manager installed as 
 the main boot loader, and LILO in the superblock of the Linux partition) : 

   1. Ctrl+Alt+Del or Power on gives control to the power-on-self-test (POST) 
      code. the POST code after hardware initializations issues an INT 19h. 
   2. INT 19h load the first sector of the first drive (the MBR), and executes 
      it. This code will load and execute the boot record of the active primary 
      partition. 
   3. This active partition is Boot Manager, it will present you a menu of 
      bootable systems. 
   4. When you select the OS/2 system, boot manager reads and executes the code 
      in the boot record of the selected partition. 
   5. This code is actually the LILO first stage loader. its role is to load 
      the LILO second stage loader. 
   6. The LILO second stage loader loads the disk map of the boot image (which 
      is actually a map of MICROFSD.FSD), then using this map it loads the 
      micro FSD as well as as command line parameters. It then gives control to 
      the micro FSD. 
   7. When the micro IFS receives control, it loads the OS/2 loader (stored in 
      OS2LDR). It also loads the ext2fs mini IFS (MINIFSD.FSD). It then passes 
      control to OS2LDR, along with the current state of memory. 
   8. OS2LDR then loads OS2KRNL from disk, applies required fixups, and gives 
      control to the kernel. 
   9. The kernel then loads the mini IFS from its memory image. The mini IFS is 
      used to load the base device drivers, then the normal device drivers and 
      IFSes, then the VDD drivers. 
  10. When the normal ext2-os2 IFS is loaded, it replaces the mini IFS. 


═══ 10.4. Installing and configuring OS/2 on a Linux ext2fs partition ═══

   1. Sample initial situation 
   2. Copying OS/2 a Linux ext2fs partition 
   3. Copying the required ext2-os2 files on the ext2fs partition 
   4. Updating CONFIG.SYS 
   5. Configuring and running LILO to boot OS/2 from your ext2fs partition 
   6. Now time to reboot ... 


═══ 10.4.1. Sample initial situation ═══

hard disk 1 (100 Mb IDE drive) : 

┌──────────┬──────────┬──────────┬──────────┬─────────────────────────────────────────────┐
│OS/2 drive│OS/2 BM   │Linux     │Linux     │Description                                  │
│letter    │name      │device    │mount     │                                             │
│          │          │name      │point     │                                             │
├──────────┼──────────┼──────────┼──────────┼─────────────────────────────────────────────┤
│          │          │/dev/hda2 │          │OS/2 boot manager partition                  │
├──────────┼──────────┼──────────┼──────────┼─────────────────────────────────────────────┤
│C:        │test      │/dev/hda3 │/disk_c   │Linux ext2fs test (type 0x07) partition      │
│          │          │          │          │(ext2flt not used to mount it)               │
├──────────┼──────────┼──────────┼──────────┼─────────────────────────────────────────────┤
│F:        │linux     │/dev/hda1 │/         │Linux ext2fs root (type 0x83) partition      │
│          │          │          │          │(ext2flt used to mount it)                   │
└──────────┴──────────┴──────────┴──────────┴─────────────────────────────────────────────┘

((See also linux and OS/2 fdisk output for drive 1) 

hard disk 2 (1.2 Gb SCSI drive) : 

┌──────────┬──────────┬──────────┬──────────┬─────────────────────────────────────────────┐
│OS/2 drive│OS/2 BM   │Linux     │Linux     │Description                                  │
│letter    │name      │device    │mount     │                                             │
│          │          │name      │point     │                                             │
├──────────┼──────────┼──────────┼──────────┼─────────────────────────────────────────────┤
│D:        │WARP      │/dev/sda1 │          │primary OS/2 HPFS with OS/2 WARP installed   │
│          │          │          │          │(normal use)                                 │
├──────────┼──────────┼──────────┼──────────┼─────────────────────────────────────────────┤
│D:        │EXT2-OS2  │/dev/sda2 │          │primary OS/2 HPFS with OS/2 and kernel       │
│          │          │          │          │debugger (to test ext2-os2)                  │
├──────────┼──────────┼──────────┼──────────┼─────────────────────────────────────────────┤
│E:        │          │/dev/sda5 │          │extended OS/2 HPFS data partition.           │
└──────────┴──────────┴──────────┴──────────┴─────────────────────────────────────────────┘

((See also linux and OS/2 fdisk output for drive 2) 

LILO is installed in the superblock of /dev/hda1, and is configured to load 
/vmlinuz from this partition. (lilo.conf used to configure LILO) 


═══ 10.4.2. Copying OS/2 a Linux ext2fs partition ═══

To install OS/2, the easiest way will be to use the IBM EWS bootos2.zip package 
(available on usual OS/2 ftp sites). This package can create a bootable system 
from an existing installed system. It can install a system on either floppies 
or hard disks. Moreover it can create a minimal system (single text mode 
session), a PM enabled system, a WPS enabled system ... 

Here's how I installed a PM enabled system on drive F:. To do this, I booted my 
normal OS/2 system, went to the bootos2 directory and typed : 

BOOTOS2 TARGET=F: TYPE=PM

Some important remarks: 

     you must have the two OS/2 installation disks handy, bootos2 will ask 
      them. 
     when BOOTOS2 asks if you want to format the target partition, you must 
      answer to skip format !! 
     UEXT2.DLL must be somewhere in you LIBPATH, otherwise bootos2 will abort. 
     Sometimes for some reason, bootos2 forgets to copy the hidden OS2DUMP 
      file in the root directory. Please ensure that it has been copied. 

 Note on how to install a WPS enabled system 


═══ 10.4.3. Copying the required ext2-os2 files on the ext2fs partition ═══

Make a directory on the ext2fs partition to store ext2-os2 files (for instance 
F:\os2\fs\ext2) and copy the following files in it : 

     ext2-os2.exe 
     ext2-os2.ifs 
     ext2-os2.sym 
     ext2_lw.exe 
     sync.exe 
     vfsapi.dll 
     uext2.dll 

 Copy the following files in F:\OS2 : or in F:\OS2\BOOT 

     ext2flt.flt 
     ext2flt.sym 
     mwdd32.sys 
     mwdd32.sym 

 Copy the following files in F:\ : 

     microfsd.fsd 
     minifsd.fsd 


═══ 10.4.4. Updating CONFIG.SYS ═══

Then add the following statements at the beginning of F:\CONFIG.SYS  : 

REM
REM this should be the beginning of CONFIG.SYS
REM
IFS=F:\OS2\FS\EXT2\EXT2-OS2.IFS -cache:256 -rw -case_retensive <add other options if necessary>
RUN=F:\OS2\FS\EXT2\EXT2_LW.EXE
BASEDEV=MWDD32.SYS
BASEDEV=EXT2FLT.FLT /W <add other options if necessary>

Note: 

     the IFS managing the boot drive - in this case ext2-os2.ifs - must always 
      be the first statement in CONFIG.SYS, or OS/2 will crash during boot. 
     write access must be enabled or OS/2 will not work properly. 
     the -case_retensive option is necessary or OS/2 won't boot 


═══ 10.4.5. Configuring and running LILO to boot OS/2 from your ext2fs partition ═══

 Now reboot under Linux. In this step we'll need to know several things : 

     the boot drive geometry (number of tracks, number of sectors per track). 
      You can know them in your drive's manual, or in your BIOS setup. 
     the boot partition geometry (starting sector, number of sectors). You can 
      know it using fdisk -l (list partitions) 
     the boot drive BIOS device code (first drive is 0x80, second is 0x81 ...) 
 Then edit the /etc/lilo.conf to add /microfsd.fsd as a new boot image : 

  #
  # Start LILO global section
  #
  boot = /dev/hda1                # installs LILO in the superblock of /dev/hda1
  vga = normal                    # force sane state
  ramdisk = 0                     # paranoia setting
  prompt
  #
  # End LILO global section
  #
  # Linux bootable partition config begins
  #
  image = /vmlinuz                # normal Linux kernel image
    root = /dev/hda1
    label = linux                 # name to enter at LILO prompt
    read-only                     # Non-UMSDOS filesystems should be mounted read-only for checking
  #
  # Linux bootable partition config ends
  #
  # OS/2 bootable partition config begins
  #
  image = /microfsd.fsd           # ext2-os2 micro-fsd file (OS2LDR loader).
    label = test_os2              # name to enter at LILO prompt
    append = "os2_bios_device=0x80 os2_bootdrive=5 os2_hidden_sectors=0x19884 os2_sect=35 os2_head=6 os2_total_sectors=100800 os2_bytes_per_sector=512"
  #
  # OS/2 bootable partition config ends
  #

 In the "append=" we've passed mandatory informations to the ext2-os2 micro FSD 
 so that it can boot OS/2 properly. Below is a description of these parameters. 

 Note:  All these parameters are required. A wrong value in one of them will 
        prevent OS/2 from booting normally. 

  ┌────────────────────┬─────────────────────────────────────────────┐
  │os2_bootdrive       │this is the OS/2 drive letter of the boot    │
  │                    │drive. It MUST REFLECT the drive letter as   │
  │                    │seen in a normal situation (in our case F:). │
  │                    │You CANNOT REORDER drive letters with this   │
  │                    │option ; use ext2flt's /M option instead and │
  │                    │update this value accordingly.               │
  ├────────────────────┼─────────────────────────────────────────────┤
  │os2_bios_device     │this is the physical drive BIOS device code. │
  │                    │Floppy drives are 0 (A:) and 1 (:). Hard     │
  │                    │drives are numbered starting at 0x80, so     │
  │                    │physical hard drive 1 is 0x80, drive 2 is    │
  │                    │0x81.                                        │
  ├────────────────────┼─────────────────────────────────────────────┤
  │os2_hidden_sectors  │this is sector number (512 bytes) of the     │
  │                    │first sector of the boot partition. Use fdisk│
  │                    │-uS -l under Linux to know it.               │
  ├────────────────────┼─────────────────────────────────────────────┤
  │os2_total_sectors   │this is the size of the boot partition in 512│
  │                    │bytes sectors. Use fdisk -uS -l to know it.  │
  ├────────────────────┼─────────────────────────────────────────────┤
  │os2_sect            │this is the number of sectors per track      │
  ├────────────────────┼─────────────────────────────────────────────┤
  │os2_bytes_per_sector│this is the hardware sector size in bytes    │
  │                    │(normally 512 bytes).                        │
  ├────────────────────┼─────────────────────────────────────────────┤
  │os2_head            │this is the number of heads                  │
  └────────────────────┴─────────────────────────────────────────────┘

 Then run the map installer to update LILO configuration as follows : 

  lilo -v


═══ 10.4.6. Now time to reboot ... ═══

Now we can shutdown the system and reboot. Once in OS/2 boot manager menu, 
select "linux". This will start LILO. Once LILO is up - it should prompt 
"lilo:" at the top left corner of the screen - type "test_os2", then <ENTER> 
then let's prey !  Here are some checkpoints : 

┌────────────────────┬─────────────────────────────────────────────┐
│ext2-os2 MICROFSD   │microfsd.fsd has been successfuly loaded by  │
│banner appears on   │LILO, and has been successfuly started.      │
│top of the screen   │                                             │
├────────────────────┼─────────────────────────────────────────────┤
│a little white box  │it means that microfsd.fsd has successfuly   │
│followed by the word│loaded OS2LDR and minifsd.fsd, and that      │
│"OS2" appears at the│OS2LDR has been successfuly started.         │
│top left corner of  │                                             │
│the screen.         │                                             │
├────────────────────┼─────────────────────────────────────────────┤
│the OS/2 logo bitmap│it means that OS2LDR has successfuly loaded  │
│appears             │OS2KRNL, successfuly started the kernel, that│
│                    │microfsd.fsd's job is over, and that         │
│                    │minifsd.fsd's job is starting. From this     │
│                    │point CONFIG.SYS base device drivers begin to│
│                    │be loaded.                                   │
├────────────────────┼─────────────────────────────────────────────┤
│the normal ext2-os2 │ext2-os2.ifs has just been loaded.           │
│banner appears.     │                                             │
├────────────────────┼─────────────────────────────────────────────┤
│PM appears (if a PM │OS/2 boot process is successful !            │
│system), or the OS/2│                                             │
│command prompt      │                                             │
│appears.            │                                             │
└────────────────────┴─────────────────────────────────────────────┘

Note: 

      The boot process will be probably slightly longer than a normal boot. 
      This is absolutely normal. 
      You can press ALT+F2 when the little white box appears : it will make 
      OS/2 tell what driver it is currently loading. 
      You can hook an ASCII terminal on the COM2 port (for instance a PC with a 
      terminal emulator, and a null-modem cable), ext2-os2 will write debug 
      information on COM2 port during the boot process. 


═══ 11. Linux ext2 file system utilities ═══

Note: 

     These utilities require the EMX runtime library (it can be found at 
      ftp://ftp-os2.nmsu.edu/os2/unix/emx09c/emxrt.zip). 
     These utilities require the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     e2fsck - check a Linux second extended file system 
     mke2fs - create a Linux second extended file system 
     badblocks - search a device for bad blocks 
     debugfs - ext2 file system debugger 
     chattr - change file attributes on a Linux second extended file system 
     lsattr - list file attributes on a Linux second extended file system 
     tune2fs - adjust tunable filesystem parameters on second extended 
      filesystems 


═══ 11.1. e2fsck - check a Linux second extended file system ═══

This program is also available through the standard CHKDSK command, with the 
same syntax described below. 

Note: 

     This chapter is a (almost) direct adaptation to IPF of the original Linux 
      e2fsck(8) man page. 
     If after running this program, you cannot access your drive and get a 
      message from OS/2 like "Cannot access the specified drive" or "The 
      specified drive is not formatted", DON'T PANIC ! It does *NOT* mean your 
      drive is lost; it only means that for some reason something went wrong 
      while trying to remount the drive after the e2fsck operation. Rebooting 
      OS/2 generally solves the problem. 
     This program requires the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     synopsis 
     description 
     options 
     exit code 
     reporting bugs 
     author 
     see also 


═══ 11.1.1. synopsis ═══

       e2fsck  [  -pacnyrdfvstFSV ] [ -b superblock ] [ -B block-
       size ] [ -l|-L bad_blocks_file ] device


═══ 11.1.2. description ═══

       e2fsck is used to check a Linux second extended file  sys-
       tem.

       device is  the  special  file  corresponding to the device
              (e.g /dev/hdXX on Linux, X: on OS/2).


═══ 11.1.3. options ═══

       -a     This option does the same thing as the  -p  option.
              It is provided for backwards compatibility only; it
              is suggested that people use -p option whever  pos-
              sible.

       -b superblock
              Instead  of  using  the  normal superblock, use the
              alternative superblock specified by superblock.

       -B blocksize
              Normally, e2fsck will search for the superblock  at
              various different block sizes in an attempt to find
              the appropriate block size.   This  search  can  be
              fooled in some cases.  This option forces e2fsck to
              only try locating the superblock  at  a  particular
              blocksize.   If the superblock is not found, e2fsck
              will terminate with a fatal error.

       -c     This option causes e2fsck to run  the  badblocks(8)
              program  to  find  any  blocks which are bad on the
              filesystem, and then marks them as  bad  by  adding
              them to the bad block inode.

       -d     Print  debugging  output  (useless  unless  you are
              debugging e2fsck ).

       -f     Force checking even if the file system seems clean.

       -F     Flush  the filesystem device's buffer caches before
              beginning.  Only really  useful  for  doing  e2fsck
              time trials.

       -l filename
              Add  the  blocks  listed  in  the file specified by
              filename to the list of bad blocks.

       -L filename
              Set the bad blocks list to be the  list  of  blocks
              specified by filename.  (This option is the same as
              the -l  option,  except  the  bad  blocks  list  is
              cleared  before  the  blocks listed in the file are
              added to the bad blocks list.)

       -n     Open the filesystem read-only, and assume an answer
              of  ``no''  to  all questions.  Allows e2fsck to be
              used non-interactively.  (Note: if the -c,  -l,  or
              -L  options  are  specified  in  addition to the -n
              option, then the filesystem will  be  opened  read-
              write, to permit the bad-blocks list to be updated.
              However, no other  changes  will  be  made  to  the
              filesystem.)

       -p     Automatically  repair  ("preen")  the  file  system
              without any questions.

       -r     This option does nothing at  all;  it  is  provided
              only for backwards compatibility.

              -s  This  option  will  byte-swap the filesystem so
              that it is using  the  normalized,  standard  byte-
              order  (which  is  i386  or little endian).  If the
              filesystem is already in the  standard  byte-order,
              e2fsck will take no action.

       -S     This  option will byte-swap the filesystem, regard-
              less of its current byte-order.

       -t     Print timing statistics for e2fsck.  If this option
              is  used  twice,  additional  timing statistics are
              printed on a pass by pass basis.

       -v     Verbose mode.

       -V     Print version information and exit.

       -y     Assume an  answer  of  ``yes''  to  all  questions;
              allows e2fsck to be used non-interactively.


═══ 11.1.4. exit code ═══

       The exit code returned by e2fsck is the sum of the follow-
       ing conditions:
            0    - No errors
            1    - File system errors corrected
            2    - File system errors corrected, system should
                   be rebooted if file system was mounted
            4    - File system errors left uncorrected
            8    - Operational error
            16   - Usage or syntax error
            128  - Shared library error


═══ 11.1.5. reporting bugs ═══

       Almost any piece of software will have bugs.  If you  man-
       age  to find a filesystem which causes e2fsck to crash, or
       which e2fsck is unable to repair, please report it to  the
       author.

       Please include as much information as possible in your bug
       report.  Ideally, include a  complete  transcript  of  the
       e2fsck  run,  so I can see exactly what error messages are
       displayed.  If you have a writeable filesystem  where  the
       transcript can be stored, the script(1) program is a handy
       way to save the output of to a file.

       It is also useful to send the output of dumpe2fs(8).  If a
       specific  inode  or inodes seems to be giving e2fsck trou-
       ble, try running the debugfs(8) command and send the  out-
       put  of the stat command run on the relevant inode(s).  If
       the inode is a directory, the debugfs  dump  command  will
       allow  you to extract the contents of the directory inode,
       which can sent to me after being first run  through  uuen-
       code(1).

       Always  include  the full version string which e2fsck dis-
       plays when it is run, so I know which version you are run-
       ning.


═══ 11.1.6. author ═══

       This  version  of  e2fsck  is  written  by  Theodore  Ts'o
       <tytso@mit.edu>.
       It has been ported to OS/2 by Matthieu Willm.


═══ 11.1.7. see also ═══

       mke2fs(8), tune2fs(8), dumpe2fs(8), debugfs(8)


═══ 11.2. mke2fs - create a Linux second extended file system ═══

This program is also available through the standard FORMAT command, with the 
same syntax described below. 

Note: 

     This chapter is a (almost) direct adaptation to IPF of the original Linux 
      mke2fs(8) man page. 
     If after running this program, you cannot access your drive and get a 
      message from OS/2 like "Cannot access the specified drive" or "The 
      specified drive is not formatted", DON'T PANIC ! It does *NOT* mean your 
      drive is lost; it only means that for some reason something went wrong 
      while trying to remount the drive after the mke2fs operation. Rebooting 
      OS/2 generally solves the problem. 
     This program requires the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     synopsis 
     description 
     options 
     author 
     bugs 
     availability 
     see also 


═══ 11.2.1. synopsis ═══

       mke2fs  [  -c | -l filename ] [ -b block-size ] [ -f frag-
       ment-size ] [ -i bytes-per-inode ] [  -m  reserved-blocks-
       percentage  ]  [ -o creator-os ] [ -q ] [ -v ] [ -F ] [ -L
       volume-label ] [ -M last-mounted-directory ] [ -S ] device
       [ blocks-count ]


═══ 11.2.2. description ═══

mke2fs is used to create a Linux second extended file system on a device 
(usually a disk partition). 

device is the special file corresponding to the device (e.g /dev/hdXX). 

blocks-count is the number of blocks on the device. If omitted, mke2fs 
automatically figures the file system size. 

Note:  On OS/2, device is actually the drive letter of the partition to be 
       formatted (e.g X:) 


═══ 11.2.3. options ═══

       -b block-size
              Specify the size of blocks in bytes.

       -c     Check the device for bad blocks before creating the
              file system, using a fast read-only test.

       -f fragment-size
              Specify the size of fragments in bytes.

       -i bytes-per-inode
              Specify the bytes/inode ratio.  mke2fs  creates  an
              inode  for  every bytes-per-inode bytes of space on
              the disk.   This  value  defaults  to  4096  bytes.
              bytes-per-inode must be at least 1024.

       -l filename
              Read the bad blocks list from filename

       -m reserved-blocks-percentage
              Specify  the  percentage of reserved blocks for the
              super-user.  This value defaults to 5%.

       -o     Manually override the default value of the "creator
              os"  field of the filesystem.  Normally the creator
              field is set by default to the  native  OS  of  the
              mke2fs executable.

       -q     Quiet  execution.   Useful  if  mke2fs  is run in a
              script.

       -v     Verbose execution.

       -F     Force mke2fs to run, even if the  specified  device
              is not a block special device.

       -L     Set the volume label for the filesystem.

       -M     Set  the last mounted directory for the filesystem.
              This might be useful for the sake of utilities that
              key  off of the last mounted directory to determine
              where the filesytem should be mounted.

       -S     Write superblock and group descriptors only.   This
              is  useful  if  all  of  the  superblock and backup
              superblocks are corrupted, and a last-ditch  recov-
              ery  method is desired.  It causes mke2fs to reini-
              tialize the superblock and group descriptors, while
              not  touching  the  inode  table  and the block and
              inode bitmaps.  The e2fsck program  should  be  run
              immediately after this option is used, and there is
              no guarantee that any data will be salvageable.


═══ 11.2.4. author ═══

       This version of mke2fs has been written by  Theodore  Ts'o
       <tytso@mit.edu>.
       It has been ported to OS/2 by Matthieu Willm.


═══ 11.2.5. bugs ═══

       mke2fs  accepts  the  -f  option  but currently ignores it
       because the second extended file system does  not  support
       fragments yet.
       There  may be some other ones.  Please, report them to the
       author.


═══ 11.2.6. availability ═══

       mke2fs is available for anonymous ftp from ftp.ibp.fr  and
       tsx-11.mit.edu in /pub/linux/packages/ext2fs.


═══ 11.2.7. see also ═══

       dumpe2fs(8), e2fsck(8), tune2fs(8)


═══ 11.3. badblocks - search a device for bad blocks ═══

Note: 

     This chapter is an (almost) direct adaptation to IPF of the original 
      Linux badblocks(8) man page. 
     To conform to the FAT 8:3 naming convention, this program has been 
      renamed badblks.exe. 
     This program requires the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     synopsis 
     description 
     options 
     warning 
     author 
     bugs 
     availability 
     see also 


═══ 11.3.1. synopsis ═══

       badblocks [ -b block-size ] [ -o output_file ] [ -v ] [ -w
       ] device blocks-count


═══ 11.3.2. description ═══

       badblocks is used to search for bad  blocks  on  a  device
       (usually a disk partition).
       device  is  the  special  file corresponding to the device
       (e.g /dev/hdXX).
       blocks-count is the number of blocks on the device.


═══ 11.3.3. options ═══

       -b block-size
              Specify the size of blocks in bytes.

       -o output_file
              Write the list of bad blocks to the specified file.
              Without this option, badblocks displays the list on
              its standard output.

       -v     Verbose mode.

       -w     Use write-mode test. With  this  option,  badblocks
              scans  for  bad  blocks  by  writing  some patterns
              (0xaa, 0x55, 0xff, 0x00)  on  every  block  of  the
              device,  reading every block and comparing the con-
              tents.

       Note:  The -w option is not yet implemented on OS/2.


═══ 11.3.4. warning ═══

       Never use the `-w'  option  on  an  device  containing  an
       existing file system.  This option erases data!


═══ 11.3.5. author ═══

       badblocks     has    been    written    by    Remy    Card
       <card@masi.ibp.fr>, the developer and  maintainer  of  the
       ext2 fs.
       It has been rewritten to OS/2 by Matthieu Willm.


═══ 11.3.6. bugs ═══

       I  had no chance to make reals tests of this program since
       I use IDE drives which remap bad blocks. I only made  some
       tests on floppies.


═══ 11.3.7. availability ═══

       badblocks  is  available for anonymous ftp from ftp.ibp.fr
       and tsx-11.mit.edu in /pub/linux/packages/ext2fs.


═══ 11.3.8. see also ═══

       e2fsck(8), mke2fs(8)


═══ 11.4. debugfs - ext2 file system debugger ═══

Note: 

     This chapter is a (almost) direct adaptation to IPF of the original Linux 
      debugfs(8) man page. 
     This program requires the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     synopsis 
     description 
     options 
     commands 
     specifying files 
     author 
     see also 


═══ 11.4.1. synopsis ═══

       debugfs [ -w ] [ -f cmd_file ] [ -R request ] [ device ]


═══ 11.4.2. description ═══

       The  debugfs  program is a file system debugger. It can be
       used to examine and change the state of an ext2 file  sys-
       tem.
       device  is  the  special  file corresponding to the device
       containing the ext2 file system (e.g /dev/hdXX).


═══ 11.4.3. options ═══

       -w     Specify that the file  system  should  be  open  in
              read-write mode. Without this option, the file sys-
              tem is open in read-only mode.

       -f cmd_file
              Causes debugfs to read in commands  from  cmd_file,
              and execute them.  When debugfs is finsihed execut-
              ing those commands, it will exit.

       -R request
              Causes  debugfs  to  execute  the  single   command
              request, and then exit.


═══ 11.4.4. commands ═══

       debugfs  is an interactive debugger. It understands a num-
       ber of commands.

       cat filespec
              Dump the contents of the inode filespec to  stdout.

       cd filespec
              Change the current working directory to filespec.

       chroot filespec
              Change the root directory to be the directory file-
              spec.

       close  Close the currently open file system.

       clri file
              Clear the contents of the inode file.

       dump [-p] filspec out_file
              Dump the contents of the inode filespec to the out-
              put  file  out_file.  If the -p option is given set
              the owner, group  and  permissions  information  on
              out_file to match filespec.

       expand_dir filespec
              Expand the directory filespec.


       find_free_block [goal]
              Find  the  first free block, starting from goal and
              allocates it.

       find_free_inode [dir [mode]]
              Find a free inode and allocates  it.   If  present,
              dir  specifies  the  inode  number of the directory
              which the inode  is  to  be  located.   The  second
              optional argument mode specifies the permissions of
              the new inode.  (If the directory bit is set on the
              mode,  the allocation routine will function differ-
              ently.)

       freeb block
              Mark the block number block as not allocated.

       freei filespec
              Free the inode specified by filespec

       help   Print a list of commands understood by  debugfs(8).

       icheck block ...
              Print  a listing of the inodes which use the one or
              more block specified on the command line.

       initialize device blocksize
              Create an ext2 file system on  device  with  device
              size blocksize.  Note that this does not fully ini-
              tialize all of the data structures; to do this, use
              the  mke2fs(8) program.  This is just a call to the
              low-level library, which sets up the superblock and
              block descriptors.

       kill_file filespec
              Dellocate  the inode filespec and its blocks.  Note
              that this does not remove any directory entries (if
              any) to this inode.  See the rm command if you wish
              to unlink a file.

       ln filespec dest_file
              Create a link named dest_file which is  a  link  to
              filespec.  Note this does not adjust the inode ref-
              erence counts.

       ls [-l] filespec
              Print a listing of the files in the directory file-
              spec.

       modify_inode filespec
              Modify  the  contents of the inode structure in the
              inode filespec.

       mkdir filespec
              Make a directory.


       mknod filespec [p|[[c|b] major minor]]
              Create a special device file (a named pipe, charac-
              ter  or  block  device).   If  a character or block
              device is to be made, the major  and  minor  device
              numbers must be specified.

       ncheck inode_num ...
              Take the requested list of inode numbers, and print
              a listing of pathnams to those inodes.

       open [-w] device
              Open a file system for editing.

       pwd    Print the current working directory.

       quit   Quit debugfs

       rm pathname
              Unlink pathname.  If this cuases the inode  pointed
              to by pathname to have no other references, deallo-
              cate the  file.   This  command  functions  as  the
              unlink() system call.

       rmdir filespec
              Remove  the  directory  filespec.  This function is
              currently not implemented.

       setb block
              Mark the block number block as allocated.

       seti filespec
              Mark inode filespec as in use in the inode  bitmap.

       show_super_stats
              List the contents of the super block.

       stat filespec
              Display  the contents of the inode structure of the
              inode filespec.

       testb block
              Test if the block number block is marked  as  allo-
              cated in the block bitmap.

       testi filespec
              Test  if  the inode filespec is marked as allocated
              in the inode bitmap.

       unlink pathname
              Remove the link specified by pathname to an  inode.
              Note  this  does  not  adjust  the  inode reference
              counts.

       write source_file out_file
              Create a file in the filesystem named out_file, and
              copy  the contents of source_file into the destina-
              tion file.


═══ 11.4.5. specifying files ═══

       Many debugfs commands take a filespec as  an  argument  to
       specify  an  inode  (as  opposed  to  a  pathname)  in the
       filesystem which is currently opened by debugfs. The file-
       spec  argument  may  be specified in two forms.  The first
       form is an inode  number  surrounded  by  angle  brackets,
       e.g., <2>.  The second form is a pathname; if the pathname
       is prefixed by a forward slash ('/'), then  it  is  inter-
       preted  relative  to  the  root of the filesystem which is
       currently opened by debugfs.   If  not,  the  pathname  is
       interpreted  relative  to the current working directory as
       maintained by debugfs.  This may be modified by using  the
       debugfs command cd.


═══ 11.4.6. author ═══

       debugfs was written by Theodore Ts'o <tytso@mit.edu>.


═══ 11.4.7. see also ═══

       dumpe2fs(8), e2fsck(8), mke2fs(8)


═══ 11.5. chattr - change file attributes on a Linux second extended file system ═══

Note: 

     This chapter is a (almost) direct adaptation to IPF of the original Linux 
      chattr(1) man page. 
     This program requires the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     synopsis 
     description 
     options 
     attributes 
     author 
     bugs and limitations 
     availability 
     see also 


═══ 11.5.1. synopsis ═══

       chattr [ -RV ] [ -v version ] [ mode ] files...


═══ 11.5.2. description ═══

       chattr changes the files attributes on an second  extended
       file system.

       The format of a symbolic mode is +-=[ASacdisu].

       The  operator  `+'  causes  the  selected attributes to be
       added to the existing attributes of the files; `-'  causes
       them  to  be  removed;  and `=' causes them to be the only
       attributes that the files have.

       The letters `ASacdisu' select the new attributes  for  the
       files:  don't  update  atime (A), synchronous updates (S),
       append only (a), compressed (c), immutable  (i),  no  dump
       (d), secure deletion (s), and undeletable (u).


═══ 11.5.3. options ═══

       -R     Recursively  change  attributes  of directories and
              their contents.

       -V     Verbosely describe changed attributes.

       -v version
              Set the files version.


═══ 11.5.4. attributes ═══

       When a file with the 'A' attribute set  is  modified,  its
       atime record is not modified.  This avoid a certain amount
       of disk I/O for laptop systems.

       A file with the `a' attribute set  can  only  be  open  in
       append mode for writing.

       A  file  with  the `c' attribute set is automatically com-
       pressed on the disk by the kernel. A read from  this  file
       returns uncompressed data. A write to this file compresses
       data before storing them on the disk.

       A file with the `d' attribute set  is  not  candidate  for
       backup when the dump(8) program is run.

       A  file with the `i' attribute cannot be modified: it can-
       not be deleted or renamed, no link can be created to  this
       file  and  no  data  can  be written to the file. Only the
       superuser can set or clear this attribute.

       When a file with the `s' attribute  set  is  deleted,  its
       blocks are zeroed and written back to the disk.

       When  a  file  with the `S' attribute set is modified, the
       changes are written synchronously on  the  disk;  this  is
       equivalent  to the `sync' mount option applied to a subset
       of the files.

       When a file with the `u' attribute  set  is  deleted,  its
       contents  is  saved.   This allows the user to ask for its
       undeletion.


═══ 11.5.5. author ═══

       chattr has been written by Remy  Card  <card@masi.ibp.fr>,
       the developer and maintainer of the ext2 fs.
       It has been ported to OS/2 by Matthieu Willm <willm@ibm.net>.


═══ 11.5.6. bugs and limitations ═══

       As of ext2 fs 0.5a, the `c' and `u' attribute are not hon-
       oured by the kernel code.  As of the Linux 2.0 kernel, the
       'A'  attribute  is  not  yet supported by the kernel code.
       (The noatime code is still in testing.)

       These attributes will be implemented in a future  ext2  fs
       version.


═══ 11.5.7. availability ═══

       chattr  is available for anonymous ftp from ftp.ibp.fr and
       tsx-11.mit.edu in /pub/linux/packages/ext2fs.


═══ 11.5.8. see also ═══

       lsattr(1)


═══ 11.6. lsattr - list file attributes on a Linux second extended file system ═══

Note: 

     This chapter is a (almost) direct adaptation to IPF of the original Linux 
      lsattr(1) man page. 
     This program requires the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     synopsis 
     description 
     options 
     author 
     bugs 
     availability 
     see also 


═══ 11.6.1. synopsis ═══

       lsattr [ -Radv ] [ files...  ]


═══ 11.6.2. description ═══

       lsattr lists the file attributes on a second extended file
       system.


═══ 11.6.3. options ═══

       -R     Recursively  list  attributes  of  directories  and
              their contents.

       -a     List all files in directories, including files that
              start with `.'.

       -d     List  directories  like  other  files,  rather than
              listing their contents.

       -v     List the files version.


═══ 11.6.4. author ═══

       lsattr has been written by Remy  Card  <card@masi.ibp.fr>,
       the developer and maintainer of the ext2 fs.
       It has been ported to OS/2 by Matthieu Willm <willm@ibm.net>


═══ 11.6.5. bugs ═══

       There are none :-).


═══ 11.6.6. availability ═══

       lsattr  is available for anonymous ftp from ftp.ibp.fr and
       tsx-11.mit.edu in /pub/linux/packages/ext2fs.


═══ 11.6.7. see also ═══

       chattr(1)


═══ 11.7. tune2fs - adjust tunable filesystem parameters on second extended filesystems ═══

Note: 

     This chapter is a (almost) direct adaptation to IPF of the original Linux 
      tune2fs(8) man page. 
     This program requires the "-errors=continue" switch to be set on the 
      ext2-os2.ifs command line in CONFIG.SYS. 

     synopsis 
     description 
     options 
     bugs 
     warning 
     author 
     availability 
     see also 


═══ 11.7.1. synopsis ═══

       tune2fs [ -l ] [ -c max-mount-counts ] [ -e  errors-behav-
       ior ] [ -i interval-between-checks ] [ -m reserved-blocks-
       percentage ] [ -r reserved-blocks-count  ]  [  -s  sparse-
       super-flag ] [ -u user ] [ -g group ] [ -C mount-count ] [
       -L volume-name ] [ -M last-mounted-directory ] [ -U UUID ]
       device


═══ 11.7.2. description ═══

       tune2fs  adjusts  tunable filesystem parameters on a Linux
       second extended filesystem.

       Never use tune2fs on a read/write  mounted  filesystem  to
       change parameters!


═══ 11.7.3. options ═══

       -c max-mount-counts
              adjust   the   maximal  mounts  count  between  two
              filesystem checks.

       -e errors-behavior
              change the behavior of the kernel code when  errors
              are  detected.   errors-behavior  can be one of the
              followings:
                   continue       Continue normal execution.
                   remount-ro     Remount  the  filesystem  read-
              only.
                   panic          Causes a kernel panic.

       -g group
              set  the  user  group  which  can  benefit from the
              reserved blocks.
              group can be a numerical gid or a group name.

       -i interval-between-checks[d|m|w]
              adjust the  maximal  time  between  two  filesystem
              checks.   No  postfix or `d' result in days, `m' in
              months, and `w' in weeks.  A  value  of  zero  will
              disable the timedependent checking.

       -l     list the contents of the filesystem superblock.

       -m reserved-blocks-percentage
              adjust  the reserved blocks percentage on the given
              device.

       -r reserved-blocks-count
              adjust the  reserved  blocks  count  on  the  given
              device.

       -s sparse_super_flag
              sets  and  resets  the sparse_superblock flag.  The
              sparse_superblock feature saves space on really big
              filesystems.   Warning:  The  Linux 2.0 kernel does
              not properly support this feature.  Neither do  all
              Linux 2.1 kernels; please don't use this unless you
              know what you're doing!

       -u user
              set the user who  can  benefit  from  the  reserved
              blocks.   user  can  be  a  numerical uid or a user
              name.

              -C mount-count set the number of times the filesys-
              tem has been mounted.

       -L volume-label
              set the volume label of the filesystem.

       -M last-mounted-directory
              set the last-mounted direcctory for the filesystem.

       -U UUID
              set the UUID of  the  filesystem.   A  sample  UUID
              looks                   like                  this:
              "c1b9d5a2-f162-11cf-9ece-0020afc76f16".   The  uuid
              may  also  be "null", which will set the filesystem
              UUID to the null UUID.  The uuid may also be  "ran-
              dom", which will generate a new random UUID for the
              filesystem.


═══ 11.7.4. bugs ═══

       We didn't find any bugs yet. Perhaps there  are  bugs  but
       it's unlikely.


═══ 11.7.5. warning ═══

       Use  this  utility  on  your  own  risk.  You're modifying
       filesystems.


═══ 11.7.6. author ═══

       tune2fs has been written by Remy Card  <card@masi.ibp.fr>,
       the developer and maintainer of the ext2 fs.
       tune2fs  uses  the ext2fs library written by Theodore Ts'o
       <tytso@mit.edu>.
       This manual page was written by Christian Kuhtz <chk@data-
       hh.Hanse.DE>.
       Timedependent    checking    was   added   by   Uwe   Ohse
       <uwe@tirka.gun.de>.
       tune2fs has been ported to OS/2 by Matthieu Willm <willm@ibm.net>


═══ 11.7.7. availability ═══

       tune2fs is available for anonymous ftp from ftp.ibp.fr and
       tsx-11.mit.edu in /pub/linux/packages/ext2fs.


═══ 11.7.8. see also ═══

       dumpe2fs(8), e2fsck(8), mke2fs(8)


═══ 12. Other useful utilities ═══

Note:  These utilities require the EMX runtime library (it can be found at 
       ftp://ftp-os2.nmsu.edu/os2/unix/emx09c/emxrt.zip). 

     sync - commits data to disk 
     umount - detaches the current file system from a disk 
     remount - attaches a file system to a disk 
     hardlink - creates a hard link to an existing file 


═══ 12.1. sync - commits data to disk ═══

sync.exe does exactly what its Unix counterpart (sync) does : it commits all 
buffered data to disks. 

Invocation : 

sync

Note:  This utility will only commit data located on Linux ext2fs partitions. 
       It does NOT work on other file systems. 


═══ 12.2. umount - detaches the current file system from a disk ═══

unmount.exe will cleanly unmount a drive by committing all data and detaching 
the file system currently attached to it. After this operation, the OS/2 drive 
letter is still present, but the drive cannot be accessed. This utility issues 
a true unmount of the file system from the drive (eg it's a real unmount, not a 
hack, by calling the IFS's FS_MOUNT entry point). 

Invocation : 

umount X:

X: is the drive to be unmounted. If successful, the drive X: can no longer be 
accessed. Fo instance typing "dir X:" should give the following message : 
"SYS0026: The specified disk or diskette cannot be accessed.". If the media is 
removable, it can be safely removed or ejected. 

Note: 

     umount.exe requires that NO files are open on the drive. 
     umount.exe only works on LOCAL file systems (FAT, HPFS, ext2, CDFS ...). 
      It does NOT work on remote file systems (network drives) which should 
      come with their own unmount utility. 

 Warning: *** NEVER *** run any programs that directly access drives on a drive 
 previously unmounted with unmount.exe. Most of these programs blindly assume 
 the drive is formatted as FAT and thus can cause massive loss of data. These 
 programs include : 

     CHKDSK 
     PMCHKDSK 
     FORMAT (... unless you want to format the drive !) 
     Disk optimizers, defrag utilities 
     File system checkers (Norton disk doctor and such) 
     File recoverers, undelete utilities 
     Disk editors (unless you know what you're doing) 
     ... 

 Remember the main purpose of unmount.exe is to remove the media right after ! 


═══ 12.3. remount - attaches a file system to a disk ═══

remount.exe will force OS/2 to reattach a file system to the specified drive. 
If a file system was already attached, it will be detached first. remount.exe 
can be called to be able to gain access to a drive previously unmounted with 
unmount.exe. 

Invocation : 

remount X:

X: is the drive to be remounted. If successful, the drive X: can then be 
accessed normally. 

Note: 

     the OS/2 kernel will automatically determine the proper file system to be 
      attached to the drive. 
     remount.exe requires that NO files are open on the drive. 
     remount.exe only works on LOCAL file systems (FAT, HPFS, ext2, CDFS ...). 
      It does NOT work on remote file systems (network drives) which should 
      come with their own mount utility. 


═══ 12.4. hardlink - creates a hard link to an existing file ═══

hardlink is a small replacement of the UNIX "ln" command. Its purpose is to 
create a hard link to an existing file. 

Invocation : 

hardlink <oldname> <newname>

This will creates <newname> linked to the file <oldname>. 


═══ 13. Frequently Asked Questions (FAQ) ═══

     I cannot write to my Linux ext2 partitions, even if I enable write access 
      in CONFIG.SYS 

      Write access must be enabled for BOTH ext2flt.flt (/W switch) AND 
      ext2-os2.ifs (-rw switch) in CONFIG.SYS. 

     When the system boots, OS/2 tells me that EXT2-OS2.IFS contains an 
      invalid device driver or file system driver. 

      ext2-os2 requires the MWDD32.SYS driver to be loaded. Ensure that you 
      CONFIG.SYS contains a line like : 

            basedev=mwdd32.sys

     When OS/2 boots, I get SYS2068 messages. 

      You can try to turn on quiet mode on the following drivers : 

           ext2-os2.ifs (-q switch) 
           mwdd32.sys (-q switch) 
           ext2flt.flt (/Q switch) 

     I cannot run OS/2 programs from a Linux partitions 

      The Linux ext2 file system is a case sensitive file system, and this 
      breaks many OS/2 or DOS programs which uppercase or lowercase file names 
      (because FAT and HPFS are NOT case sensitive). The OS/2 program loader 
      itself uppercase file names. To run OS/2 programs from a Linux partition, 
      you must use the -case_retensive switch on the ext2-os2.ifs command line 
      (see Case sensitivity considerations). 

     My system hangs during boot, and it seems to occur before ext2-os2.ifs is 
      loaded. 

      There seems to be a problem between EXT2FLT.FLT, and drivers that makes 
      removable devices appear as fixed, like Syquest's SYQLOCK.FLT, IBM's 
      LOCKDRV.FLT and so on. If you have removable media like MO drives, ZIP 
      drives, Syquest drives and so on, try to put EXT2FLT.FLT as the FIRST 
      .FLT driver in config.sys, AHEAD of your vendor's filter driver. 

     Can I backup my Linux system from OS/2 using ext2-os2 and XYZ backup 
      software ? 

      Unfortunately you won't be able to backup a full Linux system from OS/2. 
      This is not a ext2-os2 limitation : your OS/2 XYZ backup software knows 
      NOTHING about UNIX permissions or UNIX special files (symbolic links, 
      fifos, sockets, drivers in /dev), it will not save them and naturally not 
      restore them (your restored system will probably not work). 

      However, if you intend to backup only data files (for instance a 
      development source tree), and the tree you plan to backup doesn't contain 
      special files, then you can backup them from OS/2. But when you restore, 
      you will have to manually reset UNIX ownership and file permissions using 
      the UNIX commands chown and chmod. 


═══ 14. History of changes ═══

This chapter describes the changes between each release of ext2-os2 

   1. From 0.1 to 0.2 
   2. From 0.2 to 0.3 
   3. From 0.3 to 0.4 
   4. From 0.4 to 0.5 
   5. From 0.5 to 0.6 
   6. From 0.6 to 0.7 
   7. From 0.7 to 0.8 
   8. From 0.8 to 0.9 
   9. From 0.9 to 2.0 
  10. From 2.0 to 2.1 
  11. From 2.1 to 2.2 
  12. From 2.2 to 2.3 
  13. From 2.3 to 2.4 


═══ 14.1. From 0.1 to 0.2 ═══

     DOS sessions can now access ext2fs partitions. WinOS2 sessions can also, 
      but they crash randomly and I couldn't figure out why yet ! 
     The per volume disk cache size is now user configurable from 64 Kb to 60 
      * 64 Kb. However it is not evident that a huge disk cache will improve 
      performance significantly : the cache lookup algorithm is still primitive 
      (single LRU list). 
     Sources are now included 
     FS_FINDFIRST, FS_FINDNEXT and FS_FINDFROMNAME now use the same lookup 
      routine 
     The directory entry cache is now implemented (identical to the Linux one) 
     readdir() now uses the original ext2_readdir() routine instead of a quick 
      (and somewhat dirty) rewrite 
     FS_FSINFO now returns a 'true' volume serial number instead of garbage. 
      This volume serial number is a CRC checksum of sector 0. 
     FS_MOUNT now checks for a previous instance of the volume to be mounted, 
      instead of remounting blindly a previously mounted volume. 
     FS_MOUNT with MOUNT_RELEASE flag now closes the volume and frees memory 
     brelse() no longer produces a system halt (trap D) when receiving a NULL 
      buffer : it now ignores it like its Linux counterpart 
     Direct access opens (OPEN_FLAGS_DASD) are now supported in both read only 
      and readwrite modes 
     Write capabilities added : 
        -  FS_MKDIR and FS_RMDIR implemented 
        -  FS_SHUTDOWN implemented 
        -  FS_OPENCREATE in writeonly and readwrite mode implemented 
        -  FS_WRITE implemented 
        -  FS_DELETE implemented 
     FS_CLOSE now calls ext2_release_file() when the last instance of a file 
      is closed 
     "-cache:<disk cache size>" Command line parameters added to set the per 
      volume disk cache size 
     "-q" command line parameter added (quiet initialisation) 
     "-rw" command line parameter added (write access allowed) 
     The memory I-node management routines (iput() iget() ...) are now a port 
      of their Linux counterparts instead of a rewrite. 
     Linux scheduling and semaphore management routines (sleep_on(), wake_up() 
      down() up() ...) have been rewritten using ProcBlock/ProcRun device 
      helpers instead of FSH_SEMxxx : now their behaviour should be identical 
      to their Linux counterparts. 
     The mount process is now almost identical as in ext2fs. 
     When OS/2 mounts a ext2fs partition, it sets the mount count to the 
      maximum to force Linux to autocheck ext2fs partitions "touched" by OS/2. 
     Buffer head structures now *NEVER* reach segment boundaries : this avoids 
      TREMENDOUS problems with some ext2fs routines (ext2_add_entry() ...) 
      which do some pointer arithmetics which simply don't work in a 16:16 
      memory model. 


═══ 14.2. From 0.2 to 0.3 ═══

     There is now a translation from Linux error codes (those in 
      linux/errno.h) to OS/2 error codes (bseerr.h). Now for instance, cmd.exe 
      says "Directory not empty" instead of "Cannot find message 65503". This 
      translation currently occurs in FS_READ, FS_WRITE, FS_RMDIR , FS_DELETE, 
      FS_MOVE and FS_MKDIR. 
     FS_COPY() is now supported (in fact emulated by the OS/2 kernel). It was 
      just a matter of returning ERROR_CANNOT_COPY from FS_COPY() - this makes 
      the kernel emulate the copy operation - instead of ERROR_NOT_SUPPORTED. 
      This applies only to copy operations inside or between ext2fs partitions. 
     Some cleanup in the code, by removing much direct ext2fs call and 
      replacing them by Linux VFS style calls (inode->i_op->read() for instance 
      instead of ext2_file_read()). This would ease te work of people trying to 
      port other Linux file systems to OS/2. 
     FS_MOVE is now implemented. 
     Corrected a bug in FS_DevHelp_ProcBlock : register di was not preserved 
      across the call, causing some random trap D. 
     Corrected a bug in read_inode when in DASD mode (OPEN_FLAGS_DASD) : some 
      I-node fields are now properly initialized ; This was preventing FS_READ 
      to work on a direct access handle (inode->i_mode was 0 and FS_READ 
      reported "not a regular file"). 
     Suppressed strict locking of buffers : now buffers can be used more than 
      once, conforming to Linux buffers (no more mutex in getblk and brelse), 
      some ext2fs routines do rely on it (ext2_rename)  ... ext2-os2 on OS/2 
      for SMP is not for tomorrow !! 
     Added -no_auto_fsck command line switch to disable the forced e2fsck when 
      Linux mounts a partition "touched" by OS/2. 
     Added code to close remaining open files at shutdown. Normally they 
      shouldn't be any, but the WPS lets some pending search handles 
      (DosFindxxx) ; this could result in file system corruption (a search 
      handle on a directory later deleted for instance : the deletion is not 
      committed until all references to the directory are closed). 
     Synchronous files (OPEN_FLAGS_WRITE_THROUGH and OPEN_FLAGS_NO_CACHE) 
      should be now handled correctly. 


═══ 14.3. From 0.3 to 0.4 ═══

     Corrected a bug in FS_MOUNT that caused a system halt (trap D) when 
      loading the WinOS2 file manager and ext2-os2 not being the last IFS 
      statement in CONFIG.SYS 
     Corrected a bug in FS_CLOSE that caused some random file system 
      corruption : ext2_release_file() was not correctly called. 
     Now FS_OPENCREATE opens only regular files, as it seems to be the default 
      OS/2 behaviour. (FS_OPENCREATE no longer opens directories). 
     Added the -trace option to ext2-os2.exe : ext2d.exe is no longer 
      necessary. 
     Now ext2-os2.exe correctly ends conversation with the IFS when being 
      closed. 
     ext2flt has been upgraded from V1.2 to V1.2b 
     During shutdown the flush mechanism could be disturbed if other activity 
      (lw daemon ...) was taking place (resulting in potential file and/or file 
      system corruption). This has been corrected. 
     The famous random trap D at CS:EIP=<any cs>:142e occuring at boot time 
      should now be corrected ... It was in fact a trap occuring inside the 
      Microsoft Visual C++ runtime library, only under retail OS/2 kernel (no 
      trap at all under kernel debugger !). After having taken 5 or 6 system 
      dumps I realized it was because somewhere an INT 21h call was issued 
      inside the vsprintf() routine. I changed vsprintf() for the Linux kernel 
      one. 


═══ 14.4. From 0.4 to 0.5 ═══

     Added support for strategy 2 I/O for non removable medias (for block 
      devices that support it !). This results in huge perfomance improvements 
      : throughput can be 5 times better than with standard I/O in some 
      circumstances (large file read/write). 
     Improved and corrected many bugs concerning time stamping of regular 
      files (last access date/time and last write date/time should now reflect 
      the reality !) - Time stamping of directories is yet to be improved. 
     Corrected a bug in ll_rw_block() that could cause some file/file system 
      corruption in rare circumstances. 
     Some cleanup in the source code : removed some inline assembly, improved 
      init routines .... 
     Added support for the standard DosFsCtl function 1 call (get extended 
      error info). 
     Added a first version of the VFSAPI library : it currently implements the 
      stat(), fstat() and sync() calls. 
     DosResetBuffer is now supported (FS_COMMIT entry point). 
     Disk cache management has been completely rewritten : the disk cache code 
      is now a port of the Linux buffer management code (dynamic disk cache 
      size) 
     ext2-os2.exe is now a PM application, and can now monitor some data from 
      ext2-os2.ifs. 
     Changed many FSH_PROBEBUF into VMLock/VMUnlock pairs so that user buffers 
      passed to the IFS can't be freed by other threads. 
     Added -case_retensive command line option to make ext2-os2.ifs behave 
      like HPFS (case is not significant but kept within the file system, for 
      instance File and FILE are considered to be the same file). This fixes 
      MANY apps which uppercase file names, but there's a risk of loosing data 
      if 2 files with the same case exists in the same directory (see usage 
      information for details). 
     Added command line switches to trace some IFS entry points (for instance 
      -FS_MKDIR will trace FS_MKDIR calls). This makes the debug version of 
      ext2-os2.ifs (ext2d.ifs) no longer necessary. 
     Deon van der Westhuysen's ext2flt.flt has been updated to version 1.2c in 
      this package. I recompiled ext2flt.flt with MS Visual C++ simply because 
      I don't have borland C++, and I needed to temporarily include some 
      debugging code to solve problems with strategy 2 I/Os (in fact the 
      problems were in OS2DASD.DMD !). Normally this "MS Visual C++" version 
      should behave absolutely like the original Borland compiled version. But 
      in case you prefer using the original version, you can find it on 
      ftp-os2.nmsu.edu (/os2/diskutil/e2flt12c.zip) or on quark.cs.sun.ac.za 
      (/pub/ext2flt/e2flt12c.zip). See the file ext2flt/makefile.msc for 
      details. 
     File systems greater than 2 Gb should now be mounted correctly. 
     Many other changes ! 


═══ 14.5. From 0.5 to 0.6 ═══

     Corrected a bug in FS_NEWSIZE : the I-node wasn't marked dirty after the 
      size change. This resulted in random truncation failures. 
     Corrected a bug in FS_OPENCREATE : When creating a new file or replacing 
      an existing one (OPEN_ACTION_REPLACE_IF_EXISTS or 
      OPEN_ACTION_CREATE_IF_NEW) the new file size passed to the IFS was 
      ignored, this resulted in truncation failures. 
     DosSetFileInfo is now supported for level 1 : it will change the file 
      dates. Level 2 (used to store or change EAS) is ignored because ext2fs 
      doesn't support EAS. 
     DosSetPathInfo is now supported for level 1 : it will change the file 
      dates. Level 2 (used to store or change EAS) is ignored because ext2fs 
      doesn't support EAS. 
     Corrected a *STUPID* bug that caused a system halt at boot time with 
      message "EXT2-os2 : buffer_init() - Couldn't allocate disk block hash 
      table" when the max. cache size was not set on the IFS command line. If 
      not set, the cache size remained uninitialized (0) instead of getting its 
      default value...hence the system halt when trying to allocate a 0 length 
      buffer. 
     Corrected a bug in date_dos2unix ; it made FS_CLOSE write incorrect file 
      times to disk (file dates were OK). 
     The time zone is now taken into account by ext2-os2.ifs. This makes 
      ext2-os2.ifs reflects file time stamps in local time instead of UTC 
      (ext2-os2.ifs and linux times should now match). The time zone is set in 
      CONFIG.SYS through the EXT2_TZ environment variable (minutes from UTC). 
     Trace for the following entry points has been added : 
        -  FS_PATHINFO 
        -  FS_FILEATTRIBUTE 
        -  FS_FILEINFO 
        -  FS_PROCESSNAME 
     Optimized a little bit the use of iget/iput in FS_DELETE and 
      FS_OPENCREATE. 
     Corrected some bugs related to time stamping. ext2-os2.ifs no longer let 
      the kernel do the time stamping, but does it itself. For instance when 
      copying a file on an ext2fs file system, the file creation date should 
      now be the one from the original file. 
     ext2-os2.exe now traces the use of iput and iget routines. 
     changed the panic code. 
     addded the -errors=[panic|cont] command line parameter 
     Setting or clearing the DOS READONLY attribute is now supported 
      (FS_FILEATTRIBUTE, FS_FILEINFO and FS_PATHINFO). 


═══ 14.6. From 0.6 to 0.7 ═══

     changed the EXT2_TZ valid range from [-720;720] to [-1440;1440] so that 
      every countries can be supported (New Zeland has a 13 hour shift from UTC 
      ...) 
     corrected a major bug in sys_bdflush that made the system hang when 
      ext2_lw.exe was loaded before any Linux partitions are mounted. 
     corrected a time stamp problem in FS_COMMIT. 


═══ 14.7. From 0.7 to 0.8 ═══

     corrected a bug that prevented negative values to be specified for the 
      EXT2_TZ environment variable. 
     corrected a bug in FS_OPENCREATE that caused a truncation failure when 
      OPEN_ACTION_REPLACE_IF_EXISTS was set and a non zero file size was 
      passed. 
     corrected a problem in bdflush() that made IOSTONE giving ridiculously 
      small results. After that IOSTONE results on my antic IDE disk jumped 
      from 80 (no it's not a typo!) to 2500. 
     changed a compiler option : now only externally called routines load DS 
      on entry. This allows better optimization. 
     now DGROUP is compeletely fixed in memory (instead of only 
      EXT2_FIXED_DATA), because I discovered that the compiler sometimes makes 
      references to some internal data in CONST segment (it seems to create 
      selector aliases for optimization I think). This could cause a trap E in 
      routines called in interrupt context if CONST was swapped out. I 
      simplified DevHelp bindings in devhelp2.asm accordingly. 
     added debug output to COM ports (COM1 or COM2). 
     added boot capability to ext2-os2 (wrote a micro FSD and a mini FSD). 


═══ 14.8. From 0.8 to 0.9 ═══

     corrected a bug in the file table management code. I incorrectly assumed 
      that FSH_SEGALLOC returned zeroed memory and this could cause a system 
      halt in put_last_in_list saying the used list was corrupted. 
     corrected a race condition problem in the file table management code. 


═══ 14.9. From 0.9 to 2.0 ═══

     Corrected yet another  bug in FS_NEWSIZE 
     Corrected (I hope) a bug that could make ext2-os2 crash when 
      HPOFS.IFS/MO.SYS are present. 
     Implemented swapper entry points : now SWAPPER.DAT can be put on a Linux 
      ext2fs partition. 
     Rewrote the mount/unmount/shutdown mechanism (including superblock 
      allocation). 
     Ported the whole ext2-os2.ifs to 32 bits. (this introduced many small 
      changes). 


═══ 14.10. From 2.0 to 2.1 ═══

     Added support for e2fsck and mke2fs 
     Improved lookup routines a lot (rewrote open_by_name to use dir_namei 
      instead of readdir). 
     Replaced some DevHlp32_VirtToLin calls with faster VDHQueryLin calls. 
     Replaced references to TKSSBase in mwdd32.sys with _TKSSBase in MVMD.DLL 
      (this one has one indirection level less). 
     Corrected a bug in iput when freeing a "DASD" mode inode (i_dirt was not 
      cleared). 
     Overall performance greatly improved (on my machine, IOSTONE benchmark 
      jumped from 4000 to almost 30000, with 4096 Kb cache) 
     Corrected a bug in do_ext2_rename (it caused a trap D when renaming a 
      directory). 
     The lazy writer thread in ext2_lw.exe now has a time critical priority; 
      This corrects low I/O throughput problem while other processes hog the 
      CPU. 
     Now this 32 bits version of ext2-os2 has the same boot capabilities as 
      the 16 bits version. 
     Ext2flt.flt no longer write protects sector 0 of virtual linux 
      partitions. This change was necessary to make mke2fs and e2fsck work. 
     Corrected a bug in the ext2-os2 log facility (ext2-os2.exe was hogging 
      100 % of the CPU). 
     Added the -force_strat2 option to force use of strategy 2 I/O (if 
      available) on removable media. The purpose of this option is to allow 
      fast access to PCMCIA hard drives and such, which are generally seen by 
      OS/2 as large floppies (and thus can only be formatted as FAT). 


═══ 14.11. From 2.1 to 2.2 ═══

     Corrected a stupid packaging error : microfsd.fsd and minifsd.fsd were 
      not included in the distribution. 
     Corrected a thunking problem in vfs_stat 
     Corrected a thunking problem in vfs_fstat 
     Added support to create hard links from OS/2 (new VFSAPI entry point : 
      vfs_link, new utility : hardlink.exe) 


═══ 14.12. From 2.2 to 2.3 ═══

     Corrected a bug in fsh32_qsysinfo that made ext2-os2 use case insensitive 
      comparisions, even if the -case_retensive option was not set. 
     Corrected several problems in microfsd.fsd and minifsd.fsd. 
     Added the umount.exe and remount.exe utilities for better removable media 
      support. 
     Corrected a problem in the mount process that made FSH_DOVOLIO fail with 
      a weird error when formatting a drive. Now ext2fs_check_magic operates 
      with the NO_CACHE flag set. 
     mke2fs now properly tells the user to run badblocks separately, instead 
      of crashing. 
     mke2fs now writes a DOS compatible boot sector in sector 0 and the media 
      type in sector 1, to avoid problems with an old time OS/2 bug : OS/2 
      incorrectly prevents FLOPPIES with non DOS boot records to be accessed by 
      an IFS. 
     Installation default is now to use ERRORS=CONTINUE instead of 
      ERRORS=PANIC option, in order to make mke2fs and e2fsck work without 
      halting the system in case of an unexpected error. 


═══ 14.13. From 2.3 to 2.4 ═══

     Corrected a memory overflow problem in microfsd.fsd that prevented OS/2 
      to boot from huge (Gigabyte range) ext2 partitions. 
     Increased the maximum number of Linux partitions in EXT2FLT.FLT from 16 
      to 26. 
     Removed the 2 Gb limit in e2fsck, mke2fs and badblocks. 
     Rewrote badblocks in a more OS/2 like style (makes use of the 
      DSK_VERIFYTRACK IOCTL instead of a standard read() call). It can now be 
      safely invoked through mke2fs's -c switch. 
     Now uext2.dll correctly removes FORMAT.COM specific options before 
      calling mke2fs.exe (eg. FORMAT X:: /FS:ext2 no longer fails). 
     Updated e2fsck and mke2fs from version 1.06 (e2fsprogs-1.06) to 1.10 
      (e2fsprogs-1.10). 
     Added a port of debugfs, chattr, lsattr and tune2fs. 
     Added path optimization : when relevant, directory lookup is now starting 
      from current directory and not from root directory. 
     Corrected some time stamping related problems that made the "make" 
      utility fail in some cases. 
     Corrected a race condition in the buffer cache code that could cause a 
      tight lockup in heavy I/O situations. 
     Volume labels are now supported and can be modified from OS/2, they are 
      limited by the kernel to 11 characters (Linux ext2 volume names can be 16 
      characters long, ext2-os2 truncates them to 11 characters). They can be 
      queried or set either through the OS/2 label or vol commands, or throught 
      tune2fs. 
     The volume serial number is now a 32 bits CRC of the Linux ext2 UUID (the 
      uuid is a 128 bits unique identifier present in the superblock. This is 
      now a standard feature of the Linux ext2 file system). 


═══ <hidden>  ═══

		    GNU GENERAL PUBLIC LICENSE
		       Version 2, June 1991

 Copyright (C) 1989, 1991 Free Software Foundation, Inc.
                          675 Mass Ave, Cambridge, MA 02139, USA
 Everyone is permitted to copy and distribute verbatim copies
 of this license document, but changing it is not allowed.

			    Preamble

  The licenses for most software are designed to take away your
freedom to share and change it.  By contrast, the GNU General Public
License is intended to guarantee your freedom to share and change free
software--to make sure the software is free for all its users.  This
General Public License applies to most of the Free Software
Foundation's software and to any other program whose authors commit to
using it.  (Some other Free Software Foundation software is covered by
the GNU Library General Public License instead.)  You can apply it to
your programs, too.

  When we speak of free software, we are referring to freedom, not
price.  Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
this service if you wish), that you receive source code or can get it
if you want it, that you can change the software or use pieces of it
in new free programs; and that you know you can do these things.

  To protect your rights, we need to make restrictions that forbid
anyone to deny you these rights or to ask you to surrender the rights.
These restrictions translate to certain responsibilities for you if you
distribute copies of the software, or if you modify it.

  For example, if you distribute copies of such a program, whether
gratis or for a fee, you must give the recipients all the rights that
you have.  You must make sure that they, too, receive or can get the
source code.  And you must show them these terms so they know their
rights.

  We protect your rights with two steps: (1) copyright the software, and
(2) offer you this license which gives you legal permission to copy,
distribute and/or modify the software.

  Also, for each author's protection and ours, we want to make certain
that everyone understands that there is no warranty for this free
software.  If the software is modified by someone else and passed on, we
want its recipients to know that what they have is not the original, so
that any problems introduced by others will not reflect on the original
authors' reputations.

  Finally, any free program is threatened constantly by software
patents.  We wish to avoid the danger that redistributors of a free
program will individually obtain patent licenses, in effect making the
program proprietary.  To prevent this, we have made it clear that any
patent must be licensed for everyone's free use or not licensed at all.

  The precise terms and conditions for copying, distribution and
modification follow.

		    GNU GENERAL PUBLIC LICENSE
   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION

  0. This License applies to any program or other work which contains
a notice placed by the copyright holder saying it may be distributed
under the terms of this General Public License.  The "Program", below,
refers to any such program or work, and a "work based on the Program"
means either the Program or any derivative work under copyright law:
that is to say, a work containing the Program or a portion of it,
either verbatim or with modifications and/or translated into another
language.  (Hereinafter, translation is included without limitation in
the term "modification".)  Each licensee is addressed as "you".

Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope.  The act of
running the Program is not restricted, and the output from the Program
is covered only if its contents constitute a work based on the
Program (independent of having been made by running the Program).
Whether that is true depends on what the Program does.

  1. You may copy and distribute verbatim copies of the Program's
source code as you receive it, in any medium, provided that you
conspicuously and appropriately publish on each copy an appropriate
copyright notice and disclaimer of warranty; keep intact all the
notices that refer to this License and to the absence of any warranty;
and give any other recipients of the Program a copy of this License
along with the Program.

You may charge a fee for the physical act of transferring a copy, and
you may at your option offer warranty protection in exchange for a fee.

  2. You may modify your copy or copies of the Program or any portion
of it, thus forming a work based on the Program, and copy and
distribute such modifications or work under the terms of Section 1
above, provided that you also meet all of these conditions:

    a) You must cause the modified files to carry prominent notices
    stating that you changed the files and the date of any change.

    b) You must cause any work that you distribute or publish, that in
    whole or in part contains or is derived from the Program or any
    part thereof, to be licensed as a whole at no charge to all third
    parties under the terms of this License.

    c) If the modified program normally reads commands interactively
    when run, you must cause it, when started running for such
    interactive use in the most ordinary way, to print or display an
    announcement including an appropriate copyright notice and a
    notice that there is no warranty (or else, saying that you provide
    a warranty) and that users may redistribute the program under
    these conditions, and telling the user how to view a copy of this
    License.  (Exception: if the Program itself is interactive but
    does not normally print such an announcement, your work based on
    the Program is not required to print an announcement.)

These requirements apply to the modified work as a whole.  If
identifiable sections of that work are not derived from the Program,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works.  But when you
distribute the same sections as part of a whole which is a work based
on the Program, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote it.

Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.

In addition, mere aggregation of another work not based on the Program
with the Program (or with a work based on the Program) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.

  3. You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the terms of
Sections 1 and 2 above provided that you also do one of the following:

    a) Accompany it with the complete corresponding machine-readable
    source code, which must be distributed under the terms of Sections
    1 and 2 above on a medium customarily used for software interchange; or,

    b) Accompany it with a written offer, valid for at least three
    years, to give any third party, for a charge no more than your
    cost of physically performing source distribution, a complete
    machine-readable copy of the corresponding source code, to be
    distributed under the terms of Sections 1 and 2 above on a medium
    customarily used for software interchange; or,

    c) Accompany it with the information you received as to the offer
    to distribute corresponding source code.  (This alternative is
    allowed only for noncommercial distribution and only if you
    received the program in object code or executable form with such
    an offer, in accord with Subsection b above.)

The source code for a work means the preferred form of the work for
making modifications to it.  For an executable work, complete source
code means all the source code for all modules it contains, plus any
associated interface definition files, plus the scripts used to
control compilation and installation of the executable.  However, as a
special exception, the source code distributed need not include
anything that is normally distributed (in either source or binary
form) with the major components (compiler, kernel, and so on) of the
operating system on which the executable runs, unless that component
itself accompanies the executable.

If distribution of executable or object code is made by offering
access to copy from a designated place, then offering equivalent
access to copy the source code from the same place counts as
distribution of the source code, even though third parties are not
compelled to copy the source along with the object code.

  4. You may not copy, modify, sublicense, or distribute the Program
except as expressly provided under this License.  Any attempt
otherwise to copy, modify, sublicense or distribute the Program is
void, and will automatically terminate your rights under this License.
However, parties who have received copies, or rights, from you under
this License will not have their licenses terminated so long as such
parties remain in full compliance.

  5. You are not required to accept this License, since you have not
signed it.  However, nothing else grants you permission to modify or
distribute the Program or its derivative works.  These actions are
prohibited by law if you do not accept this License.  Therefore, by
modifying or distributing the Program (or any work based on the
Program), you indicate your acceptance of this License to do so, and
all its terms and conditions for copying, distributing or modifying
the Program or works based on it.

  6. Each time you redistribute the Program (or any work based on the
Program), the recipient automatically receives a license from the
original licensor to copy, distribute or modify the Program subject to
these terms and conditions.  You may not impose any further
restrictions on the recipients' exercise of the rights granted herein.
You are not responsible for enforcing compliance by third parties to
this License.

  7. If, as a consequence of a court judgment or allegation of patent
infringement or for any other reason (not limited to patent issues),
conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License.  If you cannot
distribute so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you
may not distribute the Program at all.  For example, if a patent
license would not permit royalty-free redistribution of the Program by
all those who receive copies directly or indirectly through you, then
the only way you could satisfy both it and this License would be to
refrain entirely from distribution of the Program.

If any portion of this section is held invalid or unenforceable under
any particular circumstance, the balance of the section is intended to
apply and the section as a whole is intended to apply in other
circumstances.

It is not the purpose of this section to induce you to infringe any
patents or other property right claims or to contest validity of any
such claims; this section has the sole purpose of protecting the
integrity of the free software distribution system, which is
implemented by public license practices.  Many people have made
generous contributions to the wide range of software distributed
through that system in reliance on consistent application of that
system; it is up to the author/donor to decide if he or she is willing
to distribute software through any other system and a licensee cannot
impose that choice.

This section is intended to make thoroughly clear what is believed to
be a consequence of the rest of this License.

  8. If the distribution and/or use of the Program is restricted in
certain countries either by patents or by copyrighted interfaces, the
original copyright holder who places the Program under this License
may add an explicit geographical distribution limitation excluding
those countries, so that distribution is permitted only in or among
countries not thus excluded.  In such case, this License incorporates
the limitation as if written in the body of this License.

  9. The Free Software Foundation may publish revised and/or new versions
of the General Public License from time to time.  Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.

Each version is given a distinguishing version number.  If the Program
specifies a version number of this License which applies to it and "any
later version", you have the option of following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation.  If the Program does not specify a version number of
this License, you may choose any version ever published by the Free Software
Foundation.

  10. If you wish to incorporate parts of the Program into other free
programs whose distribution conditions are different, write to the author
to ask for permission.  For software which is copyrighted by the Free
Software Foundation, write to the Free Software Foundation; we sometimes
make exceptions for this.  Our decision will be guided by the two goals
of preserving the free status of all derivatives of our free software and
of promoting the sharing and reuse of software generally.

			    NO WARRANTY

  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
REPAIR OR CORRECTION.

  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES.

		     END OF TERMS AND CONDITIONS

	Appendix: How to Apply These Terms to Your New Programs

  If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these terms.

  To do so, attach the following notices to the program.  It is safest
to attach them to the start of each source file to most effectively
convey the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.

    <one line to give the program's name and a brief idea of what it does.>
    Copyright (C) 19yy  <name of author>

    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software
    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.

Also add information on how to contact you by electronic and paper mail.

If the program is interactive, make it output a short notice like this
when it starts in an interactive mode:

    Gnomovision version 69, Copyright (C) 19yy name of author
    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
    This is free software, and you are welcome to redistribute it
    under certain conditions; type `show c' for details.

The hypothetical commands `show w' and `show c' should show the appropriate
parts of the General Public License.  Of course, the commands you use may
be called something other than `show w' and `show c'; they could even be
mouse-clicks or menu items--whatever suits your program.

You should also get your employer (if you work as a programmer) or your
school, if any, to sign a "copyright disclaimer" for the program, if
necessary.  Here is a sample; alter the names:

  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
  `Gnomovision' (which makes passes at compilers) written by James Hacker.

  <signature of Ty Coon>, 1 April 1989
  Ty Coon, President of Vice

This General Public License does not permit incorporating your program into
proprietary programs.  If your program is a subroutine library, you may
consider it more useful to permit linking proprietary applications with the
library.  If this is what you want to do, use the GNU Library General
Public License instead of this License.


═══ <hidden>  ═══

This is an example where drive letters won't be changed when you toggle the 
ext2fs partition ID from 0x83 to 0x7. ext2flt gives the same result without 
having to alter the partition table. 

Before 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       OS/2              07      C:    Extended         Bootable
  1       Linux ext2fs      83      -     Extended         Bootable   <=== We toggle it from 0x83 ...

If we toggle the ext2fs partition type 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       OS/2              07      C:    Extended         Bootable
  1       Linux ext2fs      07      D:    Extended         Bootable   <=== ... to 0x07

If we use ext2flt.flt 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       OS/2              07      C:    Extended         Bootable
  2       Linux ext2fs      07      D:    Extended

The pink partition is the one as shown by ext2flt.flt to OS/2. 


═══ <hidden>  ═══


In this case, the OS/2 drive letter is changed from C: to D: when you change 
the Linux partition ID. If you want to use the ext2 IFS driver you'll have to 
use ext2flt.flt  so that the ext2fs partition appears after the OS/2 partition. 

Before 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       Linux ext2        83      -     Primary          Bootable   <=== We toggle it from 0x83 ...
  1       OS/2 System       07      C:    Extended         Bootable

If we toggle the partiton ID, the drive letters change : 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       Linux ext2        07      C:    Primary          Bootable   <=== ... to 0x07
  1       OS/2 System       07      D:    Extended         Bootable

The solution is to use ext2flt.flt : 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       OS/2 System       07      C:    Extended         Bootable
  2       Linux ext2        07      D:    Extended

The pink partition is the one as shown by ext2flt.flt to OS/2. 


═══ <hidden>  ═══

In this case, you cannot use the ext2 IFS driver alone and you are a good 
candidate for using ext2flt.flt. This is because only ONE primary partition can 
be active. ext2flt.flt will make your ext2fs partition appear as an extended 
volume located after your OS/2 partition. 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       Linux ext2        83      -     Primary          Bootable
  1       OS/2 System       07      C:    Primary (active) Bootable

If we use ext2flt.flt : 

Physical  Description  Partition  Drive   Type             Status
  drive                     ID    letter
  1       Boot mgr          0a      -     -                Startable
  1       OS/2 System       07      C:    Primary (active) Bootable
  2       Linux ext2        07      D:    extended


═══ <hidden>  ═══

[-q] : quiet initialization of ext2-os2.ifs (no banner during boot). Default is 
verbose initialization. 


═══ <hidden>  ═══

[-cache:<max. disk cache size>] Specifies the MAXIMUM disk cache size. This 
value is the maximum amount of physical memory ext2-os2.ifs can dynamically 
allocate for disk cache buffers. The default value is 240 Kb. The minimum value 
is 240Kb. The maximum value depends on system resources (free physical memory). 

Note:  Please read the chapter "Dynamic disk cache considerations" 


═══ <hidden>  ═══

[rw] : enable write access. Default is write disabled. 

Note:  If you use ext2flt, don't forget to enable write access on the 
       ext2flt.flt command line. (/W switch) 

 WARNING - BEFORE ENABLING WRITE ACCESS, REMEMBER THAT DOING REGULAR BACKUPS IS 
 NEVER A BAD PRACTICE !!! 


═══ <hidden>  ═══

[-no_auto_fsck] : prevents ext2-os2 from forcing Linux to run e2fsck on 
partitions "touched" by OS/2 Default is to force Linux to run e2fsck. 

Note:  Please read the chapter "Automatic check disk" 


═══ <hidden>  ═══

[-case_retensive] : makes ext2-os2.ifs behave like HPFS. If this option is set, 
case will not be significant but will be retained within the file system. For 
instance File and FILE are considered to be the same file. Default is case 
sensitive. 

WARNING - PLEASE READ "Case sensitivity considerations" BEFORE USING THIS 
OPTION !!! 


═══ <hidden>  ═══

[-<IFS entry point to trace>] : traces a specific IFS entry point. Using this 
option makes ext2-os2.ifs produce an output in ext2-os2.exe whenever the 
specified IFS entry point is entered. You can specify several IFS entry points 
to trace for instance "IFS=ext2-os2.ifs -FS_MKDIR -FS_RMDIR". 

Note:  Please notice that tracing will have a performance hit (can be 
       important). 

 Currently supported IFS entry points for tracing are : 

     FS_ALLOCATEPAGESPACE 
     FS_CHDIR 
     FS_CHGFILEPTR 
     FS_CLOSE 
     FS_COMMIT 
     FS_COPY 
     FS_DELETE 
     FS_DOPAGEIO 
     FS_EXIT 
     FS_FILEATTRIBUTE 
     FS_FILEINFO 
     FS_FINDCLOSE 
     FS_FINDFIRST 
     FS_FINDFROMNAME 
     FS_FINDNEXT 
     FS_FLUSHBUF 
     FS_FSINFO 
     FS_IOCTL 
     FS_MKDIR 
     FS_MOUNT 
     FS_MOVE 
     FS_NEWSIZE 
     FS_OPENCREATE 
     FS_OPENPAGEFILE 
     FS_PATHINFO 
     FS_PROCESSNAME 
     FS_READ 
     FS_RMDIR 
     FS_SHUTDOWN 
     FS_WRITE 


═══ <hidden>  ═══

[-no_strat2] : disable use of extended I/O operations (strategy 2 calls). 
Default is to use them on fixed disks when they are supported by the block 
device driver. Normally it works well with OS2DASD.DMD, but if you encounter 
problems with extended I/O with other drivers, you can disable them. 

Note:  Using extended I/O results in HUGE performance improvements ; disabling 
       them with have a tremendous performance hit. 


═══ <hidden>  ═══

[-errors=[panic|continue]] : specifies the behaviour to adopt when the ext2fs 
file system code detects an error within the file system. "panic" halts the 
system, "continue" makes ext2-os2 only output a message in ext2-os2.exe and 
continue. Default is "continue" (like Linux). 

Note:  If a system halt occurs with a message beginning with "ext2-os2 detected 
       a file system error on drive %c", it means that the ext2fs file system 
       code found a file system corruption. You MUST run e2fsck from Linux to 
       repair your file system. This type of error is generally not specific to 
       ext2-os2, but the error would also have occured on Linux. 
       "-errors=continue" will avoid a system halt and just mark the file 
       system as bad (forcing e2fsck on next Linux boot) ; but if you continue 
       to work intensively after a file system error, you are likely to spread 
       the error and make things even worse. 


═══ <hidden>  ═══

[-output=[com1|com2]] : Makes ext2-os2 write its messages on COM1 or COM2 port 
in addition to the ext2-os2 monitor. 

Note:  Output to COM ports has a performance hit, but can be necessary for 
       tracing or debugging in some cases. 


═══ <hidden>  ═══

[-tz:<time zone in minutes from UTC>] : Sets the time zone to be taken into 
account for file dates and times. Please read Time zone considerations. 


═══ <hidden>  ═══

[-force_strat2:<removable drive number to be forced to use strat 2 I/O>] 

Force a removable media to use strategy 2 I/O (if available). Normally ext2-os2 
uses strategy 2 I/O only for fixed disks. For certain large removable media 
like MO drives or PCMCIA hard drives which are seen by OS/2 as large floppies 
(and thus cannot be formatted as HPFS, nor use lazy writes), using strategy 2 
I/O can DRAMATICALLY improve performance. But you MUST leave the drive in its 
media and NEVER remove it while running. If you want to remove the drive 
anyway, you MUST unmount it first using the unmount.exe utility. This option 
can make sense for a PCMCIA drive (for instance I have a Thinkpad with a 260 Mb 
PCMCIA hard disk. I use this option to speed up access to this drive). 

Note:  The drive number on the command line is 0 based : 0 means 'A', 1 means 
       'B' and so on. 


═══ <hidden>  ═══

/Q : Quiet install. Prevents the filter from displaying messages. 


═══ <hidden>  ═══

/V : Verbose install. Displays some diagnostic messages. 


═══ <hidden>  ═══

/W : Allows the virtual partitions to be written to. 


═══ <hidden>  ═══

/A : Virtualize all partitions found. The only good reason to use this is along 
with the /M option when you want to control the mounting order of all the 
partitions. You must use the /W option if you use the /A option; otherwise OS/2 
may crash when it tries to write to the partition. 

Note:  Using the /A option may change your drive letters causing the your 
       system to be unbootable. Use the /M option together with /A. (This may 
       change in the future. 


═══ <hidden>  ═══

/M <mount list> : Mounts the partitions in the order specified in the mount 
list, which should be a list of numbers separated by commas. The partitions are 
numbered 0 on a first come basis. (Not the OS/2 uses.) 


═══ <hidden>  ═══

IBM Visualage C++ version 3.0 for OS/2. 

Note:  The CSD CTC305 is highly recommended. 


═══ <hidden>  ═══

IBM Assembly Language Processor (ALP) version 4.00. 

Note:  Earlier versions of ALP will NOT work. 


═══ <hidden>  ═══

Microsoft Visual C++ 1.51. 

Note:  This is actually MSC version 8.0c and is available as the 16 bits 
       compiler in the MS Visual C++ 2.0 package. 


═══ <hidden>  ═══

Borland C++ version 4 


═══ <hidden>  ═══

Microsoft Assembler (MASM) 5.10A.15 

Note:  This is the version of MASM included in the DevCon DDK. 


═══ <hidden>  ═══

Borland Turbo Assembler (available with Borland C++ 4) 


═══ <hidden>  ═══

IBM Linker for OS/2 (ILink) version 1. 

Note:  This is the version of ILINK included in the IBM Visualage C++ package, 
       with CSD CTC305 installed (ILINK version 01.04.r1a_CTC305A). 


═══ <hidden>  ═══

Microsoft Linker (LINK) version 5.6. 

Note:  This is the version of LINK included in the Microsoft Visual C++ version 
       2.0 package. 


═══ <hidden>  ═══

linux fdisk 3.04 output for disk 1 (this is actually an OS/2 port of fdisk : 
1:1 means /dev/hda1 and so on) : 

fdisk -uS -l 1:
Disk 1:: 6 heads, 35 sectors, 978 cylinders
Units = sectors of 512 bytes, counting from 0
   Device Boot    Start       End  #sectors  Id  System
      1:1          2100    104579    102480  83  Linux native
      1:2   *        35      2099      2065   a  OS/2 Boot Manager or Coherent swap
      1:3        104580    205379    100800   7  OS/2 IFS (e.g., HPFS) or NTFS or QNX2 or Advanced UNIX
      1:4             0         -         0   0  Empty

OS/2 fdisk output for disk 1 : 

FDISK /QUERY /DISK:1
Drive Name  Partition Vtype FStype Status   Start     Size
    1              :    1    0a      2          0        1
    1 Linux        :    1    83      1          1       50
    1 test        C:    1    07      5         51       49
                                             **BIOS:1024MB


═══ <hidden>  ═══

linux fdisk 3.04 output for disk 2 (this is actually an OS/2 port of fdisk : 
1:1 means /dev/sda1 and so on) : 

fdisk -uS -l 2:
Disk 2:: 64 heads, 32 sectors, 1169 cylinders
Units = sectors of 512 bytes, counting from 0
   Device Boot    Start       End  #sectors  Id  System
      2:1        163840   2095103   1931264   7  OS/2 IFS (e.g., HPFS) or NTFS or QNX2 or Advanced UNIX
      2:2   *        32    163839    163808  17  OS/2 BM: hidden IFS
      2:3       2095104   2394111    299008   5  DOS Extended
      2:4             0         -         0   0  Empty
      2:5   *   2095136   2394111    298976   7  OS/2 IFS (e.g., HPFS) or NTFS or QNX2 or Advanced UNIX

OS/2 fdisk output for disk 2 : 

FDISK /QUERY /DISK:2
Drive Name  Partition Vtype FStype Status   Start     Size
    2 ext2-os2     :    1    17      3          0       80
    2 Warp        D:    1    07      1         80      943
    2 001ff820    E:    2    07      0       1023      146
                                             **BIOS:1024MB


═══ <hidden>  ═══

#
# Start LILO global section
#
boot = /dev/hda1       # install LILO in the superblock of /dev/hda1.
vga = normal           # force sane state
ramdisk = 0            # paranoia setting
prompt
#
# End LILO global section
#
# Linux bootable partition config begins
image = /vmlinuz
  root = /dev/hda1
  label = linux
  read-only            # Non-UMSDOS filesystems should be mounted read-only for checking
# Linux bootable partition config ends


═══ <hidden>  ═══

Here's a brief description of what to do to install a WPS enabled system with 
BOOTOS2 on a LInux ext2fs partition, and put the WPS desktop tree on a HPFS or 
FAT partition (assume we want to store the WPS tree in D:\tmp\desktop) : 

   1. Go to the OS2 directory of your boot drive and make a copy of the file 
      named INI.RC (for instance TEST.RC) 
   2. Edit TEST.RC as follows : 

      Locate a line of the form 

              "PM_InstallObject" "Desktop;WPDesktop;?:\" "OBJECTID=<WP_DESKTOP>"
      and replace it with 

              "PM_InstallObject" "Desktop;WPDesktop;D:\tmp\desktop" "OBJECTID=<WP_DESKTOP>"

      Locate a line of the form 

              "PM_InstallObject" "Nowhere;WPFolder;?:\"  "OBJECTID=<WP_NOWHERE>"
      and replace it with 

              "PM_InstallObject" "Nowhere;WPFolder;D:\tmp\nowhere"  "OBJECTID=<WP_NOWHERE>"

   3. Reboot OS/2 FROM FLOPPIES 
   4. Press F3 to go to a command prompt. At the command prompt, go to the OS2 
      directory of your hard drive and type : 

            MAKEINI BOS2U300.INI TEST.RC

   5. Reboot OS/2 normally 
   6. Go to your BOOTOS2 directory and replace BOS2U300.INI with the one we've 
      created. 
   7. To install a WPS enabled system type the following command : 

            BOOTOS2 TARGET=F TYPE=WPS